You are on page 1of 765

Mountain Weather Research and Forecasting

Springer Atmospheric Sciences

For further volumes:


http://www.springer.com/series/10176
Fotini Katopodes Chow • Stephan F.J. De Wekker
Bradley J. Snyder
Editors

Mountain Weather Research


and Forecasting
Recent Progress and Current Challenges

123
Editors
Fotini Katopodes Chow Stephan F.J. De Wekker
Department of Civil and Environmental Department of Environmental Sciences
Engineering, MC 1710 University of Virginia
University of California, Berkeley McCormick Rd. 291
Berkeley, CA 94720-1710 Charlottesville, Virginia
USA USA

Bradley J. Snyder
Meteorological Service of Canada
#201 401 Burrard Street
Vancouver, BC, Canada

ISBN 978-94-007-4097-6 ISBN 978-94-007-4098-3 (eBook)


DOI 10.1007/978-94-007-4098-3
Springer Dordrecht Heidelberg New York London

Library of Congress Control Number: 2012940730

© Springer Science+Business Media B.V. 2013


This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology
now known or hereafter developed. Exempted from this legal reservation are brief excerpts in connection
with reviews or scholarly analysis or material supplied specifically for the purpose of being entered
and executed on a computer system, for exclusive use by the purchaser of the work. Duplication of
this publication or parts thereof is permitted only under the provisions of the Copyright Law of the
Publisher’s location, in its current version, and permission for use must always be obtained from Springer.
Permissions for use may be obtained through RightsLink at the Copyright Clearance Center. Violations
are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for
any errors or omissions that may be made. The publisher makes no warranty, express or implied, with
respect to the material contained herein.

Printed on acid-free paper

Springer is part of Springer Science+Business Media (www.springer.com)


Preface

This book is the result of a multiyear effort that began with the organization of a
workshop designed to bring researchers and forecasters together to discuss current
progress and challenges in mountain weather. The chapters herein represent the
topics from this Mountain Weather Workshop, which took place in Whistler, British
Columbia, Canada, 5–8 August 2008. The inspiration for the workshop and book
arose under the guidance of the American Meteorological Society (AMS) Mountain
Meteorology Committee.
One of the main goals of the workshop was to bridge the gap between the
research and forecasting communities by providing a forum for extended discussion
and joint education. The workshop consisted of lectures given by 13 distinguished
speakers, several discussion opportunities in small groups, and a day of laboratory
exercises designed for forecaster training for the 2010 Winter Olympics in Vancou-
ver. The lectures provided a detailed overview of important and emerging topics
in mountain meteorology. About 100 participants attended, roughly evenly split
among forecasters, researchers, and graduate students (see Fig. 12.2 for a picture
of participants). One of the highlights of the week was a group activity to design
the best observation and modeling system to nowcast for the Olympic ski jump
event; this was an excellent opportunity for researchers and operational forecasters
to work together and “bridge the gap.” The lectures from the workshop can be
accessed online in the COMET MetEd tutorial collection (http://www.meted.ucar.
edu/training module.php?id=878).
The chapters in this book have been written with the intent to provide a thorough
overview of each topic with an emphasis on recent research and progress in the field,
especially since the last collection of topics in mountain meteorology was published
more than two decades ago (Blumen 1990). It is our hope that this new offering
will be used extensively in mountain weather courses at universities and forecast
offices and also used as a general reference book for researchers, forecasters, and
students. Readers will be provided with a broad understanding of the fundamental
principles driving flow over complex terrain, including historical context for recent
developments and future directions for researchers and forecasters. For academic

v
vi Preface

Fig. P1.1 View of the PEAK 2 PEAK gondola connecting Whistler and Blackcomb mountains,
looking across toward Blackcomb. Whistler village is on the far left (© James Dunning. Reprinted
with permission)

researchers, the book will provide some insight into issues important to the
forecasting community. For the forecasting community, we hope the book will
provide training on fundamentals of flows specific to mountainous regions which
are notoriously difficult to predict, understanding of current research challenges,
and an opportunity to learn about the latest contributions and advancements to the
field. Our goal of bridging the gap between research and forecasting with this book
is aptly captured in the image below showing Whistler and Blackcomb mountains,
connected by the new PEAK 2 PEAK gondola, built for the 2010 Winter Olympics,
bridging the gap between the two mountains (Fig. P1.1).
The first chapter provides an overview of mountain weather and forecasting
challenges specific to complex terrain. This is followed by chapters that focus on
diurnal mountain/valley flows that develop under calm conditions (Chap. 2) and
dynamically driven winds under strong forcing (Chap. 3). The focus then shifts to
other specific phenomena that are difficult to understand and predict in mountain
regions: Alpine foehn (Chap. 4) and boundary layer phenomena and air quality
(Chap. 5). The following two chapters address processes that bring wet mountain
weather, in the form of rain, snow, or other hydrometeors, with a discussion of
specific orographic precipitation processes (Chap. 6) and the details of microphysics
parameterizations (Chap. 7). Having covered the major physical processes, the book
shifts to observation and modeling techniques used in mountain regions. First, a
detailed discussion of field measurements in complex terrain is given (Chap. 8).
Preface vii

Then, the following three chapters describe the basics of mesoscale numerical
modeling (Chap. 9), model configuration and physical parameterizations such as
turbulence (Chap. 10), and model applications in operational forecasting (Chap. 11).
The book concludes with a chapter that discusses the current state of research and
forecasting in complex terrain, including a vision of how to bridge the gap in the
future (Chap. 12).
We are quite fortunate to have a set of conscientious and thorough authors who
have contributed their knowledge and expertise to create this book, largely in their
spare time. We are also extremely grateful to the many reviewers who were involved
in ensuring the quality of this book. Given the length of some of the chapters, we
were particularly impressed by the care they took to thoroughly review the chapter
content, from comments on overall structure to details on style and formatting.
Funding to support the publication of this book and for student travel to the
workshop was provided by the National Science Foundation (NSF) (award ATM-
0810090). Funding for the workshop was provided by the American Meteorological
Society (AMS), the University Corporation for Atmospheric Research (UCAR)
acting on behalf of the Cooperative Program for Operational Meteorology, Edu-
cation and Training (COMET), and the Meteorological Service of Canada (MSC).
The workshop was mainly organized by us (editors of this book) with the help of
many others on the AMS Mountain Meteorology Committee, in addition to Cara
Campbell at AMS. We thank our colleagues who were AMS Mountain Meteorology
Committee members with us over the years (Brian Colle, Lisa Darby, Mike Meyers,
Stephen Mobbs, Greg Poulos, Heather Reeves, Alex Reinecke, Simon Vosper, Doug
Wesley, and David Whiteman) and members of the AMS publications department
(Peter Lamb, Sarah Jane Shangraw, and Ken Heideman) for their support of this
effort.
Contents

1 Mountain Weather Prediction: Phenomenological


Challenges and Forecast Methodology .. . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 1
Michael P. Meyers and W. James Steenburgh
2 Diurnal Mountain Wind Systems . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 35
Dino Zardi and C. David Whiteman
3 Dynamically-Driven Winds . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 121
Peter L. Jackson, Georg Mayr, and Simon Vosper
4 Understanding and Forecasting Alpine Foehn . . . . . .. . . . . . . . . . . . . . . . . . . . 219
Hans Richner and Patrick Hächler
5 Boundary Layers and Air Quality in Mountainous Terrain . . . . . . . . . . . 261
Douw G. Steyn, Stephan F.J. De Wekker, Meinolf Kossmann,
and Alberto Martilli
6 Theory, Observations, and Predictions of Orographic
Precipitation .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 291
Brian A. Colle, Ronald B. Smith, and Douglas A. Wesley
7 Microphysical Processes Within Winter Orographic
Cloud and Precipitation Systems . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 345
Mark T. Stoelinga, Ronald E. Stewart, Gregory Thompson,
and Julie M. Thériault
8 Observational Techniques: Sampling the Mountain Atmosphere.. . . . 409
Robert M. Banta, C.M. Shun, Daniel C. Law, William Brown,
Roger F. Reinking, R. Michael Hardesty, Christoph J. Senff,
W. Alan Brewer, M.J. Post, and Lisa S. Darby

ix
x Contents

9 Mesoscale Modeling over Complex Terrain: Numerical


and Predictability Perspectives. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 531
James D. Doyle, Craig C. Epifanio, Anders Persson,
Patrick A. Reinecke, and Günther Zängl
10 Meso- and Fine-Scale Modeling over Complex Terrain:
Parameterizations and Applications.. . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 591
Shiyuan Zhong and Fotini Katopodes Chow
11 Numerical Weather Prediction and Weather Forecasting
in Complex Terrain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 655
Brad Colman, Kirby Cook, and Bradley J. Snyder
12 Bridging the Gap Between Operations and Research to
Improve Weather Prediction in Mountainous Regions .. . . . . . . . . . . . . . . . 693
W. James Steenburgh, David M. Schultz, Bradley J. Snyder,
and Michael P. Meyers

Author Index.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 717

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . 725


Contributors

Robert M. Banta NOAA Earth System Research Laboratory, Boulder, CO, USA
W. Alan Brewer NOAA Earth System Research Laboratory, Boulder, CO, USA
William Brown National Center for Atmospheric Research, Boulder, CO, USA
Fotini Katopodes Chow Department of Civil and Environmental Engineering,
University of California, Berkeley, CA, USA
Brian A. Colle School of Marine and Atmospheric Sciences, Stony Brook Univer-
sity/SUNY, Stony Brook, NY, USA
Brad Colman NOAA/National Weather Service, Seattle-Tacoma, WA, USA
Kirby Cook NOAA/National Weather Service, Seattle-Tacoma, WA, USA
Lisa S. Darby NOAA Earth System Research Laboratory, Boulder, CO, USA
Stephan F.J. De Wekker Department of Environmental Sciences, University of
Virginia, Charlottesville, VA, USA
James D. Doyle Naval Research Laboratory, Marine Meteorology Division, Mon-
terey, CA, USA
Craig C. Epifanio Texas A&M University, Texas, USA
Patrick Hächler Federal Office of Meteorology and Climatology MeteoSwiss,
Zurich, Switzerland
R. Michael Hardesty NOAA Earth System Research Laboratory, Boulder, CO,
USA
Peter L. Jackson University of Northern British Columbia, Prince George, BC,
Canada
Meinolf Kossmann Climate and Environment Consultancy, Deutscher Wetterdi-
enst, Offenbach am Main, Germany

xi
xii Contributors

Daniel C. Law NOAA Earth System Research Laboratory, Boulder, CO, USA
Alberto Martilli CIEMAT Unidad de Contaminacion Atmosferice, Madrid, Spain
Georg Mayr University of Innsbruck, Innsbruck, Austria
Michael P. Meyers NOAA/National Weather Service, Grand Junction, CO, USA
Anders Persson UK MetOffice, SMHI, Norrköping, Sweden
M.J. Post NOAA Earth System Research Laboratory, Boulder, CO, USA
Patrick A. Reinecke Naval Research Laboratory, Marine Meteorology Division,
Monterey, CA, USA
Roger F. Reinking NOAA Earth System Research Laboratory, Boulder, CO, USA
Hans Richner Institute for Atmospheric and Climate Science (IACETH), ETH,
Zurich, Switzerland
David M. Schultz Division of Atmospheric Sciences, Department of Physics,
University of Helsinki, Finland
Finnish Meteorological Institute, Helsinki, Finland
Centre for Atmospheric Science, School of Earth, Atmospheric and Environmental
Sciences, University of Manchester, Manchester, United Kingdom
Christoph J. Senff NOAA Earth System Research Laboratory, Boulder, CO, USA
Cooperative Institute for Research in the Environmental Sciences, Boulder, CO,
USA
C.M. Shun Hong Kong Observatory, Kowloon, Hong Kong
Ronald B. Smith Geology and Geophysics Department, Yale University, New
Haven, CT, USA
Bradley J. Snyder Meteorological Services of Canada, Vancouver, BC, Canada
W. James Steenburgh Department of Atmospheric Sciences, University of Utah,
Salt Lake City, UT, USA
Ronald E. Stewart Department of Environment and Geography, University of
Manitoba, Winnipeg, MB, Canada
Douw G. Steyn Department of Earth and Ocean Sciences, The University of
British Columbia, Vancouver, BC, Canada
Mark T. Stoelinga 3TIER, Inc, Seattle, WA, USA
Julie M. Thériault National Center for Atmospheric Research, Boulder, CO, USA
Gregory Thompson National Center for Atmospheric Research, Boulder, CO,
USA
Contributors xiii

Simon Vosper Met Office, Exeter, UK


Douglas A. Wesley Compass Wind, Denver, CO, USA
C. David Whiteman Atmospheric Sciences Department, University of Utah, Salt
Lake City, UT, USA
Dino Zardi Atmospheric Physics Group, Department of Civil and Environmental
Engineering, University of Trento, Italy
Günther Zängl Deutscher Wetterdienst, Offenbach, Germany
Shiyuan Zhong Department of Geography, Michigan State University, East Lans-
ing, MI, USA
Chapter 1
Mountain Weather Prediction:
Phenomenological Challenges and Forecast
Methodology

Michael P. Meyers and W. James Steenburgh

Abstract This chapter summarizes the modern practice of weather analysis and
forecasting in complex terrain with special emphasis placed on the role of humans.
Weather in areas of complex terrain affects roughly half of the world’s land
surface, population, and surface runoff, and frequently poses a threat to lives and
property. Mountain weather phenomena also impact a diverse group of users, which
may have both beneficial and detrimental implications on societal and economic
levels.
Advances in forecast skill derive not only from advances in numerical weather
prediction, geophysical observations, and cyber infrastructure, but also improve-
ments in the utilization of these advances by operational weather forecasters.
Precipitation skill scores during the past two decades, for example, show that
operational weather forecasters have maintained a consistent threat score advantage
over numerical precipitation forecasts. Although the role of human forecasters is
evolving, for many applications, the so-called “human-machine mix” continues to
provide an improved product over what can be produced by automated systems
alone. To produce the best forecasts possible for the benefit of society, it is crucial for
the mountain meteorologist to possess an in-depth knowledge of mountain weather
phenomena and the tools and techniques used for atmospheric observations and
prediction in complex terrain.

M.P. Meyers ()


National Weather Service, 2844 Aviators Way, Grand Junction, CO 81506, USA
e-mail: mike.meyers@noaa.gov
W.J. Steenburgh
Department of Atmospheric Sciences, University of Utah, Salt Lake City, UT, USA

F. Chow et al. (eds.), Mountain Weather Research and Forecasting, 1


Springer Atmospheric Sciences, DOI 10.1007/978-94-007-4098-3 1,
© Springer ScienceCBusiness Media B.V. 2013
2 M.P. Meyers and W.J. Steenburgh

1.1 Introduction

Contemporary mountain weather forecasting involves the integration of geophysical


observations, numerical and statistical analysis and modeling, and human cognition
to meet the challenges posed by a diverse range of terrain-induced phenomena.
This integration, known as the “human-machine mix” (Snellman 1977), produces
significantly better forecasts than can be produced by automated systems alone, with
the value added by human cognition representing a 5–10 year advance in numerical
weather prediction skill (Bosart 2003; Steenburgh et al. 2012, Chap. 12). The
human-machine mix is only effective, however, when operational meteorologists
possess in-depth knowledge of mountain weather phenomena and the tools and
techniques used for atmospheric observation and prediction in complex terrain.
In this chapter we provide a review of the major phenomenological challenges
confronting mountain meteorologists and qualitatively describe the contemporary
forecast process, with emphasis on the human element over complex terrain. Our
goal is to provide a foundation for subsequent chapters that focus on specific
mountain weather phenomena or forecast tools and techniques, including numerical
weather prediction, which ultimately must be integrated to produce societally
relevant forecasts. We conclude with a discussion of ongoing forecast applications
in areas of complex terrain.

1.2 Phenomenological Challenges in Complex Terrain

Mountains cover 25% of the Earth’s land surface, contain 26% of the global
population, and produce 32% of the surface runoff (Meybeck et al. 2001). Hills
and plateaus account for another 21% of the land surface, 20% of the population,
and 19% of the runoff. Thus, the weather in areas of complex terrain affects roughly
half of the world’s land surface, population, and surface runoff. The numbers are
greater if one considers the remote effects of mountains on the general circulation,
storm tracks, moisture transport, and river runoff.
The protection of lives and property from high impact events is a forecast
priority; in addition, accurate forecasts of day-to-day mountain weather variability
benefit commerce and the general public. For instance, many mountain recreation-
alists are impacted by mountain weather. In the United States, the total number of
people who participated in outdoor activities in 2007 is estimated at 217 million
(Cordell 2008). Outdoor recreation (camping, snow sports, rafting, hiking, hunting
and fishing, etc.) contributes $730 billion to the economy annually and supports
6.5 million jobs (1 in 20 U.S. jobs) according to the Outdoor Industry Association
(http://www.outdoorindustry.org). In addition to the general population, numerous
other industries are dependent on weather that occurs over complex terrain.
The primary phenomenological challenges confronting mountain meteorologists
include: (a) snow and (b) ice storms produced by orographic precipitation and/or
1 Mountain Weather Prediction: Phenomenological Challenges. . . 3

terrain-induced cold advection and cold-air damming; (c) floods, landslides, and
debris flows generated by orographic rainfall and/or terrain-induced deep con-
vection; (d) droughts; (e) extreme wildfire spread and behavior driven by fuels,
topography, and weather; (f) severe local windstorms created by high-amplitude
mountain waves and gap flows; (g) severe convective storms; and (h) cold-air
pools and associated air quality hazards. In addition to loss of life, high-impact
weather events generated by these phenomena can produce staggering economic
losses, often requiring participation from government entities to address them. To
adequately meet these phenomenological challenges, forecasters need not only a
strong foundation in mountain meteorology (also see: Whiteman 2000), but also
knowledge in areas such as climatology, hydrology, ecology, land-surface processes,
and societal vulnerability.

1.2.1 Snowstorms

Snowstorms (also see Chaps. 6 and 7) exert a heavy toll on public safety and
transportation. The average annual cost of snow removal for public roadways in
the United States exceeds US $2 billion (Doesken and Judson 1997; National
Research Council (NRC) 2004). Thornes (2000) estimated that the United Kingdom
spends over US $2 billion annually on direct and indirect costs related to winter
maintenance of roadways and road traffic delays. Airport delays and closures cost
US $3 billion annually for US carriers and produce adverse sociological impacts for
the travelers. The potential benefits from improved forecasting of snow and icing
diagnostics at U.S. airports exceed US $600 million annually (Adams et al. 2004).
In mountainous regions, the economic losses due to snow-related highway
closures are considerable (Fig. 1.1). For example, in Europe, during a prolonged
snowy period in February 1999, more than 40 tourist resorts in the Swiss Alps
were cut off from the outside world due to road closures for up to 14 consecutive
days, resulting in indirect costs of US $200 million (Nöthiger and Elsasser 2004).
In the United States, losses produced by the snow-related closure of Interstate
90, the major highway bisecting the Cascade Mountains of Washington State,
are estimated at US $700,000 per hour, and similar shut downs of Interstates 70
and 80 through Colorado and Wyoming, respectively, approach US $1 million
per hour. Poor driving conditions due to weather cause over 1.5 million vehicular
accidents in the United States annually with total economic costs of US $42 billion
dollars. These vehicular accidents result in 800,000 injuries with 7,400 fatalities
indirectly related to poor weather conditions (National Research Council 2004). A
conservative estimate of the annual costs of weather-related vehicular accidents in
Canada is US $1.1 billion dollars (Andrey et al. 2001).
Snowstorms seriously impact urban corridors adjacent to mountain locations.
Personal claims from the March 2003 Colorado Front Range Blizzard exceeded
US $93 million. However, with the bad comes the good: during the blizzard the
4 M.P. Meyers and W.J. Steenburgh

Fig. 1.1 Snow removal in the southwestern Colorado Mountains (Courtesy of the Colorado
Avalanche Information Center [CAIC] and the Colorado Department of Transportation [CDOT])

snowpack in the Colorado Rockies went from inadequate to above 100% of average.
A senior agriculturist for the Western Sugar Cooperative in eastern Colorado called
it a billion-dollar storm due to its positive impact on the snowpack (Kohler 2003).
Avalanches are another potential impact of winter storms in complex ter-
rain (Fig. 1.2). Recent major weather related avalanche disasters have killed 14
in Súðavı́k, Iceland in January 1995; 20 in Flateyri, Iceland in October 1995
(Jóhannesson and Arnalds 2001); and 55 in Swiss and Austrian villages in February
1999 (Keiler et al. 2005). Avalanche disasters are not restricted to mountainous
regions. On New Year’s Day 1999, nine were killed in Kangiqsualujjuaq, Quebec
when an avalanche ran down a steep hill and hit the local school (Branswell 1999).
Conversely, mountain snowstorms can be a winter recreationalist’s dream
(Fig. 1.3) and provide benefits for the winter recreational economy. There are
approximately 6,000 ski resorts with nearly 400 million skier days per year (i.e.,
1 day of downhill skiing or snowboarding with a pass/ticket) globally (Hudson
2002; Skistar 2009). Europe is the largest market with 200 million skier days
per year, followed by North America with 80 million skier days per year (http://
www.nsaa.org/nsaa/press). The ski industry in the United States has annual revenue
of approximately US $12 billion (Scott 2006). Additionally, in North America,
snowmobilers spend more than US $28 billion annually on equipment, clothing,
1 Mountain Weather Prediction: Phenomenological Challenges. . . 5

Fig. 1.2 Avalanche over the southwestern Colorado Mountains (Courtesy of CAIC and CDOT)

accessories and vacations (US $6 billion in Canada) (International Snowmobile


Manufacturers Association http://www.snowmobile.org). Avalanches can, however,
pose a hazard for these recreationists. From 1991 to 2001, the International
Commission of Avalanche Rescue (http://www.ikar-cisa.org) reports nearly 1,500
fatalities due to avalanches with France and the United States having the two highest
percentages of deaths.

1.2.2 Ice-Storms

The advection and/or entrenchment of cold air by cold-air damming (Forbes et al.
1987; Bell and Bosart 1988) and orographic channeling affect the locations of rain-
snow transition zones in winter storms (also see: Chaps. 6 and 7). The January 1998
ice storm that devastated the northeastern United States and southeastern Canada
6 M.P. Meyers and W.J. Steenburgh

Fig. 1.3 Recreational skiing in Colorado (Courtesy of Meyers)

produced 80–100 mm of freezing rain over a 5 day period, causing more than
US $4 billion in economic damage, including US $3 billion in Canada (Reagan
1998; Roebber and Gyakum 2003). Cold-air channeling within the St. Lawrence,
Ottawa River, and Lake Champlain valleys enabled the persistence of low-level
cold air within the precipitating region, controlled the position of the surface-
based freezing line, and enhanced precipitation rates through frontogenesis. Such
orographic channeling can also lower freezing levels and snow levels in interior
basins and mountain passes, such as the Columbia River Basin, Snake River Plain,
Columbia River Gorge, Snoqualmie Pass, and the Frazier River Valley in the
Cascade Mountains and Coast Range of western North America (e.g., Decker 1979;
Ferber et al. 1993; Steenburgh et al. 1997).

1.2.3 Floods, Flash Floods and Debris Flows

Floods are among the most common of geologic hazards worldwide. Typically, most
river systems flood (i.e., leave their confining channels and flow outward onto the
1 Mountain Weather Prediction: Phenomenological Challenges. . . 7

adjacent floodplain) every year or two. There are two types of floods: regional floods
that can last for several weeks or months, and flash floods that last for minutes to
hours. Both are dangerous and capable of adversely impacting lives and property
(National Research Council 2005; http://www.azgs.state.az.us/).
Flooding produced by orographic precipitation is responsible for many of the
weather-related natural disasters in mountainous regions. Of the 13 weather-related
disasters observed in the western United States since 1980 with damages and costs
exceeding US $1 billion, four were produced by heavy and/or persistent orographic
precipitation and associated surface snowmelt (Lott and Ross 2006). In the United
States, floods cost upwards of US $6 billion and about 140 people are killed by
floods each year (Knutson 2001). In Europe, losses from the Italian Piedmont floods
of 1994 included 64 casualties and US $9 billion in property damage (Linnerooth-
Bayer and Amendola 2003; Barredo 2007). The damage and fatalities produced
by these events frequently extend over the flood plains well removed from the
orography responsible for the precipitation enhancement.
Orography also contributes to localized but extremely hazardous flash floods
by influencing the formation and movement of deep convection and mesoscale
convective systems. Well documented examples include the Rapid City Flash Flood
of June 1972, which killed 238 and produced US $100 million in property damage
in South Dakota; the Big Thompson Canyon Flood of July 1976, which killed 145
and produced US $25.5 million in property damage in Colorado; the Vaison-La-
Romaine Flash Flood of September 1992, which killed 46 and produced US $460
million in property damage in southeastern France (Maddox et al. 1978; Sénési
et al. 1996; Barredo 2007); and the September 2002 severe flood event in the
western Mediterranean mountainous region of southern France (Nuissier et al. 2008;
Ducrocq et al. 2008) which killed 24 people and produced an economic damage
estimated at nearly US $2 billion (Huet et al. 2003).
Debris flows can occur on steep slopes where loose, unconsolidated earthen
materials, such as soils and rocks, experience gravitational acceleration during
heavy rains, glacial melt or snowmelt (Iverson 1997). Debris flows can move
downslope rapidly, at speeds of greater than 10 m s1 ; their less viscous, fine-
grained relative, mudflows, have been clocked traveling at 40 m s1 in steep
mountain canyons. Wildfires may potentially increase the risks for debris flow
development by destroying vegetation and making soils more hydrophobic (Cannon
et al. 1998, 2001, 2003; Cannon and Reneau 2000). The major hazards of debris
flows are from the impact of earthen materials, such as boulders and rocks, and
being buried or carried away by the flow. Debris flows can be devastating to
life and property. In the United States, damage estimates due to debris flows are
close to US $3 billion annually (Restrepo et al. 2008). During December of 1999
exceptionally heavy rain triggered catastrophic floods and landslides along portions
of the mountainous coastal region of northern Venezuela (Lyon 2003). Over 10,000
fatalities were reported and the cost of reconstruction was estimated at nearly US $2
billion.
8 M.P. Meyers and W.J. Steenburgh

1.2.4 Droughts

One long duration weather phenomenon which may impact short term decision
making for a mountain meteorologist, including hydrologists and fire weather
meteorologists, is drought. Droughts come in various forms, which may impact
society with varying intensities and durations. By definition, a drought is “a period
of abnormally dry weather sufficiently prolonged for the lack of water to cause
serious hydrologic imbalance (i.e. crop damage, water-supply shortage, etc.) in
the affected area” (American Meteorological Society 1986). On an annual basis,
average losses and costs in the United States due to drought are estimated to exceed
US $8 billion (Knutson 2001).
The main water source for over 30 million people in the mainly arid climate
of the southwestern United States is the Colorado River. Snowmelt from mountain
snowpack provides over 70% of the water supply for this region (Chang et al. 1987;
Christensen et al. 2004), and the estimated benefit of water storage exceeds US
$350 billion dollars annually (Adams et al. 2004). Water is also the driving force
behind the agricultural industry of the southwestern United States. In California,
the agriculture industry accounts for nearly US $150 billion annually according
to the California Department of Food and Agriculture (http://www.cdfa.ca.gov/).
The economic impacts on the agricultural community due to droughts can be
tremendous. The University of California, Davis, estimated that US $2.8 billion
in agriculture-related wages, and as many as 95,000 jobs across the valley were
potentially lost due to the 2008 drought (Howitt et al. 2009).
Droughts can adversely impact the winter recreational industry, such as skiing
and snowmobiling, which depends on mountain snowpack. For example, many
skiers and snowboarders tend to favor lower-density, abundant snow (Steenburgh
and Alcott 2008), and low quality or meager snow can affect the demand for ski days
(Englin and Moeltner 2004). Summertime recreational use is not immune from low
snowpack. Low runoff during drought years can significantly reduce white-water
rafting revenue, which in Colorado alone, attracts 540,000 visitors and generates
US $150 million annually. Recreational fishing and hunting may also be affected by
drought. The impact on fish and other aquatic life due to drought may be significant
(Matthews and Marsh-Matthews 2003). Drought-depleted ecosystems and wetlands
would have a drastic effect on other wildlife.
Drought is not as visually obvious as other weather phenomena in the mountains,
but it is perhaps the most devastating to the ecological state because it can signif-
icantly weaken a forest’s defenses against insect infestation and wildfires (Morris
and Walls 2009). Recently, various populations of bark beetles are impacting
the western United States and western Canada with unprecedented levels of tree
mortality (Fettig et al. 2007). For instance, beetles have devastated several million
acres of trees in Colorado and Wyoming by the end of 2008 (Robbins 2008).
Extreme drought conditions in mountainous regions may also lead to increased
wildfire activity, as was the case in 1988 (Yellowstone National Park Fire) and in
2002 (Hayman, Missionary Ridge and Coal Seam fires in Colorado).
1 Mountain Weather Prediction: Phenomenological Challenges. . . 9

Fig. 1.4 2002 Big Fish Wildfire in Colorado (Photo by Mike Chamberlain; courtesy of the
NOAA/NWS/GJT)

1.2.5 Wildfire Behavior

Prescribed and unplanned wildfires (Fig. 1.4) impact tens of millions of acres
annually around the world (also see Chap. 2). In the western United States, where
the wealth and development of communities has migrated to the wildlands or the
wildland-urban interface over the past few decades, there have been eight wildfires
since 1980 that have produced more than US $1 billion in damage (Lott and Ross
2006). Of particular safety concern for firefighters are the conditions that lead to
fire “blowups”, which cause rapid fire spread. Wildfires over complex terrain are
especially susceptible to these blowups, due the diurnal changes of surface wind
direction and speed, as well as the surface mixing of synoptic and convectively-
driven winds.
One of the most devastating wildfires, the Big Burn of 1910, killed at least 78
firefighters and burned millions of acres in northern Idaho and western Montana
(National Wildfire Coordination Group 1997). This wildfire raised political aware-
ness of the economic and human impacts produced by wildfires. More recently,
from 1990 to 2006, there were 310 fatalities during western US wildland fire
operations, including the 1994 Storm King wildfire (14 fatalities) (National Wildfire
Coordination Group 2007). The number one cause of death during this time period
10 M.P. Meyers and W.J. Steenburgh

was burnovers (40%), “a situation where personnel or equipment is caught in an


advancing flamefront” (National Wildfire Coordination Group 2007), and secondly
due to aircraft and vehicular accidents (30%). Wildfires often have an adverse
impact on recreation and tourism. After the devastating wildfires in Yellowstone
National Park in 1988, visits to the park dropped 15% the following year (National
Park Service 2009). High wildfire danger may result in the closure of wildland areas,
which negatively impact recreation and logging.

1.2.6 Severe Windstorms Created by High-Amplitude


Mountain Waves and Gap Flows

Terrain-forced flows such as downslope windstorms and gap winds can produce
severe winds, many of which have exotic names like the Chinook along the eastern
slopes of the Rocky Mountains, the Mistral of southeast France, the Bora of
Slovenia, Croatia, and Bosnia, the Zonda along the east slopes of the Andes in South
America, the Taku in Alaska and the Föehn of central Europe (also see: Chaps. 3
and 4). Surface winds during extreme downslope windstorms can exceed hurricane
force (>33 m s1 ) and associated turbulence, rotors, and aircraft icing present a
threat to aviation safety (Nance and Colman 2000). Downslope windstorms are also
associated with the rapid spread of wildfires (e.g. Santa Ana [southwest United
States], Föehn, Chinook), making some regions which experience these winds
particularly prone to extreme fire weather behavior. Terrain-forced flows are also
a concern for maritime travel and commerce in and near mountainous coastal zones
around the world. It even has been suggested that these downslope windstorms are
often associated with illnesses ranging from migraines to psychosis (Soyka 1983).
Less frequent downslope windstorms sometimes occur in a synoptically un-
common flow pattern and represent a difficult forecast problem given their low
frequency. In October 1997, while a blizzard was occurring over the Front Range
of Colorado (Poulos et al. 2002), an easterly downslope windstorm was occurring
over the Mt. Zirkel Wilderness in the north central Colorado Rockies (Meyers
et al. 2003). This easterly downslope wind event had devastating ecological
consequences, resulting in 13,000 acres of forest blowdown in the Routt National
Forest.
Although orography also produces thermally driven winds, they are not typically
severe. Exceptions include large-scale katabatic flows such as those that occur
along the coast of Antarctica, which can become violent, particularly if they are
accelerated through interactions with local topography or enhanced by the large-
scale pressure gradient (e.g., Parish and Bromwich 1998).
1 Mountain Weather Prediction: Phenomenological Challenges. . . 11

1.2.7 Severe Convective Storms

Mountainous terrain can have an indirect impact on tornado development by


modifying airmasses downstream over the plains. One scenario occurs when the
mountains modify the plains environment by providing an elevated mixed layer
which results in a higher severe weather potential for tornado development (Lanicci
and Warner 1991). Over the mountains, tornadoes are more infrequent than their
plains counterparts, but can be significant when they do form. Tornadoes and
funnel clouds occur occasionally over the Rocky Mountains during the late spring
and summer (Bluestein and Golden 1993). These tornadoes usually develop in
non-supercell storms because the vertical shear is usually too weak for supercell
formation. Similar type storms have been observed in Switzerland (Linder and
Schmid 1996). However, mountain tornadoes can potentially be devastating. Fujita
(1989) documented an F4 tornado that crossed the Continental Divide within
Yellowstone National Park in Wyoming in 1987. This tornado traveled 24 miles
and leveled 15,000 acres of mature pine forest. On 11 August 1999 an F2 tornado
developed southwest of downtown Salt Lake City, Utah, and moved directly through
the city. This tornado resulted in one fatality, more than 100 injuries and US $170
million in damages (Dunn and Vasiloff 2001). Bosart et al. (2006) analyzed a long-
lived supercell that became tornadic over complex terrain in Massachusetts, on 29
May 1995. The F3 tornado left a 50–1,000-m-wide damage path that stretched for
50 km. Other documented supercell tornadoes over complex terrain include those
over the hilly terrain of the upper Rhine Valley of Germany (Hannesen et al. 2000)
and over the Colorado Mountains (Bluestein 2000).
Another potentially devastating severe wind that occurs in proximity to mountain
locations is the downburst wind or the smaller scale microburst (<4 km). Downburst
winds in arid mountain areas like the western United States are frequently dry and
develop in an environment with a deep, nearly dry-adiabatic subcloud layer; a shal-
low moist mid-layer near 500 mb; weak synoptic-scale forcing with only moderate
(<25 m s1 ) winds aloft; and weak instability [lifted index (LI) usually >2 K]
(Wakimoto 1985). The exact locations of these downburst winds are difficult to
forecast since they are often formed by seemingly benign-looking clouds or reflec-
tivity signatures (Mielke and Carle 1987; Meyers et al. 2006). Downbursts can be
extremely hazardous for aircraft operations, especially during takeoffs and landings.
Hail and graupel (soft hail) is often found in convective precipitation over
mountainous regions, in part, due to the relatively low freezing level above the
ground compared to lower elevations. Severe hail can often occur adjacent to
mountain locations such as the High Plains to the lee of the Rocky Mountains
(Doswell 1980). For example, a hailstorm caused $350 million dollars in damage
over the Front Range, in Denver, Colorado on 13 June 1984 (Blanchard and Howard
1986). Numerous severe hailstorms have also been documented to the east of the
Canadian Rockies in Alberta, Canada (Wojtiw 1975; Smith and Yau 1987) and in
central Switzerland near the Jura Mountains to the north and the Alps to the south
(Houze et al. 1993).
12 M.P. Meyers and W.J. Steenburgh

1.2.8 Cold-Air Pools and Air Quality

Cold-air pools are not typically considered a “severe weather phenomenon”, but
their persistence can lead to poor air quality episodes in basins and valleys (also see
Chaps. 2 and 5). The World Health Organization recently estimated that 800,000
deaths per year worldwide could be attributed to urban outdoor air pollution and
the economic impact from air pollution-related illness is estimated at US $150
billion per year (World Health Organization 2002). Often these cold pools form
in basins or within valleys with terrain constrictions that allow cold air to build
up behind the constriction. The development of cold pools reduces the dispersion
of air pollutants and adversely affects air quality. These poor air quality episodes
are not restricted to large urban areas, but can occur anywhere where emissions
are concentrated, such as in the lower end of the Colorado Gore Valley west of
Vail or along the Mossau and Finenbach Valleys in Germany (Geiger et al. 1995;
Whiteman 2000). Some of the highest particulate matter concentrations observed in
the United States occur episodically in Logan, Utah, a mid-sized metropolitan area
with a population of about 125,000 that is in the topographically confined Cache
Valley (Malek et al. 2006). More recently, in the vicinity of the Jonah–Pinedale
Anticline natural gas field, in the rural Upper Green River Basin of Wyoming,
air quality instrumentation measured 8-h averaged ozone concentrations above the
Environmental Protection Agency’s threshold of 75 parts per billion (ppb). This
criteria, which is more typical of summertime events, was exceeded on 14 days and
resulted in the first ever wintertime ozone advisories in Wyoming (Schnell et al.
2009). The formation of diurnal cold-air pools and valley inversions has a potential
secondary effect. It can also be a complicating factor in frost events, which can be
problematic for agriculture at critical times of the growing season.
Degraded air quality due to mountain haze and pollutant transport can impact the
tourism industry including the mountain vistas in national parks. Poor air quality
degrades the majestic views on public lands, but also can negatively impact the
long-term health of plants, trees and animals.

1.3 The Contemporary Forecast Process in Complex Terrain

At its core, weather forecasting is a scientific endeavor involving hypothesis


formulation, hypothesis testing, and prediction (Roebber et al. 2004), and applying
this scientific method is especially crucial over complex terrain. The forecaster
must develop a conceptual understanding of the past and present weather, formu-
late hypotheses about why and how the atmosphere is evolving, and then seek
evidence to confirm or reject the hypotheses. This iterative process continues and
the hypothesis is refined until a prediction is made. Given the huge volume of
data, analyses, and numerical forecasts available, the forecaster must employ rapid
cognition, make quick decisions, and make judgments in the face of uncertainty
1 Mountain Weather Prediction: Phenomenological Challenges. . . 13

(Doswell 2004). When faced with a “firehose” of data, the ability of a forecaster
to skillfully determine what is important is an example of what Gladwell (2005)
described as “thin slicing”.
Klein Associates has studied the cognitive and psychological aspects of U.S.
military meteorologists (Klein 1998; Stuart et al. 2007). They found distinct
differences between inexperienced or non-engaged forecasters and the experienced
or expert forecasters. The inexperienced forecasters typically relied too much on
computer models and tended to be reactive with the forecasts. The experienced
forecasters had a more global perspective and they tended to be more flexible
with their tools and procedures, often relying on conceptual models in the forecast
process. They typically employed a recognition-primed decision model (Klein
1998) which combines both analysis and intuitive methods (Doswell 2004) that
allows the forecaster to absorb incomplete information under time constraints and
arrive at proper forecasts. The expert forecasters draw from their vast forecast
experience to arrive at their forecast decisions.
Bosart (2003) argues that the forecast process is most effective when the
forecaster addresses six critical questions: (1) What happened? (2) Why did it
happen? (3) What is happening? (4) Why is it happening? (5) What is going
to happen? (6) Why is it going to happen? In an era of increasingly skillful
numerical weather prediction models, it is easy for forecasters to concentrate only
on question 5. Nevertheless, knowledge of the antecedent conditions and underlying
physical processes is extremely valuable, particularly when the numerical guidance
“goes awry” or is unable to resolve critical orographic effects (Bosart 2003; Dunn
2003).
Although there are a variety of forecasting styles (Pliske et al. 2004), skillful
forecasting in mountainous regions typically requires: (1) a core understanding of
synoptic scale and orographic processes, (2) careful evaluation of the evolving syn-
optic setting and flow interaction with the terrain, (3) knowledge of the advantages
and limitations of the objective tools of forecasting over complex terrain, and (4)
the subjective integration of these tools by the forecaster.

1.3.1 Scale Interaction: The Forecast Funnel

Forecasters commonly use the so-called forecast funnel to evaluate the synoptic
setting and flow interaction with terrain (Snellman 1982; Horel et al. 1988;
Steenburgh 2002; Dunn 2003). As illustrated conceptually by Fig. 1.5, the forecaster
begins at the global or planetary scale, focuses attention on progressively smaller
scales, and ultimately builds in the orographic effects at the local scale. The
forecaster considers how the interaction of the large-scale or synoptic flow with the
regional and local orography will influence weather locally. An important premise
of the forecast funnel is that processes on each scale are dependent upon those at
other scales. For example, in the case of orographic precipitation, the forecaster
typically begins by evaluating the past, current, and future synoptic setting and
14 M.P. Meyers and W.J. Steenburgh

Fig. 1.5 The forecast funnel (Courtesy of COMET) (The source of this material is the COMET®
Website at http://meted.ucar.edu/ of the University Corporation for Atmospheric Research
[UCAR], sponsored in part through cooperative agreement(s) with the National Oceanic and
Atmospheric Administration [NOAA], U.S. Department of Commerce [DOC]. ©1997–2010
University Corporation for Atmospheric Research. All Rights Reserved)

the large-scale characteristics of the airmass interacting with the topography. The
forecaster will determine his/her confidence in the large scale forecast and if it
is low, it will adversely affect his/her confidence in the forecast guidance on the
smaller scales. As the forecaster examines the regional scales, parameters such
as stability, temperature, humidity and wind and their influence on the terrain-
induced flow and precipitation dynamics are considered. Finally, the forecaster will
scrutinize the local scale to determine how the topography has influenced and will
influence winds, moisture and precipitation distribution at a specific location. Local
orographic and microphysical effects are considered at this level. This scale is the
most difficult to address in an operational environment and requires reliance on
pattern recognition and knowledge of the local climatology.
The forecast funnel in complex terrain requires a sound understanding of
synoptic, mesoscale and mountain weather processes. It also requires the forecaster
to recognize the advantages and limitations of the available observational and
numerical tools and apply the proper forecasting techniques. Ultimately, the forecast
funnel enables the forecaster to prioritize and assimilate the massive volume of geo-
physical and numerical data and to identify what is important on a day-to-day basis.
1 Mountain Weather Prediction: Phenomenological Challenges. . . 15

1.3.2 Objective Tools

The forecaster utilizes a number of objective analysis and forecast tools including
in situ and remotely-sensed observations (also see Chap. 8) and numerical weather
prediction models (also see Chaps. 9, 10, 11). Unique aspects of the use of these
tools in complex terrain are described below.

1.3.2.1 Surface-Based Observations

A major challenge for weather analysis over complex terrain is the need for
high-density surface observations to resolve the fine-scale gradients in surface
weather produced by topographic forcing. As a result, forecasters desire as much
observational data as possible and are willing to compromise, often relying on
observations from heterogeneous networks with differing sensor types, biases and
reliabilities (e.g., Horel et al. 2002). Some of the advantages of this approach
include higher data density, higher frequency observations at some locations,
and observations from non-conventional locations. Some of the disadvantages
include non-uniform data and siting characteristics, stations that do not report the
full suite of data, an increased need for quality control by the forecaster, and
varying instrumentation with inconsistent averaging intervals. Another critical issue
regarding surface observations in the mountains is the siting or microclimate where
instrumentation is placed. For example, many Remote Automated Weather Systems
(RAWS) (Myrick and Horel 2008) which are geared to fire weather observations
are sited on south aspects to capture worst case scenarios with regard to fire
weather applications. Geiger et al. (1995) showed that temperatures on south aspects
in the midlatitudes are typically 3ı C warmer than northern slopes due to sun
angle. Most Snowpack Telemetry (SNOTEL) instrumentation, run by the USDA’s
National Resources Conservation Service, is sited on aspects more conducive for
deeper snowpack and hydrological applications (Dressler et al. 2006). A failure to
understand these types of differences can lead to unrealistic biases in analyses and
forecasts. Additionally, incorrect or imprecise specification of the location of the
instrumentation can negatively impact resolution of fine-scale features and model
verification scores in complex terrain (Ludwig et al. 2006). The bottom line is that
forecasters need to make sure their observational data is quality controlled and siting
biases are taken into account.
Surface-based observations provide an added benefit after the event as well.
Observations allow the forecaster to do post-analysis for verification and to conduct
post-mortems of the event. This process is vital to understanding the physical
controls of a particular storm or event. Observational data are also used as a “first
forecast” to produce weather fields based on the previous day’s observations through
bias-corrected statistics.
16 M.P. Meyers and W.J. Steenburgh

1.3.2.2 Radar

Radars (also see Chap. 8) are an important tool in weather analysis and forecasting
over complex terrain. However, one must understand some of the limitations of
radars in the mountains such as strong ground clutter and beam blockage by terrain
(Germann and Joss 2004). MeteoSwiss found that one algorithm improvement
may lead to a degradation of other algorithm statistics. They took a systematic
approach with a combination of refinements to eliminate residual ground clutter and
shielding from mountains, hardware calibration, and minimize biases to improve
quantitative precipitation estimates from radar (Germann and Joss 2004). Typically,
mountain radars in the western United States are situated at higher elevations to
optimize scan strategies and to limit shielding or beam blockage (Wood et al. 2003)
which potentially creates another problem due to overshooting of precipitation. This
problem is more apparent in wintertime orographic precipitation events and also
shallow monsoonal or tropical precipitation. Breidenbach et al. (1999, 2001) showed
the effective radar coverage for summer and winter seasons in the Pacific Northwest
(Fig. 1.6). These coverages are based on a climatology analysis of radar-derived
precipitation computed from WSR-88D data (1996–2000) during the warm and
cold seasons. It shows that radars located at high elevations in mountainous terrain
are impacted by range effects (beam overshooting precipitation) and severe beam
blockage issues and these impacts are exacerbated during the winter season. Rea-
sonable precipitation estimates can be obtained only within the effective coverage
area. Wetzel et al. (2004) has documented cases where the Grand Junction, Colorado
WSR-88D radar (GJX) detected little or no precipitation signal over the Park Range
(elevations 2,150–3,550 m) of north-central Colorado (200 km from GJX and
7,300 m MSL at 0.5ı ) when, in fact, shallow orographically-forced heavy snowfall
was occurring over these mountains. Additionally, the spacing between WSR-88D
radars across the western third of the United States (Fig. 1.6) is much greater than
the remainder of the country which increases issues with detection (Westrick et al.
1999; Maddox et al. 2002). Vasiloff et al. (2011) has examined the benefits of using
mobile dual-polarization radar over the beam blocked complex terrain areas of the
central Colorado Mountains. These deficiencies of the radar coverage over complex
terrain must be considered by the forecaster.

1.3.2.3 Satellite

Satellite data provides a significant benefit over complex terrain for the mountain
meteorologist, providing a diversity of data for analysis and forecast applications.
Satellite fog products are very useful in valley and basin locations to detect and
forecast low level clouds and fog which have important implications for aviation
and ground transportation. Water vapor imagery is an excellent tool for diagnosing
mountain wave signatures. As was discussed in the previous section, radar coverage
can be seriously compromised over complex terrain. However, satellite products
supplement analysis of data sparse regions which may be neglected by the radar
1 Mountain Weather Prediction: Phenomenological Challenges. . . 17

Fig. 1.6 WSR-88D coverage


over the northwestern United
States during summer (top)
and winter (bottom) months
(Breidenbach et al. 1999,
2001). These coverages are
based on a climatology
analysis of radar-derived
precipitation computed from
WSR-88D data (1996–2000)
during the warm and cold
seasons (© American
Meteorological Society.
Reprinted with permission)
18 M.P. Meyers and W.J. Steenburgh

coverage, especially at farther distances from the radar. Visible satellite imagery
is especially useful during wintertime orographic snowfall where radar detection
may overshoot the shallow orographic clouds. Similarly, visible satellite imagery
and other satellite applications provide supplemental detection of the shallower
convective events such as monsoonal surges in the mountainous terrain, as well as
wildfire detection.

1.3.2.4 Lightning

Lightning data is also crucial for the forecaster in the mountains (Holle and Lopez
1993). Lightning observations are often used by forecasters in complex terrain to
provide another layer of detection of convection cells, particularly in areas with poor
radar coverage. Lightning analysis is often overlaid on radar or satellite imagery to
complement these observational tools. Lightning detection adds in the diagnosis
of location, storm strength, and ice processes and the potential for hail. In some
situations, lightning may provide a secondary method of storm detection when radar
analysis is inadequate.

1.3.2.5 Numerical Weather Prediction (NWP)

Numerical models (also see Chap. 11) are arguably the most used tool in a
forecaster’s toolbox. However, it is imperative that the forecaster understands the
limitations of numerical models over complex terrain. After all, as well stated
by Doswell (1986), “when guidance is needed the most, it is generally the least
useful.” Topography may enhance predictability of certain flow types, but can also
exacerbate large-scale forecast errors.
Evaluating the value added utility of a high-resolution NWP forecast is the
ultimate challenge for the forecaster. Application of the forecast funnel helps with
this component of the forecast process, is especially important when the large-scale
forecast is not representative of reality (McMurdie and Mass 2004), and is essential
for predicting phenomena that are unresolved or poorly simulated with existing
models and physics parameterizations. If the large scale is not predicted well, the
forecaster may have low confidence in the mesoscale NWP forecast and must adjust
the numerical guidance substantially. On the other hand, forecasters can lean toward
the high-resolution model solution when confidence in both the large-scale forecast
and the mesoscale model forecast are high.
Ensemble forecast systems are now widely utilized by weather forecasters
(Roebber et al. 2004; Novak et al. 2008). These ensemble forecast systems may
incorporate different models, systematically varying initial conditions, physical
schemes, and/or boundary conditions to produce separate runs which are valid at
a specific time (Brooks and Doswell 1993; Sivillo et al. 1997). Ensemble forecast
systems have recently been used for short term prediction which is useful in
topographic settings (Arribas et al. 2005; Yuan et al. 2007; Reinecke and Durran
1 Mountain Weather Prediction: Phenomenological Challenges. . . 19

2009). These ensemble solutions help determine pattern evolution and model
consistency.
Mountain meteorologists often rely on post processing techniques, such as Model
Output Statistics (MOS) or other statistical applications as an objective tool for
forecasting specific surface parameters at site specific locations (Glahn and Lowry
1972; Cheng and Steenburgh 2007). Although it is a very useful tool, mountain
weather forecasters need to be careful when applying these statistical methods over
gridded forecasts of varying degrees of complex terrain. The forecaster can possibly
fall into the trap of using MOS too much without using other available forecasting
tools. MOS is a convenient and quick way to provide a forecast; however, it may
come with potentially dire consequences. As Bosart (2003) mentioned, “Forecasters
who grow accustomed to letting MOS and the models do their thinking for them on
a regular basis during the course of their daily activities are at high risk of ‘going
down in flames’ when the atmosphere is in an outlier mode”.

1.3.3 Subjective Tools: The Human Element

1.3.3.1 Geographic Familiarization

Forecasting in complex terrain requires a familiarization of the geography to provide


a basis for climatology and conceptual models that are useful in the mountains.
Roebber et al. (1996) speculated that regional knowledge is especially useful for
mountain forecasting skill. Forecasters who do not have a strong knowledge of the
local topography will have a difficult time determining forcing at the local scale of
the forecast funnel. The best way for the forecaster to achieve this understanding
is to travel around the forecast area and familiarize themselves with the geography
and local topography. If this is not possible, using a Geographic Information System
(GIS) or Google Earth application to evaluate the terrain is helpful. The populations
of mountain towns and resorts have a strong seasonal dependency, and it is important
to know the potential weather impacts on these communities. Furthermore, it is vital
to know the potentially problematic regions for hazardous weather. For instance, in
the southwestern United States, narrow slotted canyon regions will have a much
higher flash flood potential than other topographic areas. It is critical for the
forecaster to have a firm understanding of these diverse geographic regions, as well
as the potential vulnerabilities for hazardous weather in the areas.

1.3.3.2 Rules of Thumb

A “rule of thumb” is an easy-to-remember forecasting method based on experience


or statistics. Ideally, a “rule of thumb” is validated over a statistically significant
sample to be useful for the forecaster. Two examples: 3ı C at 700 hPa corresponds
to a 7,500 ft (2,286 m) rain-snow line or an inch (2.5 cm) of precipitable water
20 M.P. Meyers and W.J. Steenburgh

is required before any flash flooding will occur. A rule of thumb usually provides
the forecaster with a first guess, but it may be fine-tuned based on the conditions
of the given day. For instance, in the rain-snow example above, the snow level
may be adjusted lower in a convective environment. “Rules of thumb” have
drawbacks which can limit the forecaster. They can be limited by a particular
model configuration, where the effectiveness is lost after a model upgrade or other
modification. Often “rules of thumb” are a black box to the user, where the physical
basis is not known or is non-existent. “Rules of thumb” can be helpful tools, but the
forecaster needs to understand the limitations and the need for sound scientific basis
behind these approaches.

1.3.3.3 Pattern Recognition and Climatology

Despite dramatic advances in numerical weather prediction over the past few
decades, pattern recognition and climatology continue to play an essential role in the
forecast process. Knowledge of the precipitation climatology is particularly valuable
in complex terrain. Despite the limited density of observing stations, generally reli-
able mean precipitation climatologies are now available for the contiguous United
States (Fig. 1.7; Daly et al. 1994, 2008; http://prism.oregonstate.edu/; Johnson
et al. 2000) and have been developed for other mountainous regions, including the
European Alps (e.g., Frei and Schär 1998; Schwarb 2000; Schwarb et al. 2001).
These climatologies are useful for downscaling or refining precipitation forecasts
from numerical models that are unable to fully resolve the local topographic effects.
Forecasters use conditional climatologies, numerical model forecasts, and pattern
recognition as tools to identify situations where the precipitation distribution can
potentially deviate significantly from the mean, such as when (1) the precipitation-
altitude relationship is unusually large or small, (2) orographic effects alter the
mesoscale distribution of precipitation, or (3) the synoptic forcing is not uniform
(e.g., frontal precipitation is expected to the south but not the north). For example,
the distribution and intensity of orographically-driven precipitation is dependent
on flow direction, which controls where orographic ascent and descent occur. Even
a relatively small shift in large-scale wind direction can have dramatic effects on
precipitation distribution (Ralph et al. 2003).
Pattern recognition is also useful for adjusting systematic model biases. For
example, the mesoscale modeling system used for the 2002 Olympic Winter Games,
which was based on the fifth-generation Pennsylvania State University–National
Center for Atmospheric Research (PSU–NCAR) Mesoscale Model (MM5), featured
a prominent wet bias on the windward slopes of the Wasatch Mountains and a dry
bias in the leeward valleys (Fig. 1.8; Hart et al. 2005). Similar biases have been
identified in MM5 simulations over the Cascade Mountains by Colle et al. (2000).
Knowledge of these biases enables forecasters to make appropriate adjustments.
Furthermore, model biases are often dependent on the synoptic regime (and flow
direction). With this knowledge, forecasters should be able to make adjustments
based on experience.
1 Mountain Weather Prediction: Phenomenological Challenges. . . 21

Fig. 1.7 Annual precipitation climatology for the United States generated by the PRISM model
([Daly et al. 1994, 2008]. Copyright © 2011, PRISM Climate Group, Oregon State University,
Reprinted with permission. http://prism.oregonstate.edu. Map created Jun 16 2006)

Fig. 1.8 Analysis of the 24-h precipitation PB scores for the (a) 12- and (b) 4-km domains.
Contours drawn at 40% intervals from 80% to 200%. Bold lines indicate maxima (solid) and
minima (dashed). Inset shows the 24-h precipitation PB scores versus model elevation bias (m).
Elevation scale at lower left (From Hart et al. 2005. © American Meteorological Society. Reprinted
with permission)
22 M.P. Meyers and W.J. Steenburgh

Fig. 1.9 Conceptual model based on the 100-h snowstorm over the Wasatch Mountains (From
Steenburgh 2003. © American Meteorological Society. Reprinted with permission)

1.3.3.4 Conceptual Models

A conceptual model is a physically-based model detailing a simplified represen-


tation of the physical phenomena. Conceptual models are a higher-order hybrid
approach using both the objective and the human element. These models are
based on information received through experimental and observational data, theory,
and climatology to ensure a proper physical basis. Conceptual models are widely
used by meteorologists as an important diagnostic tool and learned through training.
The conceptual model may help the forecaster determine whether a critical threat
exists, or may enhance the situational awareness of an event already occurring
or imminent. An example of a conceptual model over western United States is
shown in Fig. 1.9. In this conceptual model, forecasters are provided a basis for
this type of storm evolution. An example from the Alps is shown in Fig. 1.10, which
details a blocked-stable case versus an unblocked-unstable precipitation case. In
both these scenarios the forecaster is provided with a forecast foundation based on
the conceptual model. Often the storm does not duplicate the details of the model,
however a subjective assessment based on proper training, past experience and an
intuitive forecast process (Doswell 2004; Stuart et al. 2007) will enable him/her to
improve upon the conceptual model and anticipate the future outcome.
1 Mountain Weather Prediction: Phenomenological Challenges. . . 23

Fig. 1.10 Conceptual model of (a) an unblocked and unstable case, (b) a blocked and stable
case over the Alps (From Medina and Houze 2003; Rotunno and Houze 2007. © American
Meteorological Society. Reprinted with permission)

The ability to effectively “thin slice” as described by Gladwell (2005) and formu-
late and test hypotheses about the evolving atmosphere requires accurate conceptual
models that describe the structure and physical processes controlling synoptic,
mesoscale, and mountain weather systems. Conceptual models with application
to complex terrain are especially lacking compared to existing conceptual models
developed for the Plains and for severe weather (Andra et al. 2002). Conceptual
models are one application which would benefit from input of both the research and
operational forecasting community, providing the forecasters a mental model of the
theory. This interaction would require the forecasters and researchers to bridge the
gap and work together to improve the science of meteorology over complex terrain
as discussed later in this book (see Chap. 12).

1.4 Forecasting in Complex Terrain

1.4.1 Complexity of Mountain Weather Forecasting

The main priority for any forecaster, including the mountain meteorologist is to
protect lives and property during severe weather events. The phenomenological
challenges which confront the mountain weather forecaster are detailed in Sect. 1.2.
It is critical that the mountain meteorologist have a strong foundation in mountain
meteorology and understand the advantages and limitations of the subjective and
24 M.P. Meyers and W.J. Steenburgh

objective forecasting tools used in mountainous terrain as discussed in the previous


section. The mountain meteorologist must understand that hazardous weather may
occur during synoptically-benign weather. For instance, hot and dry conditions may
lead to snowmelt flooding during the spring or significant fire weather issues during
the summer. In severe weather events, the situational awareness of the forecaster
must be dynamic, especially since severe weather over complex terrain may be
more isolated and develop in more benign environments than at lower elevations.
In complex terrain mountain weather forecasters need to be flexible and multi-task
oriented, often forecasting for multiple severe weather types of events (e.g. winter
weather, convective weather) that may occur simultaneously across the forecast area
of responsibility (Meyers et al. 2006). The forecaster must also understand how their
users, such as the media and emergency managers, will assimilate the hazardous
forecasts (Morss et al. 2005; Morss and Ralph 2007).
Accurate predictions of day-to-day non-severe weather variability in regions of
complex terrain are essential for commerce and the public good. The daily challenge
for the mountain weather forecaster is to produce temporally and spatially detailed
forecasts that are accurate, reliable, and well communicated for a diverse user com-
munity in regions with extreme topographic, meteorological, and in many instances
socio-economic contrasts (American Meteorological Society (AMS) 2007). Further,
a dilemma exists which fuels our forecast challenge; a good impact for some may
have an adverse impact on another.
People who live in the mountains understand the complexity of mountain
weather. There is a saying, “if you don’t like the weather in the mountains, just
wait an hour”. However, they do not seem to be as forgiving, when it comes to
missed forecasts. Typically, the majority of users are not adversely affected by false
alarms (Barnes et al. 2007). However, there is a group of users in the mountain
snow community that relies on adverse weather for their business or pleasure and
false alarms may not be viewed favorably by this sector. A forecaster must be
cognizant of this community that relies on snow to make a living or to satisfy their
recreational desires. For example, there may be situations when heavy snowfall is
forecasted and it occurs at one ski area while another ski area 20 km away receives
little or none. The forecast may be a hit for one area but a miss for the other and
the area that did not receive the snow will often let you know about the missed
forecast. Mountain weather forecasters cannot be overly-sensitive when dealing
with these forecast inconsistencies and the potentially negative feedback from the
customer.

1.4.2 Specialized Forecast Examples

Many forecast consumers need a level of specificity beyond that of the general
weather forecast (AMS 2007). Businesses need very specific information relevant
to their operations, while the public is interested in how forecasts impact their
1 Mountain Weather Prediction: Phenomenological Challenges. . . 25

activities. For example, the Tyrolean ski industry in Austria invests nearly US $100
million a year in snowmaking (Steiger and Mayer 2008). This industry relies on
accurate temperature and humidity forecasts to determine wet bulb temperatures
below 5ı C when good snow quality can be produced. Accurate forecasts also
impact the winter Olympic Nordic ski performance. The correct choice of ski wax
used by the athletes depends on accurate forecasts of the snow surface conditions
on the race course. The collection of snow measuring and forecasting techniques for
the 2010 winter Olympics are detailed by Wagner and Horel (2008). Several more
examples of specialized forecasts in complex terrain are discussed next.

1.4.2.1 Fire Weather Forecasts

Fire weather forecasters have unique concerns that typically focus on diverse
microscale interactions over complex terrain. For this reason, fire weather forecasts
are very demanding and stressful due to the high impact of the forecast on lives or
property. A fire weather forecaster must understand the fire environment, which is
dependent on the fuel moisture, topography and weather. How these components
interact determines the fire behavior. A fire weather forecaster should be well-
informed of the fuel conditions in a particular forecast area. To understand fuel
conditions the forecaster must understand the antecedent seasonal conditions that
have occurred for the forecast area, including how drought may have impacted a
particular region. The forecaster must also understand the fine-scale topography
and how it can modify the diurnal evolution of the surface and boundary layer
structures. This knowledge aids in temperature, wind, and humidity forecasts, which
can vary tremendously over small horizontal distances in mountainous regions. The
fire weather forecaster scrutinizes local observations in the complex terrain and
must also predict thunderstorm activity, which is a critical source for new fire starts
when fuels are dry. Additionally, the fire weather forecaster must understand smoke
management issues which are complicated along slopes. The parameters associated
with a smoke management forecast are dependent on an accurate forecast of the
mixed boundary layer, as well as transport winds which are complicated along
slopes where in-situ measurements may not be representative of the entire area.
Fire weather forecasters must clearly understand the potential risks of smoke on
population centers near these fires.
A critical element of fire weather forecasts includes vital site-specific spot
forecasts for wildfire and prescribed fire projects. Requests for spot forecasts have
been increasing dramatically, with the NWS alone providing nearly 20,000 spots
every year. Wildfire mitigation through prescribed burns can be an efficient and
effective tool in land management. Accurate weather forecasts are essential for the
effective and safe use of prescribed fires, and also to ensure the safety of firefighters
who frequently operate in areas of rugged and remote terrain during wildfires and
prescribed fire projects.
26 M.P. Meyers and W.J. Steenburgh

1.4.2.2 Hydrological Runoff Forecasts

Hydrologists have a variety of responsibilities over complex terrain. One of their


main concerns, especially during the winter and spring months is to determine the
condition of the snowpack for runoff prediction. This task is especially difficult
to determine due to the sparse network of observations over the mountains. In the
western United States hydrologists depend on SNOTEL observations which provide
a network of snow water equivalent (SWE) measurements over the higher terrain.
SWE is a better indicator of the amount of liquid within the snowpack than snow
depth or accumulated liquid equivalent precipitation, a portion of which potentially
can fall as rain. SWE also will account for losses due to sublimation or melting.
They must determine the representativeness of these observations, which may be
compromised by anomalous precipitation-elevation distributions, wind and drifting
issues, and other measurement inconsistencies. The forecasters employ a variety of
hydrologic and hydraulic models to determine how the snowpack will runoff. These
models consider subsurface soil moisture conditions, soil type and permeability,
losses due to evapotranspiration, stream and river channel characteristics, and the
current state of the snowpack. Equally important, the hydrologists must maintain
a daily weather watch of the current weather conditions where snowpack runoff
is critically dependent on temperature, wind, moisture and precipitation. Rain-on-
snow events, where significant storm rainfall combined with snowmelt can lead to
devastating flooding and damage in mountain locations (Kattelmann 1997; Graybeal
and Leathers 2006). There have also been recent studies (Barrett et al. 2008)
documenting the impact of dust on snow events. These dust layers decrease the
albedo of the snowpack and accelerate snowmelt which has significant implications
for snowpack runoff and avalanche potential.
Another responsibility of the hydrologists is to maintain a heightened awareness
of the dams and reservoirs. This awareness includes an assessment of the integrity of
the dam as well as the constant monitoring of critical levels of dam capacity, inflow
rates, and other factors which may lead to a dam break. These dam breaks (e.g. Teton
Dam Break) can potentially have devastating human and economic consequences.

1.4.2.3 Avalanche Forecasts

Similar to hydrologists, avalanche forecasters must maintain an active weather


watch and understanding of the current surface conditions. Avalanche forecasters
are interested in the amount of snowfall, but pay close attention to the snow density
or the percentage of water in the snowfall. They are concerned with periods when
atmospheric conditions lead to snow metamorphism or wind transport. Other issues
such as ‘rain on snow’ events can dramatically alter the structure and stability of the
snowpack.
Avalanche forecasters spend a great deal of time outside evaluating the snow-
pack. They need to have a strong understanding of the topographic variability
and knowledge of the avalanche-prone regions. Avalanche forecasters often ski
1 Mountain Weather Prediction: Phenomenological Challenges. . . 27

to representative locations and dig a snow pit to evaluate the thermal, mois-
ture and snow characteristics of the snow core to determine its integrity for
avalanche potential. Avalanche forecast users include governmental transportation
agencies which use their expertise to keep transportation corridors safe. Another
group of users include outdoor recreationalists such as skiers and snowboarders,
snowmobilers, and snowshoers. These groups sometimes seek adventure during
dangerous situations. The avalanche forecasters mitigate the potential danger by
educating these user groups, and issuing forecasts including warnings of critical
conditions which potentially could lead to avalanches.

1.4.2.4 Aviation Forecasts

Aviation customers in the mountains include commercial industry, private pilots,


hot air balloonists, glider pilots, hang-glider pilots, and others. Aviation forecasters
in complex terrain have a multitude of issues which they must deal with to
provide an accurate and timely forecast for these customers. In the absence of
synoptic or convectively driven winds, diurnal winds in the vicinity of mountains
dominates the typical wind structure with upslope/upvalley winds developing during
the late morning hours and downslope/downvalley winds developing after sunset.
Climatology of the wind evolution for specific airport locations is a useful tool for
the forecaster. It is also beneficial for the forecaster to visit the airport to understand
the complexity of the terrain surrounding the airport. Even though the diurnal
evolution of the winds of each airport location is easily known, it is important to
forecast the magnitude and transition times of this diurnal evolution for aviation
purposes. Typically the environment is complicated by synoptic or convective winds
that can overwhelm the diurnal winds. The aviation forecaster needs to keep an
active weather watch since even subtle changes in the synoptic or convective forcing
can have a dramatic impact on the surface wind patterns in the mountains.
Aviation forecasters must know airport thresholds and minimums for takeoffs
and landings. Each airport has different criteria based on their local terrain and
requirements. The forecaster must know these thresholds when forecasting ceiling
and visibility degradation due to weather obscuration. A forecast with ceilings
and visibilities in the Marginal Visual Flight Rules (MVFR) or Instrument Flight
Rules (IFR) categories could require commercial aircraft to carry more fuel or
possibly restrict flights until conditions improve. The fact that the precipitation
may be showery, with varying intensities of precipitation over short periods of time,
exacerbates this problem.
The aviation industry is especially concerned with wind shear, turbulence, and
icing potential at the ground and during flights through the mountains. These
potential hazards are prevalent due to the complex winds and areas of local lifting
that occur in the vicinity of mountains. In extreme cases, with certain stability and
flow, downslope windstorms may develop. These extreme wind events can produce
strong gradients, extreme wind shear, rotors, (also see Chap. 3) turbulence and icing
conditions.
28 M.P. Meyers and W.J. Steenburgh

One additional issue which the aviation industry must address especially during
the warm season is density altitude. Density altitude is the corrected altitude for non-
standard atmospheric conditions. Pressure, humidity and temperature affect density
altitude, potentially resulting in a higher altitude than the actual altitude. In these
situations, aircraft performance is potentially compromised and density altitude
is especially sensitive at higher elevations. The implication of density altitude is
especially critical to smaller aircraft which are used in fire weather and emergency
response.

1.5 Conclusions

This chapter has summarized some of the phenomenological challenges of mountain


weather and the societal and economic impacts associated with them. These
challenges demonstrate the importance of forecasting in mountainous regions. The
core forecast process includes a careful evaluation of the evolving scales and flow
interactions on the local terrain as well as an understanding of the objective and
subjective forecasting tools and techniques used in mountain weather analysis
and forecasting. In summary, the mountain meteorologist should have a core
understanding of the synoptic scale and orographic processes that are detailed in
the subsequent chapters of this book.

Acknowledgments We thank the participants in the 2008 AMS/COMET/MSC Mountain Weather


Workshop: Bridging the Gap between Research and Forecasting for 4 days of lectures and
discussion that stimulated this chapter, as well as our editors and chapter coauthors and their
contributions to this book. Thanks are also extended to the Grand Junction NWS staff, Jeffrey
Manion, and three anonymous reviewers for their contributions to the manuscript. Contributing
author Steenburgh acknowledges the support of the National Science Foundation under grant
ATM-0627937 and the National Weather Service under a series of grants provided by the C-STAR
program. Any opinions, findings, and conclusions or recommendations expressed in this material
are those of the authors and do not necessarily reflect the views of the National Science Foundation
or National Weather Service.

References

Adams, R. M., L. J. Houston, R. F. Weiher, 2004: The value of snow and snow information
services. Report prepared for NOAA’s National Operational Hydrological Remote Sensing
Center, Chanhassen, MN, under contract DG1330-03-SE-1097.
American Meteorology Society, 1986: Glossary of Meteorology. Edited by Ralph E. Huschke.
Boston, MA, AMS. 638 pp.
American Meteorological Society, 2007: Weather analysis and forecasting: An information
statement of the American Meteorological Society. Bull. Amer. Meteor. Soc., 88, 1655–1659.
Andra, D.L., E.M. Quoetone, and W.F. Bunting, 2002: Warning Decision Making: The Relative
Roles of Conceptual Models, Technology, Strategy, and Forecaster Expertise on 3 May 1999.
Wea. Forecasting, 17, 559–566.
1 Mountain Weather Prediction: Phenomenological Challenges. . . 29

Andrey, J., B. Mills, and J. Vandermolen, 2001: Weather information and road safety. ICLR Paper
Series No. 15, Institute for Catastrophic Loss Reduction, London, Ontario. 36 pp.
Arribas, A., K.B. Robertson, and K.R. Mylne, 2005: Test of a Poor Man’s Ensemble Prediction
System for Short-Range Probability Forecasting. Mon. Wea. Rev., 133, 1825–1839.
Barnes, L.R., E.C. Gruntfest, M.H. Hayden, D.M. Schultz, and C. Benight, 2007: False Alarms
and Close Calls: A Conceptual Model of Warning Accuracy. Wea. Forecasting, 22, 1140–1147.
Barredo, J. I., 2007: Major flood disasters in Europe: 1950–2005. Nat. Hazards, 42, 125–148.
Barrett, A. P., T. H. Painter and C. C. Landry, 2008: Desert dust enhancement of moun-
tain snowmelt. Intermountain West Climate Summary. Boulder, CO, Western Water
Assessment. 4:2.
Bell, G.D., and L.F. Bosart, 1988: Appalachian Cold-Air Damming. Mon. Wea. Rev., 116, 137–161.
Blanchard, D.O., and K.W. Howard, 1986: The Denver Hailstorm of 13 June 1984. Bull. Amer.
Meteor. Soc., 67, 1123–1131.
Bluestein, H. B. and J. H. Golden, 1993: A review of tornado observations. The Tornado: Its
Structure, Dynamics, Prediction, and Hazards, Geophys. Monogr., No. 79, Amer. Geophys.
Union, 319–352.
Bluestein, H. B., 2000: A tornadic supercell over elevated, complex terrain: The Divide, Colorado,
storm of 12 July 1996. Mon. Wea. Rev.,128, 795–809.
Bosart, L.F., 2003: Whither the Weather Analysis and Forecasting Process? Wea. Forecasting, 18,
520–529.
Bosart, L.F., A. Seimon, K.D. LaPenta, and M.J. Dickinson, 2006: Supercell Tornadogenesis
over Complex Terrain: The Great Barrington, Massachusetts, Tornado on 29 May 1995. Wea.
Forecasting, 21, 897–922.
Branswell, B., 1999: Avalanche in Quebec. Maclean’s Magazine. Maclean Hunter Publishing Ltd.
Breidenbach, J. P., D.J. Seo, P. Tilles and K. Roy, 1999: Accounting for radar beam blockage
patterns in radar-derived precipitation mosaics for River Forecast Centers. Preprints, 15th Int.
Conference on Interactive Information Processing Systems (IIPS) for Meteorology, Oceanog-
raphy, and Hydrology, Dallas, TX, AMS 5.22. [Available online at http://ams.confex.com/ams/
older/99annual/abstracts/1699.htm.].
Breidenbach, J.P., D.J. Seo, P. Tilles and C. Pham, 2001: Seasonal variation in multi-radar
coverage for WSR-88D precipitation estimation in mountainous regions. Preprints. Symposium
on Precipitation extremes: Prediction, Impacts, and Responses, Albuquerque, NM, AMS.
[Available online at http://ams.confex.com/ams/annual2001/techprogram/paper 18668.htm.].
Brooks, H. E., and C. A. Doswell III 1993: New technology and numerical weather prediction: A
wasted opportunity? Weather, 48, 173–177.
Cannon, S. H., P. S. Powers, and W. Z. Savage, 1998: Fire-related hyperconcentrated and debris
flows on Storm King Mountain, Glenwood Springs, Colorado, USA. Environ. Geol., 35 (2–3),
210–218.
Cannon, S. H., and S. L. Reneau, 2000: Conditions for generation of fire-related debris flows,
Capulin Canyon, New Mexico. Earth Surf. Processes Landforms, 25, 1103–1121.
Cannon, S. H., E. R. Bigio, and E. Mine, 2001: A process for fire-related debris flow initiation,
Cerro Grande fire, New Mexico. Hydrol. Processes, 15, 3011–3023.
Cannon, S. H., J. E. Gartner, A. Holland-Sears, B. M. Thurston, and J. A. Gleason, 2003: Debris-
flow response of basins burned by the 2002 Coal Seam and Missionary Ridge fires, Colorado.
Engineering Geology in Colorado—Contributions, Trends, and Case Histories, D. D. Boyer,
P. M. Santi, and W. P. Rogers, Eds., Association of Engineering Geologists Special Publication
14, 1–31.
Chang, A. T. C., J. L. Foster, P. Gloersen, W. J. Campbell, E. G. Josberger, A. Rango and Z. F.
Danes. 1987. Estimating snowpack parameters in the Colorado River basin. In: Proc. Large
Scale Effects of Seasonal Snow Cover, IAHS Pub. No. 166, 343–353.
Cheng, W.Y.Y., and W.J. Steenburgh, 2007: Strengths and Weaknesses of MOS, Running-Mean
Bias Removal, and Kalman Filter Techniques for Improving Model Forecasts over the Western
United States. Wea. Forecasting, 22, 1304–1318.
30 M.P. Meyers and W.J. Steenburgh

Christensen, N.S., A.W. Wood, N. Voisin, D.P. Lettenmaier, and R.N. Palmer. 2004: The Effects of
Climate Change on the Hydrology and Water Resources of the Colorado River Basin. Climatic
Change, 62, 377–363.
Colle, B. A., C. F. Mass and K. J. Westrick, 2000: MM5 precipitation verification over the Pacific
Northwest during the1997–1999 cool seasons. Wea. Forecasting, 15, 730–744.
Cordell, H.K., 2008: The latest on trends in nature-based outdoor recreation. Forest History
Today, Spring, pp 4–10. Last retrieved on October 20, 2008, from http://www.foresthistory.
org/Publications/FHT/FHTSpring2008/Cordell.pdf.
Daly, C., R.P. Neilson, and D.L. Phillips, 1994: A Statistical-Topographic Model for Mapping
Climatological Precipitation over Mountainous Terrain. J. Appl. Meteor., 33, 140–158.
Daly, C., Halbleib, M., Smith, J.I., Gibson, W.P., Doggett, M.K., Taylor, G.H., Curtis, J., and
Pasteris, P.A. 2008. Physiographically-sensitive mapping of temperature and precipitation
across the conterminous United States. Int. J. Climatol., 28, 2031–2064.
Decker, F. W., 1979: Oregon’s silver thaw. Weatherwise, 32, 76–78.
Doesken, N. J. and A. Judson, 1997. “The Snow Booklet: A Guide to the Science, Climatology,
and Measurement of Snow in the United States”. Colorado Climate Center, Department of
Atmospheric Science, Solorado State University, Fort Collins, CO. p. 86.
Doswell, C.A., 1980: Synoptic-Scale Environments Associated with High Plains Severe Thunder-
storms. Bull. Amer. Meteor. Soc., 61, 1388–1400.
Doswell, C.A. III 1986: The Human Element in Weather Forecasting, Nat. Wea. Dig., 11, 6–17.
Doswell, C.A. III, 2004: Weather Forecasting by Humans—Heuristics and Decision Making. Wea.
Forecasting, 19, 1115–1126.
Dressler, K.A., S.R. Fassnacht, and R.C. Bales, 2006: A Comparison of Snow Telemetry and Snow
Course Measurements in the Colorado River Basin. J. Hydrometeor., 7, 705–712.
Ducrocq, V., O. Nuissier, D. Ricard, C. Lebeaupin, and T. Thouvenin, 2008: A numerical study
of three catastrophic precipitating events over southern France. II: Mesoscale triggering and
stationarity factors. Quart. J. Roy. Meteor. Soc., 134, 131–145.
Dunn, L. B., and S. V. Vasiloff, 2001: Tornadogenesis and operational considerations of the
11 August 1999 Salt Lake City Tornado as seen from two different doppler radars. Wea.
Forecasting, 16, 377–398.
Dunn, L. B., 2003: Modern weather forecasting. Handbook of Weather, Climate, and Water, T. D.
Potter and B. R. Colman, Eds., Wiley-Interscience, 677–688.
Englin, J., and K. Moeltner. 2004. The Value of Snowfall to Skiers and Boarders. Environmental
and Resource Economics,29,123–136.
Ferber, G. K., C. F. Mass, G. M. Lackmann, and M. W. Patnoe, 1993: Snowstorms over the Puget
Sound lowlands. Wea. Forecasting, 8, 481–504.
Forbes, G.S., R.A. Anthes, and D.W. Thomson, 1987: Synoptic and Mesoscale Aspects of an
Appalachian Ice Storm Associated with Cold-Air Damming. Mon. Wea. Rev., 115, 564–591.
Fettig, C.J., K.D. Klepzig, R.F. Billings, A.S. Munson, T.E. Nebeker, J.F. Negron, and J.T. Nowak.
2007: The Effectiveness of Vegetation Management Practices for Prevention and Control of
Bark Beetle Infestations in Coniferous Forests of the Western and Southern United States.
Forest Ecology and Management, 238, 24–53.
Frei, C., and C. Schär, 1998: A precipitation climatology of the Alps from high-resolution rain-
gauge observations. Int. J. Climatol., 18, 873–900.
Fujita, T.T., 1989: The Teton–Yellowstone tornado of 21 July 1987. Mon. Wea. Rev., 117,
1913–1940.
Geiger, R., R. H. Aron, and P. Todhunter: 1995, The Climate Near the Ground, Vieweg Publishing,
528 pp.
Germann, U. and J. Joss, 2004: Operational measurement of precipitation in mountainous terrain.
In P.Meischner, editor, Weather Radar: Principles and Advanced Applications, vol XVII of
Physics of Earth and Space Environment, chap. 2, 52–77, Springer Verlag.
Gladwell, M., 2005: Blink: The Power of Thinking without Thinking. Little, Brown and Company,
288 pp.
1 Mountain Weather Prediction: Phenomenological Challenges. . . 31

Glahn, H.R., and D.A. Lowry, 1972: The Use of Model Output Statistics (MOS) in Objective
Weather Forecasting. J. Appl. Meteor., 11, 1203–1211.
Graybeal, D.Y. and D.J. Leathers, 2006: Snowmelt-related flood risk in Appalachia: First estimates
from a historical snow climatology. J. Appl. Meteor. Climatol., 45, 178–193.
Hannesen, R., N. Dotzek, and J. Handwerker, 2000: Radar analysis of a tornado over hilly terrain
on 23 July 1996. Phys. Chem. Earth B 25, 1079–1084.
Hart, K. A., W. J. Steenburgh, and D. J. Onton, 2005: Model forecast improvements with decreased
horizontal grid spacing over fine-scale Intermountain orography during the 2002 Olympic
Winter Games. Wea. Forecasting, 20, 558–576.
Holle, R.L, and R.E. Lopez, 1993: Overview of Real-Time Lighting Detection Systems and Their
Meteorological Uses, NOAA Tech. Memo. ERL NSSL-102, 68 pp.
Horel, John D., Staley, Lloyd R., Barker, Timothy W. 1988: The University of Utah Interactive
Dynamics Program-One Approach to Interactive Access and Storage of Meteorological Data.
Bull. Amer. Meteor. Soc., 69, 1321–1327.
Horel, J., M. Splitt, L. Dunn, J. Pechmann, B. White, C. Ciliberti, S. Lazarus, J. Slemmer, D. Zaff,
J. Burks, 2002: MesoWest: Cooperative Mesonets in the Western United States. Bull. Amer.
Meteor. Soc., 83, 211–226.
Houze, R., W. Schmid, R. Fovell, and H.H. Schiesser, 1993: Hailstorms in Switzerland: Left
Movers, Right Movers, and False Hooks. Mon. Wea. Rev., 121, 3345–3370.
Howitt, R. E., D. MacEwan, and J. Medellin-Azuara, 2009: Economic impacts of reductions
in delta exports on Central Valley agriculture. Agriculture and Resource Economics update.
Giannini Foundation of Agricultural Economics, University of California. 4 pp.
Hudson, S., 2002: Sport and adventure tourism. Binghamton, NY, Haworth Press. 324 pp.
Huet, P., X. Martin, J.-L. Prime, P. Foin, C. Laurain, and P. Cannard, 2003: Retour d’expérience
des crues de septembre 2002 dans les departments du Gard, de l’Hèrault, du Vaucluse, des
Bouches-du-Rhône, de l’Ardèche et de la Drôme. Technical report, Ministère del’Ecologie et
du Développement Durable, Republique Franc¸aise, 133 pp.
Iverson, R.M., 1997: The physics of debris flows. Reviews of Geophysics, 35, 245–296.
Jóhannesson, T., and Þ. Arnalds, 2001: Accidents and economic damage due to snow avalanches
and landslides in Iceland. Jökull, 50, 81–94.
Johnson, G. L., C. Daly, G.H. Taylor, and C.L. Hanson, 2000: Spatial Variability and Interpolation
of Stochastic Weather Simulation Model Parameters. J. Appl. Meteor., 39, 778–796.
Kattelmann, R., 1997: Flooding from rain-on-snow events in the Sierra Nevada. Destructive Water:
Water-Caused Natural Disasters, Their Abatement and Control, G. H. Leavesley et al., Eds.,
IAHS Publication 239, 59–65.
Klein, G., 1998: Sources of Power: How People Make Decisions. The MIT Press, 330 pp.
Keiler, M., A. Zischg, S. Fuchs, M. Hama, and J. Stötter, 2005: Avalanche related damage
potential – changes of persons and mobile values since the mid-twentieth century, case study
Galtür. Natural Hazards and Earth System Sciences, 5, 49–58.
Knutson, C. 2001. “A comparison of droughts, floods and hurricanes in the U.S”. National Drought
Mitigation Center, Lincoln, NE, USA.
Kohler, J., 2003: Colorado, Wyoming recovering from worst blizzard in 90 years. Associated Press.
21 March 2003.
Lanicci, J.M., and T. T. Warner, 1991: A Synoptic Climatology of the Elevated Mixed-Layer
Inversion over the Southern Great Plains in Spring. Part I: Structure, Dynamics, and Seasonal
Evolution. Wea. Forecasting, 6, 198–213.
Linder, W., and W. Schmid, 1996: A tornadic thunderstorm in Switzerland exhibiting a radar-
detectable low-level vortex. Preprints, 12th Conf. on Clouds and Precipitation, Zurich,
Switzerland, International Commission on Cloud Physics and WMO, 577–580.
Linnerooth-Bayer, J. and A. Amendola, 2003: Introduction to special issue on flood risks in Europe.
Risk Analysis, 23, 537–543.
Lott, N., and T. Ross, 2006: Tracking and evaluating U.S. billion dollar weather disasters,
1980–2005. Preprints, AMS Forum: Environmental Risk and Impacts on Society: Successes
and Challenges. Atlanta, GA, Amer. Meteor. Soc.
32 M.P. Meyers and W.J. Steenburgh

Ludwig, F. L., D. K. Miller and S. G. Gallaher, 2006: Evaluating a hybrid prognostic/diagnostic


model that improves wind forecast resolution in complex coastal topography. J. Appl. Meteorol.
and Climatol., 45, 155–177.
Lyon, B., 2003: Enhanced Seasonal Rainfall in Northern Venezuela and the Extreme Events of
December 1999. J. Climate, 16, 2302–2306.
Maddox, R.A., L.R. Hoxit, C.F. Chappell, and F. Caracena, 1978: Comparison of Meteorological
Aspects of the Big Thompson and Rapid City Flash Floods. Mon. Wea. Rev., 106, 375–389.
Maddox, R. A., J. Z. Zhang, J. J. Gourley, and K. W. Howard, 2002: Weather radar coverage over
the contiguous United States. Wea. Forecasting, 17, 927–934.
Malek, E., T. Davis, R. S. Martin, and P. J. Silva, 2006: Meteorological and environmental aspects
of one of the worst national air pollution episodes (January 2004) in Logan, Cache Valley, Utah,
USA. Atmos. Res., 79, 108–122.
Matthews, W.J. and E. Marsh-Matthews. 2003. Effects of drought on fish across axes of space,
time, and ecological complexity. Freshwater Biology, 48, 1232–1253.
McMurdie, L. and C. Mass, 2004: Major numerical forecast failures over the northeast Pacific.
Wea. Forecasting, 19, 338–356.
Medina, S., and R. A. Jr. Houze, 2003: Air motions and precipitation growth in Alpine storms.
Q. J. R. Meteorol. Soc., 129, 345–372.
Meybeck, M., P. Green, and C. Vörösmarty, 2001: A new typology for mountains and other
relief classes: an application to global continental water resources and population distribution.
Mountain Research and Development 21, 34–45.
Meyers, M.P., J.S. Snook, D.A. Wesley, and G.S. Poulos, 2003: A Rocky Mountain storm. Part II:
The forest blowdown over the West Slope of the northern Colorado mountains – Observations,
analysis, and modeling. Wea. Forecasting, 18, 662–674.
Meyers, M. P., J. Ladue, J. Pringle, and C. Cuoco, 2006: A Forecaster challenge: convectively
enhanced winds in a strongly forced synoptic environment in complex terrain, 12th Conference
on Mountain Meteorology, Santa Fe, NM, AMS.
Mielke, K.B., and E.R. Carle, 1987: An Early Morning Dry Microburst in the Great Basin. Wea.
Forecasting, 2, 169–174.
Morris, D. and M. Walls, 2009: Climate change and outdoor recreation resources. Washington,
DC, Resources for the future.
Morss, R. E., O. V. Wilhelmi, M. W. Downton, and E. Gruntfest, 2005: Flood risk, uncertainty,
and scientific information for decision making: Lessons from an interdisciplinary project. Bull.
Amer. Meteor. Soc., 86, 1593–1601.
Morss, R.E., and F.M. Ralph, 2007: Use of Information by National Weather Service Fore-
casters and Emergency Managers during CALJET and PACJET-2001. Wea. Forecasting, 22,
539–555.
Myrick, D.T., and J.D. Horel, 2008: Sensitivity of Surface Analyses over the Western United States
to RAWS Observations. Wea. Forecasting, 23, 145–158.
Nance, L. B., and B. R. Colman, 2000: Evaluating the use of a nonlinear two-dimensional model
in downslope wind forecasts. Wea. Forecasting, 15, 715–729.
National Park Service. 2009. NPS Stats. National Park Service Public Use Statistics Office. http://
www.nature.nps.gov/stats/park.cfm.
National Research Council, 2004: Where the Weather Meets the Road: A Research Agenda for
Improving Road Weather Services. National Academies Press, Washington, D. C., 188 pp.
National Research Council, 2005: Flash Flood Forecasting over Complex Terrain. National
Academies Press, Washington, D. C., 206 pp.
National Wildfire Coordinating Group. 1997: Historical wildland firefighter fatalities: 1910–1996.
NFES 1849. Boise, ID: National Interagency Fire Center. 29 pp.
National Wildfire Coordinating Group, 2007: Wildland Firefighter Fatalities in the United States:
1990–2006. Boise, ID, National Interagency Fire Center. 30 pp.
Nöthiger, C., and H. Elsasser, 2004: Natural hazards and tourism: New findings on the European
Alps. Mountain Research and Development, 24, 24–27. Publisher: International Mountain
Society.
1 Mountain Weather Prediction: Phenomenological Challenges. . . 33

Novak, D.R., D.R. Bright, and M.J. Brennan, 2008: Operational Forecaster Uncertainty Needs and
Future Roles. Wea. Forecasting, 23, 1069–1084.
Nuissier, O., V. Ducrocq, D. Ricard, C. Lebeaupin, and S. Anquetin, 2008: A numerical study
of three catastrophic precipitating events over southern France. I: Numerical framework and
synoptic ingredients. Quart. J. Roy. Meteor. Soc., 134, 111–130.
Parish, T. R., and D. H. Bromwich, 1998: A case study of Antarctic katabatic wind interaction with
large-scale forcing. Mon. Wea. Rev., 126, 199–209.
Pliske, R., B. Crandall, and G. Klein, 2004: Competence in weather forecasting. Psychological
Investigations of Competent Decision Making. J. Schanteau, P. Johnson, and K. Smith, Eds.,
Cambridge University Press.
Poulos, G. S., D. A. Wesley, J. S. Snook, M. P. Meyers, 2002: A Rocky Mountain Storm. Part I:
The Blizzard – kinematic evolution and the potential for high-resolution numerical forecasting
of snowfall. Wea. Forecasting, 17, 955–970.
Ralph, F. M., P. J. Neiman, D. E. Kingsmill, P. O. G. Persson, A. B. White, E. T. Strem, E. D.
Andrews, R. C. Antweiler, 2003: The impact of a prominent rain shadow on flooding in
California’s Santa Cruz Mountains: A CALJET case study and sensitivity to the ENSO cycle.
Journal of Hydrometeorology, 4, 1243–1264.
Reagan , M., 1998: Canadian ice storm 1998. World Meteorological Organization Bulletin, Vol. 47,
No. 3, 250–256.
Reinecke P.A., D. R. Durran, 2009: Initial-condition sensitivities and the predictability of
downslope winds. J. Atmos. Sci., 66, 3401–3418.
Restrepo, P., D.P. Jorgensen, S.H. Cannon, J. Costa, J. Laber, J. Major, B. Martner, J. Purpura,
and K. Werner, 2008: Joint NOAA/NWS/USGS Prototype Debris Flow Warning System for
Recently Burned Areas in Southern California. Bull. Amer. Meteor. Soc., 89, 1845–1851.
Robbins, J., 2008: Bark Beetles Kill Millions of Acres of Trees in West. New York Times, November
18, D3.
Roebber, P.J., L.F. Bosart, and G.S. Forbes, 1996: Does Distance from the Forecast Site Affect
Skill? Wea. Forecasting, 11, 582–589.
Roebber, P. J., and J. R. Gyakum, 2003: Orographic influences on the mesoscale structure of the
1998 ice storm. Mon. Wea. Rev., 131, 27–50.
Roebber, P. J., D. M. Schultz, B. A. Colle, and D. J. Stensrud, 2004: Toward improved prediction:
High-resolution and ensemble modeling in operations. Wea. Forecasting, 19, 936–949.
Rotunno, R., Houze, R. A., 2007: Lessons on orographic precipitation from the Mesoscale Alpine
Programme. Quarterly Journal of the Royal Meteorological Society, 133, 811–830.
Schnell, R. C., S. J. Oltmans, R. R. Neely, M. S. Endres, J. V. Molenar & A. B. White, 2009: Rapid
photochemical production of ozone at high concentrations in a rural site during winter. Nature
Geoscience, 2(2), 120–122.
Schwarb, M., 2000: The Alpine precipitation climate: Evaluation of a high-resolution analysis
scheme using comprehensive rain-gauge data. Dissertation, Swiss Federal Institute of Technol-
ogy Zurich, 131 pp.
Schwarb, M., C. Daly, C. Frey, and C. Schär, 2001: Mean annual precipitation throughout the
European Alps 1971–1990. Hydrological Atlas of Switzerland, plate 2.6.
Scott, D. 2006. Ski Industry Adaptation to Climate Change. In: Tourism and Global Environmental
Change. S. Gossling and M. Hall (eds). London: Routledge. 262–285.
Sénési, S., P. Bougeault, J.-L. Chèze, P. Cosentino, and R.-M. Thepenier, 1996: The Vaison-
La Romaine Flash Flood: Mesoscale analysis and predictability issues. Wea. Forecasting, 4,
417–442.
Sivillo, J.K., J.E. Ahlquist, and Z. Toth, 1997: An Ensemble Forecasting Primer. Wea. Forecasting,
12, 809–818.
Skistar, 2009. The Alpine world market. http://corporate.skistar.com.
Smith, S. B. and M. K. Yau, 1987: The mesoscale effect of topography on the genesis of Alberta
hailstorms. Beitr. Phys, Atmos., 60, 371–392.
Snellman, L.W., 1977: Operational Forecasting Using Automated Guidance. Bull. Amer. Meteor.
Soc., 58, 1036–1044.
34 M.P. Meyers and W.J. Steenburgh

Snellman, L. W., 1982: Impact of AFOS on Operational Forecasting, Reprints, 9th Conference on
Weather Forecasting and Analysis, AMS, 13–16.
Soyka, F., 1983: The ion effect. New York, NY. Bantam Books.
Steiger, R. and M. Mayer, 2008: Snowmaking and climate change: Future options for snow
production in Tyrolean ski resorts. Mountain Research and Development, 28, 292–298.
Steenburgh, W. J., 2002: Using real-time mesoscale modeling in undergraduate education. Bull.
Amer. Meteor. Soc., 83, 1447–1451.
Steenburgh, W. J., 2003: One hundred inches in one hundred hours: Evolution of a Wasatch
Mountain winter storm cycle. Wea. Forecasting, 18, 1018–1036.
Steenburgh W. J., C. F. Mass, and S. A. Ferguson, 1997: The influence of terrain-induced
circulations on wintertime temperature and snow level in the Washington Cascades. Wea.
Forecasting, 12, 208–227.
Steenburgh, W.J., and T.I. Alcott, 2008: Secrets of the “Greatest Snow on Earth”. Bull. Amer.
Meteor. Soc., 89, 1285–1293.
Stuart, N.A., D.M. Schultz, and G. Klein, 2007: Maintaining the Role of Humans in the Forecast
Process: Analyzing the Psyche of Expert Forecasters. Bull. Amer. Meteor. Soc., 88, 1893–1898.
Thornes, J.E., 2000: Road Salting – an international benefit/cost review. Proceedings of the 8th
World Salt Symposium, 787–792, Elsevier, Amsterdam.
Vasiloff, S. K., J. Busto, M. P. Meyers, D. J. Gochis, and K. Friedrich, 2011: Flood hazard
monitoring in Colorado mountains with mobile Doppler radars [INVITED]. 25th Conference
on Hydrology. Seattle, WA. AMS.
Wagner, W. and J. D. Horel, 2008: Observing and forecasting snow surface temperatures for Nordic
ski race courses at the 2010 Winter Olympic games in Vancouver, BC. Proceedings for the
International Snow Science Workshop, Whistler, BC.
Wakimoto, R. M., 1985: Forecasting dry microburst activity over the high plains. Mon. Wea. Rev.,
113, 1131–1143.
Westrick, K. L., C. F. Mass, and B. A. Colle, 1999: The limitations of the WSR-88D radar network
for quanititive precipitation measurement over the coastal western United States. Bull. Amer.
Meteor. Soc., 80, 2289–2298.
Wetzel, M., M. Meyers, R. Borys, R. McAnelly, W. Cotton, A. Rossi, P. Frisbie, D.Nadler, D.
Lowenthal, S. Cohn, and W. Brown, 2004: Mesoscale snowfall prediction and verification in
mountainous terrain. Wea. Forecasting, 19, 806–828.
Whiteman, C. D., 2000: Mountain meteorology: Fundamentals and applications. Oxford University
Press, New York and Oxford. 355 pp.
Wojtiw, L., 1975: Climate summaries of hailfall in central Alberta (1957–73). Alberta Research
Council, 102 pp. [Available from Alberta Research Council, Edmonton, Alberta, T6H 5X2.]
Wood, T. V., R. A. Brown, and S. V. Vasiloff, 2003: Improved detection using negative elevation
angles for mountaintop WSR-88Ds. Part II: Simulations of the three radars covering Utah. Wea.
Forecasting, 18, 393–403.
World Health Organization, 2002: The world health report: Reducing risks, promoting healthy life.
Geneva, Switzerland. World Health Organization.
Yuan, H., S.L. Mullen, X. Gao, S. Sorooshian, J. Du, and H.M.H. Juang, 2007: Short-Range
Probabilistic Quantitative Precipitation Forecasts over the Southwest United States by the RSM
Ensemble System. Mon. Wea. Rev., 135, 1685–1698.
Chapter 2
Diurnal Mountain Wind Systems

Dino Zardi and C. David Whiteman

Abstract Diurnal mountain wind systems are local thermally driven wind systems
that form over mountainous terrain and are produced by the buoyancy effects
associated with the diurnal cycle of heating and cooling of the lower atmospheric
layers. This chapter reviews the present scientific understanding of diurnal mountain
wind systems, focusing on research findings published since 1988. Slope flows
are examined first, as they provide a good introduction to the many factors
affecting diurnal mountain wind systems. The energy budgets governing slope
flows; the effects of turbulence, slope angle, ambient stability, background flows
and slope inhomogeneities on slope flows; and the methods used to simulate slope
flows are examined. Then, valley winds are reviewed in a similar manner and the
diurnal phases of valley and slope winds and their interactions are summarized.
Recent research on large-scale mountain-plain wind systems is reviewed, with an
emphasis on the Rocky Mountains and the Alps. Winds occurring in closed basins
and over plateaus are then discussed, and analogies between the two wind systems
are outlined. This is followed by a discussion of forecasting considerations for
diurnal mountain wind systems. Finally, the chapter concludes with a summary of
open questions and productive areas for further research.

D. Zardi ()
Atmospheric Physics Group, Department of Civil and Environmental Engineering,
University of Trento, Via Mesiano, 77, Trento I-38123, Italy
e-mail: dino.zardi@ing.unitn.it
C.D. Whiteman
Atmospheric Sciences Department, University of Utah, Salt Lake City, UT, USA

F. Chow et al. (eds.), Mountain Weather Research and Forecasting, 35


Springer Atmospheric Sciences, DOI 10.1007/978-94-007-4098-3 2,
© Springer ScienceCBusiness Media B.V. 2013
36 D. Zardi and C.D. Whiteman

2.1 Introduction

Diurnal mountain winds develop, typically under fair weather conditions, over
complex topography of all scales, from small hills to large mountain massifs, and
are characterized by a reversal of wind direction twice per day. As a rule, upslope,
up-valley and plain-to-mountain flows occur during daytime and downslope, down-
valley and mountain-to-plain flows occur during nighttime.
Because diurnal mountain winds are produced by heating of atmospheric layers
during daytime and cooling during nighttime, they are also called thermally driven
winds. Horizontal temperature differences develop daily in complex terrain when
cold or warm boundary layers form immediately above sloping surfaces. Air at
a given point in the sloping boundary layer is colder or warmer than the air just
outside the boundary layer at the same elevation. These temperature differences
result in pressure differences and thus winds that blow from areas with lower
temperatures and higher pressures toward areas with higher temperatures and lower
pressures. The boundary layer flows and return, or compensation, flows higher in
the atmosphere form closed circulations.
The atmosphere above and around a mountain massif is composed of three
distinct regions or sloping layers in which the thermal structure undergoes diurnal
variations and in which diurnal winds develop: the slope atmosphere, the valley
atmosphere and the mountain atmosphere (Ekhart 1948; Fig. 2.1). The slope
atmosphere is the domain of the slope flows, the valley atmosphere is the domain
of the valley flows and the mountain atmosphere is the domain of the mountain-
plain flows. It is difficult to observe any one component of the diurnal mountain
wind system in its pure form, because each interacts with the others and each can
be affected by larger scale flows aloft.
Indeed well-organized thermally driven flows can be identified over a broad spec-
trum of spatial scales, ranging from the dimension of the largest mountain chains
to the smallest local landforms, such as, for instance, “farm-scale” topographic
features (Dixit and Chen 2011; Bodine et al. 2009). All of these flows are induced
by local factors, and in particular by the combination of landforms and surface
energy budgets, under synoptic scale situations allowing for their development, and
eventually conditioning them “top-down” by acting as boundary conditions. On the
other hand the same flows may in turn affect large-scale phenomena. This is the case,
for instance, of diurnal flows over sloping terrain forcing the Great Plains low-level
jet (cf. Holton 1966; Parish and Oolman 2010) and of some organized convective
systems triggered by thermally induced flows in the southern part of the Himalayas
(Egger 1987; Yang et al. 2004; Bhatt and Nakamura 2006; Liu et al. 2009).
The regular evolution of the diurnal mountain wind systems exhibits four
phases that are closely connected to the formation and dissipation of temperature
inversions. The nighttime phase, characterized by the presence of a ground-based
temperature inversion, or stable layer, and winds flowing down the terrain, is
followed by a morning transition phase in which the inversion is destroyed and the
2 Diurnal Mountain Wind Systems 37

Fig. 2.1 Diagram of the structure of the atmosphere above a mountain ridge (Adapted from Ekhart
1948, © Société Météorologique de France. Used with permission)

winds undergo a reversal to the daytime winds flowing up the terrain. The evening
transition phase completes the cycle as the inversion rebuilds and the winds once
again reverse to the nighttime direction.
Diurnal mountain winds are a key feature of the climatology of mountainous
regions (Whiteman 1990, 2000; Sturman et al. 1999). In fact, they are so consistent
that they often appear prominently in long-term climatic averages (Martı́nez et al.
2008). They are particularly prevalent in anticyclonic synoptic weather conditions
where background winds are weak and skies are clear, allowing maximum incoming
solar radiation during daytime and maximum outgoing longwave emission from
the ground during nighttime. Diurnal wind systems are usually better developed in
summer than in winter, because of the stronger day-night heating contrasts. The
local wind field patterns and their timing are usually quite similar from day to day
under anticyclonic weather conditions (see, e.g., Guardans and Palomino 1995). The
characteristic diurnal reversal of the slope, valley, and mountain-plain wind systems
is also seen under partly cloudy or wind-disturbed conditions, though it is sometimes
weakened by the reduced energy input or modified by winds aloft.
The speed, depth, duration and onset times of diurnal wind systems vary from
place to place, depending on many factors including terrain characteristics, ground
cover, soil moisture, exposure to insolation, local shading and surface energy budget
(Zängl 2004). Many of these factors have a strong seasonal dependence. Indeed, the
amplitude of horizontal pressure gradients driving valley winds displays appreciable
seasonal variations (see e.g. Cogliati and Mazzeo 2005).
38 D. Zardi and C.D. Whiteman

The scientific study of diurnal mountain winds has important practical


applications, ranging from operational weather forecasting in the mountains, to the
climatological characterization of mountain areas (e.g. for agricultural purposes),
to the impact assessments of proposed new settlements or infrastructures. Indeed
diurnal winds affect the distribution of air temperature in complex terrain (Lindkvist
and Lindqvist 1997; Gustavsson et al. 1998; Lindkvist et al. 2000; Mahrt 2006),
the transport and diffusion of smoke and other air pollutants, the formation and
dissipation of fogs (Cuxart and Jiménez 2011) and low clouds, frost damage in
vineyards and orchards, surface transportation safety from risks due to frost and
fog, flight assistance (e.g. through the issuance of Terminal Aerodrome Forecasts
and the evaluation of flying conditions for soaring or motorized flights), long-term
weathering of structures, and the behavior of wildland and prescribed fires.
Diurnal mountain winds have been studied scientifically since the nineteenth
century. Wagner (1938) laid the foundation for our current understanding of diurnal
mountain wind systems. He provided a systematic overview and synthesis of the
findings from field measurements and theoretical investigations gained at that time.
Indeed much of the early research was published in German or French. This
is evident in the first comprehensive English-language review of early research
provided by Hawkes (1947). Whiteman and Dreiseitl (1984) published English
translations of seminal papers by Wagner, Ekhart and F. Defant. Additional reviews
of research on diurnal mountain wind systems include those of Defant (1951),
Vergeiner and Dreiseitl (1987), Whiteman (1990), Egger (1990), Simpson (1999),
and Poulos and Zhong (2008). Additionally, chapters on diurnal mountain wind
systems are available in textbooks by Yoshino (1975), Atkinson (1981), Stull (1988),
Whiteman (2000), Barry (2008), and Geiger et al. (2009).
A fundamental obstacle to rapid progress in mountain meteorology is that
there are almost infinitely many possible terrain configurations. So, any field
measurement or numerical experiment that is valid for a specific situation does not
automatically have greater significance beyond that case. Nonetheless, significant
progress has been achieved by gathering real cases into groups according to similar
landforms, and generalizing results from these groups with respect to the inherent
important physical processes or mechanisms. The organization of this chapter
benefits from this approach, with a special section on basin and plateau circulations.
Another useful approach has been to focus conceptual or numerical simulations
on simple or idealized topographies to investigate the essential meteorological
processes. Other obstacles to progress should also be mentioned. Because most
studies of diurnal mountain wind systems have been conducted on fine weather days
with weak synoptic forcing, relatively little is known about other conditions. Further,
the effects of forest canopies and other types of varying and patchy surface cover
on diurnal wind systems have been largely neglected in much of the experimental
and modeling work to date. Finally, while diurnal mountain wind systems are
thought to be complete or closed circulations, little experimental evidence has been
accumulated on the difficult-to-measure “return” or upper-branch circulations.
This review aims to provide information on the basic physics of diurnal
mountain wind systems and to summarize new research findings since the American
2 Diurnal Mountain Wind Systems 39

Meteorological Society’s 1988 workshop on mountain meteorology, as later


published in an AMS book (Blumen 1990). Priority in the referenced material
is given to peer-reviewed papers published in English since 1988. Other chapters
in the present book also deal peripherally with diurnal mountain wind systems.
Chapters 3 and 4 discuss the effects on valley winds of dynamically induced
channeling from larger scale flows. Chapter 5 treats atmospheric boundary layer
phenomena in mountain areas that affect pollutant transport. Chapters 6 and 7 deal
with orographic precipitation, including triggering from thermally forced flows.
Chapter 8 treats observational resources and strategies for mountain atmosphere
processes, including diurnal mountain winds. Chapters 9, 10, and 11 discuss
numerical model simulations and operational weather forecasting.

2.2 The Slope Wind System

The slope wind system is a diurnal thermally driven wind system that blows up
or down the slopes of a valley sidewall or an isolated hill or mountain, with
upslope flows during daytime and downslope flows during nighttime. The up- and
downslope flows are the lower branch of a closed circulation produced by inclined
cold or warm boundary layers that form above the slopes. During daytime, the
heated air in the boundary layer above a slope rises up the slope while continuing to
gain heat from the underlying surface. During nighttime, air in the cooled boundary
layer flows down the slope while continuing to lose heat to the underlying surface.
Upslope and downslope winds are alternatively called anabatic and katabatic winds,
respectively. These terms, however, are also used more comprehensively to refer to
any flows that run up or down the terrain. Further, the term katabatic is typically used
to describe the large-scale slope flows on the ice domes of Antarctica and Greenland.
The term drainage wind may be used to refer collectively to downslope and down-
valley winds, and may also include mountain-to-plain flows. To reduce confusion,
we will use the terms upslope and downslope, and will omit from consideration
the flows over extensive ice surfaces which, in contrast to the smaller-scale mid-
latitude flows discussed here, are affected by Coriolis forces (see, e.g., Kavčič and
Grisogono 2007; Shapiro and Fedorovich 2008). An example of typical nighttime
and daytime wind and temperature profiles through the slope flow layer is given
in Fig. 2.2. The biggest daytime temperature excesses and nighttime temperature
deficits occur near the ground, where radiative processes heat and cool the surface.
It is here that the buoyancy forces that drive the slope flows are strongest. The wind
profiles, however, have their peak speeds above the surface, since surface friction
causes the speeds to decrease to zero at the ground. This gives both the daytime and
nighttime wind speed profiles their characteristic jet-like shape in which the peak
wind speed occurs above the surface.
Typical characteristics of the equilibrium steady-state wind and temperature
profiles for mid-latitude downslope winds (Fig. 2.2a) include the strength (3–7ı C)
and depth (1–20 m) of the temperature deficit layer, the maximum wind speed
40 D. Zardi and C.D. Whiteman

Fig. 2.2 Typical wind and a


temperature profiles through
(a) downslope and (b)
upslope flows on valley
sidewalls. Shown are the
typical depths and strengths
of the temperature deficits or
excesses, the heights and flow
strengths of the wind
maxima, and the slope flow
depths. The TKE profile in
(a) follows that of
Skyllingstad (2003) (Adapted
from Whiteman 2000)

(1–4 m s1 ), its height above ground (1–15 m), and the depth of the downslope
flow layer (3–100 m). Downslope flows on uniform slopes often increase in both
depth and strength with distance down the slope. A useful guideline for field studies
is that the depth of the temperature deficit layer at a point on a slope is about 5% of
the vertical drop from the top of the slope (Horst and Doran 1986). Characteristic
values for upslope flows (Fig. 2.2b) include the strength (2–7ıC) and depth (10–
100 m) of the temperature excess layer, the maximum wind speed (1–5 m s1 ),
its height above ground (10–50 m), and the depth of the upslope flow layer (20–
200 m). Upslope flows on uniform slopes often increase in both depth and speed
with distance up the slope.
Not covered in this review are mid-latitude “skin flows” (Manins and Sawford
1979a) – shallow (<1 or 2 m), weak (perhaps <1 m s1 ), non-turbulent katabatic
2 Diurnal Mountain Wind Systems 41

flows that have been reported by several investigators (Thompson 1986; Manins
1992; Mahrt et al. 2001; Soler et al. 2002; Clements et al. 2003). These have not yet
been studied systematically.
Our discussion of slope flows begins with a simplified set of Reynolds-
averaged Navier–Stokes (RANS) equations to guide understanding. Then different
approaches to modeling slope flows are discussed and the state of scientific
knowledge on upslope and downslope flows, as supported by both field and
modeling investigations, are summarized.

2.2.1 Slope Flow Models

2.2.1.1 Reynolds-Averaged Navier–Stokes Equations

A reduced set of RANS equations can be used heuristically to gain an understanding


of the key physics of slope flows. This is conveniently done using a coordinate
system that is oriented along (s) and perpendicular (n) to an infinitely extended
slope of uniform inclination angle ˛. n is perpendicular to the slope and increases
upwards; s is parallel to the slope and increases in the downslope direction. Slope
flows result from the cooling or heating of an air layer over the slope relative to air
at the same elevation away from the slope. The potential temperature profile over
the slope is expressed as

 D 0 C  z C d.s; n; t/ (2.1)

where 0 is a reference value for the potential temperature at the ground in the
absence of surface cooling or heating,  is the ambient vertical gradient of potential
temperature, z is the vertical coordinate (z D n cos ˛  s sin ˛), and d is the
potential temperature perturbation in the vicinity of the slope, negative when the air
is colder than in the unperturbed state. The along-slope and slope perpendicular
equations of motion, the thermodynamic energy equation, and the continuity
equation are then written as

@u @u @u 1 @.p  pa / d @u0 w0
Cu Cw D  g sin ˛  (2.2)
@t @s @n 0 @s 0 @n
@w @w @w 1 @.p  pa / d
Cu Cw D C g cos ˛ (2.3)
@t @s @n 0 @n 0

@ @ @ 1 @R @w0  0
Cu Cw D  (2.4)
@t @s @n 0 cp @n @n
42 D. Zardi and C.D. Whiteman

and

@u @w
C D0 (2.5)
@s @n
where u and w are the velocity components parallel to s and n, respectively. These
equations, from Horst and Doran (1986), were derived by Manins and Sawford
(1979b) using the Boussinesq approximation. They considered the flow to be two-
dimensional, with wind and temperature independent of the cross-slope direction,
with the ambient air at rest, and with negligible Coriolis force. 0 is a reference value
of density in the absence of surface cooling, g is the gravitational acceleration, cp
is the specific heat of air at constant pressure, R is the upward radiative flux, and
u0 w0 and w0  0 are the kinematic turbulent Reynolds fluxes of momentum and heat.
Note that, consistent with the assumption of a shallow layer, this set of reduced
equations includes only the slope-normal turbulent fluxes in (2.2) and (2.4); other
turbulent fluxes are ignored. For simplicity, latent heat fluxes are not included here,
although they can sometimes play a crucial role in the overall energy budgets. The
ambient pressure pa is determined from hydrostatic balance @pa =@z D a g, and,
through the ideal gas law and the definition of potential temperature, is a function of
the ambient potential temperature distribution a D 0 C  z. Accelerations normal
to the slope are generally considered negligible so that the atmosphere normal to
the slope is in quasi-hydrostatic balance and the RHS of (2.3) is practically zero.
The coordinate system rotation needed to get to this set of equations can lead to
interpretation errors if not done properly, as explained by Haiden (2003). Mahrt
(1982) provided further details on the coordinate system and an overview of the
classification of gravity flows in terms of the relative magnitudes of the individual
terms of the along-slope momentum equation (2.2).
This system of equations can be solved numerically after the turbulent fluxes
are parameterized in terms of mean flow variables. The parameterization of the
momentum flux normal to the slope is usually expressed in terms of the wind
shear and turbulence length and velocity scales. This type of parameterization was
originally developed using knowledge of turbulence profiles over flat ground in
flows that were driven by large-scale pressure gradients that changed little with
height outside the surface layer. In contrast, slope flows are shallow flows in which
the largest forcing is at the ground. The suitability of such parameterizations is thus
called into question (Grisogono et al. 2007).
It is instructive to consider which terms in these equations are most important
for understanding slope flow physics. The key term that drives slope flows is the
buoyancy force caused by the temperature excess or deficit that forms over the slope
(second term on the right hand side of (2.2)). The temperature excess or deficit itself
is produced by radiative and sensible heat flux convergences or divergences (the first
and second terms on the RHS of (2.4)) that warm or heat a layer of air above the
slope. Considering the along-slope momentum equation (2.2) for the downslope
flow situation, a steady state flow may be retarded by advection of slower moving
air into the flow (two advection terms on the LHS), by the thickening or cooling of
2 Diurnal Mountain Wind Systems 43

the layer with downslope distance which causes an adverse pressure gradient (first
term on the RHS, see Princevac et al. 2008), and by the divergence of turbulent
momentum flux (last term on the RHS).
Mesoscale numerical models that use the full set of RANS equations (see
Chaps. 9 and 10) have been used to simulate a variety of meteorological phenomena.
These models, where turbulence is parameterized with closure assumptions, have
been applied to complex terrain and have proven very useful in gaining an under-
standing of diurnal mountain wind systems (see, e.g., Luhar and Rao 1993; Denby
1999). Extensive experience with mesoscale models of this type, however, has
shown that shallow slope flows cannot be resolved adequately in a mesoscale setting
for which the turbulence parameterizations are most appropriate. Simulated downs-
lope flows are often too deep and strong, temperature deficits over the slope are too
deep and weak, and it has been a challenge to get both the wind and temperature
fields to agree with observations (see Chap. 10). In instances where the wind and
temperature fields are in rough agreement with observations, these models have been
used to investigate the relative roles of individual terms in the momentum and heat
budget equations. Comparisons of slope flow simulations with multiple mesoscale
models have provided valuable information on the effects of different parameteriza-
tions, resolutions, coordinate systems, and boundary conditions on model outputs.
Zhong and Fast (2003) compared simulations of flows in Utah’s Salt Lake Valley
using the RAMS, MM5 and Meso-Eta models and compared the simulations to
slope flow data. Forecast errors with these models were surprisingly similar, despite
different coordinate systems, numerical algorithms and physical parameterizations.
The models exhibited a cold bias, with weaker than observed inversion strengths,
especially near the surface. All models successfully simulated the diurnally revers-
ing slope flows but there were significant discrepancies in depths and strengths of
wind circulations and temperature structure. Surprisingly, it is only recently that
mesoscale numerical models have added parameterizations of slope shadowing from
surrounding topography (Colette et al. 2003; Hauge and Hole 2003).

2.2.1.2 Large-Eddy Simulation (LES) Models

Large-eddy simulation (LES) of turbulent flows is a numerical alternative to the


RANS equations. It builds on the observation that most of the turbulence energy
is contained in the largest eddies which, as they break down, cascade their energy
into the smaller isotropic turbulent eddies where energy is dissipated. Accordingly,
the LES approach explicitly resolves the larger energy-containing eddies and uses
turbulence theory to parameterize the effects on the turbulent flow of the smallest
scale eddies in the inertial sub-range of atmospheric motions. Because energy-
containing eddies are much smaller in size in stable boundary layers than in
convective boundary layers, the first LES simulations for inclined slopes were made
for upslope flows (Schumann 1990). A more detailed account on these findings
will be given in Sect. 2.4 below. LES simulations of downslope flows in the
stable boundary layer, with its much smaller turbulence length scales, requires finer
44 D. Zardi and C.D. Whiteman

grids and more extensive computational resources. Thus, numerical simulations of


downslope flows have only recently become possible on highly idealized slopes with
recent increases in computer power (Skyllingstad 2003; Axelsen and van Dop 2008,
2009). Research in this area looks very promising, but is still at an early stage. Some
applications of LES in complex terrain are provided in Chap. 10.

2.2.1.3 Analytical Models

Useful slope flow models can sometimes be developed by further simplifying the
reduced set of RANS equations (2.1) through (2.5), retaining only the key terms,
and solving the simplified set analytically. Such analytical models are of two types –
profile and hydraulic flow models.

Profile Models

Profile models consider local equilibria in the buoyancy and momentum equations
and obtain analytical solutions for equilibrium vertical profiles of velocity and
temperature in the slope flow layer by neglecting or simplifying some of the terms
in the equations. The classical Prandtl (1942) model is an equilibrium gravity
flow solution in which constant eddy diffusivities are used for the friction term
and a simple thermodynamic relationship balances the diffusion of heat with the
temperature advection associated with the basic state stratification. The Prandtl
model agreed qualitatively with observations of both upslope and downslope
flows, successfully producing the typical jet-like wind profiles and temperature
deficit profiles seen in slope flows (Fig. 2.3). Prandtl’s model was modified by
Grisogono and Oerlemans (2001a, b) to allow eddy diffusivities to vary with
height, improving its agreement with observations. Extension of Prandtl’s (1942)
approach to reproduce nonstationary flows was accomplished by Defant (1949) and
Grisogono (2003). Other investigators have developed alternative models of slope
flows by finding analytical solutions to reduced sets of equations (Zammett and
Fowler 2007; Shapiro and Fedorovich 2009; see also in Barry 2008).

Hydraulic Flow Models

Hydraulic models start from the RANS equations, average the terms over the
slope flow depth and treat the drainage flow as a single layer interacting with an
overlying stationary fluid. They focus on layer-averaged quantities such as mass
and momentum fluxes and can be used to determine how bulk quantities vary with
distance up or down the slope. Manins and Sawford (1979a, b), Brehm (1986), and
Horst and Doran (1986) were early developers of this approach for slope flows.
They, Papadopoulos et al. (1997), Papadopoulos and Helmis (1999), Haiden and
Whiteman (2005) and Martı́nez and Cuxart (2009) have used observations from
2 Diurnal Mountain Wind Systems 45

a b

Fig. 2.3 Observed (O) and theoretical (T) (a) downslope and (b) upslope wind profiles. Obser-
vations are from a slope on the Nordkette near Innsbruck, Austria. The theoretical curves come
from Prandtl’s (1942) model. Differences above the jet maximum between theory and observations
are indicated by the dotted lines (Adapted from Defant 1949, © Springer-Verlag Publishing Co.,
Vienna, Austria. Used with permission)

lines of instrumented towers or tethersondes running down a slope, or the output of


numerical model simulations, to estimate the magnitudes of the different terms in
the hydraulic flow equations.

2.2.1.4 Laboratory Tank Models

Laboratory water tanks have been used to investigate both up- and downslope
flows on sloping surfaces in basins containing stratified fluid. A bottom-heated,
salt-stratified water tank over a slope with adjacent plain and plateau was used by
Reuten et al. (2007) and Reuten (2008) to investigate layering, venting and trapping
of pollutants that are carried in upslope flows. Conditions conducive to trapping
include weak large-scale flows, strong sensible heat flux, weak stratification, a
short or no plateau, symmetric geometry, a low ridge, and inhomogeneities in
surface heat flux. Reuten et al. (2007) speculate that inhomogeneities in slope
angle and surface roughness can also produce trapping. In a similar laboratory
experiment by Princevac and Fernando (2008), but this time with a V-shaped tank
containing thermally stratified water that was heated uniformly on the sloping
surfaces, upslope flows were found to leave the sidewalls under certain conditions,
intruding horizontally into the stratified fluid over the valley center.
46 D. Zardi and C.D. Whiteman

Upslope winds flow up the slope as a bent-over plume rather than rising
vertically, even in a neutral environment. These flows become turbulent when the
heat flux becomes large enough. Laboratory simulations by Princevac and Fernando
(2007) show that, for a certain range of Prandtl numbers,1 there is a minimum
slope angle above which upslope flows can be sustained. For average atmospheric
conditions the critical angle is only 0.1ı . Princevac and Fernando (2007) used this
result to explain Hunt et al.’s (2003) observations of a dominant upslope flow on a
0.18ı slope in the Phoenix valley.
Downslope flows into a stratified environment have been studied in laboratory
tanks, as well, by introducing a continuous source of negatively buoyant fluid at
the top of a constant-angle slope (Baines 2001). In this case, the flow maintains
a uniform thickness, with a distinct boundary at its top, until it approaches its
level of neutral buoyancy, where it leaves the slope. Turbulent transfers of mass
and momentum occur across the interface, causing a continuous loss of fluid
(detrainment) from the downslope flow. Flows of negatively buoyant fluid over steep
slopes, on the other hand, are in the form of entraining plumes (Baines 2005). It
should be mentioned that realistic downslope flows over valley sidewalls are not
produced by a continuous source of negatively buoyant fluid at the top of the slope.
Rather, they must lose heat continuously to the underlying surface to maintain their
negative buoyancy. Nonetheless, this laboratory tank analog provides important
information on interactions that occur at the top of the downslope flow layer.

2.2.2 Surface Radiative and Energy Budgets: The Driving


Force for Slope Flows

Thermally driven winds in complex terrain are produced by the formation of


inclined layers of temperature excess or deficit (relative to the ambient environment)
along terrain slopes of different scale. It is thus of paramount importance to
understand the physical processes that produce temperature changes above slopes.
The key principle is that of energy conservation, which relates local changes in
potential air temperature at any point to the budget of heat fluxes at that point. The
potential temperature tendency equation, neglecting thermal conduction and latent
heat release, can be written as

@   
 cp  D r  A  r  H  r  R (2.6)
@t T

1
The Prandtl number Pr is a nondimensional number defined, for any fluid, as the ratio of the
kinematic viscosity  to the thermal diffusivity ›, i.e. Pr D /›. As an extension, for turbulent
flows the turbulent Prandtl number Prt is defined as the ratio of the eddy viscosity Km to the eddy
heat diffusivity Kh : Prt D Km /Kh . Both of these numbers provide, for laminar and turbulent flows
respectively, an estimate of the relative importance of convection versus diffusion in heat transfer
processes involved with the flow.
2 Diurnal Mountain Wind Systems 47

where A D  cp u is the sensible heat flux associated with local advection of


warmer or colder air by the mean three-dimensional wind velocity u, H D cp u0  0
is the turbulent sensible heat flux produced by the coupling between turbulent
fluctuations of wind velocity and potential temperature, R is the radiative heat flux
(including both the shortwave and longwave spectra). In the absence of mean wind,
as for instance over flat terrain under anticyclonic synoptic-scale conditions, the
mean advection term A is negligible. In contrast, over complex terrain identical
weather conditions produce thermally driven flows, and thus the advection term is
usually a relevant one. While latent heat release is not included in (2.6), it may be
very important over vegetated areas and in mountainous regions adjacent to large
water bodies, and even in some arid environments where condensation of moisture
advected from elsewhere produces dew or fog (cf. Khodayar et al. 2008). Also
not considered explicitly in (2.6) is local heat storage in forest canopies and other
materials. During the course of a day all of the terms in (2.6) undergo changes with
time, and display strongly varying distributions in space.
Slope flows, the smallest-scale diurnal wind system, respond rapidly to temporal
and spatial variations in sensible heat fluxes caused by variations in the surface
energy budget. The surface energy budget is affected by changes in net radiation,
ground heat flux, soil moisture and its influence on the partitioning of energy
between sensible and latent heat fluxes, by changes in surface cover or vegetation,
cloud cover (Ye et al. 1989), break-in of background winds or turbulent episodes,
advection, etc. Surface radiation and energy budget principles, measurements in
mountain valleys, and examples of spatial variations across sample topographies
have been discussed by Whiteman (1990), Matzinger et al. (2003) and Oncley
et al. (2007).
Radiative and turbulent sensible heat flux divergences (convergences) cool (heat)
a layer of air above a slope (see [2.4]), creating a buoyancy force that has an along-
slope gravitational component (2.2) that is the basic driving force for downslope
(upslope) flows. In this section we will consider topographic effects on radiative
and sensible heat flux divergences, a topic that has received increasing attention in
the last 20 years.

2.2.2.1 Radiative Flux Divergence, r  R

Radiative flux divergence contributes significantly to the cooling of clear air in


the nocturnal boundary layer over flat terrain (Garratt and Brost 1981; André and
Mahrt 1982; Ha and Mahrt 2003) and can be expected to play an important role in
producing the cooling over sloping surfaces that drives downslope flows. Despite
this expectation, slope flow models are generally driven by assumed or measured
rates of surface sensible heat flux or prescribed surface temperatures, with radiative
flux divergence in the atmosphere above the slope assumed negligible. While no
direct measurements of radiative flux divergence have yet been made over slopes,
simple radiative transfer (RT) models (e.g., Manins 1992) suggest that radiative
flux divergence cannot be neglected as a source of cooling or heating of the slope
boundary layer.
48 D. Zardi and C.D. Whiteman

Because of recent improvements in RT models and increases in computer speeds,


it is now possible to treat both long- and shortwave RT in a realistic way for
complex terrain areas for clear sky conditions using Monte Carlo simulations
(Mayer 2009). Monte Carlo RT models track the paths of individual photons,
accounting in a stochastic way for emission, scattering, reflection and absorption
interactions between photons and air molecules, aerosols and ground surfaces. The
simulations, which often involve the tracking of millions of individual photons, can
be accomplished for situations where required input data are available, including a
digital terrain model, vertical atmospheric profiles of temperature, radiatively active
gases (e.g. water vapor and carbon dioxide) and aerosols, and spatially resolved
albedos and surface radiating temperatures. Improvements in complex terrain RT
modeling will come from the further development of these models to handle key
three-dimensional aspects of complex terrain atmospheres including the existence
of shallow cold and warm air layers on the slopes. Monte Carlo models can, in
principle, be used to simulate not only radiative fluxes (see, e.g., Chen et al. 2006)
but also radiative flux divergences, even over shallow slope layers. In the future,
such simulations can perhaps be tested against measurements and used to develop
better parameterizations for simpler models.
The MYSTIC Monte Carlo RT model (Mayer and Kylling 2005) has recently
been applied to the complex terrain of Arizona’s Meteor Crater and successfully
evaluated against longwave and shortwave irradiance measurements made on slope-
parallel broadband radiometers on the floor, sidewalls and rim of the crater on one
October day (Mayer et al. 2010). Figure 2.4 shows Monte Carlo simulations of
direct and diffuse irradiance at noon over the entire crater domain for the same day.
Following this successful evaluation of the model performance, further simulations
were performed to determine the cooling rates caused by radiative flux divergence
in idealized valley and basin atmospheres and in a realistic simulation of the Meteor
Crater (Hoch et al. 2011). In this parametric study the radiative flux divergence-
induced cooling rates for different temperature profiles representative of different
times of day were compared to those over flat terrain for valleys and basins of
different widths. Longwave radiative cooling in topographic depressions is generally
weaker than over flat terrain because of the counter-radiation from surrounding
terrain, but strong temperature gradients near the surface associated with nighttime
inversions and temperature deficit layers over slopes significantly increase longwave
cooling rates. The effects of the near-surface temperature gradients extend tens of
meters into the overlying atmosphere and can produce cooling rates on the order
of 30 K day1 that play an important role in the in-situ cooling which produces
and maintains drainage flows. Hoch et al. (2011) attributed nearly 30% of the total
nighttime cooling observed in the Meteor Crater on a calm October night to radiative
flux divergence. They approximated near-slope temperature gradients by specifying
the ground radiating temperature and the air temperature at the first model level at
5 m. Since model results for flat terrain show that radiative flux divergence depends
strongly on the detailed air temperature profile in the lowest meters above a surface
(Räisänen 1996), it will be important in future RT work to gain better resolution of
slope temperature structure profiles.
2 Diurnal Mountain Wind Systems 49

Fig. 2.4 Monte Carlo radiative transfer model simulations of direct (left) and diffuse (right)
shortwave radiation fluxes under clear sky conditions over a 4  4 km domain encompassing
Arizona’s Meteor Crater at noon, 15 October 2006 (From Sebastian Hoch)

2.2.2.2 Turbulent Sensible Heat Fluxes and Their Divergence, r  H

Several field experiments (e.g., Manins and Sawford 1979a; Horst and Doran
1988; Doran et al. 1989, 1990) have measured turbulent sensible heat fluxes at
multiple heights on slope towers, but vertical gradients of these fluxes (i.e., the
vertical divergence of sensible heat fluxes) have generally not been reported because
of concerns regarding accuracy, with radiative flux divergence being the small
difference between two large quantities. An accurate measurement of the total
divergence through the slope flow layer would require a measurement very close to
the surface and a measurement at the top of the slope flow layer. Both measurements
are problematic: the upper measurement would have to be at the top of the slope
flow – which depends on background weather conditions, varies with time during
the day, and is not easily detected – and the lower measurement would suffer from
underreporting of fluxes caused by instrumental errors associated with near-ground
eddies being smaller than the sonic anemometer sampling volume (Kaimal and
Finnigan 1994). Moreover, many slopes are inhomogeneous, covered with patches
of forest, bushes, rocks, or uneven ground making it difficult to find representative
measurement sites (Van Gorsel et al. 2003a). Further requirements regarding the
appropriate coordinate system for the measurements, the representativeness of
the measurement location in terms of the homogeneity of the surface flux in the
upwind direction (i.e., flux footprint), and the post-processing of turbulence data
collected over sloping surfaces make turbulence measurements in slope flows prone
to errors. The reader will find further detailed discussions on these issues in papers
by Andretta et al. (2002), de Franceschi and Zardi (2003), Hiller et al. (2008) and
de Franceschi et al. (2009).
A surprising finding from evidence that has accumulated since the mid 1990s is
that significant inaccuracies are occurring with the measurement of surface heat
fluxes with existing research equipment. When individual terms of the surface
50 D. Zardi and C.D. Whiteman

energy budget (see (2.7) below) are measured independently, an imbalance is often
found. This finding comes primarily from field experiments over flat, homogeneous
terrain in a variety of climate settings (see e.g., Oncley et al. 2007). The surface
energy budget imbalance during daytime is characterized by a sum of ground,
sensible and latent heat fluxes that is too small to balance the measured net radiation,
undershooting this value by 10–40%. Imbalances have also been found during
nighttime, again with the sum of the ground, sensible and latent heat flux terms
being smaller than the nighttime longwave loss. Since the Law of Conservation
of Energy must be satisfied, this suggests that experimental errors are the culprit.
In complex terrain, additional sources of error may arise because of the neglect
of budget terms under assumptions of shallow flows and horizontal homogeneity.
Ground heat flux is often poorly measured, but is expected to be relatively small
and thus incapable of explaining the error magnitude. Net radiation, on the other
hand, is relatively well measured, especially when individual components of net
radiation are measured with high-quality broadband radiometers. This suggests that
the turbulent heat fluxes are being underestimated, and attention has been focused
on advective effects, the differing flux footprints at different measurement heights,
and unmeasured flux convergences/divergences below the typical 5- to 10-m eddy
correlation measurement heights. Recent complex terrain research programs have
also encountered these measurement problems (e.g., Rotach et al. 2008).
Because sensible heat flux divergence measurements are unavailable over slopes,
most slope flow models have been driven by an assumed value of surface sensible
heat flux. Most models additionally assume that sensible heat flux is constant along
a slope, a useful approximation for a uniform slope on an isolated mountainside,
but questionable on valley sidewalls. For simple surfaces with uniform surface
cover, the surface heat flux QH can be expressed in terms of the surface energy
and radiation budgets as follows

QH D cp w0  0 s D .S C D C K " CL # CL " CQG C QE / (2.7)

where S is the incoming direct radiation, D is the incoming diffuse radiation, K " is
the reflected shortwave radiation, L # is the incoming longwave radiation, L "
is the outgoing longwave radiation, QG is the ground heat flux, and QE is the
latent heat flux. The individual terms represent the transfer of energy per unit time
through a sloping unit surface area, as measured in W m2 . In this equation, the sign
convention is that fluxes toward the surface are considered positive and fluxes away
from the surface are considered negative. Equation 2.7 illustrates the large number
of processes on which sensible heat flux depends.
During daytime, in complex, three-dimensional valley topography, the time
varying sensible heat flux is driven primarily by the input of direct radiation on a
slope (S in (2.7)). This input depends on the path of the sun’s movement through the
sky, on the azimuth and inclination angles of the slopes, the surface albedo (Pielke
et al. 1993), the physical and vegetative properties of the surface, and especially
near sunrise and sunset, on whether shadows are cast on the slope from surrounding
topography (Oliver 1992; Sun et al. 2003; Zoumakis et al. 2006). The complicated
2 Diurnal Mountain Wind Systems 51

Fig. 2.5 Modeled propagation of shadows and extraterrestrial insolation across the Meteor Crater
on October 15 at different times of day (MST). Shades of gray are insolation, with black indicating
0 and white indicating 1,364 W m2 . The horizontal scale is as in Fig. 2.4

time evolution of shadows and insolation on sloping surfaces in Arizona’s Meteor


Crater are shown as an example of this factor in Fig. 2.5. Matzinger et al. (2003) and
Rotach and Zardi (2007) suggest that the daytime spatial variation of sensible heat
fluxes over complex terrain can be reasonably well estimated by relating sensible
heat fluxes to global insolation on slopes. This approach is promising, but it is
clear from (2.7) that many other factors affect the sensible heat flux, including both
radiative and surface energy budget variables.
During nighttime, solar radiative fluxes become zero, but additional factors lead
to significant spatial variations in sensible heat flux on the sidewalls. Strong vertical
variations in temperature are superimposed on the slopes once an inversion forms
and grows in the valley, affecting outgoing longwave radiation. At the same time,
incoming longwave fluxes vary along the slope due to their varying views of the sky
and surrounding terrain, with significant amounts of back radiation received at low
elevation sites from the surrounding terrain.
A review paper by Duguay (1993) summarized some of the interactions between
the radiation field and topography, focusing on the challenge of modeling radiative
fluxes in complex topography. Key questions were the parameterization of diffuse
sky radiation and of the terrain-reflected shortwave radiation, and the difficulty
of dealing with anisotropic radiation fields. Other investigators have addressed
individual aspects of the radiative transfer problem. Olyphant (1986) investigated
the influence of longwave radiation from surrounding topography on the energy
balance of snowfields in the Colorado Front Range. Longwave loss is typically
reduced by about 50% when compared to the ridge tops. Plüss and Ohmura (1997)
52 D. Zardi and C.D. Whiteman

addressed the influence of longwave counter-radiation from elevated snow-covered


terrain in mountainous terrain in a modeling study. They stress the importance of the
surface temperature of the sky-obstructing terrain, and also of the air temperature in
between the point of interest and the surrounding terrain, especially for inclined
surfaces. In snow-covered environments, neglecting the effects of the usually
warmer air temperatures leads to an underestimation of the radiation emitted from
the obstructed parts of the hemisphere. Height dependence of fluxes and of other
influences such as seasonal dependencies, cloud radiative forcing, the Greenhouse
Effect, etc. were investigated by Marty et al. (2002), Philipona et al. (2004) and
Iziomon et al. (2001).
The effects of the smaller scale topography on the radiation field arising
from variations in exposure, shading and sky view effects as well as surface
properties like albedo were investigated by Whiteman et al. (1989a) and Matzinger
et al. (2003) by measuring radiative fluxes on surfaces parallel to the underlying
topography. Oliphant et al. (2003) addressed these terrain effects by combining
simple radiometric measurements with modeling efforts. Whiteman et al. (1989a)
reported instantaneous values and daily totals of radiation budget components for
five sites in different physiographic regions of Colorado’s semiarid, northwest-
southeast-oriented Brush Creek Valley for a clear-sky day near the fall equinox,
finding significant differences in the radiation budgets between sites. A higher
daytime net radiation gain and a higher nighttime net radiation loss at a ridge top
site were attributed to its unobstructed view of the sky. Oliphant et al. (2003) found
slope aspect and slope angle to be the dominant features in the radiation budget of
a complex terrain area in New Zealand, followed by elevation, albedo, shading,
sky view factor and leaf area index. Matzinger et al. (2003), in a cross-section
through Switzerland’s Riviera Valley, found a decrease in downward longwave
radiation and an increase of net radiation with height. Müller and Scherer (2005)
recognized the importance of including topographic influences on the radiation
balance in mesoscale weather forecast models. They introduced a subgrid scale
parameterization of the topographic effects and managed to improve 2-m temper-
ature forecasts for areas where the interaction between topography and radiation
was most apparent. These include areas dominated by wintertime shading, areas
with increased daytime sun exposure, and deeper valleys where nighttime counter-
radiation is important. Others have addressed more specific topographic effects and
their impacts on slope boundary layer evolution. Colette et al. (2003) focused on the
effect of terrain shading on the modeled morning break-up of the nocturnal stable
boundary layer in an idealized valley, finding that shading can have a significant
impact on the timing of inversion break-up.
Soil moisture and spatial variations of soil moisture also affect slope flows
(Banta and Gannon 1995). Using a numerical model, they found that higher soil
moisture values during daytime partition more energy into latent heat flux (QE in
(2.7)), reducing sensible heat flux and decreasing upslope flow strength. The main
effect of high soil moisture during nighttime, in contrast, is an increase in soil heat
conductivity so that heat diffuses upward from a deeper layer of soil (QG in (2.7))
to replace the energy lost by longwave radiation at the surface.
2 Diurnal Mountain Wind Systems 53

2.2.3 Role of Turbulence in Slope Flows

The bulk of the research on slope flows to date has been concerned with the mean
characteristics of the slope flows, such as their depths, strengths, wind profiles,
relationships to temperature structure, etc. The important role of turbulence in
producing these characteristics has received much less attention. In contrast to flat
terrain, where calm or very weak winds often accompany fair weather conditions,
marked up- and downslope flows usually occur over slopes on fair weather days.
Turbulent kinetic energy (TKE) in slope flows is always produced, as in neutrally
stratified turbulent shear layers, by the coupling between vertical wind shear that
develops in the mean slope flow profile and the momentum flux. However buoyancy
contributes too, either to produce turbulent energy, when the sensible heat flux is
upwards, or to suppress it, when the flux is downward. This turbulence is critical
to the vertical mixing of heat and momentum that governs the height of the jet,
the depth of the flow and the ultimate equilibrium temperature deficit (Cuxart
et al. 2011).
Turbulence in upslope flows is generally considered to be a more tractable
problem than turbulence in downslope flows. This comes from parallels with
the study of turbulence over flat, homogeneous terrain, where understanding of
turbulence has progressed much faster for convective boundary layers than for stable
boundary layers. Towers on slopes are generally too short to fully encompass the
rapidly growing upslope flow layer, however. Thus, unless additional observational
tools can be developed for measuring turbulence over deeper layers above slopes,
future improvements in understanding of upslope flows may come primarily from
modeling studies.
The continuous turbulence production by shear in the wind profile in fully
developed shallow downslope flows is expected to be a much more manageable
problem than turbulence in the deeper stable boundary layers that form over both
flat and complex terrain areas, where the turbulence is intermittent and discontin-
uous (Mahrt 1999; Van de Wiel et al. 2003). The shallow downslope boundary
layers are more accessible to tower measurements, but appropriate measurement
techniques and post-processing methods (e.g. filtering, wind component rotations,
and alignment to an appropriate reference frame) are still problematical. Turbulence
data are generally obtained from high-rate sampling of the three components of
the wind. Additional sampling of scalar variables (temperature, humidity, CO2
concentrations, etc.) on the same time scale can be used to correlate fluctuations of
the wind components and scalars to determine fluxes. Representative measurements
require that the upwind fetch be homogeneous with respect to sources and sinks of
scalars and surface properties. Following the collection of wind data on the slope,
the wind measurements are usually rotated to align the coordinate system to the
mean wind, which is generally parallel to the underlying ground (at least, near the
surface). Flaws in the typical double- or triple-rotation method used over flat terrain
are encountered in sloping terrain (see Finnigan et al. 2003), so that the “planar-
fit” approach developed by Wilczak et al. (2001) is considered preferable. But an
54 D. Zardi and C.D. Whiteman

appropriate post-processing also requires carefully designed filtering and averaging


procedures to extract turbulent quantities from the often non-stationary mean flows
encountered over slopes (de Franceschi and Zardi 2003; Weigel et al. 2007; de
Franceschi et al. 2009).
Few model simulations of turbulent quantities are yet available for downslope
flows since most of the RANS numerical models to date have been developed
for mesoscale simulations on much larger scales than the slope flows. The lack
of vertical resolution of the slope flows in these models, and uncertainty in the
applicability of the turbulence parameterizations have limited their utility. Recent
improvements in LES model formulations (also see Chap. 10), and corresponding
increases in computer power, are now leading to continuously improved simulations
as to grid resolution, numerical schemes and appropriateness of turbulence closures
(Skyllingstad 2003; Chow et al. 2006; Axelsen and Van Dop 2008, 2009; Serafin
and Zardi 2010a).

2.2.4 Upslope Flows

The radiative heating of the ground and the resulting upward turbulent sensible
heat fluxes, typically occurring in connection with upslope flows, increase rapidly
after sunrise, reaching magnitudes that are much higher than those associated with
downslope flows. As a consequence of increasing heat flux, upslope flows cannot
approach a steady state, but rather grow continuously with time in strength and
depth. Upslope flows are confined to a layer over the heated slope in the morning by
the highly stable air that formed above the slope overnight (the confining stability
can be especially strong over valley sidewalls, compared to slopes on isolated
mountainsides). Like the situation over flat ground, heating of the slopes causes
an unstable or convective boundary layer to grow upward into the remnants of
the nighttime stable layer. The daytime boundary layer over slopes, in contrast to
over flat ground, is a convective-advective boundary layer, with bent-over plumes
and convective elements moving up the slope. While the upslope flow grows in
depth and strengthens with time after sunrise, the volume flux of the flow increases
with upslope distance as air is entrained into the flow at its upper boundary. The
growing depth of the convective-advective upslope flow layer eventually allows the
convection to break through the stable layer. Following the breakup of the confining
stable layer, one might expect the convective currents to simply rise vertically from
the heated ground. In fact, the flow continues to rise up the slope. This is apparently
caused by a horizontal pressure gradient associated with the inclined superadiabatic
sub-layer that forms over the slope. The upslope flows are often disturbed, however,
by convective mixing that brings down the generally stronger background winds
from higher layers of the atmosphere.
Our understanding of upslope flows has suffered from the lack of observational
tools for making continuous measurements of the mean and turbulent profiles
2 Diurnal Mountain Wind Systems 55

through the full depth of the growing upslope flow layer. Towers are generally
not tall enough to penetrate this layer and the non-stationarity and inhomogeneity
of the flows has made observations particularly difficult. Some field observations
(Kuwagata and Kondo 1989; Van Gorsel et al. 2003a; Reuten et al. 2005; Geerts
et al. 2008; De Wekker 2008), laboratory simulations (Hunt et al. 2003; Princevac
and Fernando 2007; Reuten et al. 2007; Reuten 2008) and model simulations
(Kuwagata and Kondo 1989; Schumann 1990; Atkinson and Shahub 1994; Axelsen
and van Dop 2008) have improved this situation. Two interesting observations on
upslope flows came from a field experiment over an isolated slope northeast of
Vancouver, BC, in 2001. The observed depression of a mixed layer at the base of the
mountain slope is thought to have been caused by subsidence in the return branch of
the upslope circulation (De Wekker 2008; Serafin and Zardi 2010a). On a different
day, scanning Doppler lidar observations up the slope appeared to show a closed
slope flow circulation within the inclined convective boundary layer (Reuten et al.
2005), although these findings are controversial.
Schumann (1990) simulated upslope flows over an infinite, uniform, heated,
sloping surface for a range of slope angles and surface roughness. The simulations
were revolutionary, providing new insight into the role of turbulence in upslope
flows and the utility of LES modeling to improve understanding of geophysical
phenomena on small length scales. Schumann’s work, as a parametric study, was
limited to a constant ambient stability (3 K km1 ) and sensible heating rate (about
100 Wm2 ), with a quiescent atmosphere above the slope flow layer. The simulated
steady-state profiles of upslope wind speed and excess temperature (with respect to
the unperturbed thermal structure) for slopes of different inclination angle from 2ı to
30ı are shown in Fig. 2.6, where the axis variables are normalized by a scale height
H of 58 m, a wind speed scale v of 0.58 m s1 , and a temperature scale  of 0.17 K.
A dry adiabatic lapse rate curve is provided on the temperature profile plot for
comparison with the simulated profiles. Several interesting features are seen in the
simulations. The steady state was attained by a balance between the buoyancy term
and surface friction in (2.2) and between the surface heating and upslope advection
of heat against the mean temperature gradient in (2.4), just like in Prandtl’s model.
As expected, the temperature excess present at the ground decayed with height.
Temperature profiles, except those over the steepest slopes, were characterized
by a near-adiabatic temperature decrease with altitude, indicating good mixing in
the turbulent flows. Strong local minima occurred in the velocity and temperature
profiles at the upper edge of the mixed layer. Upslope flow depths and volume fluxes
were greatest over low-angle slopes. Shallow and weak upslope flows occurred on
steeper slopes. Wind maxima occurred within about 50 m of the surface over slopes
of all inclination angles. The turbulence kinetic energy profile in the upslope flow
was generated by vertical shear, with high positive values of shear just above the
ground but below the height of the jet maximum and high negative values of shear
above the jet maximum. This resulted in a slope-normal transport of momentum
both upward and downward from the jet level (see Fig. 2.2).
56 D. Zardi and C.D. Whiteman

Fig. 2.6 Normalized upslope mean wind component and temperature profiles over heated slopes,
with indicated slope angles from 2ı to 30ı , as simulated using an LES model. H, v* and
 * are respectively the scaling values adopted to normalize the slope-normal coordinate n,
the wind component <u> and the temperature <T>. Error bars indicate standard deviations
(From Schumann 1990, © Quarterly Journal of the Royal Meteorological Society. Reprinted with
permission)

LES simulations such as this have the advantage of being able to produce profiles
of all turbulence quantities above the slope and to determine all contributions to
the turbulence kinetic energy budget. One caveat, however, is that subgrid scale
TKE contributions are parameterized. The contributions of these terms near the
surface, where eddies are small, have not been verified. It is important to note that
detailed observations are, so far, unavailable to properly test LES simulations of the
upslope boundary layer. The idealized nature of the LES simulations should also be
emphasized. In real, three-dimensional topography, slopes are very inhomogeneous
in terms of slope angle, roughness, vegetative cover, sensible heat flux, non-
planarity, etc. and the flows are distinctly non-stationary. In contrast, the LES
simulations were conducted on uniform slopes and the equations were integrated
until steady-state solutions were reached. The simulations certainly have interesting
implications for upslope flows on sidewalls with varying inclination angles. In these
cases, along-slope convergences or divergences of mass in the upslope layer will
cause detrainment and entrainment from the slope flow layer (see, also, Vergeiner
and Dreiseitl 1987), with implications for the air mass over the valley center. There
are ample opportunities to extend Schumann’s (1990) initial work.
2 Diurnal Mountain Wind Systems 57

2.2.5 Downslope Flows

Downslope flows on isolated mountainsides and on valley sidewalls have been stud-
ied extensively over the years. In this section we will summarize current knowledge
about the mean and turbulence structure of downslope flows using observations and
simulations, concentrating on research conducted since about 1990.
For downslope flows, the terms of importance in the governing equations
(2.2–2.4), have been evaluated mainly with hydraulic models (Manins and Sawford
1979b; Horst and Doran 1988; Papadopoulos et al. 1997; Haiden and Whiteman
2005). Steady-state flow on mid-latitude slopes is typically an “equilibrium flow”
balance between buoyancy and turbulent momentum flux divergence (i.e., friction),
in agreement with the local equilibrium assumption made in Prandtl’s (1942)
analytical model. The surface stress is thought to contribute more friction than
the entrainment of slower moving air from above the downslope layer. The
balance can sometimes include a significant along-slope advective contribution in
the momentum balance near the height of the wind maximum, where shear changes
sign (Skyllingstad 2003), in agreement with Horst and Doran’s (1988) results. The
slope-perpendicular momentum equation, since the RHS of (2.3) is near zero, is
a balance between the two advection terms. The thermodynamic energy equation,
(2.4), is a balance between the radiative and sensible heat flux divergences and
temperature advection.
Horst and Doran (1988) found that the turbulence structure over slopes was
consistent with local shear production of turbulence. Turbulence, which is generated
primarily by vertical wind shear and destroyed by viscous dissipation, is critical for
the vertical mixing of momentum and buoyancy in the mean flow. The jet-like wind
profile associated with downslope flows produces a turbulence structure (Fig. 2.2)
that is quite different from that over flat terrain. Vertical shear is positive below the
jet, disappears at the height of the jet and is negative above the jet. This vertical
distribution of wind shear produces TKE profiles with a maximum near the surface,
and a local minimum at the height of the wind maximum. Above that height, TKE
is highly dependent on the speed and direction of the ambient wind. Turbulence
above the wind maximum is decoupled from the surface so that normal surface
similarity theory is not expected to be applicable above the jet. Vertical velocity
variance profiles depend on cooling rates and external wind speeds (Coulter and
Martin 1996).
In the remainder of this section we will highlight the important influences of
slope angle, ambient stability, external flows and other disturbances on downslope
flows. We begin with slope angle.

2.2.5.1 Impact of Slope Angle on Downslope Flows

If a continuous source of cold air were introduced at the top of a slope (as
in Burkholder et al. 2009), the downslope flow would be stronger on steeper
58 D. Zardi and C.D. Whiteman

slopes where the along-slope component of the gravity vector is maximized. But,
downslope flows in valleys and on mountainsides are not generally fed by a source
of cold air at the top of the slope. Rather, cold air forms in place on the slope through
a downward flux of sensible heat to the underlying radiatively cooled surface. The
air moves down the slope only as long as it is negatively buoyant with respect to
the ambient environment. The air in the flow is heated adiabatically at 9.8ı C km1
as it moves down the slope. To continue its descent it must be continuously cooled
by a downward sensible heat flux to the cold underlying surface. The faster the
flow, the lower the rate of cooling per unit mass of flow. On low angle slopes, the
air travels a greater horizontal distance along a slope to descend the same vertical
distance compared to a steeper slope. The air loses heat through downward sensible
heat flux along its entire trajectory, which is longer on a shallow slope than on a
steep slope.
Slope flows will not form on horizontal surfaces, where radiative loss to the
sky and downward sensible heat flux occur but the along-surface component of
the gravity vector is zero. At the other limit, a vertical surface will have a strong
component of the gravity vector along the slope, but a vertical surface will not
radiate well to the sky and, because of the small distance that the air would travel
along the slope to lose a unit of altitude, the rate of cooling of the flow to the
underlying surface would be small. The strongest downslope flows will thus develop
on slopes of intermediate angle where there is a combination of good outgoing
longwave radiative loss to the sky that drives a slope-normal sensible heat flux and
a component of the gravity vector acting along the slope.
Observations (Table 2.1) on simple slopes show that downslope flows on low-
angle slopes are deeper and stronger than those on steeper slopes. Slope flows are
often strongest in the early evening when the ambient stability is weak. For slopes
on valley sidewalls, this is the time before the overlying valley flows strengthen to
produce a cross-slope shear that disturbs the slope flows. The temperature deficit
layer over the slope (i.e., inversion depth) often has a depth corresponding to
about 5% of the vertical drop from the top of the slope and the height of the
wind maximum is usually within the temperature inversion layer, often at heights
that are 30–60% of the inversion depth. The downslope flow itself usually extends
much higher than the height of the wind maximum or the height of the inversion
top because of the entrainment of ambient air by the downslope flow (Princevac
et al. 2005). But, this height is much more variable than the inversion height and
is very dependent on the speed and direction of the ambient flow above the slope
flow layer. In general, the downslope flow extends to heights of several times the
inversion depth.
The speed and depth of downslope flows (and, thus mass or volume flux)
increase with downslope distance. This has been seen in experiments on Rattlesnake
Mountain, Washington (Horst and Doran 1986), on a low angle slope in Utah’s
Salt Lake Valley (Haiden and Whiteman 2005; Whiteman and Zhong 2008), and on
Greece’s Mt. Hymettos (Amanatidis et al. 1992; Papadopoulos and Helmis 1999).
An example illustrating this increase in depth and speed is shown in Fig. 2.7.
Downslope flows accelerate with downslope distance while increasing their depth
2 Diurnal Mountain Wind Systems 59

Table 2.1 Observations of downslope flow on slopes of different inclination


Angle Inversion Inversion Max speed Height of max
(ı ) depth (m) strength (K) (m s1 ) speed (m) References
21 4–8 5 1–2 0.6–2 Horst and Doran (1986)
9 16 4 1–2 6 Papadopoulos et al. (1997)
6.5 40–50 10 1.5 25 Horst and Doran (1986)
4 50 3 3.5–4 40 Monti et al. (2002)
1.6 25 7 4–6 15–20 Whiteman and Zhong (2008),
Haiden and Whiteman (2005)

1700
a b c
1650
Height (m MSL)

1600

1550

1500

1450
290 295 300 305 -3 -2 -1 0 1 2 3 4 5 -3 -2 -1 0 1 2 3 4 5
-1)
Potential temp (K) Along-slope wind (m s Cross-slope wind (m s-1)

Fig. 2.7 Vertical profiles of (a) potential temperature, (b) downslope wind speed component
(m s–1 ) and (c) cross-slope wind speed component (m s1 ) at 2217 MST 8 October 2000 from
four tethersondes running down a 1.6ı slope at the foot of the Oquirrh Mountains at various
distances from the top of the slope, namely 4,200 m (light gray), 5,100 m (dark gray), and 6,200 m
(black) (From Whiteman and Zhong 2008. © American Meteorological Society. Reprinted with
permission)

as additional mass is entrained into the flowing layer at its upper boundary. Some of
the simple steady-state analytical slope flow models (see, e.g., those reviewed
by Barry 2008) are in conflict with observations and with predictions from full
physics models. Because the models are often run using the same sensible heat
flux forcing on slopes of all angles, they have led to the mistaken concept that
downslope flows on upper steeper slopes are stronger and deeper than flows on
lower slopes, and will overrun the flows on lower slopes or will travel down the
slope as frontal disturbances. While such frontal propagation episodes can occur in
special circumstances (e.g., when the upper slopes go first into shadow while the
lower slopes are still in sunlight), they appear to be atypical.
60 D. Zardi and C.D. Whiteman

RANS and LES models are in agreement with field data, predicting weaker
downslope flows (i.e. displaying both smaller maximum jet speed and smaller
mass flux) on steeper slopes (Skyllingstad 2003; Zhong and Whiteman 2008;
Axelsen and Van Dop 2009). The effect of stepwise changes in slope angle partway
down a mountainside were investigated using both RANS and LES models by Smith
and Skyllingstad (2005) for an 11.6ı slope that changed suddenly to a 1.6ı slope
at a lower elevation. The slope flow depth increased rapidly over a distance of
500 m at the slope angle juncture and the height of the maximum velocity moved
upward about 5–10 m. Below the juncture, a new equilibrium was established with
a decreased rate of growth in slope flow depth and more rapid cooling of the
near-surface temperatures. Near-surface downslope velocity decreased, producing
a distinct detached and elevated jet structure. Potential energy generated at the top
of the slope by continued cooling was transported downslope and converted into
kinetic energy at the slope base.

2.2.5.2 Impact of Ambient Stability on Downslope Flows

The effect of ambient stability on downslope flows is now well known from
both simulations on simple slopes and from field experiments. The downslope
flow must adjust to any changes in ambient stability of the atmosphere adjacent
to the slope. If the flow, for example, encounters a sudden increase in ambient
stability, it must be cooled at a higher rate to be advected along the surface at
the same speed. Downslope flow intensity is thus inversely proportional to ambient
stability (Rao and Snodgrass 1981; King et al. 1987; Ye et al. 1990; Helmis and
Papadopoulos 1996; Zhong and Whiteman 2008). The nighttime buildup in ambient
stability within valleys reduces downslope flow speed and may even suppress
downslope currents. In enclosed basins, downslope flows die in the late evening as
ambient stability increases and the atmosphere becomes increasingly horizontally
stratified and quiescent (Clements et al. 2003; Whiteman et al. 2004a, c; Steinacker
et al. 2007).

2.2.5.3 Impact of Background Flow on Downslope Flows

Downslope flows, which are usually rather weak and shallow air currents, are quite
sensitive to disturbance by background or ambient flows (Fitzjarrald 1984; Barr and
Orgill 1989; Doran 1991) that may oppose or follow the downslope flows or may
have a significant cross-slope component. Background flows can originate as valley
flows, larger mesoscale circulations, synoptic flows, sea or lake breezes, seiches,
gravity waves, wakes from flow over ridges, etc. The effects of these differing
phenomena on downslope flows are difficult to separate using field data, but it is
generally understood that background flows with strengths of only 2–3 m s1 can
have far reaching effects on the structure of the downslope flows and even their
existence.
2 Diurnal Mountain Wind Systems 61

Downslope flows on valley sidewalls are routinely subject to the development


of overlying valley flows. These valley flows, which have jet-like vertical and
horizontal profiles within the valley cross-section, produce horizontal shears above
the slope flows that vary with elevation. Because these flows are oriented along the
valley axis, while the slope flows are perpendicular to the valley axis, they produce a
strong directional shear through the shallow slope flow layer. When the valley winds
get strong enough, horizontal and vertical shears and the associated turbulence can
greatly modify the downslope flows or remove them entirely from valley sidewalls
(Doran et al. 1990).
As cross-slope winds increase, slope flows deepen and behave more like a
weakly stratified sheared boundary layer (Skyllingstad 2003). The cross-slope flow
produces an extra source of mixing from the additional vertical shear, which weak-
ens the vertical stratification. Because the velocity near the surface becomes lower
there is less loss of momentum to surface drag. But, as cross-slope flow increases,
the downslope jet becomes weaker and deeper and is eventually overpowered by the
cross-slope-generated turbulence.

2.2.5.4 The Effects of Slope Inhomogeneity on Downslope Flows

Vergeiner and Dreiseitl (1987) pointed out that real mountain slopes are much
more complicated than those assumed in existing models. Since, as emphasized
above, radiative and turbulent fluxes vary in time and space, upslope and downslope
flows must vary continuously in space and time in response to small-scale slope
inhomogeneities. This variation makes it difficult to find representative measure-
ment locations on slopes (e.g., sites with uniform slope angles and surface cover)
to compare with idealized models. Slopes on mountain sides vary not only in
topography and surface cover, but they are also subject to time- and space-varying
interactions with larger scale flows and to varying ambient stabilities.
Only a few observational and modeling studies have investigated the response
of slope flows to slope inhomogeneities. As mentioned, several studies have
investigated the effects of changes in slope angle. Others have looked into the
effects on downslope flows of forests (e.g., Lee and Hu 2002; van Gorsel 2003);
differentially cooled sloping surfaces (Shapiro and Fedorovich 2007; Burkholder
et al. 2009) and sunset time variations on complex slopes (Papadopoulos and
Helmis 1999). Haiden and Whiteman’s (2005) investigation of downslope flows
on a visually rather uniform slope at the foot of Utah’s Oquirrh Mountains found
imbalances in both the momentum and heat budget equations that were attributed to
small scale convexities and concavities on the slope that channeled the downslope
currents. Monti et al. (2002) observed multi-layered stability above a slope that
appeared to come from elevated downslope currents. They suggested that slope
flows can develop features similar to hydraulic flows, such as jump conditions or
turbulent transitions.
62 D. Zardi and C.D. Whiteman

2.2.5.5 Drainage Flow Oscillations

Oscillations in drainage (i.e., downslope and down-valley) flow strengths with


periodicities between about 5 and 90 min have been frequently observed. They can
be produced by a variety of different mechanisms, including interactions between
valley flows and tributary flows (Porch et al. 1989), variation of cold air sources
from upstream (Allwine et al. 1992), interactions with mountain waves (Stone and
Hoard 1989; Poulos et al. 2000, 2007), modulation of slope and valley flows by
interactions of synoptic and mesoscale flows with surrounding topography (Mahrt
and Larsen 1982; Doran et al. 1990; De Wekker 2002), interactions between slope
flows and sea breezes (Bastin and Drobinski 2005), meandering motions in stratified
flows and, in the case of downslope flows, by internal waves impinging normal to
the slope (Princevac et al. 2008) and by overshooting and oscillations caused by the
resulting buoyancy disequilibrium (Fleagle 1950; Doran and Horst 1981; McNider
1982; Helmis and Papadopoulos 1996; van Gorsel et al. 2003b; Chemel et al. 2009;
Viana et al. 2010). Additionally, cold air forming over plateaus or other elevated
terrain source areas sometimes cascades down the slopes at intervals as cold air
avalanches (Küttner 1949). Atmospheric analogs to the seiches that form in ocean
and lake basins have also been observed in basin temperature inversions, where it
is common to see smoke plumes from chimneys and open fires that exhibit weak
oscillatory motions in which the plume is carried back and forth quasi-horizontally
relative to its source. The back-and-forth sloshing, if produced by seiches, should
have a characteristic frequency that is dependent on basin topography, atmospheric
stability and wind speed. Seiche-like oscillations with a periodicity of about 15 min
were recently documented on the floor of Arizona’s Meteor Crater (Whiteman et al.
2008), and a spectral element model (Fritts et al. 2010) has produced seiches in
idealized topography similar to the crater.

2.2.5.6 Glacier Winds

Slope flows are generally diurnally varying flows, but over a snow surface where
the surface energy budget is consistently negative the resulting downward sensible
heat flux from the adjacent atmosphere produces a shallow cold air layer with a
continuous downslope flow. Over cold, high albedo glaciers or snow surfaces, this
downslope flow has been called a glacier wind. A special feature of downslope
winds over snow surfaces is the constrained upper limit value of surface tempera-
ture, 0ı C. In daytime during the melt season when ambient air temperatures undergo
their normal diurnal variation, this upper limit of surface temperature results in a
temperature inversion above the melting surface whose maximum strength occurs
in mid- to late-afternoon in the ablation zone, and at night at higher elevations where
radiative cooling of the surface is more important (Oerlemans 1998).
Glacier winds, which occur on a variety of scales, have received increased
attention in the last 20 years. Oerlemans et al. (1999) found temperature inversions
of about 20 m depth and 8ı C strength over a large maritime ice cap in Iceland.
2 Diurnal Mountain Wind Systems 63

Near-steady-state glacier winds were nearly always present, except for periods with
traveling intense storms, suggesting that glacier winds are the key factor shaping
the microclimate of glaciers, at least in summer. Maximum downslope wind speeds
of 3–10 m s1 were found at heights from several to a few tens of meters above
the surface, although the winds often extended to heights exceeding 100 m. The
momentum balance that produces the glacier wind appears to be a balance between
buoyancy force and friction, as expressed in the Prandtl (1942) and Oerlemans
and Grisogono (2002) models. In contrast, for a small, mid-latitude glacier in
the Alps the buoyancy force was balanced by both friction and the mesoscale
pressure gradient that drives the valley wind above the downslope layer (Van den
Broeke 1997a, b).

2.3 The Valley Wind System

Diurnal valley winds are thermally driven winds that blow along the axis of a
valley, with up-valley flows during daytime and down-valley flows during nighttime.
Valley winds are the lower branch of a closed circulation that arises when air in
a valley is colder or warmer than air that is farther down-valley or over the adjacent
plain at the same altitude. Unlike slope winds, valley winds are not primarily a
function of the slope of the underlying valley floor. In fact, they have been observed
even in valleys with horizontal floors (Egger 1990; Rampanelli et al. 2004).
Instead, they depend on other geometrical factors, such as the shape and aspect
of the valley cross-section and their along-valley variations, including tributaries
(Steinacker 1984), as discussed below. The diurnal reversal of the flow requires
a larger daily temperature range within the valley than over the adjacent plain
(Nickus and Vergeiner 1984; Vergeiner and Dreiseitl 1987). Horizontal pressure
gradients that develop as a function of height between air columns with different
vertical temperature structures over the valley and the adjacent plain drive the valley
wind (Fig. 2.8). These pressure gradients can be observed directly (Khodayar et al.
2008; Cogliati and Mazzeo 2005). Along the Inn Valley, for example, thermally
induced pressure differences of up to 5 hPa over a distance of 100 km have been
observed (Vergeiner and Dreiseitl 1987). As a consequence of the changing pressure
distribution in the vertical, the vertical profile of along valley winds undergoes a
cyclic evolution during the day (Gross 1990). Valley winds evolve gradually over
the daily cycle and produce weak to moderate wind speeds. Peak wind speeds over
the valley center are frequently in the range from 3 to 10 m s1 .
Ideally, the upper branch of the closed circulation consists of an elevated
horizontal flow running in the opposite direction from the winds in the valley below.
Because the upper branch is unconfined by topography and thus broader in horizon-
tal extent, it is generally quite weak and can be obscured by stronger synoptic-scale
flows. The existence of these return flows or compensation currents was debated at
the early stage of the scientific investigation of valley winds (cf. Wagner 1938) but
is today considered a well-documented feature (McGowan 2004; Zängl and Vogt
64 D. Zardi and C.D. Whiteman

Fig. 2.8 Idealized picture of the development of daytime up-valley winds (upper panel) and
nighttime down-valley winds (lower panel) in a valley-plain system with a horizontal floor. The red
and blue curves are vertical profiles of the horizontal valley wind component at a location close to
the valley inlet. Two columns of air are shown – one over the valley floor and one above the plain.
Red and blue sections of the columns indicate layers where potential temperature is relatively warm
(W) or cold (C). The free atmosphere is assumed to be unperturbed by the daily cycle at the tops
of the columns (Adapted from Whiteman 2000)

2006). Figure 2.9, for example, compares modeled temperature profiles within a
valley and over the adjacent plain to illustrate the reversing temperature gradients
with height that drive the elevated return circulations (from Rampanelli et al. 2004).
Investigations by Buettner and Thyer (1966) in valleys that radiate out from the
isolated volcano of Mt. Rainier, Washington (USA), found that the upper branches
of the valley circulations there could even be observed within the terrain-confined
2 Diurnal Mountain Wind Systems 65

Fig. 2.9 (a) Comparison of different vertical profiles of potential temperature in an idealized
valley-plain system (sketched in the insert). The profiles are taken along the valley axis (solid
line) and over the plain (line of circles) along the dashed vertical lines indicated in the insert at
t D 6 h after simulated sunrise. (b) Difference between the two profiles at various heights. The
horizontal dashed line at 1,000 m indicates the sidewall ridge height (Adapted from Rampanelli
et al. 2004. © American Meteorological Society. Reprinted with permission)

upper altitudes of the valleys. They coined the term anti-winds for these confined
currents, but the deeper currents above the ridgetops are more frequently called
return or compensating flows.
The evening reversal of the daytime up-valley wind begins in late afternoon or
early evening when the daytime wind system begins to lose strength. The gradual
loss of strength, taking place over several hours, is caused by a decrease in the
along-valley pressure gradient as temperatures along the valley axis are equilibrated
by advection and as the sensible heat flux that maintained the temperature gradient
decreases. Once the sensible heat flux reverses on the sidewalls or in the shadowed
parts of the valley, downslope flows begin, transferring the cooling experienced in
a shallow layer above the slopes to the valley atmosphere through compensatory
rising and cooling motions. Radiative processes and turbulent sensible heat flux
divergences also play a role. As the valley atmosphere cools, the horizontal pressure
66 D. Zardi and C.D. Whiteman

gradient between the valley and adjacent plain reverses and a down-valley flow
begins. The valley wind reversal lags that of the slope flows because the large
mass of the valley atmosphere must be cooled before the pressure gradient can
reverse (Vergeiner 1987). The reversal typically begins at the floor of the valley,
where cooling is most intense, with the formation of a temperature inversion. As
the cooling progresses upwards, the down-valley flow deepens and strengthens.
Within the nighttime inversion, the coldest temperatures are at the valley floor, and
temperatures increase with height. This results in relatively warmer temperatures on
the valley sidewalls than on the valley floor, i.e. a warm slope zone on the valley
sidewalls (Koch 1961).
The morning reversal of the down-valley wind, which occurs after sunrise,
similarly requires a reversal of the along-valley pressure gradient as the valley
atmosphere warms. The reversal occurs several hours after upslope flows are
initiated when the heat transfer processes become effective in warming the entire
valley atmosphere. Further details on the heat transfer mechanisms are provided
in the next section that focuses on interactions between the slope and valley wind
circulations.
The warmer temperatures during daytime and colder temperatures during night-
time in the valley atmosphere compared to the atmosphere over the adjacent plain
(or farther down the valley) are mainly caused by two factors. First, under low
background wind conditions the valley atmosphere is protected by the surrounding
topography from mass and heat exchange with the atmosphere above the valley so
that heating will be more effectively concentrated in the valley during daytime and
cooling will be enhanced in the valley during nighttime. A second, more important,
factor has been termed the “area-height relationship” or “topographic amplification
factor”, TAF. The concept (see Whiteman 1990 or 2000 for fuller discussions), is
that the daily temperature range in the valley will be amplified by the smaller mass
of air that is heated or cooled within the confined valley volume than within the
larger volume of the same depth and surface area at its top over the adjacent plain.
The higher mean temperature inside the valley during daytime causes a pressure
gradient to develop between valley and plain that drives an up-valley wind. The
lower mean temperature and higher pressure in the valley during nighttime drive a
down-valley wind.
The mechanisms by which valley geometry might affect cooling and warming
of the valley atmosphere have been explored by various authors. Wagner (1932)
was the first to express the TAF concept. Steinacker (1984) carried the concept
further, computing the distribution of drainage area as a function of height in the Inn
Valley and comparing it to that over the adjacent plain in southern Germany. McKee
and O’Neal (1989) then suggested that along-valley pressure gradient differences
produced by changes in terrain cross-section along a valley’s axis could explain
along-valley wind speed variations. While the intuitive TAF concept, which focuses
on valley geometry, helps to conceptualize the important thermodynamics leading
to up- and down-valley flows, the simple concept involves inherent assumptions that
have made it difficult to apply in practice.
2 Diurnal Mountain Wind Systems 67

There are inherent difficulties in evaluating precisely both the valley volume,
and the heat fluxes across its boundaries. As to the volume, its top boundary is
difficult to define in practice. An intuitive choice would be to locate it at crest
height, yet in real valleys the crest height varies with down-valley distance and
differs between the opposing sidewalls. Further, it appears that the volume of the
valley tributaries has to be taken into account (Steinacker 1984). As to the fluxes,
Gauss’s divergence theorem can be used with a volume integral of (2.7) to determine
the rate of change of the volume-average potential temperature in the valley. This
rate of change is driven by the surface integral of the advective, turbulent and
radiative heat fluxes across the surfaces bounding the volume, including the valley
floor, sidewalls and the top boundary. However the flux distributions along the
volume boundaries are generally not well known (e.g. on the valley sidewall slopes:
see Sect. 2.2). Yet boundary conditions allow some simplifications: for instance,
advection through the ground surface vanishes, since wind speed vanishes there.
However this is not the case on the top surface, where heat advection through
slope flows on the sidewalls and compensating subsidence at the valley core will
play a crucial role in the daytime heating of the valley atmosphere (Rampanelli
et al. 2004; Weigel et al. 2007; Schmidli and Rotunno 2010; Serafin and Zardi
2010b, c). Because the top boundary is not a solid surface, heat gain or loss in the
volume is not normally confined to the volume. This is especially the case during
daytime, when convective boundary layers can grow above ridge height. Similarly,
the down-and up-valley ends of the volume are open: once an along-valley flow
begins, advection acts to reduce temperature differences between valley and plain,
or between segments of a valley. Further, the TAF concept in its simplest form does
not consider that the rate of heat gain or loss may vary with height in the valley or
plains volumes, and assumes that rates of heat gain and loss are equal in the two
volumes. In fact, the surface energy budgets and heat transfer processes may be
quite different in the two volumes. Heat transfer processes in the valley volume, for
example, involve heat transfer over the slopes associated with up- and downslope
flows, compensatory rising and sinking motions over the valley center, and radiative
transfer processes that include back-radiation from the sidewalls (cf. Haiden 1998).
Further information on these and other factors that lead to variations in the thermally
driven pressure gradient along a valley including the partitioning between sensible
and latent heat fluxes in the surface energy budget and the confinement of heated
or cooled air within a valley can be found in publications by Zängl et al. (2001),
Rampanelli (2004), Zängl (2004), Zängl and Vogt (2006), Serafin (2006), Rucker
et al. (2008), Schmidli and Rotunno (2010), and Serafin and Zardi (2010a, b, c).
In the early evening, down-valley flows often converge into sub-basins or build
up cold-air pools or lakes behind terrain constrictions (cf. Bodine et al. 2009). In
general, though, if the valley does not have major constrictions along its length or
major changes in surface microclimates, the nighttime down-valley flows accelerate
and increase in depth with down-valley distance, resulting in an along-valley mass
or volume flux divergence. Through the law of mass conservation, the existence
of along-valley divergence requires the import of potentially warmer air into the
valley from aloft or from tributaries. Interestingly, recent field measurements and
68 D. Zardi and C.D. Whiteman

Fig. 2.10 Average up- and down-valley wind profiles from four valleys using all daytime and
nighttime data from the locations indicated (Ekhart 1944) (© E. Schweizerbart Science Publishers
www.borntraeger.cramer.de. Used with permission)

model simulations have also shown a divergence of along-valley mass flux in


the daytime up-valley flows for the Inn Valley (Vergeiner 1983; Freytag 1988;
Zängl 2004), the Kali Gandaki Valley (Egger et al. 2000; Zängl et al. 2001) and
in the Wipp Valley (Rucker et al. 2008). This along-valley mass flux divergence
is somewhat counterintuitive, since one might expect an up-valley flow in the
main valley to dissipate with up-valley distance as the flows turn up the slopes
and tributaries. Various mechanisms have been invoked to explain the mass flux
divergence, including compensating subsidence over the main valley, transition to
supercritical flow in the widening part of a valley (possibly favored by gravity waves
produced by protruding ridges), and intra-valley change in the horizontal pressure
gradient induced by differential heating rates.
In different valleys (Fig. 2.10), the vertical structure, strength and duration of
up- and down-valley flows differ, depending on climatic and/or local factors (Ekhart
1948; Löffler-Mang et al. 1997). One cannot assume that valleys with strong up-
valley flows will also have strong down-valley flows. In the Kali Gandaki valley
of Nepal, the daytime up-valley winds are much stronger than the nocturnal down-
valley winds (Egger et al. 2000, 2002). In California’s San Joaquin Valley, the up-
valley flows can persist all night (Zhong et al. 2004). In other cases, such as in
Chile’s Elqui Valley, nocturnal down-valley winds are not observed at all (Kalthoff
et al. 2002), while down-valley winds may persist all day in snow-covered valleys
(Whiteman 1990).
Down-valley winds are often enhanced where valleys issue onto the adjacent
plains or into wider valleys. Drainage flows in the valley decrease in depth,
2 Diurnal Mountain Wind Systems 69

accelerate, and fan out onto the plain at the valley exit, producing a valley exit jet.
The valley exit jet at the mouth of Austria’s Inn Valley reaches speeds of 13 m s1
at heights only 200 m above ground (Pamperin and Stilke 1985; Zängl 2004). Valley
exit jets have also been observed at the mouths of tributary canyons that issue from
the Wasatch Mountains into the wide Salt Lake Valley (Banta et al. 1995, 2004;
Fast and Darby 2004; Darby and Banta 2006; Darby et al. 2006) and at canyons that
flow onto the plains south of Boulder, Colorado (Coulter and Gudiksen 1995; Doran
1996; Varvayanni et al. 1997). They have also been observed at the exit of valleys
issuing onto the Snake River Plain in eastern Idaho (Stewart et al. 2002) and at the
exit of the Lech Valley into the Bavarian Plain (Spengler et al. 2009). Valley exit jets,
which are often a source of non-polluted or relatively clean air, can have important
air pollution consequences (Darby et al. 2006) and can bring nighttime relief from
high temperatures at canyon exits. The outflows of valley exit jets can be modulated
by larger-scale influences including the thermally driven flows beyond the valley
exit (Darby and Banta 2006).

2.3.1 The Effects of Boundary Layer Processes


and Turbulence on Valley Flows

Only recently have field research programs investigated the impact of turbulence
on valley winds (Rao and Schaub 1990; Poulos et al. 2002; Rotach et al. 2004;
Rotach and Zardi 2007). Turbulence measurements have been made primarily from
instrumented towers using eddy correlation techniques, although some airborne
measurements were made in the Riviera Valley during the Mesoscale Alpine
Programme (Weigel et al. 2006, 2007; Rotach and Zardi 2007). Many good methods
and techniques have been developed for processing and analyzing turbulence
measurements from towers (see, e.g., Kaimal and Finnigan 1994), but airborne
turbulence measurements are more difficult to make and analyze. Wind velocity
components are retrieved as a difference between absolute airplane motion with
respect to a fixed frame of reference and airspeed as indicated by the aircraft
(Crawford and Dobosy 1992). These two terms are at least an order of magnitude
larger than the difference between them. Thus, high accuracy in the wind velocity
components requires quite high accuracy in the original measurements. Airborne
measurements are, nonetheless, the most promising tool for analysis of turbulence
properties far from the surface in the middle of valley volumes (Rotach et al. 2004;
Weigel et al. 2007).
The dynamics of diurnal valley winds results from the combination of slope
flows on the sidewalls and boundary layer processes on and above the valley floor,
including such factors as free convection, shear, and compensatory subsidence
in the valley core atmosphere. As a consequence, turbulent processes associated
with valley winds occur on different space scales (height above the slope, slope
length, boundary layer depth over the valley floor, valley width) and time scales
(hour, morning, afternoon, daytime, nighttime). As a consequence it is difficult to
70 D. Zardi and C.D. Whiteman

ascertain their individual roles and interactions. A high site-to-site variability must
be considered in designing measurement strategies (Doran et al. 1989). Even the
application of standard turbulence analysis procedures, such as eddy covariance,
requires consideration of site-specific characteristics. (Kaimal and Finnigan 1994;
de Franceschi and Zardi 2003; Hiller et al. 2008; de Franceschi et al. 2009).
Over a homogeneous plain under fair weather conditions with weak synoptic
winds, shear will be minimized and convection will be the dominant process
producing turbulence. Under the same synoptic conditions, thermally driven flows
will develop in valleys, and the dominant process leading to turbulence will be shear
production, as turbulence in valleys is rarely caused by convection alone (Weigel
et al. 2007; de Franceschi et al. 2009).
Some observations suggest that the classical Monin-Obukhov surface layer
scaling for velocity and temperature variances holds in the middle of a wide valley
with a flat floor (Moraes et al. 2004; de Franceschi 2004; de Franceschi et al. 2009).
However, at other locations (e.g. sloping sidewalls) scaling approaches that are
valid for flat horizontally homogeneous terrain need to be extended and/or modified
(Andretta et al. 2001, 2002; van Gorsel et al. 2003a; Grisogono et al. 2007; Catalano
and Moeng 2010).
In Switzerland’s Riviera Valley during daytime in summer, the vertical TKE
distribution throughout the valley core atmosphere scales with the convective
velocity scale
 1=3
g 0 0
w D w  s zi (2.8)
0

defined, as usual, with gravity acceleration g, a temperature reference value  0 ,


the convective boundary layer (CBL) height zi , and the surface temperature flux
w0  0 s , provided the latter is determined at the slope rather than at the valley center
(Weigel and Rotach 2004; Weigel 2005; Weigel et al. 2006, 2007; Rotach et al.
2008). Despite highly convective conditions, it is mainly shear (as opposed to
buoyancy) that is responsible for TKE production. However the maximum of shear
production occurs not at the valley surface, but rather at higher levels above the
valley. This suggests that wind shear is mainly produced by the interaction of the
up-valley wind with the background winds aloft, rather than by surface friction. As a
consequence shear can be related to the wind strength in the core of the valley wind
and investigators have found that the scaling velocity determined from the surface
heat flux on the slope is closely related to the valley wind strength.
Surface fluxes of momentum, heat and water vapor are strongly affected by
topography and land surface cover (the reader will find various examples of such
variability in Whiteman et al. 1989b; Grimmond et al. 1996; Wotawa and Kromp-
Kolb 2000; Grossi and Falappi 2003; Antonacci and Tubino 2005; Bertoldi et al.
2006; Kalthoff et al. 2006; Serafin 2006; Martı́nez et al. 2010). In particular, vege-
tation can strongly affect turbulence production and surface layer structure, e.g. by
imposing displacement heights and roughness lengths (cf. de Franceschi et al. 2009),
2 Diurnal Mountain Wind Systems 71

and affecting fluxes and budgets of momentum, energy and mass. Additional
complications arise when boundary layer processes occur in urban areas located
in mountain valleys (Kuttler et al. 1996; Piringer and Baumann 1999; Savijärvi and
Liya 2001; Kolev et al. 2007; Giovannini et al. 2011; Fernando 2010). Transport
and mixing of pollutants in complex terrain have been addressed by many authors
(see, e.g., Salerno 1992; Prévôt et al. 1993, 2000; Wong and Hage 1995; Prabha and
Mursch-Radlgruber 1999; Grell et al. 2000; Doran et al. 2002; Gohm et al. 2009;
De Wekker et al. 2009; de Franceschi and Zardi 2009; Aryal et al. 2009; Kitada and
Regmi 2003; Regmi et al. 2003; Lehner and Gohm 2010; Panday 2006; Panday and
Prinn 2009; Panday et al. 2009; Szintai et al. 2010). Boundary layer processes are
also addressed in Chap. 5.

2.3.2 The Effects of Other Winds on Valley Flows

Valley wind systems are often modified, or even overpowered, by larger-scale


thermally driven flows or by dynamically driven flows above a valley produced by
mesoscale or synoptic-scale processes. Dynamically driven winds, including those
in valleys, are discussed in Chap. 3, so we give here only a short summary.
Background flows above mountainous terrain are driven by large-scale pressure
gradients that are superimposed on the along-valley pressure gradients produced
locally by intra-valley temperature differences. When the valley atmosphere has a
high static stability, the background flows will often be channeled along the valley
axis (Whiteman and Doran 1993). These flows, are not, however, thermally driven
flows. They occur under more active synoptic conditions and do not reverse twice
per day at the normal morning and evening transition times. They are also generally
stronger than the diurnal valley flows and, since they are imposed from aloft, do
not necessarily exhibit the typical jet-like vertical profile usually seen with valley
flows. Nonetheless their interaction with the valley atmosphere may be significantly
affected by changes in the surface heat flux during the diurnal cycle (see e.g. Jiang
and Doyle 2008).
The direction of the flow in a valley (either up- or down-valley) may be driven by
the pressure gradient aloft, rather than by locally generated temperature differences
along the length of the valley. As valleys are often curved or bent, the along-valley
component of the synoptic-scale pressure gradient may differ from one segment of
the valley to another. Accordingly, changes in wind speed and direction will occur
along the axis of a bent valley, depending on factors such as valley orientation, width
and depth (Kossmann and Sturman 2003). This phenomenon in which winds within
the valley are driven by pressure differences between the two ends of the valley is
called pressure-driven channeling. The flow direction is from the high pressure end
toward the low pressure end of the valley. This type of channeling is in contrast to
forced channeling in which momentum from the winds aloft is carried down into
the valley, producing winds along the valley axis whose direction is determined by
the component along the valley axis of the wind velocity aloft. Forced channeling
72 D. Zardi and C.D. Whiteman

is often seen near mountain passes. In addition to these major types of channeling,
smaller scale airflow separations, eddies, and dynamic instabilities can be produced
by background flow/terrain interactions that will affect thermally driven flows within
a valley (Orgill et al. 1992).
When static stability is low, downward momentum transport, associated with
convection, can bring winds aloft into a valley with wind directions not necessarily
aligned along the valley axis. These cross-valley flows are more likely to be present
in the upper altitudes of a valley, and are unlikely to be confused with thermally
driven diurnal valley winds. Flows over valley ridges can have smaller-scale
dynamic effects on valley winds. In addition to the channeling processes mentioned
above, other effects on the structure and evolution of diurnal wind circulations
within a valley come from large-scale background flows (see, e.g., Barr and Orgill
1989; Doran 1991; Orgill et al. 1992; Mursch-Radlgruber 1995; Zängl 2008; and
Schmidli et al. 2009a). During daytime, ridge-level winds may significantly modify
the subsidence over the valley center and produce asymmetries in the cross-valley
circulation. A weak ridgetop wind may flush out the upper part of the valley and
modify the valley circulation, whereas a stronger one may intrude into the valley and
even reach the valley bottom. During nighttime, upper winds can penetrate into the
valley and “erode” the existing ground based inversion. However, valley drainage,
once established, is rather resistant to erosion from above, because of atmospheric
stability, especially in valleys having large drainage areas. As an example, Orgill
et al. (1992) investigated the mesoscale external factors affecting nocturnal drainage
winds in four different valleys in western Colorado during the ASCOT project. They
found that in such valleys wind erosion can reduce drainage flows to less than half of
the valley depth, the principal contributing factors to wind erosion processes being
an along-valley wind component above the valley that opposes the drainage, a weak
stability in the valley, and winds aloft with speeds exceeding a threshold around
5 m s1 and accelerating at rates greater than 0.0004 m s2 (Orgill et al. 1992). In
the Rosalian mountain range in eastern Austria, Mursch-Radlgruber (1995) found
that an opposing ambient flow decreased the down-valley flow depth in the shallow
valleys studied, but did not reduce the jet speed maximum. In contrast, an ambient
flow in the same direction as the drainage flow suppressed the drainage flow and
led to stagnation or pooling on one night. This effect was explained by observing
that the ambient wind produced a pressure low in the upper valley in the lee of
the Rosalian range, thus counteracting the normally existing hydrostatic pressure
gradient responsible for classical drainage flow.
The contrast in temperature between air over a water surface and air over the
adjacent land typically drives diurnal sea/lake breezes that can interact with valley
winds, significantly modifying both phenomena. The sea/lake breeze and the up-
valley wind in a nearby valley may merge in a connected wind field called the
Extended Sea Breeze (Kondo 1990) or Extended Lake Breeze (McGowan et al.
1995), producing long-range moisture transport (Sturman and McGowan 1995).
This transport may provide significant moisture through dew deposition to arid
valleys, as seen at the Elqui Valley in the Andes (Khodayar et al. 2008). Small
bodies of water on valley floors do not develop appreciable lake breezes, although
2 Diurnal Mountain Wind Systems 73

they can significantly affect surface energy budgets, valley wind strength and extent,
and moisture distribution (Sturman et al. 2003a, b). Relatively large lakes on valley
floors can develop lake breezes that interact with slope and valley wind systems,
with internal boundary layers forming at the lake-land boundary (McGowan and
Sturman 1996; Zardi et al. 1999; de Franceschi et al. 2002). Bergström and Juuso
(2006) used a numerical simulation to compare up-valley winds occurring with and
without a water body at the bottom of the valley, finding that the lower lake surface
temperatures lead to higher daily average wind speeds.

2.3.3 Topographic Effects on Valley Flows

2.3.3.1 Tributary Valleys

In valleys with tributaries, upstream flows divide at each confluence of a side valley
with the main valley, and downstream flows merge into a single down-valley wind.
Bifurcating flows and merging flows have not been extensively researched, but
one would expect that the topography in the vicinity of the merging valleys and
wind system characteristics in the main and tributary valleys would be determining
factors for the structure of the valley wind system. Because of the many different
forms of topography and the range of wind system characteristics, it is not clear if
investigation of a specific instance would provide information general enough to be
applicable to other valleys. However detailed studies of the structure and dynamics
of flow mergings or bifurcations at single situations are crucial to provide a better
understanding of local pollutant concentrations, as shown by Banta et al. (1997),
Fiedler et al. (2000) and Zängl and Vogt (2006). The requirement of mass flux
balance may result in a closed circulation in which return flows produce enhanced
subsidence or rising motions over the valley juncture, affecting temperature and
wind structures there (Freytag 1988; Zängl 2004).
The bulk effect on nighttime along-valley mass flux in a main valley by
inflow from a tributary box canyon was investigated in the Atmospheric Studies
in Complex Terrain (ASCOT) program in the 1980s (Coulter et al. 1989, 1991;
Porch et al. 1991). The tributary provided more mass flux to the main canyon
than would have come from a simple valley sidewall. Simulations of ideal valley-
tributary systems by O’Steen (2000) supported this conclusion. They also showed
that the increase in down-valley mass flux in the main valley begins well before
the tributary is encountered, suggesting that the mechanism by which a tributary
provides additional mass flow to the main valley is more complicated than suggested
by the intuitive notion of a simple side-stream contribution at the confluence with a
concomitant increase in valley flux downstream of the tributary. In some cases the
flow down the main valley actually decreases in a layer corresponding to the outflow
height of the tributary flow. This leads to a negative correlation of flow down the
main valley and flow out of the tributary. Cold air flowing out of the tributary is
apparently blocked by cold air flow coming down the main valley. Cold air thus
74 D. Zardi and C.D. Whiteman

tends to accumulate in the tributary and then enter the main valley through surges.
Alternatively, the tributary contribution may glide over the main valley’s core flow.
Conservation of momentum requires that the increased mass of cold air from the
tributary slows the main valley flow, allowing additional flow to exit the tributary.
The latter may assume the form of a regularly pulsating outflow, and modulate low-
frequency oscillations in the main valley wind, as shown by Porch et al. (1991).
Drainage flow from tributaries depends not only on cooling rates, but also
on ambient winds. Unexpectedly, an external wind blowing in opposition to the
drainage wind can strengthen the drainage wind. The mechanism is not fully
understood: perhaps circulation cells introduce mass from the ambient wind into the
tributary system (Coulter et al. 1989). In any case the orientation of the tributaries
to the main valley appears to be a controlling factor in the interactions between
tributary and valley flows, and periodic flow out of the tributaries can modulate
oscillations in the main valley drainage flow (Coulter et al. 1991).

2.3.3.2 Canyons

Canyons, valleys having relatively high depth-to-width ratios and steep sidewalls,
are subject to the same up- and down-valley flow regimes as valleys. A canyon
that is a segment of a longer valley may have one of two effects on drainage
flows. If the canyon is narrow or tortuous, it may restrict the valley flow. During
nighttime this may cause a cold-air pool or cold air lake to develop above the
canyon constriction (Whiteman 2000). If, on the other hand, the canyon is not
too narrow, the air may accelerate through the constricted canyon to conserve the
total momentum of the downvalley flow. The flow and temperature structure in a
canyon are affected by several physical processes. Deep narrow canyons have weak
downslope flows on the steep slopes (see Sect. 2.5.1) because of a reduced longwave
radiation loss caused by the restricted view of the sky and by enhanced radiative
interactions between the sidewalls that tend to drive the canyon atmosphere toward
isothermalcy. Turbulence generated at the ridgeline by terrain interactions with
above-canyon flows, or turbulence generated in valley flows by terrain elements,
tend to mix the canyon atmosphere (Start et al. 1975). Simulations of nocturnal
drainage flows in small rugged canyons (Lee et al. 1995) have shown that coupling
with upper winds may produce multi-vortex structures that are very sensitive to
the radiation budget. Canyons sometimes act as conduits between segments of a
valley having quite different climates (e.g., the Columbia Gorge, a canyon between
Oregon and Washington, USA). Strong temperature contrasts between the segments
can accelerate the normal canyon flows (e.g., Sharp and Mass 2004).
Arizona’s Grand Canyon of the Colorado River, one of only a few deep canyons
that have been investigated, maintains only weak stability during both day and night
in winter (Whiteman et al. 1999c). This may be due to terrain-enhanced interactions
between above-canyon flows and air within the canyon or to overturning of the
canyon atmosphere by cold air generated on snow-covered mesas to the north and
south of the canyon. Temperature jumps form at the top of the canyon when warm
2 Diurnal Mountain Wind Systems 75

air advection occurs across the canyon rim. Because stability in the canyon is near-
neutral, wintertime valley flows are weak and the flow along the canyon is driven by
pressure gradients that are superimposed on the canyon by traveling synoptic-scale
weather disturbances, with flow directed from the high pressure to the low pressure
end of the canyon. Nonetheless, weak diurnal valley flows are sometimes present
at the north end of the canyon in the vicinity of the Little Colorado River (Banta
et al. 1999).

2.3.4 Modeling of the Valley Flows

The state of the art on modeling of thermally driven flows was reviewed by Egger
(1990). Contributions to the understanding of diurnal valley flows have come from
a range of modeling approaches including conceptual, laboratory tank, Navier–
Stokes, analytical and large-eddy simulation models. By means of a simplified
mixed layer model over idealized topography, Kimura and Kuwagata (1995) showed
how the sensible and latent heat that accumulate in an atmospheric column are
dependent on the scale of the topography. Whiteman and McKee (1982) used
a simple thermodynamic model to explain the basic mechanisms governing the
morning breakup of nocturnal inversions in valleys. Zoumakis and Efstathiou
(2006a, b) extended this approach to obtain a criterion for estimating the time
required for the complete breakup and development of a CBL. Conceptual and
analytical models relating the energy budget and valley geometry to the rate of
atmospheric cooling in a valley were developed by McKee and O’Neal (1989).
Vergeiner (1987) used an analytical model to study basic relationships between
the strength and response time of a valley flow as a function of topographic
characteristics such as valley length, width, or depth. A hydraulic model reproducing
the development of diurnal up-valley wind, based on concepts of mixed layer
schemes, was proposed by Zardi (2000).
Similarly to what has been discussed above for slope winds, a more complete
understanding of the atmospheric processes leading to diurnal valley flows has been
provided by numerical models. In the last two decades the achievements in this field
have benefitted from the increasing availability of suitable computers and from the
impressive progress in the development of numerical weather prediction models.
Further information on numerical modeling is provided in Chaps. 9, 10 and 11.
Diurnal mesoscale circulations such as slope and valleys flows are not resolved
by the coarse grids in general circulation and climate models, and must accordingly
be parameterized as turbulent motions. However, only smaller-scale turbulence and
convection are presently represented in boundary-layer parameterizations, based on
similarity arguments and data from flat terrain experiments. Even though vertical
heat fluxes induced by the diurnal slope and valley flows are as large as or even
larger than turbulent fluxes produced by smaller-scale turbulence and convection,
they are not yet considered in these parameterizations (Noppel and Fiedler 2002;
Rotach and Zardi 2007).
76 D. Zardi and C.D. Whiteman

Simulations of idealized valleys, with simplified initial and boundary conditions,


have led to a deeper insight into individual mechanisms affecting the dynamics of
diurnal valley winds. These simulations are particularly valuable in identifying the
relative roles of processes that are difficult to isolate in more realistic topography and
weather situations. While many idealized simulations have already been mentioned
in this chapter, space is unavailable here to provide a detailed account of all of
them. Representative articles include those of Enger et al. (1993); Atkinson (1995);
Anquetin et al. (1998); Li and Atkinson (1999); O’Steen (2000); Rampanelli et al.
(2004); Bergström and Juuso (2006); Bitencourt and Acevedo (2008); Schmidli
et al. (2010); Schmidli and Rotunno (2010); and Serafin and Zardi (2010a, b, c).
The complexity of real mountain valleys provides a stimulating challenge for
numerical simulations. The spatial resolution of present models is not yet fine
enough to reproduce all the small-scale features of valley winds. High resolution
digital elevation databases have to be significantly smoothed before they are used
for model runs, and information on land cover, land usage, and soil moisture is
often approximate or incomplete. Nevertheless, many simulations of real cases have
proven useful, as the reader may appreciate by inspecting the reference papers
by Leone and Lee 1989; Gross 1990; Koračin and Enger 1994; Fast 1995, 2003;
Ramanathan and Srinivasan 1998; Colette et al. 2003; Zhong and Fast 2003;
De Wekker et al. 2005; Chow et al. 2006; Lee et al. 2006; Weigel et al. 2006;
and Bischoff-Gauß et al. 2006, 2008. Model outputs must be carefully compared
with available observations to make sure that the simulations capture the relevant
mechanisms. Understanding of the model outputs benefits from good visualization
tools and from physical insight and conceptual models. Additional confidence is
gained when comprehensive field observations are available for comparison. Thus,
simulations are usually done in conjunction with large observational programs.

2.3.5 Exceptions and Anomalies

The occurrence of anomalous valley winds, i.e. developing in apparent contradiction


to established concepts and theories, has stimulated broad debates and deeper
analyses since earlier investigations in the field. Wagner’s (1938) seminal paper, for
example, mentioned two Alpine wind systems that were known to contradict valley
wind theory. The Maloja wind in Switzerland’s Upper Engadine Valley northeast of
the Maloja Pass and the Ora del Garda in the Adige (in German, Etsch) Valley near
Trento, Italy, are both known to blow down-valley during daytime.
The Maloja wind was investigated in a modeling study by Gross (1984, 1985),
who also focused on the development of a narrow, elongated cloud (the Maloja
Schlange, English: Maloja snake) that sometimes forms in conjunction with the
wind and stretches northeastward from the pass. The anomalous Maloja wind is
an up-valley wind from the Bergell Valley to the southwest of the pass that intrudes
into the Upper Engadine Valley as a down-valley wind. The explanation for the
anomalous wind is the peculiar topography in the Maloja Pass region, where the
Bergell ridgelines extend beyond the pass into the Upper Engadine Valley.
2 Diurnal Mountain Wind Systems 77

Fig. 2.11 A well known example of confluence between valleys (Wagner 1938): the area north
of the city of Trento, where the “Ora del Garda” wind, blowing from the Valle dei Laghi (a)
through an elevated saddle (), interacts with the up-valley wind blowing in the Adige Valley
from South (b) (From de Franceschi et al. 2002. © American Meteorological Society. Reprinted
with permission)

The outflow of the Ora del Garda in the Adige Valley during the early afternoon
appears as an “abnormal flow development” (Wagner 1938) over a whole segment
of the valley including the city of Trento. There, the wind near the surface flows
down-valley for the whole afternoon, rather than blowing up-valley as expected.
The down-valley flow is caused by the spillover into the Adige Valley of relatively
cold air through an elevated saddle from the upper Valle dei Laghi, which joins
the Adige Valley just up-valley from Trento (Fig. 2.11). In the late morning a
lake breeze develops at Lake Garda at the southern inlet of the Valle dei Laghi.
The cold lake air is carried up-valley and over the elevated saddle. The airflow
blowing down from this saddle onto the Adige valley floor in the early afternoon
on fair weather days brings potentially cooler air, which then flows underneath the
up-valley wind blowing at that time in the Adige Valley. The flow splits into an up-
valley ground-level flow north of the junction and a down-valley flow south of the
junction. Earlier researchers, relying only on surface observations, could not explain
the phenomenon. Subsequent field campaigns (reported in Wagner 1938) and recent
ones (de Franceschi et al. 2002) identified the vertical structure of the valley
atmosphere and demonstrated that the flows are compatible with known dynamics.
78 D. Zardi and C.D. Whiteman

In recent years, other cases of anomalous winds have been observed and
explained. In most cases the aberrant winds are produced by a large source of
cold or warm air at one end of the valley or the other. For example, nighttime
down-valley winds often fail to reverse during daytime in winter when the valleys
are covered with snow (Emeis et al. 2007; Schicker and Seibert 2009). Austria’s
Inn Valley (Whiteman 1990) and Germany’s Loisach Valley (Whiteman 2000) are
examples. Similarly, valleys with large sources of cold air at their lower end (e.g.,
an ocean or a long lake) can have up-valley winds that fail to reverse at night. The
strong and continuous cold air advection can cause the valley atmosphere to remain
stably stratified during daytime with only weak development of a shallow convective
boundary layer at the valley floor. Examples of valleys exhibiting this behavior
include Chile’s Elqui Valley (Kalthoff et al. 2002; Khodayar et al. 2008), which
ends at the Pacific Ocean, and California’s San Joaquin Valley (Zhong et al. 2004;
Bianco et al. 2011). Up-valley flows in the San Joaquin Valley are also promoted by
a large source of warm air in a desert area at its upper end.
Anomalous nighttime up-valley and daytime down-valley winds occur at several
sites along the Colorado River upstream of the Grand Canyon (Whiteman et al.
1999a). The valley floor in this area is a series of basins connected by narrow
canyons. The anomalous winds represent flows that converge at the lower ends of
the basins during nighttime and diverge out at the upper ends of the basins during
daytime. Another anomaly can be observed in the Marble Canyon area north of the
Grand Canyon where nighttime winds flow up-valley. This has been attributed to
the presence of a major geological formation that tilts downward in the up-valley
direction (Whiteman et al. 1999a).

2.4 Phases of and Interactions Between the Slope


and Valley Winds

The evolution of the coupled slope and valley systems during a typical day
progresses through four distinct phases, described below.

2.4.1 Daytime Phase

This phase typically begins in mid to late morning, after the nighttime valley
temperature inversion has been removed by daytime heating. The air in the valley is
warmer than over the adjacent plain, producing a lower pressure in the valley than
over the plain, and the resulting horizontal pressure gradient causes air to flow up
the valley axis (Fig. 2.8, upper panel).
Once the nocturnal temperature inversion is destroyed, convection from the
heated valley surfaces extends above the valley ridges. Vertical mixing associated
2 Diurnal Mountain Wind Systems 79

with the convection couples the flow in the valley to the background or synoptic-
scale flows aloft. The pressure gradients producing these flows are superimposed on
the local, thermally produced pressure gradient within the valley. Winds within the
valley can be affected by this coupling, although the background flows are generally
channeled along the valley axis. It is typical, in fact, that wind speeds, both in a
valley and over a plain, increase in the afternoon due to the coupling with the usually
stronger winds aloft (Whiteman 2000).
Maintenance of the daytime up-valley flow requires that temperatures in the
valley cross-section be greater than at the same level over the adjacent plain (or
at the same altitude farther down the valley). The TAF contributes to additional
heating in the valley, as solar radiation coming across the top of the valley at
ridgetop level, once converted to sensible heat flux at the valley floor and sidewalls,
overheats the air in the valley relative to the adjacent plain because of the smaller
volume in the valley. Radiative processes are also more effective within the valley
than over the surrounding plain because the valley air mass is partially enclosed
by radiating surfaces. In addition, advective heat transfer to the valley atmosphere
is more effective than the convective heat transfer to the same elevation over the
lower-altitude plain, again due to the smaller volume of air to be heated. The extra
heating in the valley invigorates the convection from the valley after the inversion is
destroyed, but this effect decreases as the mixed layer grows deeper above the ridges.

2.4.2 Evening Transition Phase

The surface energy budget reverses above valley surfaces when outgoing longwave
radiation first exceeds incoming shortwave radiation. This usually occurs initially in
late afternoon in areas of the valley that are shadowed by terrain. With this energy
budget reversal, the ground begins to cool and a downward turbulent sensible heat
flux begins to remove heat from a layer of air above the surface. A shallow layer
of cooled air thus forms over the sidewalls and floor of the valley. Over the slopes,
the air in this layer becomes negatively buoyant compared to the air adjacent to the
slope, and begins to move down the slope. The downslope flows converge onto
the valley floor or, after a temperature inversion forms on the valley floor, into
the elevated portion of the inversion. Rising motions over the valley center that
compensate for downslope flows on the sidewalls produce upward cold air advection
that cools the entire valley cross-section (Whiteman 2000; Fast and Darby 2004;
Brazel et al. 2005). Other processes play a role, including radiative transfer and
the continuing cold air advection associated with the daytime up-valley flow. Up-
valley winds continue to blow for several hours after the energy budget reversal
because of inertial effects and because it takes some time for the cooling over the
valley slopes to be transferred to the entire valley volume through compensatory
rising motions and other processes. The temporary post-sunset continuance of the
up-valley flow generates shear that increases turbulent sensible heat flux in the
surface layer (Kuwagata and Kimura 1995, 1997; de Franceschi et al. 2002, 2009;
80 D. Zardi and C.D. Whiteman

de Franceschi 2004). Once the air in the valley cross-section becomes colder
than the air over the adjacent plain, the pressure gradient reverses and the winds
reverse from up-valley to down-valley. The reversal occurs first in the ground-based
stable layer (temperature inversion) that forms on the valley floor. The stable layer
containing down-valley winds grows rapidly in depth in the early evening, causing
the down-valley winds to increase in strength and depth. The rate of growth of the
inversion gradually decreases as it fills the valley (Whiteman 1986) and the cooling
rate decreases exponentially for the remainder of the night. The bulk of the total
nighttime cooling occurs in the first few hours following sunset (De Wekker and
Whiteman 2006). The residual air layer, previously flowing up-valley, is then lifted
above the growing stable layer. This up-valley flow gradually decreases in strength
until it is overcome by background flows above the valley, which are often channeled
along the valley axis.
During the evening transition period the downslope flows that defined the
beginning of the period gradually decay as a strong stable layer builds up within the
valley and as down-valley flows in the stable layer expand to fill the valley volume.
The downslope flows on the valley sidewalls within the growing stable layer respond
to the increasing ambient stability by decreasing their strength. They continue on the
upper sidewalls of the valley above the growing stable layer, but leave the sidewall
and converge into the main valley volume once encountering the growing stable
layer (cf. Catalano and Cenedese 2010).
During the period of weakening up-valley flow and reversal to down-valley
flow, strong radiative cooling along the uneven valley floor and slopes results in
the formation of shallow cold air layers, that preferentially fill low spots on the
valley floor and extend up the sidewalls. These near-surface stable layers decouple
the surface from the flow aloft. During this period of light and variable winds
the surface-based temperature inversion often strengthens (Fast and Darby 2004).
The formation of the shallow cold air layer on the floor and sidewalls tends to
decouple the valley winds (whether up- or down-valley) from the friction of the
earth’s surface. Sometimes, an acceleration in valley wind caused by this decoupling
increases the shear across the shallow layers and briefly mixes them out, producing
transient evening temperature rises at the valley floor or on the slopes (Whiteman
et al. 2009). Oscillations may be generated in the valley and slope winds if the cold
air layer repeatedly reforms and decays.

2.4.3 Nighttime Phase

The beginning of the nighttime phase is hard to define because it is tied to the
gradual buildup of a stable layer containing down-valley winds rather than to a
discrete event. Typically, in clear undisturbed conditions the nighttime phase starts
sometime during the period from several hours after sunset until about midnight.
The main characteristic of the nighttime phase is the predominance of down-
valley flows throughout the valley volume. The maintenance of the down-valley
2 Diurnal Mountain Wind Systems 81

wind requires that the atmospheric cross-section continues to cool relative to the
adjacent plain. Because the downslope flows become weaker and shallower during
the latter part of the evening transition phase and are more and more affected by
(i.e., turned into the direction of) the down-valley flows (Whiteman and Zhong
2008) the importance of compensatory rising motions over the valley center to
this cooling is reduced. Instead, horizontal and vertical shears at the valley surfaces
produced by the increasing strength of the down-valley flow throughout the cross-
section become increasingly responsible for turbulent transport of heat toward the
radiatively cooling surfaces and the continued cooling of the valley cross-section.
The pressure gradient that drives the valley flow is at a maximum at the valley
floor. The vertical wind profile, however, takes a jet-like shape with the wind
maximum displaced 10–100 s of meters above the floor. The wind speed is reduced
at the surface by friction (Khodayar et al. 2008), and the winds above the jet
gradually give way to the gradient flow above (Zhong et al. 2004). Wind profiles
on a horizontal valley cross-section also show a jet-like shape with the peak wind
over the valley center where the frictional influence of the sidewalls is minimized.
The nighttime down-valley wind current is thus sometimes called a down-valley
jet (DVJ), and vertical and horizontal cross-sections through the current in a deep
valley can be described in terms of Prandtl’s well-known mathematical expression
and a parabolic profile, respectively (Clements et al. 1989; Coulter and Gudiksen
1995; Fast and Darby 2004). Because the peak winds generally occur over the
valley center, measurements taken at one location on a valley cross-section may
be not representative of the overall wind structure (King 1989). The establishment
of the DVJ is sometimes associated with abrupt warming at low levels as a result
of downward mixing and vertical transport of warm air from the inversion layer
above (Pinto et al. 2006). Time-periodic phenomena, such as pulses in the strength
of the DVJ (possibly contributing to vertical transport by creating localized areas
of low-level convergence) as well as gravity waves and Kelvin–Helmholtz waves,
which facilitate vertical mixing near the surface and atop the DVJ, are commonly
observed. These are reflected in the spectra of velocity and temperature fluctuations
(Stone and Hoard 1989). Typical periods of these oscillations are around 20 min
(Porch et al. 1991).
Nocturnal drainage flow along the valley floor displays very irregular behavior,
its structure being strongly determined by the local orography (Trachte et al. 2010).
Experimental evidence confirms the intuitive idea that drainage winds “slide” along
the ground and strongly adhere to its shape (Clements et al. 1989; Eckman 1998;
Fiedler et al. 2000; Haiden and Whiteman 2005). The down-valley flow sloshes
toward the outside of valley bends due to inertial forces (Sakiyama 1990).

2.4.4 Morning Transition Phase

The morning transition begins shortly after sunrise, when incoming solar radiation
exceeds longwave radiation loss and the surface energy budget reverses on the
82 D. Zardi and C.D. Whiteman

sidewalls. The end of the morning transition is marked by the break-up of the
ground-based inversion that fills the valley at night. Inversions break up occurs in
a fundamentally different way in valleys (Whiteman 2000) than over plains (Stull
1988). Whereas the plains atmosphere is heated primarily by the upward growth of
an unstable or convective boundary layer, heating of the valley atmosphere results
not only from the (limited) growth of an unstable boundary layer (cf. Weigel et al.
2007), but also from subsiding motions in the valley core that compensate for
upslope flows on the sidewalls.
After sunrise, as the slopes and valley floor are illuminated by the sun, the
surface energy budget reverses, and heat is transferred upward from the surface to
an initially shallow but progressively growing layer of air above the heated surfaces.
The layer of warming air over the sidewalls is on an inclined surface. Air at any point
within this layer is warmer than the air outside the layer at the same height, resulting
in an upslope flow. Differences in insolation on slopes – due to slope exposure,
shadows, ground cover, etc. – cause the incipient upslope flows to be somewhat
intermittent and inhomogeneous. Convection above the heated slopes and valley
floor entrains air from the elevated remnant of the nocturnal inversion (“stable core”)
into the upslope flow, which carries it to the valley ridgelines and vents it into the
free atmosphere. During this process, some horizontal transport of mass and heat can
contribute to the heating of the valley atmosphere, especially when mid-level stable
layers are encountered, or when topography projects into the valley atmosphere.
The transport of heated air up the sidewalls is compensated by sinking of the stable
core over the valley center. Because the stable core is statically stable, subsidence
causes downward advection of potentially warmer air, heating the atmosphere over
the valley center, and eventually causing the valley atmosphere to become warmer
than the air over the plain. The resulting horizontal pressure difference between the
valley atmosphere and the atmosphere over the adjacent plain initiates an up-valley
flow.
The break-up of nocturnal inversions in valleys was investigated in western
Colorado valleys by Whiteman (1982), resulting in a conceptual model and a simple
thermodynamic model (Whiteman and McKee 1982), and by Brehm and Freytag
(1982) in the Alps. Break-up requires that a valley be warmed from its initial
inversion temperature profile to a neutral profile having the potential temperature
of the inversion top at sunrise. The two main processes accomplishing the breakup
are the warming associated with convective boundary layer growth upward from
the heated ground and the warming caused by subsidence that compensates for
mass removal from the valley atmosphere by upslope flows. The warming can be
amplified by valley geometry as explained by the TAF principle. The time required
to destroy the Colorado inversions depended on inversion strength at sunrise and
the rate of input of sensible heat from the valley surfaces after sunrise, which
varies with season and snow cover. The Colorado valley inversions typically broke
up within 3–5 h after sunrise. Bader and McKee (1983), in a numerical model
study, found that a third mechanism, horizontal heat transfer from the sidewalls,
could also heat the valley atmosphere and contribute to inversion breakup. This
process is particularly effective in the short period of time between sunrise and
2 Diurnal Mountain Wind Systems 83

the time when upslope flows extend continuously up the sidewalls. Upslope flows
may detach from the slopes and intrude horizontally into the valley atmosphere at
elevations where sub-layers of high stability occur within the stable core (Harnisch
et al. 2009). Eckman (1998) and Rampanelli et al. (2004) performed simulations
in idealized valleys to investigate along-valley aspects of the inversion breakup.
They found that if the plain and the valley floor are perfectly horizontal, effects
of valley-plain contrasts concentrate in a region close to the mouth of the valley
and weaken with distance up the valley. Accordingly, the valley winds develop only
in a limited region upstream and downstream of the mouth of the valley. However,
when the valley floor is even weakly inclined, buoyancy effects can drive valley
winds farther up the valley (Rampanelli et al. 2004). By modifying Whiteman
and McKee’s thermodynamic model of valley inversion breakup, Zoumakis and
Efstathiou (2006a, b) focused on the relative partitioning of energy into convective
boundary layer growth and subsidence warming, but without considering horizontal
heat transfer. The occurrence of warm or cold air advection over the valley during
the inversion breakup period can change the heating requirement (Whiteman 1982;
Whiteman and McKee 1982). Warm air advection increases the time required to
destroy the inversion, while cold air advection has the opposite effect. Further, cold
air advection within the valley atmosphere increases the heating requirement, and
thus the time required to break the inversion.
The warming effect of compensatory sinking motions decreases in wider valleys
(Serafin and Zardi 2010b) and simulations show that compensatory sinking plays
only a small role in inversion breakup if the valley’s width to depth ratio exceeds
about 24 (Bader and McKee 1985). Inversions take longer to break up in wintertime
and in deep valleys (Whiteman 1982; Kelly 1988; Colette et al. 2003). The
asymmetric distribution of surface sensible heat flux between the opposing sidewalls
of a valley may result in a nonuniform inversion breakup (Urfer-Henneberger 1970;
Kelly 1988; Ramanathan and Srinivasan 1998; Anquetin et al. 1998; Matzinger
et al. 2003; Whiteman et al. 2004b). Gravity waves and horizontal mixing within
the stable layer effectively redistribute warm air and reduce temperature differences
due to differential sidewall heating (Bader and McKee 1983). Cross-valley flows
may develop and be superimposed on the valley circulations when sensible heat
fluxes differ greatly on the opposing sidewalls of a valley (Gleeson 1951; Bader and
Whiteman 1989; Atkinson 1995).
Especially in narrow valleys, the stable core inhibits the growth of a convective
boundary layer above the valley floor and sidewalls and delays the development
of deep convection (Furger et al. 2000; Bischoff-Gauß et al. 2006; Chemel and
Chollet 2006). Indeed, convective boundary layers over a valley floor are typically
not as deep as over an adjacent plain (Dosio et al. 2001; de Franceschi et al.
2003; Rampanelli and Zardi 2004; Rampanelli et al. 2004; Weigel and Rotach
2004; Bergström and Juuso 2006; Chemel and Chollet 2006; Weigel et al. 2006;
Bischoff-Gauß et al. 2008). Convective eddies in the relatively shallow convective
boundary layer over the valley floor efficiently mix and horizontally homogenize the
convective boundary layer. It thus exhibits weak cross- and along-valley variability
84 D. Zardi and C.D. Whiteman

over the valley floor (Zängl et al. 2001; De Wekker et al. 2005; Chemel and Chollet
2006) unless strong terrain heterogeneities exist (Ramanathan and Srinivasan 1998).
Switzerland’s Riviera Valley (Weigel and Rotach 2004; De Wekker et al. 2005;
Weigel et al. 2006) exhibits some hitherto unknown meteorological characteristics
that are produced by a bend in the valley. During the morning transition period
inertia causes the up-valley flow that enters the valley from the Magadino plain to
drift toward the outside of a bend at the lower end of the valley and develop a cross-
valley wind component. This produces a cross-valley tilt of the inversion layer and
anomalous downslope flows on the sunlit western sidewall that transport warm air
downward toward the valley floor, stabilizing the valley atmosphere. Cold-air ad-
vection in the up-valley flow, subsidence of warm air from the free atmosphere aloft
(compensating for up-slope flows), and curvature-induced secondary circulations
in the southern half of the valley produce a tendency for the valley atmosphere to
remain rather stable all day, although a shallow convective boundary layer does
grows slowly upward from the valley floor during daytime. The valley circulation
thus has a prolonged morning transition phase, rather than a daytime phase. Other
investigators (Cox 2005; Bergström and Juuso 2006; Bischoff-Gauß et al. 2006)
have also reported valleys that maintain persistent stable layers during daytime up-
valley flows. A common characteristic of several such valleys is a large source of
cold air at the valley mouth, such as a long lake or an ocean. During daytime, a stable
stratification and negative sensible heat fluxes can be maintained over lake and ocean
surfaces while the nearby land surfaces maintain strong positive fluxes (McGowan
and Sturman 1996; Bischoff-Gauß et al. 2006). These cases provide examples of
another mechanism by which cold air can be advected continuously up the valley,
reducing temperatures, maintaining stability and suppressing CBL growth.

2.5 The Mountain-Plain Wind System

The daytime heating and nighttime cooling contrast between the mountain atmo-
sphere over the outer slopes of the mountain massif and the free atmosphere over the
surrounding plain produces the horizontal pressure differences that drive this wind
system. The mountain-plain wind system brings low level air into the mountain
massif during daytime (the plain-to-mountain wind), with a weak return circulation
aloft. During nighttime, the circulation reverses, bringing air out of the mountain
massif at lower levels (the mountain-to-plain wind), with a weak return circulation
aloft. The daytime circulations are naturally somewhat deeper and more energetic
than the nighttime circulations because of the larger daytime heat fluxes and the
correspondingly deeper daytime boundary layer. The reversal of this large-scale
wind system is somewhat delayed relative to the slope and valley wind systems
because of its larger mass.
The mountain-plain circulation is associated with a diurnal pressure oscillation
between the mountainous region and the adjacent plains (for the Alps, cf. Frei and
2 Diurnal Mountain Wind Systems 85

Davies 1993; for the Rocky Mountains, cf. Li and Smith 2010). For the largest
mountain, this oscillation can be detected in one or more modes of the atmospheric
tides observed at global scale (Dai and Wang 1999).
According to Ekhart (1948), the mountain-plain wind system occupies a separate
outer layer from the valley and slope wind systems (see Fig. 2.1). An alternative way
of thinking of the mountain-plain wind would be to consider it as encompassing both
the slope and valley wind systems, as all three wind systems interact to transport
mass to and from a mountain massif from its surroundings.
Mountain-plain circulations play a key role in moisture transport and initiation of
moist convection (see also Chaps. 6 and 7) as well as in long-range transport of air
pollutants and their precursors (see also Chap. 5). Convergence of the daytime flows
over the mountains produces afternoon clouds and air mass thunderstorms, and the
divergent nighttime sinking motions produce late afternoon and evening clearing.
Pollutants associated with population centers on the plains adjacent to mountains are
carried toward the mountains during daytime, where they are transported vertically
to higher altitudes (Henne et al. 2005).
Due to the large spatial extent of mountain-plain circulations, their basic features
can be monitored by conventional networks of surface and upper air observing
stations (Kalthoff et al. 2002; Lugauer and Winkler 2005; Henne et al. 2005, 2008;
Taylor et al. 2005; Steinacker et al. 2006; Bica et al. 2007; Li et al. 2009). Mountain-
plain flows are generally weak, usually below 2 m s1 , with much lower speeds
in the broader and often deeper return flows aloft, and can easily be overpowered
by prevailing large scale flows. The mountain-plain circulation is diurnal, with its
evolution occurring in phases (Bossert and Cotton 1994a, b) and may be influenced
by large-scale conditions (Lugauer and Winkler 2005). The characteristic weather
associated with the plain-to-mountain circulation depends on the moisture levels
in the surroundings. Where dry air masses are present, such as in the Colorado
Plateau (Bossert et al. 1989; Bossert and Cotton 1994a, b), the circulations are
marked by strong winds, low cloudiness, and isolated thunderstorms. Where the
mountains are surrounded by oceans or other sources of low-level moisture, as in
Western Sumatra (Sasaki et al. 2004; Wu et al. 2009) or the Alps (Ekhart 1948;
Weissmann et al. 2005), the daytime rising motions transport moisture to higher
levels of the atmosphere and produce widespread clouds and precipitation (Kalthoff
et al. 2009). In some cases mountain-plain circulations also interact with sea breezes
(Taylor et al. 2005; Hughes et al. 2007).
Persistent temperature differences between the mountain atmosphere and the
surroundings occur seasonally in some mountain ranges. Seasonal thermal contrasts
are the primary driver of monsoonal circulations, that bring cold air out the
mountain massif during winter and into the mountain massif during summer. These
large-scale seasonal circulations are a type of mountain-plain circulation, often with
diurnal oscillations of wind strength superimposed on them. The initiation of the
summer monsoon caused by the heating of the mountain atmosphere in spring
is responsible for the initiation, for example, of the Himalayan Monsoon (Barros
and Lang 2003) and the variously named North American, Southwest, Mexican or
Arizona Monsoon that affects the southwestern United States (Higgins et al. 2006).
86 D. Zardi and C.D. Whiteman

The remainder of this section will discuss mountain-plain circulations that


develop around the east–west oriented Alps and around the north–south oriented
Rocky Mountains of North America. The mountain-plain system develops regard-
less of the orientation of a mountain range. The mountain-plain circulations that
develop between plains and elevated basins or plateaus are discussed in Sect. 2.6.

2.5.1 Mountain-Plain Winds in the Alps

The theory of diurnally reversing mountain wind circulations published by Prof.


A. Wagner (1932, 1938) included a postulated “equalizing flow” between the
plains and mountains. The evidence for these flows was collected by Wagner and
his students, most notably E. Ekhart and A. Burger. Burger and Ekhart (1937)
analyzed summer wind data obtained by the optical tracking of pilot balloons
released twice-daily from 20 stations located in and around the Alps to find that
a daytime circulation brought air into the Alps from the surrounding plains, with
a weak compensation current (“Ausgleichströmung”), or return circulation, aloft.
The thermal origin of this phenomenon and the associated horizontal pressure
differences that drive the mountain-plain circulation were investigated further by
Ekhart (1948). In recent years new analysis tools such as the Vienna Enhanced
Resolution Analysis (VERA; Steinacker et al. 2006; Bica et al. 2007) have been
used with the dense network of surface stations in the Alps to provide additional
information on the thermal forcing of mountain-plain circulations.
New information on the Alpine mountain-plain circulation has come from a
research program called Vertikaler Austausch und Orographie (VERTIKATOR,
or Vertical Exchange and Orography), which supported a major field experiment
in July 2002 on the north side of the Alps in the Bavarian foreland (Lugauer
et al. 2003). On summer days radiative heating of the Alps produces a horizontal
transport of air from the foreland into the Alps, and a vertical transport from the
boundary layer into the free troposphere above the mountains. The daily vertical
transport of air and moisture over the Alps by the mountain-plain circulation has
been termed “Alpine pumping” (Lugauer et al. 2003). Climatological investigations,
using the extensive network of surface stations in the Alpine foreland, have provided
additional information on the spatial and temporal extent of the circulation (Lugauer
and Winkler 2005; Winkler et al. 2006). The circulation is strongest on days when
background upper-air flows are weak and when daily incoming solar radiation
exceeds about 20 MJ m2 . Because incoming radiation is strongest in summer,
the flows are most common and best developed between April and September.
The diurnally reversing mountain-plain circulation extends approximately 100 km
north of the Alps to the Danube River. The low-level flow coming into the Alps
during daytime, with speeds generally less than 2 m s1 , extends gradually farther
into the plains during daytime and extends vertically to reach about 1–2 km
depths at the edge of the Alps by mid-afternoon. Rising motions over the Alps
produce cloudiness and precipitation, while the descending return branch of the
2 Diurnal Mountain Wind Systems 87

circulation suppresses convective cloud formation north of the Alps. An intensive


VERTIKATOR field study on 12 July 2002 used an airborne Doppler lidar, a wind-
temperature radar, dropsondes, and rawinsondes to gain additional information on
the daytime circulation (Weissmann et al. 2005). Vertical velocities in the rising
branch of the circulation were on the order of 0.05–0.10 m s1 . Despite its large
scale, the plain-to-mountain circulation responded readily to local geographical and
dynamic features. High-resolution simulations with numerical models were able
to reproduce the general flow structure and, when combined with vertical motion
data from the lidar, were able to estimate the daytime mass fluxes into the Alps
(Weissmann et al. 2005).

2.5.2 Mountain-Plain Winds in the Rocky Mountains

A robust mountain-plain circulation, forced by the diurnal cycle of solar heating and
longwave cooling, is a key feature of flows around the Rocky Mountains (Reiter and
Tang 1984). The circulation is strongest on clear days and nights, when background
synoptic-scale flows are weak. These conditions occur most frequently in summer
and fall. The diurnal geopotential height differences between the Rocky Mountains
and the surrounding plains, and the resulting near-surface diurnal inflows and
outflows, cause moisture convergence and divergence (Reiter and Tang 1984). The
daytime convergence over the mountain crests produces afternoon thunderstorms
and affects precipitation patterns. The nighttime divergence causes sinking motions
over the mountains that tend to reduce cloudiness and precipitation. The inflow
and outflow cycle can be detected in wind observations at mountaintop locations,
with inflows and outflows at speeds of several meters per second (Bossert et al.
1989). Transition periods tend to be long. The transition from the daytime inflow to
nighttime outflow occurs from 1900 to 0200 MST, the outflow to inflow transition
takes place from 0600 to 1100 MST. The transition from inflow to outflow is much
faster on highly convective days when evaporation of precipitation cools the heated
boundary layer over the mountains.
Harmonic analysis of summer wind data from the network of radar wind profilers
in the central United States shows that diurnal circulations extend up to 1,200 km
eastward from the Rocky Mountains over the gently sloping Great Plains to the
Mississippi River (Whiteman and Bian 1998). The strongest mountain-plain flows
occur in the lowest kilometer of the atmosphere where speeds, as averaged over the
100-m range gates of the radars, reach up to 3.5 m s1 just east of the mountains.
The return flow aloft is distributed through a layer extending above the highest
mountains (4,000 m) at speeds of a few tenths of a meter per second. The
mountain-plain circulation decays with distance from the mountains.
Much of the research on the mountain-plain circulation has focused on the
extensive plains to the east of the Rocky Mountains, where observational networks
are fairly dense, but the wind system also is present on the western slopes of the
Rockies and on the periphery of the mountain massif (Reiter and Tang 1984).
88 D. Zardi and C.D. Whiteman

2.6 Diurnal Wind Systems in Basins and Over Plateaus

Basins and plateaus, common landforms in many parts of the world, influence
atmospheric motions on a variety of scales, from tens of meters to thousands of
kilometers. Basins are hollows or depressions in the earth’s surface, wholly or partly
surrounded by higher land. An example of a basin that is only partly surrounded
by mountains is the Los Angeles basin (Lu and Turco 1995), which is open to
the Pacific Ocean on its western side. Closed basins with level unbroken ridgetops
are rarely encountered. Plateaus are elevated, comparatively level expanses of land.
Large plateaus include the Tibetan Plateau and the Bolivian Altiplano. A plateau
partially or completely surrounded by mountains can also be referred to as an
elevated basin.
Diurnal wind systems in basins and over plateaus are driven by horizontal pres-
sure gradients that develop between the atmosphere within or over the topography
and the free atmosphere of its surroundings. Basins and plateaus experience diurnal
slope and mountain-plain wind systems but, if there is no well-defined channel,
they lack a valley wind system. The large-scale mountain-plain wind system that
carries air between an elevated basin and the surrounding plains is called a basin-
plain wind. The daytime branch is the plain-to-basin wind and the nighttime branch
is the basin-to-plain wind (Kimura and Kuwagata 1993). Similarly, the mountain-
plain wind system over a plateau is called a plateau-plain wind. The daytime branch
is the plain-to-plateau wind and the nighttime branch is the plateau-to-plain wind
(Mannouji 1982).

2.6.1 Diurnal Wind Systems in Basins

Meteorological experiments have been conducted in basins all around the world,
including the Aizu (Kondo et al. 1989) and Ina Basins (Kimura and Kuwagata
1993) and a small frost hollow (Iijima and Shinoda 2000) in Japan, the McKenzie
Lake basin (Kossman et al. 2002; Zawar-Reza et al. 2004; Zawar-Reza and Sturman
2006a, b) in New Zealand, the Grünloch (Pospichal 2004; Whiteman et al. 2004a,
b, c; Steinacker et al. 2007) and Danube Valley (Zängl 2005) basins in Austria,
other small Alpine basins or dolines (Vertacnik et al. 2007), the Duero Basin in
Spain (Cuxart 2008); a basin on the island of Majorca (Cuxart et al. 2007; Martı́nez
2011), in the Columbia (Whiteman et al. 2001; Zhong et al. 2001) and Colorado
Plateaus basins (Whiteman et al. 1999b) of the western United States, and in
Colorado’s Sinbad Basin (Whiteman et al. 1996; Fast et al. 1996), Utah’s Peter
Sinks basin (Clements et al. 2003), and Arizona’s Meteor Crater (Savage et al. 2008;
Whiteman et al. 2008; Yao and Zhong 2009; Fritts et al. 2010; Lehner et al. 2010;
Whiteman et al. 2010; Haiden et al. 2011).
Slope winds are a feature of basin meteorology, just as they are for valley
and plateau meteorology. Basins, however, if surrounded by mountains, have no
2 Diurnal Mountain Wind Systems 89

Table 2.2 Extreme minimum temperatures in selected basins (Data from most U.S. locations
come from the U. S. National Climatic Data Center [NCDC])
Location Temperature minimum Date Reference
ı ı
Peter Sink, Utah 56.3 C (69.3 F) Feb 1, 1985 Pope and Brough (1996)
West Yellowstone, Montana 54ı C (66ı F) Feb 1933 NCDC
Taylor Park, Colorado 51ı C (60ı F) Feb 1951 NCDC
Fraser, Colorado 47ı C (53ı F) Jan 1962 NCDC
Stanley, Idaho 48ı C (54ı F) Dec 1983 NCDC
Gruenloch Basin, Austria 52.6ı C (63ı F) 19 Feb– Aigner (1952)
4 Mar 1932

equivalent to the valley wind. The absence of this wind system is responsible for the
unusually cold nighttime minimum temperatures often experienced in basins. The
low nighttime temperature minima (see examples in Table 2.2) can be explained by
contrasting the situation in a closed basin to that in a valley. Air in a valley at night,
once it becomes colder than air at the same level over the plains, begins to flow
down the valley axis and is replaced by air that sinks into the valley from above.
The air that sinks into the valley is statically stable and also warms adiabatically
as it sinks, so that it is warmer than the air it is replacing. The down-valley flow
thus causes a continuous import of warm air into a valley from aloft, reducing the
potential for cold nighttime minima. Basins, in contrast, trap air in place and cool it
continuously throughout the night. Extremes of minimum temperature are attained
by radiative cooling on clear, dry, winter nights following synoptic incursions of
cold air masses into a region. Cooling is increased on long winter nights and can be
especially strong if a fresh cover of new snow insulates the basin air from ground
heat fluxes. When winds aloft are weak, buoyancy forces tend to cause isentropes
to become horizontal within the inversion in a confined basin. Thus, temperature
measurements on the sidewalls become good proxies for free air temperatures over
the basin center, allowing temperature “soundings” to be made from surface-based
instruments (Whiteman et al. 2004a).
Nocturnal cooling is often confined within a basin during the night, although in
shallow basins the cooling may extend above the ridgeline and produce basin-to-
plain outflows over the lower passes (Pospichal et al. 2003). Because TAF causes
additional cooling in basins compared to the surrounding plains, a temperature jump
may develop at night above the basin between the air cooled within the basin and
the still-warm free atmosphere above (see e.g. Clements et al. 2003).

2.6.1.1 The Diurnal Cycle in Basins

In fair weather conditions, undisturbed by clouds or background flows, the diurnal


temperature structure in basins evolves similarly from day to day. During daytime,
the basin atmosphere becomes well mixed, with a neutral temperature profile. In
late afternoon, as shadows are cast on the sidewalls, the net longwave loss exceeds
the net solar input, and the local energy budget reverses, with downward turbulent
90 D. Zardi and C.D. Whiteman

sensible heat flux removing heat from a shallow air layer above the slopes. As the
shadows progress, the energy budget reverses over more and more of the surface
area of the basin and downslope flows begin as cold air layers form over the slopes.
A common misperception, which arises by making an analogy between air and
water, is that the nighttime cooling of the basin air mass is produced solely by the
convergence of downslope flows onto the basin floor and the cooling induced by
compensatory rising motions over the basin center. If this were the case, a series of
soundings of the basin atmosphere would show the continuous rising of the top of a
temperature inversion as the cold pool became deeper and deeper. In fact, as a layer
of cold air builds up on the floor of a basin, the downslope flows no longer have
the temperature deficit to reach all the way to the basin floor, and they converge
near the top of the growing inversion layer. Observations show that much of the
cooling in a basin takes place continuously throughout the basin’s full depth, and
this cooling continues throughout the night, even after the downslope flows slow
and die as the ambient stability builds up within the basin. The radiative and sensible
heat flux divergences that produce the cooling responsible for inversion buildup over
flat plains, where no downslope flows are involved, also play a major role in basins.
De Wekker and Whiteman (2006) have compared the cooling rates of valley, plain
and basin atmospheres and have found that basin atmospheres cool more rapidly
than those over plains or valleys. This is, no doubt, due to the TAF, the sheltering
of the basin from background winds by surrounding terrain (Vosper and Brown
2008), the divergence of heat from the confined volume, and the exclusion of diurnal
valley flows (which advect warm air into valleys from above and thus reduce their
cooling rates). The mechanisms by which the basin atmosphere cools have been dis-
cussed by Magono et al. (1982), Maki et al. (1986), Maki and Harimaya (1988), Neff
and King (1989), Kondo and Okusa (1990), and Vrhovec (1991). The cooling rates
of basin (and valley) atmospheres usually decrease during the night, but the relative
roles of turbulent and radiative flux divergences and advection are still unresolved.
Cooling rates in basins and valleys can be reduced by cloudiness or when heat is
released at the ground with the formation of dew or frost (Whiteman et al. 2007).
Nighttime temperature inversions in basins break up after sunrise (Triantafyllou
et al. 1995; Whiteman et al. 2004b) following the same patterns that have been
observed in valleys. The warming of the basin atmosphere and the ultimate
destruction of the nocturnal inversion is accomplished by convective currents from
the heated floor and sidewalls, the transport of mass from the basin by advection in
upslope flows over the sidewalls, and warming due to subsiding motions over the
basin center that compensate for mass carried up the sidewalls. Subsiding motions
are weaker in wider basins. Cross-basin asymmetries in the warming and in the
convective boundary layer and wind structure can occur because of shadows cast
into the basin by the surrounding terrain and by spatial variations in the receipt of
solar radiation on sloping surfaces of different aspect (Hoch and Whiteman 2010).
Cross-basin differences in insolation produce cross-valley temperature and pressure
differences that produce cross-basin flows toward the warmer sidewall (Lehner et al.
2010). Inversions in small basins in Austria and the Western US, despite their
great strength, break up somewhat earlier than inversions in western U.S. valleys
(Whiteman et al. 2004b).
2 Diurnal Mountain Wind Systems 91

10

0
Temperature (°C)

-5

-10

-15

-20

-25

-30
12 00 12 00 12 00 12 00 12
30 Nov 1 Dec 2 Dec 3 Dec 4 Dec
Time (MST)

Fig. 2.12 Temperature time series for surface-based temperature sensors at different heights on
the sidewall of the Gruenloch Basin, Austria. Because the coldest air settles into the bottom of the
basin, the lowest elevation sensors report the lowest temperatures and the highest elevation sensors
report the highest temperatures

The daytime cycle of warming often extends high above a basin, because of
strong daytime sensible heat fluxes and convection. The horizontal pressure gra-
dients between the air over the basin and the air over the surrounding plain produce
a plain-to-basin pressure gradient during daytime. In large basins the inflows are
delayed by upslope circulation cells from inside and outside the basin that become
anchored to the ridgeline surrounding the basin. This mechanism is explained
further in the plateaus section below. A modeling study by De Wekker et al. (1998)
shows that flows into an idealized basin (one with consistent ridge height) come
over the basin ridgetop when horizontal pressure differences are present above the
ridgeline, but the pressure gradient between air confined inside the basin and air
outside the basin at the same height plays no role in initiating or maintaining the
plain-basin flow unless there are gaps in the ridgeline.
Temperature inversions in basins often develop quite regularly on undisturbed
nights, but there are a variety of perturbations that can occur in their development
under disturbed conditions. A typical time series of temperature measurements at
different heights on the sidewalls of the 150-m deep Gruenloch Basin are shown in
Fig. 2.12 for a period of four consecutive winter days following new snow on 30
November. The night of 1–2 December was a typical undisturbed night with clear
skies and weak background winds. On disturbed nights, temperature inversions in
this basin exhibit a variety of disturbances to the normal evolution of temperature
structure, as shown in Fig. 2.13. These schematic illustrations of selected cold-pool
disturbances were observed using the same sensors as for Fig. 2.12, but on disturbed
nights in the 9-month period from October 2001 to early June 2002. Inversion
structure is sensitive to cloudiness, to background flows (especially Föhn or strong
92 D. Zardi and C.D. Whiteman

Fig. 2.13 Schematic depiction of disturbances to nocturnal temperature inversions in Austria’s


Gruenloch Basin as revealed in temperature traces from instruments located at different heights
on the sidewalls (compare Fig. 2.12). The different forms of disturbances in the sub-figures can
be compared to the first sub-figure in which an undisturbed temperature evolution is shown and a
rough indication of the time and temperature scales for this and subsequent sub-figures is given.
The disturbances are produced by strong background winds, shear at the top of the inversion, and
variations in cloudiness and surface energy budget (From Dorninger et al. 2011)

winds above the basin), turbulent erosion at the top of the cold pool (Petkovšek
1992; Rakovec et al. 2002; Zhong et al. 2003), to cold-air intrusions coming over
the basin ridgeline (Whiteman et al. 2010; Haiden et al. 2011) and to variations in
the surface energy budget. A numerical modeling study by Zängl (2003) determined
the effects of upstream blocking, drainage flows and geostrophic pressure gradient
on cold-air pool persistence for a basin similar in shape to the Gruenloch Basin.

2.6.1.2 Persistent Wintertime Inversions in Basins

Persistent or multi-day temperature inversions or cold-air pools have generally


adverse effects on valley and basin populations (Smith et al. 1997). Temperatures
remain cold during such events, suppressing the normal diurnal temperature range.
The strong stability in the inversion causes air pollution and moisture to build up
from sources within the pool. Poor air quality may persist for days (cf. Cuxart and
2 Diurnal Mountain Wind Systems 93

Jiménez 2011). Fog may form and lift to produce stratus. Light rain, snow or drizzle
may be produced if clouds persist or become thicker, or if they are seeded by higher
clouds. Rain or drizzle may freeze when reaching the cold ground. Low cloud bases,
visibility restrictions in fog, rain, snow or drizzle, and icy runway conditions may
close airports and cause transportation problems.
Multi-day temperature inversions are primarily a wintertime phenomenon. They
are produced when incoming energy during daytime is insufficient to remove cold
air that has settled into the basin or valley. The cold air may form initially through
the usual nocturnal inversion formation processes or may be left in low-lying terrain
following the passage of a cold front. Inversions, whether at the incipient or fully
formed stages, can be strengthened by differential advection when relatively warm
air is advected over the basin or can be weakened when cold air is advected over the
basin (Whiteman et al. 1999b). The shorter days of winter, lower sun angles, and the
presence of snow cover are usually responsible for the changes in the surface energy
budget at the basin floor that reduce the daytime turbulent sensible heat flux and
convection below that required to break the inversion. Additionally, fog or stratus
clouds that form in the cold-air pool may reduce insolation and sensible heat flux at
the cold pool base.
Both the onset and cessation times of persistent cold-air pools are difficult to
forecast. Forecast “busts” lead to especially large errors in maximum and minimum
temperature forecasts. Forecasting the cessation of cold-air pools is particularly
problematic because the many different processes that can affect breakup occur on
different scales. Local processes such as evaporation or melting at the basin surface
or changes in cloud cover within the pool have an effect, as do synoptic-scale warm
or cold advection above the pool, turbulent erosion at the top of the cold pool, or
drag or pressure forces that may be strong enough to push the cold pool to one side
of a basin, to cause seiches within the pool, or to produce vertical oscillations at
the top of the cold-air pool. Forecasts of cold pool breakup usually are based on
the forecast passage of troughs or low-pressure weather systems (cf. Reeves and
Stensrud 2009). These passages are accompanied by advection of colder air above
the pool, destabilizing the cold-air pool, and by stronger winds that may assist in
breakup. Cold air remnants are likely to persist at the lowest altitudes of a valley
or basin even when the cold pool is thought to be fully removed. Studies of cold-
air pools in the Colorado Plateaus Basin show that the final breakup, including any
remnants at the lowest elevations, usually occurs in the afternoon with assistance
from a final burst of convection (Whiteman et al. 1999b).
While most of the research on persistent cold-air pools to date has been on non-
cloudy pools, cloudy cold-air pools sometimes form as moisture builds up within the
inversion layer due to evaporation from the underlying ground. These pools have
a weaker, near-moist-adiabatic stability. This is apparently caused by the sinking
of cold air produced at cloud top by longwave radiation loss and the resultant
overturning and vertical mixing between the cloud top and the ground. Because
of the weaker stability, air pollutants generated inside cloudy cold-air pools will be
better mixed in the vertical than for non-cloudy pools.
94 D. Zardi and C.D. Whiteman

2.6.2 Diurnal Wind Systems over Plateaus

Diurnal wind systems over plateaus have received much less research attention than
diurnal wind systems in valleys and basins. Nonetheless, simulations of idealized
plateaus have appeared (Egger 1987; De Wekker et al. 1998; Zängl and Chico 2006)
and field studies have been conducted on the Mexican Plateau (Bossert 1997; Doran
et al. 1998; Fast and Zhong 1998; Doran and Zhong 2000; Whiteman et al. 2000),
especially in connection with air pollution problems (de Foy et al. 2008), in a valley
feeding the Himalayan Plateau (Egger et al. 2000, 2002; Zängl et al. 2001) and
on the Altiplano of Bolivia (Zängl and Egger 2005; Egger et al. 2005; Reuder and
Egger 2006).
Plateaus provide elevated surfaces on which nighttime cooling and daytime
heating can take place. During nighttime, air is cooled above the plateau. When
the plateau is a tabletop, much of the cold air that forms (especially if the tabletop is
narrow or has a small area) drains off the plateau sides, leaving only a shallow and
relatively weak inversion on the plateau. If the plateau is surrounded by mountains
the cooling can be confined in the elevated basin and produce stronger and deeper
inversions than over plateaus with no surrounding mountains. During daytime, a
convective boundary layer grows above the elevated heating surface. The daytime
destruction of the nocturnal inversions often involves interactions with flows aloft
(Banta and Cotton 1981; Banta 1984), which are generally stronger at higher
elevations. If there is a weak nocturnal inversion on the plateau, the convection
is able to grow quickly upwards through the inversion and can often reach quite
high levels of the atmosphere. The warm air above a plateau can be advected off
the plateau into the surrounding atmosphere if background winds are present. For
example, the easterly advection of warm air from the South Park plateau in the
central Rocky Mountains of Colorado can sometimes be detected east of the plateau
where it affects the development of afternoon thunderstorms over the adjacent plains
(Arritt et al. 1992).
An interesting latency occurs in the development of the daytime wind over
plateau regions. There is often a very late break-in of plain-to-plateau or plain-to-
basin winds that occurs in the late afternoon or early evening rather than in the
late morning when the temperature difference between the air over the plateau
and its surroundings starts to build up, as illustrated in a conceptual model in
Fig. 2.14 based on observations and modeling of the Mexican Plateau (Whiteman
et al. 2000). During daytime, convection builds up a deep boundary layer above
an elevated plateau that is warmer than the air surrounding the plateau. Upslope
flows form on the outer periphery of the plateau and on the plateau itself, with
return circulations aloft. The circulation cells converge at the plateau rim or edge,
and the convergence anchors the circulation cells there during daytime (Bossert and
Cotton 1994a, b). The normal tendency of the atmosphere to produce circulations
to equilibrate temperature differences that develop between the air over the plateau
and its surroundings is therefore restrained. Cold air can intrude onto the plateau,
2 Diurnal Mountain Wind Systems 95

Fig. 2.14 Schematic diagram of daytime circulations and temperature profiles above and along-
side an elevated circular plateau, illustrating the daytime buildup of a baroclinic zone at the edges
of the plateau that leads to a break-in of cold winds onto the plateau in the evening. The dry
adiabatic lapse rate d is provided as a reference

however, through passes and gaps at the plateau edge and through valleys that cut
into the plateau (Egger 1987; Doran and Zhong 2000; Egger et al. 2000, 2002, 2005;
Zängl et al. 2001; Zängl and Egger 2005). Because the circulations are anchored to
the plateau edge, a baroclinic zone forms during daytime around the periphery of
the plateau, where the temperature gradient becomes strong. As the sun starts to
set, the slope flows weaken, the circulation cell boundaries are no longer anchored
to the plateau edge, and a strong flow of cold air comes across the baroclinic
zone onto the plateau. This sudden break-in of cold air onto the plateau from all
sides often occurs in early evening, and its progression across the plateau depends
on plateau width. Model simulations show that gravity waves are an important
mechanism that allows the warm air in the convective column above the plateau
to equilibrate with the surrounding air during the nighttime (Whiteman et al. 2000;
Zängl and Chico 2006).
As an example, the afternoon or early evening break-in of flows from outside a
plateau or elevated basin is depicted in the conceptual model of Fig. 2.15 for the
South, Middle and North Park basins of Colorado, as based on observations and
numerical modeling by Bossert and Cotton (1994a, b).
96 D. Zardi and C.D. Whiteman

Fig. 2.15 Diurnal flow patterns in the Central Colorado Rocky Mountains, looking northward
from southern Colorado to the elevated basins of South Park, Middle Park and North Park. (a)
During daytime, upslope flows from inside and outside the basins converge over the basin rims. (b)
During late afternoon, cold air from the surroundings breaks into the basins across the rims (From
Bossert and Cotton 1994a. © American Meteorological Society. Reprinted with permission)
2 Diurnal Mountain Wind Systems 97

2.7 Forecasting

The many practical applications of knowledge concerning diurnal mountain winds


were enumerated in the introduction to this chapter. Because diurnal mountain winds
affect so many aspects of human life and the environment, there is a need to make
accurate forecasts of diurnal mountain winds.
Forecasting of diurnal mountain winds is, in some respects, relatively simple.
Climatological investigations have told us much about the regular development and
characteristics of diurnal mountain winds. They form regularly and reliably under
high-pressure synoptic conditions when skies are clear and background winds are
weak. The underlying topography provides the channels for these flows. Airflows
are down the terrain (slopes, valleys and mountains) during nighttime and up
the terrain during daytime, with lags that correspond to the mass of the wind
system and the corresponding rate of heat transfer required to reverse it. The larger
the temperature contrasts that form inside the terrain or between the terrain and
the surrounding plain, the stronger the flows. Stronger temperature contrasts and
stronger flows occur when the diurnal cycle of sensible heat flux is strongest,
in the summer half-year. Increasing cloudiness and interactions with strong flows
aloft decrease the temperature and pressure contrasts that drive the wind systems.
A decreased amplitude of the sensible heat flux or failure of the sensible heat
flux to change sign diurnally decreases the intensity of the diurnal wind systems
or eliminates the diurnal reversal that is their chief characteristic. The decreased
amplitude could be caused by rain, high soil moisture, increasing cloudiness, or
an increased albedo due to snow. Changes in climate along a valley can affect
the amplitude of temperature contrast along a valley’s axis, affecting wind system
strength. The diurnal wind system is often present, but weaker and less regular in its
time of reversal, on cloudy and windy days.
The more difficult aspect of forecasting comes with the need to predict wind
system strength and characteristics for specific locations and at specific times.
Present-day operational weather models are presently able to provide accurate
simulations of general synoptic weather conditions and are now reaching the
horizontal resolution (1–2 km) necessary to resolve large valleys such as the
Inn, Rhine and Rhone Valleys of the Alps. These models, however, need further
improvements to provide accurate forecasts for individual valleys. Horizontal and
vertical grid and terrain resolutions should be improved, as well as boundary layer
and surface property parameterizations. We know that some valleys in a region
have stronger wind systems than others and that wind systems in some valleys are
more susceptible to interference from external conditions. The increasingly high
resolution attained by numerical research models is expected to provide more and
more precise indications of the relative strengths of wind systems in different valleys
under synoptic conditions, or even of different segments of an individual valley.
Such simulations, however, will need to be evaluated with field data, to verify that
the models are performing adequately. Some simpler models, such as the German
98 D. Zardi and C.D. Whiteman

Weather Service’s proprietary KLAM 21 drainage flow model (Sievers 2005) have
proven useful in simulations on a range of scales from an individual vineyard to
a domain of several tens of kilometers. A simpler approach (McKee and O’Neal
1989) using topographic maps investigates the along-valley variation of TAF (i.e.,
terrain cross-sections) and may have some utility, although the results to date have
been mixed.
A number of pattern recognition approaches have been used to determine
connections between synoptic-scale patterns and observed features of local winds
in multi-valley target areas where wind and supporting data are available (Guardans
and Palomino 1995; Weber and Kaufmann 1995, 1998; Kaufmann and Weber
1996, 1998; Kastendeuch and Kaufmann 1997; Kaufmann and Whiteman 1999;
Kastendeuch et al. 2000; Ludwig et al. 2004). In the Basel, Switzerland area, for
example, smaller valleys in the mountain complex have a high frequency of drainage
flows, while larger valleys more often experience interactions with synoptic-scale
winds that produce channeled flows. Up-valley flows are generally stronger in this
valley complex than down-valley flows, and the frequency of days with thermally
driven flows varies little with season (Weber and Kaufmann 1998). Although the
theoretical number of possible wind patterns is very large in a complicated terrain
area, complex topography tends to constrain the observed fields to a small number
of typical patterns. Recurrent wind patterns can also be determined by means of
conditional sampling (Zaremba and Carroll 1999). Such information should prove
useful for forecasting.
To overcome the inherent complexity and highly resource-demanding dynamical
modeling, an alternative approach is offered by the increasing performance and
flexibility of Geographic Information Systems (GIS). The latter can easily represent
small-scale local features strongly affecting diurnal flows such as topography,
land cover, surface budgets, etc. Supported by statistical analysis of long term
observations, and with an input of ambient conditions from either station networks
or large-scale numerical models, they can provide a first guess of wind and
temperature fields at local scale, as well as a basis for the estimate of related
quantities such as surface fluxes of heat, water vapor and passive scalars (Carrega
1995; Fast 1995; Ciolli et al. 2004; Deng and Stull 2005, 2007; Bertoldi et al. 2006;
Schicker and Seibert 2009).
When forecasts are needed for a mountainous region, both real time and clima-
tological data from weather stations can be analyzed to get a basic understanding
of the strength and timing of the diurnal wind systems. Where quantitative data
are available, general conceptual models of the phenomena can be used along
with advice from individuals with local experience. Major federal projects, such as
fighting a wildland fire, can sometimes justify the on-site assignment of a weather
observer/forecaster, who can assist with interpretation and provide useful daily
feedback on forecast accuracy.
2 Diurnal Mountain Wind Systems 99

2.8 Past Progress, Future Objectives

Significant progress has been made in the understanding of diurnal mountain winds
since the last AMS book on mountain meteorology was published (Blumen 1990).
Field experiments, both large international efforts and single investigator projects,
have been conducted taking advantage of new instrumentation and data platforms.
Modeling of diurnal mountain winds has improved as advances have been made in
hardware and software and as more detailed field data have become available.
Over the past two decades, field studies and modeling of diurnal mountain
winds have worked in tandem. This collaboration should continue: field studies
provide the data required to evaluate and refine models, and modeling work is
invaluable in conceptualizing the processes observed in the field. Improvements can
be expected in both areas. The introduction of higher horizontal and vertical grid
and terrain resolution, better numerical approaches, more adequate boundary layer
parameterizations, better modeling of turbulent processes and better handling of
non-uniform surface cover and irregular terrain into models will bring significant
advances to our understanding of mountain and valley atmospheres. The use
of remote sensing instruments capable of scanning larger target volumes (e.g.,
De Wekker and Mayor 2009) is expected to produce more comprehensive and
highly resolved observational data sets, which will allow the proper evaluation of
increasingly sophisticated models.
In recent years, international cooperation has become a hallmark of research
of diurnal mountain winds. Yearly international conferences alternating between
the AMS Mountain Meteorology Conference and the International Conference
on Alpine Meteorology, periodic short courses and large, international research
programs will continue to strengthen international ties.
Future research into diurnal mountain winds will benefit not only from broad
cooperation within the global mountain meteorology community but also from
interdisciplinary cooperation. Researchers in related fields – including mathematics,
numerical analysis, physics, fluid mechanics, turbulence dynamics, chemistry, and
hydrology – will contribute to a clearer understanding of the mountain and valley
atmosphere.
In the coming decades, an effort must be made to make research findings
available to the applications community. Weather forecasters, environmental and
civil engineers, land use planners, agricultural experts and wildland fire managers,
among others, would benefit from the knowledge gained and the tools developed by
basic research.
To date, research on diurnal mountain winds has mostly focused on individual
phases of the wind systems and individual processes, including the morning
transition, the evening transition, slope winds, surface energy budgets and so on.
This research will continue to be valuable. A number of interesting questions invite
further investigation.
• With respect to slope flows, additional research is needed on upslope flows, on
turbulent and radiative processes, on the external influences of ambient stability
100 D. Zardi and C.D. Whiteman

and ambient flows, and on the effects on slope flows of inhomogeneities of


surface properties and heat budget components along a slope. Models should
be developed to apply to realistic slopes and should be tested against new, higher
resolution data sets. Large eddy simulation models should be applied to gain
knowledge of mean and turbulent processes in slope flows.
• Understanding of valley flows would benefit from a comprehensive investigation
of heat transport processes within the valley atmosphere, of the factors affecting
inversion development and characteristics, and of the interactions between flows
from multiple valleys.
• Research programs on mountain-plain circulations occurring in the major moun-
tain ranges in the world could build on the VERTIKATOR results from Europe,
with many scientific and practical benefits. For instance, concerning the Rocky
Mountains, such a program could improve understanding of the role of the
mountain-plain circulations on the transport of moisture and pollutants to the
Rocky Mountains in diurnal and monsoonal wind systems. These circulations
are closely connected to the Great Plains Low-Level Jet and the North American
Monsoon, and play a role in the timing and movement of thunderstorm and
mesoscale convective systems over the Great Plains.
• Research is needed on persistent wintertime cold-air pools to address increasing
air pollution problems being experienced in urban basins and to improve the
forecasting of their buildup and breakup. Because wintertime cold-air pools are
experienced in many countries, an international research program on this topic
would be warranted.
Additionally, and perhaps more importantly, there is a clear need to take
a comprehensive look at thermally driven circulations in complex terrain, the
interactions of the component parts, the interactions of these winds with winds
aloft and the role of turbulence in wind characteristics. The initiatives investigating
these processes should be multi-national, interdisciplinary and include weather
forecasters and applications scientists.

Acknowledgments We thank the book editors for the opportunity to contribute to this book and
the associated 2008 Mountain Weather Workshop. Special thanks go to authors and organizations
which provided figures for this chapter. The University of Trento and Prof. Dino Zardi are thanked
for providing the financial support and warm hospitality for Dr. Whiteman’s 2-month visit to
the University of Trento in 2008 to begin the collaborative work on this chapter. Moreover
D. Zardi is grateful to the University of Trento for granting him a sabbatical leave during the
academic year 2009/2010. We greatly appreciate the editorial assistance of Johanna Whiteman,
the initial reviews of portions of this chapter by Thomas Haiden, Thomas W. Horst, Stephan
De Wekker, Tina Katopodes Chow, and Stefano Serafin and the valuable comments provided by
anonymous reviewers. Dr. Whiteman wishes to acknowledge partial funding support from National
Science Foundation grants ATM-0837870 and ATM-0444205 and from the National Oceanic and
Atmospheric Administration’s National Weather Service CSTAR grant NA07NWS4680003.
2 Diurnal Mountain Wind Systems 101

References

Aigner, S., 1952: Die Temperaturminima im Gstettnerboden bei Lunz am See, Niederösterreich
[The minimum temperatures in the Gstettner basin near Lunz, Lower Austria]. Wetter und
Leben, Sonderheft, 34–37.
Allwine, K. J., B. K. Lamb, and R. Eskridge, 1992: Wintertime dispersion in a mountainous basin
at Roanoke, Virginia: Tracer study. J. Appl. Meteor., 31, 1295–1311.
Amanatidis, G. T., K. H. Papadopoulos, J. G. Bartzis, and C. G. Helmis, 1992: Evidence of
katabatic flows deduced from a 84 m meteorological tower in Athens, Greece. Bound.-Layer
Metor., 58, 117–132.
André, J. C., and L. Mahrt, 1982: The nocturnal surface inversion and influence of clear-air
radiative cooling. J. Atmos. Sci., 39, 864–878.
Andretta M., A. Weiss, N. Kljun, and M. W. Rotach, 2001: Near-surface turbulent momentum flux
in an Alpine valley: Observational results. MAP newsletter, 15, 122–125.
Andretta, M., A. H. Weigel, and M. W. Rotach, 2002: Eddy correlation flux measurements in
an Alpine valley under different mesoscale circulations. Preprints 10th AMS Conf. Mountain
Meteor., 17–21 June 2002, Park City, UT, 109–111.
Anquetin, S., C. Guilbaud, and J.-P. Chollet, 1998: The formation and destruction of inversion
layers within a deep valley. J. Appl. Meteor., 37, 1547–1560.
Antonacci, A., and M. Tubino, 2005: An estimate of day-time turbulent diffusivity over complex
terrain from standard weather data. Theor. Appl. Climatol., 80, 205–212.
Arritt, R. W., J. M. Wilczak, and G. S. Young, 1992: Observations and numerical modeling of an
elevated mixed layer. Mon. Wea. Rev., 120, 2869–2880.
Aryal, R. K., B.-K. Leeb, R. Karki, A. Gurung, B. Baral, and S.-H. Byeonb, 2009: Dynamics of
PM2.5 concentrations in Kathmandu Valley, Nepal. J. Haz. Mat., 168, 732–738.
Atkinson, B. W., 1981: Meso-scale Atmospheric Circulations. Academic Press, 495pp.
Atkinson, B. W., 1995: Orographic and stability effects on valley-side drainage flows. Bound.-
Layer Meteor., 75, 403–428.
Atkinson, B. W., and A. N. Shahub, 1994: Orographic and stability effects on day-time, valley-side
slope flows. Bound.-Layer Meteor., 68, 275–300.
Axelsen, S. L., and H. van Dop, 2008: Large-eddy simulation and observations of slope flow. Acta
Geophysica. 57, 803–836.
Axelsen, S. L., and H. van Dop, 2009: Large-eddy simulation of katabatic flow over an infinite
slope. Acta Geophysica. 57, 837–856.
Bader, D. C., and T. B. McKee, 1983: Dynamical model simulation of the morning boundary layer
development in deep mountain valleys. J. Climate Appl. Meteor., 22, 341–351.
Bader, D. C., and T. B. McKee, 1985: Effects of shear, stability and valley characteristics on the
destruction of temperature inversions. J. Climate Appl. Meteor., 24, 822–832.
Bader, D. C., and C. D. Whiteman, 1989: Numerical simulation of cross-valley plume dispersion
during the morning transition period. J. Appl. Meteor., 28, 652–664.
Baines, P. G., 2001: Mixing in flows down gentle slopes into stratified environments. J. Fluid
Mech., 443, 237–270.
Baines, P. G., 2005: Mixing regimes for the flow of dense fluid down slopes into stratified
environments. J. Fluid Mech., 538, 245–267.
Banta, R. M., 1984: Daytime boundary-layer evolution over mountainous terrain. Part I: Observa-
tions of the dry circulations. Mon. Wea. Rev., 112, 340–356.
Banta, R., and W. R. Cotton, 1981: An analysis of the structure of local wind systems in a broad
mountain basin. J. Appl. Meteor., 20, 1255–1266.
Banta, R. M., and P. T. Gannon, 1995: Influence of soil moisture on simulations of katabatic flow.
Theor. Appl. Climatol., 52, 85–94.
Banta, R. M., L. D. Olivier, W. D. Neff, D. H. Levinson, and D. Ruffieux, 1995: Influence of
canyon-induced flows on flow and dispersion over adjacent plains. Theor. Appl. Climatol., 52,
27–42.
102 D. Zardi and C.D. Whiteman

Banta, R. M., and Coauthors, 1997: Nocturnal cleansing flows in a tributary valley. Atmos.
Environ., 31, 2147–2162.
Banta, R. M., L. S. Darby, P. Kaufmann, D. H. Levinson, and C.-J. Zhu, 1999: Wind-flow patterns
in the Grand Canyon as revealed by Doppler lidar. J. Appl. Meteor., 38, 1069–1083.
Banta, R. M., L. S. Darby, J. D. Fast, J. Pinto, C. D. Whiteman, W. J. Shaw, and B. D. Orr, 2004:
Nocturnal low-level jet in a mountain basin complex. I: Evolution and effects on local flows.
J. Appl. Meteor., 43, 1348–1365.
Barr, S., and M. M. Orgill, 1989: Influence of external meteorology on nocturnal valley drainage
winds. J. Appl. Meteor., 28, 497–517.
Barros, A. P., and T. J. Lang, 2003: Monitoring the monsoon in the Himalayas: Observations in
Central Nepal, June 2001. Mon. Wea. Rev., 131, 1408–1427.
Barry, R. G., 2008: Mountain Weather And Climate, 3rd Edition. Cambridge University Press,
506pp.
Bastin, S., and P. Drobinski, 2005: Temperature and wind velocity oscillations along a gentle slope
during sea-breeze events. Bound.-Layer Meteor., 114, 573–594.
Bergström, H., and N. Juuso, 2006: A study of valley winds using the MIUU mesoscale model.
Wind Energy, 9, 109–129.
Bertoldi, G., R. Rigon, and T. M. Over, 2006: Impact of watershed geomorphic characteristics on
the energy and water budgets. J. Hydrometeor., 7, 389–403.
Bhatt, B. C., and K. Nakamura, 2006: A climatological-dynamical analysis associated with
precipitation around the southern part of the Himalayas. J. Geophys. Res., 111, D02115.
Bianco, L., I. V. Djalalova, C. W. King, and J. M. Wilczak, 2011: Diurnal evolution and annual
variability of boundary-layer height and its correlation to other meteorological variables in
California’s Central Valley. Bound.-Layer Meteor., 140, 491–511.
Bica, B., and Coauthors, 2007: Thermally and dynamically induced pressure features over complex
terrain from high-resolution analyses. J. Appl. Meteor. Climatol., 46, 50–65.
Bischoff-Gauß, I., N. Kalthoff, and M. Fiebig-Wittmaack, 2006: The influence of a storage lake in
the Arid Elqui valley in Chile on local climate. Theor. Appl. Climatol., 85, 227–241.
Bischoff-Gauß, I., N. Kalthoff, S. Khodayar, M. Fiebig-Wittmaack, and S. Montecinos, 2008:
Model simulations of the boundary-layer evolution over an arid Andes valley. Bound.-Layer
Meteor., 128, 357–379.
Bitencourt, D. P., and O. C. Acevedo, 2008: Modelling the interaction between a river surface and
the atmosphere at the bottom of a valley. Bound.-Layer Meteor., 129, 309–321.
Blumen, W. (Ed.), 1990: Atmospheric Processes over Complex Terrain. Meteor. Monogr., 23
(no. 45), Amer. Meteor. Soc., Boston, Massachusetts.
Bodine, D., P. M. Klein, S. C. Arms, and A. Shapiro, 2009: Variability of surface air temperature
over gently sloped terrain. J. Appl. Meteor. Climatol., 48, 1117–1141.
Bossert, J. E., 1997: An investigation of flow regimes affecting the Mexico City region. J. Appl.
Meteor., 36, 119–140.
Bossert, J. E., and W. R. Cotton, 1994a: Regional-scale flows in mountainous terrain. Part I: A
numerical and observational comparison. Mon. Wea. Rev., 122, 1449–1471.
Bossert, J. E., and W. R. Cotton, 1994b: Regional-scale flows in mountainous terrain. Part II:
Simplified numerical experiments. Mon. Wea. Rev., 122, 1472–1489.
Bossert, J. E., J. D. Sheaffer, and E. R. Reiter, 1989: Aspects of regional-scale flows in mountainous
terrain. J. Appl. Meteorol., 28, 590–601.
Brazel, A.J., H. J. S. Fernando, J. C. R. Hunt, N. Selover, B. C. Hedquist, and E. Pardyjak, 2005:
Evening transition observations in Phoenix, Arizona. J. Applied Meteor., 44, 99–112.
Brehm, M. 1986: Experimentelle und numerische Untersuchungen der Hangwindschicht und ihrer
Rolle bei der Erwärmung von Tälern. [Experimental and numerical investigations of the slope
wind layer and its role in the warming of valleys]. Wiss. Mitt. Nr. 54, dissertation, Universität
München – Meteorologisches Institut, München, FRG.
Brehm, M., and C. Freytag, 1982: Erosion of the night-time thermal circulation in an Alpine valley.
Arch. Meteor. Geophys. Bioclimatol., Ser. B, 31, 331–352.
2 Diurnal Mountain Wind Systems 103

Buettner, K. J. K., and N. Thyer, 1966: Valley winds in the Mount Rainier area. Arch. Meteor.
Geophys. Bioclimatol., Ser. B, 14, 125–147.
Burger, A., and E. Ekhart, 1937: Über die tägliche Zirkulation der Atmosphäre im Bereiche der
Alpen [The daily atmospheric circulation in the Alpine region]. Gerl. Beitr., 49, 341–367.
Burkholder, B. A., A. Shapiro and E. Fedorovich, 2009: Katabatic flow induced by a cross-slope
band of surface cooling. Acta Geophysica, 57, 923–949.
Carrega, P., 1995: A method for the reconstruction of mountain air temperatures with cartographic
applications. Theor. Appl. Climatol., 52, 69–84.
Catalano, F., and A. Cenedese, 2010: High-resolution numerical modeling of thermally driven
slope winds in a valley with strong capping. J. Appl. Meteor. Climatol., 49, 1859–1880.
Catalano, F., and C.-H. Moeng, 2010: Large-eddy simulation of the daytime boundary layer in an
idealized valley using the Weather Research and Forecasting numerical model. Bound.-Layer
Meteor., 137, 49–75.
Chemel, C., and J. P. Chollet, 2006: Observations of the daytime boundary layer in deep Alpine
valleys. Bound.-Layer Meteor., 119, 239–262.
Chemel, C., C. Staquet, and Y. Largeron, 2009: Generation of internal gravity waves by a katabatic
wind in an idealized alpine valley. Meteor. Atmos Phys., 103, 187–194.
Chen, Y., A. Hall, and K. N. Liou, 2006: Application of 3D solar radiative transfer to mountains.
J. Geophys. Res., 111, D21111.
Chow, F. K., A. P. Weigel, R. L. Street, M. W. Rotach, and M. Xue. 2006. High-resolution
large-eddy simulations of flow in a steep Alpine valley. Part I: Methodology, verification, and
sensitivity experiments. J. Appl. Meteor. Climatol., 45, 63–86.
Ciolli, M., M. de Franceschi, R. Rea, A. Vitti, D. Zardi, and P. Zatelli, 2004: Development and
application of 2D and 3D GRASS modules for simulation of thermally driven slope winds.
Transactions in GIS, 8, 191–209.
Clements, W. E., J. A. Archuleta, and D. E. Hoard, 1989: Mean structure of the nocturnal drainage
flow in a deep valley. J. Appl. Meteor., 28, 457–462.
Clements, C. B., C. D. Whiteman, and J. D. Horel, 2003: Cold-air-pool structure and evolution in
a mountain basin: Peter Sinks, Utah. J. Appl. Meteor., 42, 752–768.
Cogliati, M. G., and N. A. Mazzeo, 2005: Air flow analysis in the upper Rio Negro Valley
(Argentina). Atmos. Res., 80, 263–279.
Colette, A., F. K. Chow, and R. L. Street, 2003: A numerical study of inversion-layer breakup and
the effects of topographic shading in idealized valleys. J. Appl. Meteor., 42, 1255–1272.
Coulter, R. L., and P. Gudiksen, 1995: The dependence of canyon winds on surface cooling and
external forcing in Colorado’s Front Range. J. Appl. Meteor., 34, 1419–1429.
Coulter, R. L., and T. J. Martin, 1996: Effects of stability on the profiles of vertical velocity and its
variance in katabatic flow. Bound.-Layer Meteor., 81, 23–33.
Coulter, R. L., M. Orgill, and W. Porch, 1989: Tributary fluxes into Brush Creek Valley. J. Appl.
Meteor., 28, 555–568.
Coulter, R. L., T. J. Martin, and W. M. Porch, 1991: A comparison of nocturnal drainage flow in
three tributaries. J. Appl. Meteor., 30, 157–169.
Cox, J. A. W., 2005: The sensitivity of thermally driven mountain flows to land-cover change.
Dissertation, Univ. Utah, Meteorology Department. 109pp.
Crawford, T. L., and R. J. Dobosy, 1992: A sensitive fast-response probe to measure turbulence
and heat flux from any airplane. Bound.-Layer Meteor., 59, 257–278.
Cuxart, J., 2008: Nocturnal basin low-level jets: An integrated study. Acta Geophysica, 56,
100–113.
Cuxart, J., and M. A. Jiménez, 2011: Deep radiation fog in a wide closed valley:
Study by numerical modeling and remote sensing. Pure Appl. Geophys. In press. DOI:
10.1007/s00024-011-0365-4.
Cuxart, J., M. A. Jiménez, and D. Martı́nez, 2007: Nocturnal meso-beta basin and katabatic flows
on a midlatitude island. Mon. Wea. Rev., 135, 918–932.
104 D. Zardi and C.D. Whiteman

Cuxart, J., J. Cunillera, M. A. Jiménez, D. Martı́nez, F. Molinos, and J. L. Palau, 2011:


Study of mesobeta basin flows by remote sensing. Bound.-Layer Meteor. In press. DOI:
10.1007/s10546-011-9655-8.
Dai, A., and J. Wang, 1999: Diurnal and semidiurnal tides in global surface pressure fields.
J. Atmos. Sci., 56, 3874–3891.
Darby, L. S., and R. M. Banta, 2006: The modulation of canyon flows by larger-scale influences.
Ext. Abstr., 12th Conf. Mountain Meteor., Santa Fe, NM. Amer. Meteor. Soc., Boston, MA.
Darby, L. S., K. J. Allwine, and R. M. Banta, 2006: Nocturnal low-level jet in a mountain basin
complex. Part II: Transport and diffusion of tracer under stable conditions. J. Appl. Meteor.
Climatol., 45, 740–753.
Defant, F., 1949: Zur Theorie der Hangwinde, nebst Bemerkungen zur Theorie der Berg- und
Talwinde. [A theory of slope winds, along with remarks on the theory of mountain winds
and valley winds]. Arch. Meteor. Geophys. Bioclimatol., Ser. A, 1, 421–450 (Theoretical and
Applied Climatology). [English translation: Whiteman, C.D., and E. Dreiseitl, 1984: Alpine
meteorology: Translations of classic contributions by A. Wagner, E. Ekhart and F. Defant. PNL-
5141/ASCOT-84-3. Pacific Northwest Laboratory, Richland, Washington, 121 pp].
Defant, F., 1951: Local Winds. In: Compendium of Meteorology, T. M. Malone (Ed.), Boston,
Amer. Meteor. Soc., 655–672.
de Foy, B., and Coauthors, 2008: Basin-scale wind transport during the MILAGRO field campaign
and comparison to climatology using cluster analysis. Atmos. Chem. Phys., 8, 1209–1224.
de Franceschi, M., 2004: Investigation of atmospheric boundary layer dynamics in Alpine valleys.
Ph.D. thesis in Environmental Engineering, University of Trento, Italy, 136pp. http://www.ing.
unitn.it/dica/eng/monographs/index.php.
de Franceschi, M., and D. Zardi, 2003: Evaluation of cut-off frequency and correction of filter-
induced phase lag and attenuation in eddy covariance analysis of turbulence data. Bound.-Layer
Meteor., 108, 289–303.
de Franceschi, M., and D. Zardi, 2009: Study of wintertime high pollution episodes during the
Brenner-South ALPNAP measurement campaign. Meteor. Atmos. Phys., 103, 237–250.
de Franceschi, M., G. Rampanelli, and D. Zardi, 2002: Further investigations of the “Ora del
Garda” valley wind. 10th Conf. Mount. Meteor. and MAP Meeting, 13–21 June 2002, Park
City, UT, Amer. Meteor. Soc., Boston, MA. Available at http://ams.confex.com/ams/pdfpapers/
39949.pdf.
de Franceschi, M., G. Rampanelli, D. Sguerso, D. Zardi, and P. Zatelli, 2003: Development of
a measurement platform on a light airplane and analysis of airborne measurements in the
atmospheric boundary layer. Ann. Geophys., 46, 269–283.
de Franceschi, M., D. Zardi, M. Tagliazucca, and F. Tampieri, 2009: Analysis of second order
moments in the surface layer turbulence in an Alpine valley. Quart. J. Roy. Meteor. Soc., 135,
1750–1765.
De Wekker, S. F. J., 2002: Structure and morphology of the convective boundary layer in
mountainous terrain. Dissertation, University of British Columbia, Department of Earth and
Ocean Sciences, Vancouver, Canada, 191pp.
De Wekker, S. F. J., 2008: Observational and numerical evidence of depressed convective boundary
layer heights near a mountain base. J. Appl. Meteor., 47, 1017–1026.
De Wekker, S. F. J., and S. D. Mayor, 2009: Observations of atmospheric structure and dynamics in
the Owens Valley of California with a ground-based, eye-safe, scanning aerosol lidar. J. Appl.
Meteor. Climatol., 48, 1483–1499.
De Wekker, S. F. J., and C. D. Whiteman, 2006: On the time scale of nocturnal boundary layer
cooling in valleys and basins and over plains. J. Appl. Meteor., 45, 813–820.
De Wekker, S. F. J., S. Zhong, J. D. Fast, and C. D. Whiteman, 1998: A numerical study of the
thermally driven plain-to-basin wind over idealized basin topographies. J. Appl. Meteor., 37,
606–622.
De Wekker, S. F. J., D. G. Steyn, J. D. Fast, M. W. Rotach, and S. Zhong, 2005: The performance of
RAMS in representing the convective boundary layer structure in a very steep valley. Environ.
Fluid Mech., 5, 35–62.
2 Diurnal Mountain Wind Systems 105

De Wekker S. F. J., A. Ameen, G. Song, B. B. Stephens, A. G. Hallar and I. B. McCubbin,


2009: A preliminary investigation of boundary layer effects on daytime atmospheric CO2
concentrations at a mountaintop location in the Rocky Mountains Acta Geophysica, 57,
904–922.
Denby, B., 1999: Second-order modelling of turbulence in katabatic flows. Bound.-Layer Meteor.,
92, 67–100.
Deng, X., and R. Stull, 2005: A mesoscale analysis method for surface potential temperature in
mountainous and coastal terrain. Mon. Wea. Rev., 133, 389–408.
Deng, X., and R. Stull, 2007: Assimilating surface weather observations from complex terrain into
a high-resolution numerical weather prediction model. Mon. Wea. Rev., 135, 1037–1054.
Dixit, P. N., and D. Chen, 2011: Effect of topography on farm-scale spatial variation in extreme
temperatures in the Southern Mallee of Victoria, Australia. Theor. Appl. Climatol., 103,
533–542.
Doran, J. C., 1991: The effect of ambient winds on valley drainage flows. Bound.-Layer Meteor.,
55, 177–189.
Doran, J. C., 1996: The influence of canyon winds on flow fields near Colorado’s Front Range.
J. Appl. Meteor., 35, 587–600.
Doran, J. C., and T. W. Horst, 1981: Velocity and temperature oscillations in drainage winds.
J. Appl. Meteor., 20, 361–364.
Doran, J. C., and S. Zhong, 2000: Thermally driven gap winds into the Mexico City basin. J. Appl.
Meteor., 39, 1330–1340.
Doran, J. C., M. L. Wesley, R. T. McMillen, and W. D. Neff, 1989: Measurements of turbulent heat
and momentum fluxes in a mountain valley. J. Appl. Meteor., 28, 438–444.
Doran, J. C., T. W. Horst, and C. D. Whiteman, 1990: The development and structure of nocturnal
slope winds in a simple valley. Bound.-Layer Meteor., 52, 41–68.
Doran, J. C., and Coauthors, 1998: The IMADA-AVER boundary-layer experiment in the Mexico
City area. Bull. Amer. Meteor. Soc., 79, 2497–2508.
Doran, J. C., J. D. Fast, and J. Horel, 2002: The VTMX 2000 campaign. Bull. Amer. Meteor. Soc.,
83, 537–551.
Dorninger, M., Whiteman, C. D., B. Bica, S. Eisenbach, B. Pospichal, and R. Steinacker, 2011:
Meteorological events affecting cold-air pools in a small basin. J. Appl. Meteor. Climatol.
Accepted.
Dosio, A., S. Emeis, G. Graziani, W. Junkermann, and A. Levy, 2001: Assessing the meteorological
conditions of a deep Italian Alpine valley system by means of a measuring campaign and
simulations with two models during a summer smog episode. Atmos. Environ., 35, 5441–5454.
Duguay, C. R., 1993: Radiation modeling in mountainous terrain: Review and status. Mount. Res.
Dev., 13, 339–357.
Eckman, R. M., 1998: Observations and numerical simulations of winds in a broad forested valley.
J. Appl. Meteor., 37, 206–219.
Egger, J., 1987: Valley winds and the diurnal circulation over plateaus. Mon. Wea. Rev., 115,
2177–2186.
Egger, J., 1990: Thermally forced flows: Theory. In W. Blumen (Ed.), Atmospheric Processes over
Complex Terrain. Amer. Meteor. Soc., Boston, 43–57.
Egger, J., S. Bajrachaya, U. Egger, R. Heinrich, J. Reuder, P. Shayka, H. Wendt, and V. Wirth,
2000: Diurnal winds in the Himalayan Kali Gandaki Valley. Part I: Observations. Mon. Wea.
Rev., 128, 1106–1122.
Egger, J., and Coauthors, 2002: Diurnal winds in the Himalayan Kali Gandaki valley. Part III:
Remotely piloted aircraft soundings. Mon. Wea. Rev., 130, 2042–2058.
Egger, J., and Coauthors, 2005: Diurnal circulation of the Bolivian Altiplano. Part I: Observations.
Mon. Wea. Rev., 133, 911–924.
Ekhart, E., 1944: Beiträge zur alpinen Meteorologie. [Contributions to Alpine Meteorology].
Meteor. Z., 61, 217–231. [English translation: Whiteman, C.D., and E. Dreiseitl, 1984: Alpine
Meteorology: Translations of Classic Contributions by A. Wagner, E. Ekhart and F. Defant.
PNL-5141/ASCOT-84-3. Pacific Northwest Laboratory, Richland, Washington, 121 pp].
106 D. Zardi and C.D. Whiteman

Ekhart, E., 1948: De la structure thermique de l’atmosphere dans la montagne [On the thermal
structure of the mountain atmosphere]. La Meteorologie, 4, 3–26. [English translation: White-
man, C.D., and E. Dreiseitl: 1984. Alpine Meteorology: Translations of Classic Contributions
by A. Wagner, E. Ekhart and F. Defant. PNL-5141/ASCOT-84-3. Pacific Northwest Laboratory,
Richland, Washington, 121 pp].
Emeis, S., C. Jahn, C. Münkel, C. Münsterer, and K. Schäfer, 2007: Multiple atmospheric layering
and mixing-layer height in the Inn valley observed by remote sensing. Meteor. Z., 16, 415–424.
Enger, L., D. Koračin, and X. Yang, 1993: A numerical study of boundary-layer dynamics in a
mountain valley. Part 1: Model validation and sensitivity experiment. Bound.-Layer Meteor.,
66, 357–394.
Fast, J. D., 1995: Mesoscale modeling and four-dimensional data assimilation in areas of highly
complex terrain. J. Appl. Meteor., 34, 2762–2782.
Fast, J. D., 2003: Forecasts of valley circulations using the terrain-following and step-mountain
vertical coordinates in the Meso-Eta model. Wea. Forecasting, 18, 1192–1206.
Fast, J. D., and S. Zhong, 1998: Meteorological factors associated with inhomogeneous ozone
concentrations within the Mexico City basin. J. Geophys. Res., 103 (D15), 18,927-18,946.
Fast, J. D., and L. S. Darby, 2004: An evaluation of mesoscale model predictions of down-valley
and canyon flows and their consequences using Doppler lidar measurements during VTMX
2000. J. Appl. Meteor., 43, 420–436.
Fast, J., S. Zhong, and C. D. Whiteman, 1996: Boundary layer evolution within a canyonland
basin. Part II. Numerical simulations of nocturnal flows and heat budgets. J. Appl. Meteor., 35,
2162–2178.
Fernando, H. J. S., 2010: Fluid dynamics of urban atmospheres in complex terrain. Ann. Rev. Fluid
Mech., 42, 365–89.
Fiedler, F., I. Bischoff-Gauß, N. Kalthoff, and G. Adrian, 2000: Modeling of the transport and
diffusion of a tracer in the Freiburg-Schauinsland area. J. Geophys. Res., 105, D1, 1599–1610.
Finnigan, J. J., R. Clement, Y. Malhi, R. Leuning, and H. A. Cleugh, 2003: A re-evaluation of long-
term flux measurement techniques. Part I. Averaging and coordinate rotation. Bound.-Layer
Meteor., 107, 1–48.
Fitzjarrald, D. R., 1984: Katabatic winds in opposing flow. J. Atmos. Sci., 41, 1143–1158.
Fleagle, R. G., 1950: A theory of air drainage. J. Meteor., 7, 227–232.
Frei, C., and H. C. Davies, 1993: Anomaly in the Alpine diurnal pressure signal: Observations and
theory. Quart. J. Roy. Meteor. Soc., 119, 1269–1289.
Freytag, C., 1988: Atmosphärische Grenzschicht in einem Gebirgstal bei Berg- und Talwind.
[Atmospheric boundary layer in a mountain valley during mountain and valley winds]. Wiss.
Mitteil. Nr. 60, Münchener Universitätsschriften, Facultät für Physik, Universität München –
Meteorologisches Institut, 197pp.
Fritts, D. C., D. Goldstein, and T. Lund, 2010: High-resolution numerical studies of stable
boundary layer flows in a closed basin: Evolution of steady and oscillatory flows in an
axisymmetric Arizona Meteor Crater. J. Geophys. Res., 115, D18109.
Furger, M., and Coauthors, 2000: The VOTALP Mesolcina Valley campaign 1996 – Concept,
background and some highlights. Atmos. Environ., 34, 1395–1412.
Garratt, J. R., and R. A. Brost, 1981: Radiative cooling effects within and above the nocturnal
boundary layer. J. Atmos. Sci., 38, 2730–2746.
Geerts, B., Q. Miao, and J. C. Demko, 2008: Pressure perturbations and upslope flow over a heated,
isolated mountain. Mon. Wea. Rev., 136, 4272–4288.
Geiger, R., R. H. Aron, and P. Todhunter, 2009: The Climate Near the Ground, 7th Edition. Roman
and Littlefield Publishers, Maryland, 523pp.
Giovannini, L., D. Zardi, and M. de Franceschi, 2011: Analysis of the urban thermal fingerprint of
the city of Trento in the Alps. J. Appl. Meteor. Climatol., 50, 1145–1162.
Gleeson, T. A., 1951: On the theory of cross-valley winds arising from differential heating of the
slopes. J. Meteor., 8, 398–405.
2 Diurnal Mountain Wind Systems 107

Gohm, A., F. Harnisch, J. Vergeiner, F. Obleitner, R. Schnitzhofer, A. Hansel, A. Fix, B. Neininger,


S. Emeis, and K. Schäfer, 2009: Air pollution transport in an Alpine valley: Results from
airborne and ground-based observations. Bound.-Layer Meteor., 131, 441–463.
Grell, G. A., S. Emeis, W. R. Stockwell, T. Schoenemey, R. Forkel, J. Michalakes, R. Knoche, and
W. Seidl, 2000: Application of a multiscale, coupled MM5/chemistry model to the complex
terrain of the VOTALP valley campaign. Atmos. Environ., 34, 1435–1453.
Grimmond, C. S. B., C. Souch, and M. D. Hubble, 1996: Influence of tree cover on summertime
surface energy balance fluxes, San Gabriel Valley, Los Angeles. Climate Research, 6, 45–57.
Grisogono, B., 2003: Post-onset behaviour of the pure katabatic flow. Bound.-Layer Meteor., 107,
157–175.
Grisogono, B., and J. Oerlemans, 2001a: A theory for the estimation of surface fluxes in simple
katabatic flow. Quart. J. Roy. Meteor. Soc., 127, 2125–2139.
Grisogono, B., and J. Oerlemans, 2001b: Katabatic flow: Analytic solution for gradually varying
eddy diffusivities. J. Atmos. Sci., 58, 3349–3354.
Grisogono, B., L. Kraljevič, and A. Jeričevič, 2007: The low-level katabatic jet height versus
Monin-Obukhov height. Quart. J. Roy. Meteor. Soc., 133, 2133–2136.
Gross, G., 1984: Eine Erklärung des Phänomens Maloja-Schlange Mittels Numerischer Simula-
tion. [An Explanation of the Maloja Snake (Cloud) Phenomenon by Means of a Numerical
Simulation]. Dissertation, Technischen Hochschule Darmstadt, Darmstadt, FRG.
Gross, G., 1985: An explanation of the Maloja-serpent by numerical simulation, Beitr. Phys.
Atmos., 58, 441–457.
Gross, G., 1990: On the wind field in the Loisach Valley–Numerical simulation and comparison
with the LOWEX III data. Meteor. Atmos. Phys., 42, 231–247.
Grossi, G., and L. Falappi, 2003: Comparison of energy fluxes at the land surface-atmosphere
interface in an Alpine valley as simulated with different models. Hydrology and Earth System
Sciences, 7, 920–936.
Guardans, R., and I. Palomino, 1995: Description of wind field dynamic patterns in a valley and
their relation to mesoscale and synoptic-scale situations. J. Appl. Meteor., 34, 49–67.
Gustavsson, T., M. Karlsson, J. Bogren, and S. Lindqvist, 1998: Development of temperature
patterns during clear nights. J. Appl. Meteor., 37, 559–571.
Ha, K.-J., and L. Mahrt, 2003: Radiative and turbulent fluxes in the nocturnal boundary layer.
Tellus, 55A, 317–327.
Haiden, T., 1998: Analytical aspects of mixed-layer growth in complex terrain. Preprints, Eighth
Conf. Mountain Meteorology, Amer. Meteor. Soc., Flagstaff, Arizona, 368–372.
Haiden, T., 2003: On the pressure field in the slope wind layer. J. Atmos. Sci., 60, 1632–1635.
Haiden, T., and C. D. Whiteman, 2005: Katabatic flow mechanisms on a low-angle slope. J. Appl.
Meteor., 44, 113–126.
Haiden, T., C. D. Whiteman, S. W. Hoch, and M. Lehner, 2011: A mass-flux model of nocturnal
cold air intrusions into a closed basin. J. Appl. Meteor. Climatol., 50, 933–943.
Harnisch, F., A. Gohm, A. Fix, R. Schnitzhofer, A. Hansel, and B. Neininger, 2009: Spatial
distribution of aerosols in the Inn Valley atmosphere during wintertime. Meteor. Atmos. Phys.,
103, 215–228.
Hauge, G., and L. R. Hole, 2003: Implementation of slope irradiance in Mesoscale Model version
5 and its effect on temperature and wind fields during the breakup of a temperature inversion.
J. Geophys. Res., 108, D2, 4058.
Hawkes, H. B., 1947: Mountain and valley winds with special reference to the diurnal mountain
winds of the Great Salt Lake region. Dissertation, Ohio State University, 312pp.
Helmis, C. G., and K. H. Papadopoulos, 1996: Some aspects of the variation with time of katabatic
flows over a simple slope. Quart. J. Roy. Meteor. Soc., 122, 595–610.
Henne, S., M. Furger, and A. S. H. Prévôt, 2005: Climatology of mountain venting-induced
elevated moisture layers in the lee of the Alps. J. Appl. Meteor., 44, 620–633.
Henne, S., W. Junkermann, J. M. Kariuki, J. Aseyo, and J. Klausen, 2008: The establishment of
the Mt. Kenya GAW station: Installation and meteorological characterization. J. Appl. Meteor.
Climatol., 47, 2946–2962.
108 D. Zardi and C.D. Whiteman

Higgins, W., and Coauthors, 2006: The NAME 2004 field campaign and modeling strategy. Bull.
Amer. Meteor. Soc., 87, 79–94.
Hiller, R., M. Zeeman, and W. Eugster, 2008: Eddy-covariance flux measurements in the complex
terrain of an Alpine valley in Switzerland. Bound.-Layer Meteor., 127, 449–467.
Hoch, S. W., and C. D. Whiteman, 2010: Topographic effects on the surface radiation balance in
and around Arizona’s Meteor Crater. J. Appl. Meteor. Climat., 49, 1894–1905.
Hoch, S. W., C. D. Whiteman, and B. Mayer, 2011: A systematic study of longwave radiative
heating and cooling within valleys and basins using a three-dimensional radiative transfer
model. J. Appl. Meteor. Climatol. In press.
Holton, J. R., 1966: The diurnal boundary layer wind oscillation above sloping terrain. Tellus, 19,
199–205.
Horst, T. W., and J. C. Doran, 1986: Nocturnal drainage flow on simple slopes. Bound.-Layer
Meteor., 34, 263–286.
Horst, T. W., and J. C. Doran, 1988: The turbulence structure of nocturnal slope flows. J. Atmos.
Sci., 45, 605–616.
Hughes, M., A. Hall, and G. Fovell, 2007: Dynamical controls on the diurnal cycle of temperature
in complex topography, Clim. Dyn., 29, 277–292.
Hunt, J. C. R., H. J. S. Fernando, and M. Princevac, 2003: Unsteady thermally driven flows on
gentle slopes. J. Atmos. Sci., 60, 2169–2182.
IIjima, Y., and M. Shinoda, 2000: Seasonal changes in the cold-air pool formation in a subalpine
hollow, central Japan. Int. J. Climatol., 20, 1471–1483.
Iziomon, M. G., H. Mayer, W. Wicke, and A. Matzarakis, 2001: Radiation balance over low-lying
and mountainous areas in southwest Germany. Theor. Appl. Climatol., 68, 219–231.
Jiang, Q., and J. D. Doyle, 2008: Diurnal variation of downslope winds in Owens Valley during
the Sierra Rotor experiment. Mon. Wea. Rev., 136, 3760–3780.
Kaimal, J. C., and J. J. Finnigan, 1994: Atmospheric boundary-layer flows: Their structure and
measurement. Oxford University Press, New York, 289pp.
Kalthoff, N., and Coauthors, 2002: Mesoscale wind regimes in Chile at 30ı S. J. Appl. Meteor., 41,
953–970.
Kalthoff, N., M. Fiebig-Wittmaack, C. Meißner, M. Kohler, M. Uriarte, I. Bischoff-Gauß, and E.
Gonzales, 2006: The energy balance, evapo-transpiration and nocturnal dew deposition of an
arid valley in the Andes. J. Arid Environments, 65, 420–443.
Kalthoff, N., and Coauthors, 2009: The impact of convergence zones on the initiation of deep
convection: A case study from COPS. Atmos. Res., 93, 680 – 694.
Kastendeuch, P. P., and P. Kaufmann, 1997: Classification of summer wind fields over complex
terrain. Int. J. Climatol., 17, 521–534.
Kastendeuch, P. P., P. Lacarrere, G. Najjar, J. Noilhan, F. Gassmann, and P. Paul, 2000: Mesoscale
simulations of thermodynamic fluxes over complex terrain. Int. J. Climatol., 20, 1249–1264.
Kaufmann, P., and R. O. Weber, 1996: Classification of mesoscale wind fields in the MISTRAL
field experiment. J. Appl. Meteor., 35, 1963–1979.
Kaufmann, P., and R. O. Weber, 1998: Directional correlation coefficient for channeled flow and
application to wind data over complex terrain. J. Atmos. Oceanic. Technol., 15, 89–97.
Kaufmann, P., and C. D. Whiteman, 1999: Cluster-analysis classification of wintertime wind
patterns in the Grand Canyon region. J. Appl. Meteor., 38, 1131–1147.
Kavčič, I., and B. Grisogono, 2007: Katabatic flow with Coriolis effect and gradually varying eddy
diffusivity. Bound.-Layer Meteor., 125,377–387.
Kelly, R. D., 1988: Asymmetric removal of temperature inversions in a high mountain valley.
J. Appl. Meteor., 27, 664–673.
Khodayar, S., N. Kalthoff, M. Fiebig-Wittmaack, and M. Kohler, 2008: Evolution of the atmo-
spheric boundary-layer structure of an arid Andes valley. Meteor. Atmos. Phys., 99, 181–198.
Kimura, F., and T. Kuwagata, 1993: Thermally induced wind passing from plain to basin over a
mountain range. J. Appl. Meteor., 32, 1538–1547.
Kimura, F., and T. Kuwagata, 1995: Horizontal heat fluxes over complex terrain computed using a
simple mixed-layer model and a numerical model. J. Appl. Meteor., 34, 549–558.
2 Diurnal Mountain Wind Systems 109

King, C. W., 1989: Representativeness of single vertical wind profiles for determining volume
fluxes in valleys. J. Appl. Meteor., 28, 463–466.
King, J. A., F. H. Shair, and D. D. Reible, 1987: The influence of atmospheric stability on pollutant
transport by slope winds. Atmos. Environ., 21, 53–59.
Kitada, T., and R. P. Regmi, 2003: Dynamics of air pollution transport in late wintertime over
Kathmandu Valley, Nepal: As revealed with numerical simulation. J. Appl. Meteor., 42,
1770–1798.
Koch, H. G., 1961: Die warme Hangzone. Neue Anschauungen zur nächtlichen Kaltluftschichtung
in Tälern und an Hängen. [The warm slope zone. New views of the nocturnal cold air layer in
valleys and on slopes]. Z. Meteor., 15, 151–171.
Kolev, N., I. Grigorov, I. Kolev, P. C. S. Devara, P. E. Raj, and K. K. Dan, 2007: Lidar and sun
photometer observations of atmospheric boundary-layer characteristics over an urban area in a
mountain valley. Bound.-Layer Meteor., 124, 99–115.
Kondo, H., 1990: A numerical experiment on the interaction between sea breeze and valley wind
to generate the so-called “extended sea breeze.” J. Meteor. Soc. Japan, 68, 435–446.
Kondo, J., and N. Okusa, 1990: A simple numerical prediction model of nocturnal cooling in a
basin with various topographic parameters. J. Appl. Meteor., 29, 604–619.
Kondo, J., T. Kuwagata, and S. Haginoya, 1989: Heat budget analysis of nocturnal cooling and
daytime heating in a basin. J. Atmos. Sci., 19, 2917–2933.
Koračin, D., and L. Enger, 1994: A numerical study of boundary-layer dyamics in a mountain
valley. Part 2: Observed and simulated characteristics of atmospheric stability and local flows.
Bound.-Layer Meteor., 69, 249–283.
Kossmann, M., and A. P. Sturman, 2003: Pressure-driven channeling effects in bent valleys. J. Appl.
Meteor., 42, 151–158.
Kossman, M., A. P. Sturman, P. Zawar-Reza, H. A. McGowan, A. J. Oliphant, I. F. Owens, and R.
A. Spronken-Smith, 2002: Analysis of the wind field and heat budget in an alpine lake basin
during summertime fair weather conditions. Meteor. Atmos. Phys., 81, 27–52.
Küttner, J., 1949: Periodische Luftlawinen. [Periodic air avalanches]. Meteor. Rundsch., 2,
183–184.
Kuttler, W., A.-B. Barlag, and F. Roßmann, 1996: Study of the thermal structure of a town in a
narrow valley. Atmos. Environ., 30, 365–378.
Kuwagata, T., and J. Kondo, 1989: Observation and modeling of thermally induced upslope flow.
Bound.-Layer Meteor., 49, 265–293.
Kuwagata, T., and F. Kimura, 1995: Daytime boundary layer evolution in a deep valley. Part I:
Observations in the Ina Valley. J. Appl. Meteor., 34, 1082–1091.
Kuwagata, T., and F. Kimura, 1997: Daytime boundary layer evolution in a deep valley. Part II:
Numerical simulation of the cross-valley circulation. J. Appl. Meteor., 36, 883–895.
Lee, I. Y., R. L. Coulter, H. M. Park, and J.-H. Oh, 1995: Numerical simulation of nocturnal
drainage flow properties in a rugged canyon. Bound.-Layer Meteor., 72, 305–321.
Lee, S.-M., W. Giori, M. Princevac, and H. J. S. Fernando, 2006: Implementation of a stable PBL
turbulence parameterization for the mesoscale model MM5: Nocturnal flow in complex terrain.
Bound.-Layer Meteor., 119, 109–134.
Lee, X., and X. Hu, 2002: Forest-air fluxes of carbon, water and energy over non-flat terrain.
Bound.-Layer Meteor, 103, 277–301.
Lehner, M., and A. Gohm, 2010: Idealised simulations of daytime pollution transport in a steep
valley and its sensitivity to thermal stratification and surface albedo. Bound.-Layer Meteor.,
134, 327–351.
Lehner, M., C. D. Whiteman, and S. W. Hoch, 2010: Diurnal cycle of thermally driven cross-basin
winds in Arizona’s Meteor Crater. J. Appl. Meteor. Climatol., 50, 729–744.
Leone, J. M., Jr., and R. L. Lee, 1989: Numerical simulation of drainage flow in Brush Creek,
Colorado. J. Appl. Meteor., 28, 530–542.
Li, J.-G., and B. W. Atkinson, 1999: Transition regimes in valley airflows. Bound.-Layer Meteor.,
91, 385–411.
110 D. Zardi and C.D. Whiteman

Li, Y., and R. B. Smith, 2010: The detection and significance of diurnal pressure and potential
vorticity anomalies east of the Rockies. J. Atmos. Sci., 67, 2734–2751.
Li, Y., R. B. Smith, and V. Grubišić, 2009: Using surface pressure variations to categorize the
diurnal valley circulations: Experiments in Owens Valley. Mon. Wea. Rev., 137, 1753–1769.
Lindkvist, L., and S. Lindqvist, 1997: Spatial and temporal variability of nocturnal summer frost
in elevated complex terrain. Agric. Forest Meteor., 87, 139–153.
Lindkvist, L., T. Gustavsson, and J. Bogren, 2000: A frost assessment method for mountainous
areas. Agric. Forest Meteor., 102, 51–67.
Liu, X., A. Bai, and C. Liu, 2009: Diurnal variations of summertime precipitation over the
Tibetan Plateau in relation to orographically-induced regional circulations. Environ. Res. Lett.,
4, 045203.
Löffler-Mang, M., M. Kossmann, R. Vögtlin, and F. Fiedler, 1997: Valley wind systems and their
influences on nocturnal ozone concentrations. Beitr. Phys. Atmosph., 70, 1–14.
Lu, R., and R. P. Turco, 1995: Air pollutant transport in a coastal environment – II. Three-
dimensional simulations over Los Angeles basin. Atmos. Environ., 29, 1499–1518.
Ludwig, F. L., J. Horel, and C. D. Whiteman, 2004: Using EOF analysis to identify important
surface wind patterns in mountain valleys. J. Appl. Meteor., 43, 969–983.
Lugauer, M., and P. Winkler, 2005: Thermal circulation in South Bavaria – climatology and
synoptic aspects. Meteor. Z., 14, 15–30.
Lugauer, M., and Coauthors, 2003: An overview of the VERTIKATOR project and results of
Alpine pumping. Extended Abstracts, Vol. A, International Conf. on Alpine Meteor. and MAP-
Meeting 2003, May 19–23, 2003, Brig, Switzerland. Publications of MeteoSwiss, No. 66,
129–132.
Luhar, A. K., and S. K. Rao, 1993: Random-walk model studies of the transport and diffusion of
pollutants in katabatic flows. Bound.-Layer Meteor., 66, 395–412.
Magono, C., C. Nakamura, and Y. Yoshida, 1982: Nocturnal cooling of the Moshiri Basin,
Hokkaido in midwinter. J. Meteor. Soc. Japan, 60, 1106–1116.
Mahrt, L., 1982: Momentum balance of gravity flows. J. Atmos. Sci., 39, 2701–2711.
Mahrt, L., 1999: Stratified atmospheric boundary layers. Bound.-Layer Meteor., 90, 375–396.
Mahrt, L., 2006: Variation of surface air temperature in complex terrain. J. Appl. Meteor. Climatol.,
45, 1481–1493.
Mahrt, L. J., and S. Larsen, 1982: Small scale drainage flow. Tellus, 34, 579–587.
Mahrt, L., D. Vickers, R. Nakamura, M. R. Soler, J. Sun, S. Burns, and D. H. Lenschow, 2001:
Shallow drainage flows. Bound.-Layer Meteor., 101, 243–260.
Maki, M., and T. Harimaya, 1988: The effect of advection and accumulation of downslope cold air
on nocturnal cooling in basins. J. Meteor. Soc. Japan, 66, 581–597.
Maki, M., T. Harimaya, and K. Kikuchi, 1986: Heat budget studies on nocturnal cooling in a basin.
J. Meteor. Soc. Japan, 64, 727–740.
Manins, P. C., 1992: Vertical fluxes in katabatic flows. Bound.-Layer Meteor., 60, 169–178.
Manins, P. C., and B. L. Sawford, 1979a: Katabatic winds: A field case study. Quart. J. Roy. Meteor.
Soc., 105, 1011–1025.
Manins, P. C., and B. L. Sawford, 1979b: A model of katabatic winds. J. Atmos. Sci., 36, 619–630.
Mannouji, N., 1982: A numerical experiment on the mountain and valley winds. J. Meteorol. Soc.
Jap., 60, 1085–1105.
Martı́nez, D., 2011: Topographically induced flows and nocturnal cooling in the atmospheric
boundary layer. Doctoral dissertation. Department of Physics, University of the Balearic
Islands.
Martı́nez, D., and J. Cuxart, 2009: Assessment of the hydraulic slope flow approach using a
mesoscale model. Acta Geophys., 57, 882–903.
Martı́nez, D., J. Cuxart, and J. Cunillera, 2008: Conditioned climatology for stably stratified nights
in the Lleida area. Tethys, 5, 13–24.
Martı́nez, D., M. A. Jiménez, J. Cuxart, and L. Mahrt, 2010: Heterogeneous nocturnal cooling in a
large basin under very stable conditions. Bound.-Layer Meteor., 137, 97–113.
2 Diurnal Mountain Wind Systems 111

Marty, C., R. Philipona, C. Fröhlich, and A. Ohmura, 2002: Altitude dependence of surface
radiation fluxes and cloud forcing in the alps: results from the alpine surface radiation budget
network. Theor. Appl. Climatol., 72, 137–155.
Matzinger, N., M. Andretta, E. van Gorsel, R. Vogt, A. Ohmura, and M. W. Rotach, 2003: Surface
radiation budget of an Alpine valley. Quart. J. Roy. Meteor. Soc., 129, 877–895.
Mayer, B., 2009: Radiative transfer in the cloudy atmosphere. Eur. Phys. J. Conf., 1, 75–99.
Mayer, B., and A. Kylling, 2005: Technical note: The libRadtran software package for radiative
transfer calculations – Description and examples of use. Atmos. Chem. Phys., 5, 1855–1877.
Mayer, B., S. W. Hoch, and C. D. Whiteman, 2010: Validating the MYSTIC three-dimensional
radiative transfer model with observations from the complex topography of Arizona’s Meteor
Crater. Atmos. Chem. Phys. Discuss., 10, 13373–13405.
McGowan, H. A., 2004: Observations of anti-winds in a deep Alpine valley, Lake Tekapo, New
Zealand. Arctic, Antarctic, and Alpine Res., 36, 495–501.
McGowan, H. A., and A. P. Sturman, 1996: Interacting multi-scale wind systems within an alpine
basin, Lake Tekapo. New Zealand. Meteor. Atmos. Phys., 58, 165–177.
McGowan, H. A., I. F. Owens, and A. P. Sturman, 1995: Thermal and dynamic characteristics of
alpine lake breezes, Lake Tekapo, New Zealand. Bound.-Layer Meteor., 76, 3–24.
McKee, T. B., and R. D. O’Neal, 1989: The role of valley geometry and energy budget in the
formation of nocturnal valley winds. J. Appl. Meteor., 28, 445–456.
McNider, R. T., 1982: A note on velocity fluctuations in drainage flows. J. Atmos. Sci., 39,
1658–1660.
Monti, P., H. J. S. Fernando, M. Princevac, W. C. Chan, T. A. Kowalewski, and E. R. Pardyjak,
2002: Observations of flow and turbulence in the nocturnal boundary layer over a slope.
J. Atmos. Sci., 59, 2513–2534.
Moraes, O. L. L., O. C. Acevedo, R. Da Silva, R. Magnago, and A. C. Siqueira, 2004: Nocturnal
surface-layer characteristics at the bottom of a valley. Bound.-Layer Meteor., 112, 159–177.
Müller, M. D., and D. Scherer, 2005: A grid- and subgrid-scale radiation parameterization of
topographic effects for mesoscale weather forecast models. Mon. Wea. Rev., 133, 1431–1442.
Mursch-Radlgruber, E., 1995: Observations of flow structure in a small forested valley system.
Theor. Appl. Climatol., 52, 3–17.
Neff, W. D., and C. W. King, 1989: The accumulation and pooling of drainage flows in a large
basin. J. Appl. Meteor., 28, 518–529.
Nickus, U., and I. Vergeiner, 1984: The thermal structure of the Inn Valley atmosphere. Arch.
Meteor. Geophys. Bioclimatol., Ser. A, 33, 199–215.
Noppel, H., and F. Fiedler, 2002: Mesoscale heat transport over complex terrain by slope winds –
A conceptual model and numerical simulations. Bound.-Layer Meteor., 104, 73–97.
O’Steen, L. B., 2000: Numerical simulation of nocturnal drainage flows in idealized valley-
tributary systems. J. Appl. Meteor., 39, 1845–1860.
Oerlemans, J., 1998: The atmospheric boundary layer over melting glaciers. Chapter 6 in: Clear
and Cloudy Boundary Layers (Eds: A. A. M. Holtslag and P. G. Duynkerke), Royal Netherlands
Academy of Arts and Sciences,129–153.
Oerlemans, J., and B. Grisogono, 2002: Glacier winds and parameterisation of the related surface
heat fluxes. Tellus A, 54, 440–452.
Oerlemans, J., H. Björnsson, M. Kuhn, F. Obleitner, F. Palsson, C. J. P. P. Smeets, H. F. Vugts,
and J. De Wolde, 1999: Glacio-meteorological investigations on Vatnajökull, Iceland, summer
1996: An overview. Bound.-Layer Meteor., 92, 3–26.
Oliphant, A. J., R. A. Spronken-Smith, A. P. Sturman, and I. F. Owens, 2003: Spatial variability of
surface radiation fluxes in mountainous terrain. J. Appl. Meteor., 42, 113–128.
Oliver, H. R., 1992: Studies of the energy budget of sloping terrain. Int. J. Climatol., 12, 55–68.
Olyphant, G. A., 1986: Longwave radiation in mountainous areas and its influence on the energy
balance of alpine snowfields. Water Resources Res., 22, 62–66.
Oncley, S. P., and Coauthors, 2007: The Energy Balance Experiment EBEX-2000. Part I: Overview
and energy balance. Bound.-Layer Meteor., 123, 1–28.
112 D. Zardi and C.D. Whiteman

Orgill, M. M., J. D. Kincheloe, and R. A. Sutherland, 1992: Influences on nocturnal valley drainage
winds in Western Colorado valleys. J. Appl. Meteor., 31, 121–141.
Pamperin, H., and G. Stilke, 1985: Nächtliche Grenzschicht und LLJ im Alpenvorland nahe dem
Inntalausgang. [Nocturnal boundary layer and low level jet near the Inn Valley exit]. Meteor.
Rundsch., 38, 145–156.
Panday, A. K., 2006: The diurnal cycle of air pollution in the Kathmandu Valley, Nepal.
Dissertation No. 37361. Massachusetts Institute of Technology. [http://hdl.handle.net/1721.1/
37361]
Panday, A. K., and R. G. Prinn, 2009: Diurnal cycle of air pollution in the Kathmandu Valley,
Nepal: Observations, J. Geophys. Res., 114, D09305.
Panday, A. K., R. G. Prinn, and C. Schär, 2009: Diurnal cycle of air pollution in the Kathmandu
Valley, Nepal: 2. Modeling results. J. Geophys. Res., 114, D21308.
Papadopoulos, K. H., and C. G. Helmis, 1999: Evening and morning transition of katabatic flows.
Bound.-Layer Meteor., 92, 195–227.
Papadopoulos, K. H., C. G. Helmis, A. T. Soilemes, J. Kalogiros, P. G. Papageorgas, and
D. N. Asimakopoulos, 1997: The structure of katabatic flows down a simple slope. Quart.
J. Roy. Meteor. Soc., 123, 1581–1601.
Parish, T. R., and L. D. Oolman, 2010: On the role of sloping terrain in the forcing of the Great
Plains low-level jet. J. Atmos. Sci., 67, 2690–2699.
Petkovšek, Z., 1992: Turbulent dissipation of cold air lake in a basin. Meteor. Atmos. Phys., 47,
237–245.
Philipona, R., B. Dürr, and C. Marty, 2004: Greenhouse effect and altitude gradients over the
Alps – by surface longwave radiation measurements and model calculated LOR. Theor. Appl.
Climatol., 77, 1–7.
Pielke, R. A., J. H. Rodriguez, J. L. Eastman, R. L. Walko, and R. A. Stocker, 1993: Influence of
albedo variability in complex terrain on mesoscale systems. J. Climate, 6, 1798–1806.
Pinto, J. O., D. B. Parsons, W. O. J. Brown, S. Cohn, N. Chamberlain, and B. Morley, 2006:
Coevolution of down-valley flow and the nocturnal boundary layer in complex terrain. J. Appl.
Meteor. Climatol., 45, 1429–1449.
Piringer, M., and K. Baumann, 1999: Modifications of a valley wind system by an urban area –
Experimental results. Meteor. Atmos. Phys., 71, 117–125.
Plüss, C., and A. Ohmura, 1997: Longwave radiation on snow-covered mountainous surfaces.
J. Appl. Meteor., 36, 818–824.
Porch, W. M., R. B. Fritz, R. L. Coulter, and P. H. Gudiksen, 1989: Tributary, valley and sidewall
air flow interactions in a deep valley. J. Appl. Meteor., 28, 578–589.
Porch, W. M., W. E. Clements, and R. L. Coulter, 1991: Nighttime valley waves. J. Appl. Meteor.,
30, 145–156.
Pope, D., and C. Brough, 1996: Utah’s Weather and Climate. Publishers Press, 245pp.
Pospichal, B., 2004: Struktur und Auflösung von Temperaturinversionen in Dolinen am Beispiel
Grünloch [Structure and breakup of temperature inversions in Dolines, with an example from
the Grünloch]. Diplomarbeit, Instit. Meteor. u. Geophysik, Univ. Wien, 69pp.
Pospichal, B., S. Eisenbach, C. D. Whiteman, R. Steinacker, and M. Dorninger, 2003: Observations
of the cold air outflow from a basin cold pool through a low pass. Ext. Abstr., Vol. A, Intl Conf.
Alpine Meteor. and the MAP-Meeting 2003, Brig, Switzerland, 19–23 May 2003, 153–156.
Available as MeteoSwiss Publication 66, MeteoSwiss, Zurich, Switzerland.
Poulos, G. S., and S. Zhong, 2008. The observational history of small-scale katabatic winds in
mid-latitudes. Geography Compass, 2, 1–24.
Poulos, G. S., J. E. Bossert, T. B. McKee, and R. A. Pielke, 2000: The interaction of katabatic
flow and mountain waves. Part I: Observations and idealized simulations. J. Atmos. Sci., 57,
1919–1936.
Poulos, G. S., and Coauthors, 2002: CASES-99: A comprehensive investigation of the stable
nocturnal boundary layer. Bull. Amer. Meteor. Soc., 83, 555–581.
2 Diurnal Mountain Wind Systems 113

Poulos, G. S., J. E. Bossert, T. B. McKee, and R. A. Pielke, Sr., 2007: The interaction of katabatic
flow and mountain waves. Part II: Case study analysis and conceptual model. J. Atmos. Sci.,
64, 1857–1879.
Prabha, T. V., and E. Mursch-Radlgruber, 1999: Modeling of diffusion in a wide Alpine valley.
Theor. Appl. Climatol. 64, 93–103.
Prandtl, L., 1942: Strömungslehre [Flow Studies]. Vieweg und Sohn, Braunschweig, 382pp.
Prévôt, A. S. H., J. Staehelin, H. Richner, and T. Griesser, 1993: A thermally driven wind system
influencing concentrations of ozone precursors and photo-oxidants at a receptor site in the
Alpine foothills. Meteor. Z., N.F., 2, 167–177.
Prévôt, A. S. H., J. Dommen, M. Bäumle, and M. Furger, 2000: Diurnal variations of volatile
organic compounds and local circulation systems in an Alpine valley. Atmos. Environ., 34,
1413–1423.
Princevac, M., and H. J. S. Fernando, 2007: A criterion for the generation of turbulent anabatic
flows. Phys. Fluids, 19, 105102.
Princevac, M., and H. J. S. Fernando, 2008: Morning breakup of cold pools in complex terrain.
J. Fluid Mech., 616, 99–109.
Princevac, M., H. J. S. Fernando, and C. D. Whiteman, 2005: Turbulent entrainment into natural
gravity-driven flows. J. Fluid Mech., 533, 259–268.
Princevac, M., J. C. R. Hunt, and H. J. S. Fernando, 2008: Quasi-steady katabatic winds on slopes
in wide valleys: Hydraulic theory and observations. J. Atmos. Sci., 65, 627–643.
Räisänen, P., 1996: The effect of vertical resolution on clear-sky radiation calculations: tests with
two schemes. Tellus, 48A, 403–423.
Rakovec, J., J. Merše, S. Jernej, and B. Paradiž, 2002: Turbulent dissipation of the cold-air pool
in a basin: Comparison of observed and simulated development. Meteor. Atmos. Phys., 79,
195–213.
Ramanathan, N., and K. Srinivasan, 1998: Simulation of airflow in Kashmir valley for a summer
day. J. Appl. Meteor., 37, 497–508.
Rampanelli, G., 2004: Diurnal thermally-induced flows in Alpine valleys. Ph.D. thesis in Environ-
mental Engineering, University of Trento, Italy.
Rampanelli, G., and D. Zardi, 2004: A method to determine the capping inversion of the convective
boundary layer. J. Appl. Meteor., 43, 925–933.
Rampanelli, G., D. Zardi, and R. Rotunno, 2004: Mechanisms of up-valley winds. J. Atmos. Sci.,
61, 3097–3111.
Rao, S. K., and M. A. Schaub, 1990: Observed variations of ¢ ™ and ¢ ¥ in the nocturnal drainage
flow in a deep valley. Bound.-Layer Meteor., 51, 31–48.
Rao, K. S., and H. F. Snodgrass, 1981: A nonstationary nocturnal drainage flow model. Bound.-
Layer Meteor., 20, 309–320.
Reeves, H. D., and D. J. Stensrud, 2009: Synoptic-scale flow and valley cold pool evolution in the
Western United States. Wea. Forecasting, 24, 1625–1643.
Regmi, R. P., T. Kitada, and G. Kurata, 2003: Numerical simulation of late wintertime local flows in
Kathmandu Valley, Nepal: Implication for air pollution transport. J. Appl. Meteor., 42, 389–403.
Reiter, E. R., and M. Tang, 1984: Plateau effects on diurnal circulation patterns. Mon. Wea. Rev.,
112, 638–651.
Reuder, J., and J. Egger, 2006: Diurnal circulation of the South American Altiplano: Observations
in a valley and at a pass. Tellus A, 58, 254–262.
Reuten, C., 2008: Uplsope flow systems: Scaling, structure, and kinematics in tank and atmosphere.
VDM Verlag Dr. Mueller e.K., Saarbrücken, Germany, 200pp.
Reuten, C., D. G. Steyn, K. B. Strawbridge, and P. Bovis, 2005: Observations of the relation
between upslope flows and the convective boundary layer in steep terrain. Bound.-Layer
Meteor., 116, 37–61.
Reuten, C., D. G. Steyn, and S. E. Allen, 2007: Water tank studies of atmospheric boundary layer
structure and air pollution transport in upslope flow systems. J. Geophys. Res., 112, D11114.
Rotach, M. W., and D. Zardi, 2007: On the boundary layer structure over complex terrain: Key
findings from MAP. Quart. J. Roy. Meteor. Soc., 133, 937–948.
114 D. Zardi and C.D. Whiteman

Rotach M. W., and Coauthors, 2004: Turbulence structure and exchange processes in an Alpine
valley: The Riviera project. Bull. Amer. Meteor. Soc., 85, 1367–1385.
Rotach, M. W., M. Andretta, P. Calanca, A. P. Weigel, and A. Weiss, 2008: Turbulence character-
istics and exchange mechanisms in highly complex terrain. Acta Geophysicae, 56, 194–219.
Rucker, M., R. M. Banta, and D. G. Steyn, 2008: Along-valley structure of daytime thermally
driven flows in the Wipp Valley. J. Appl. Meteor. Climatol., 47, 733–751.
Sakiyama, S. K., 1990: Drainage flow characteristics and inversion breakup in two Alberta
mountain valleys. J. Appl. Meteor., 29, 1015–1030.
Salerno, R., 1992: Analysis of flow and pollutant dispersion by tracer experiments in south Alpine
valleys. Theor. Appl. Climatol. 45, 19–35.
Sasaki, T., and Coauthors, 2004: Vertical moisture transport above the mixed layer around the
mountains in western Sumatra. Geophys. Res. Lett., 31, L08106.
Savage, L. C., III, S. Zhong, W. Yao, W. J. O. Brown, T. W. Horst, and C. D. Whiteman, 2008:
An observational and numerical study of a regional-scale downslope flow in northern Arizona.
J. Geophys. Res., 113, D14114.
Savijärvi H., and J. Liya, 2001: Local winds in a valley city. Bound.-Layer Meteor., 100, 301–319.
Schicker, I., and P. Seibert, 2009. Simulation of the meteorological conditions during a winter smog
episode in the Inn Valley. Meteor. Atmos. Phys., 103, 203–214.
Schmidli, J., and R. Rotunno, 2010: Mechanisms of along-valley winds and heat exchange over
mountainous terrain. J. Atmos. Sci., 67, 3033–3047.
Schmidli, J., G. S. Poulos, M. H. Daniels, and F. K. Chow, 2009a: External Influences on nocturnal
thermally driven flows in a deep valley. J. Appl. Meteor. Climatol., 48, 3–23.
Schmidli, J., B. Billings, F. K. Chow, J. Doyle, V. Grubišić, T. Holt, Q. Jiang, K. A. Lundquist,
P. Sheridan, S. Vosper, S. F. J. de Wekker, C. D. Whiteman, A. A. Wyszogrodzki, and G. Zängl,
2010: Intercomparison of mesoscale model simulations of the daytime valley wind system.
Mon. Wea. Rev., 139, 1389–1409.
Schumann, U., 1990: Large-eddy simulation of the up-slope boundary layer. Quart. J. Roy. Meteor.
Soc., 116, 637–670.
Serafin, S., 2006: Boundary-layer processes and thermally driven flows over complex terrain. Ph.D.
thesis in Environmental Engineering, Univ. of Trento, Italy, 200pp. http://www.ing.unitn.it/
dica/eng/monographs/index.php.
Serafin, S., and D. Zardi, 2010a: Structure of the atmospheric boundary layer in the vicinity of a
developing upslope flow system: A numerical model study. J Atmos. Sci., 67, 1171–1185.
Serafin, S., and D. Zardi, 2010b: Daytime heat transfer processes related to slope flows and
turbulent convection in an idealized mountain valley. J Atmos. Sci., 67, 3739–3756.
Serafin, S., and D. Zardi, 2010c: Daytime development of the boundary layer over a plain and in a
valley under fair weather conditions: a comparison by means of idealized numerical simulations
J Atmos. Sci., 68, 2128–2141.
Shapiro, A., and E. Fedorovich, 2007: Katabatic flow along a differentially cooled sloping surface.
J. Fluid Mech., 571, 149–175.
Shapiro, A., and E. Fedorovich, 2008: Coriolis effects in homogeneous and inhomogeneous
katabatic flows. Quart. J. Roy. Meteor. Soc., 134, 353–370.
Shapiro, A., and E. Fedorovich, 2009: Nocturnal low-level jet over a shallow slope. Acta
Geophysica, 57, 950–980.
Sharp, J., and C. F. Mass, 2004: Columbia Gorge gap winds: Their climatological influence and
synoptic evolution. Wea. Forecasting, 19, 970–992.
Sievers, U., 2005: Das Kaltluftabflussmodell KLAM 21 – Theoretische Grundlagen, Anwendung
und Handhabung des PC-Modells [The cold air drainage model KLAM 21 – Theoretical basis,
application, and operation of the PC model]. Berichte des Deutschen Wetterdienstes no. 227,
Offenbach am Main, Germany, 101pp.
Simpson, J. E., 1999: Gravity currents, 2nd Edition. Cambridge Univ. Press, 259pp.
Skyllingstad, E. D., 2003: Large-eddy simulation of katabatic flows. Bound.-Layer Meteor., 106,
217–243.
2 Diurnal Mountain Wind Systems 115

Smith, C. M., and E. D. Skyllingstad, 2005: Numerical simulation of katabatic flow with changing
slope angle. Mon. Wea. Rev., 133, 3065–3080.
Smith, R., and Coauthors, 1997: Local and remote effects of mountains on weather: research needs
and opportunities. Bull. Amer. Meteor. Soc., 78, 877–892.
Soler, M. R., C. Infante, P. Buenestado, and L. Mahrt, 2002: Observations of nocturnal drainage
flow in a shallow gully. Bound.-Layer Meteor., 105, 253–273.
Spengler, T., J. H. Schween, M. Ablinger, G. Zängl, and J. Egger, 2009: Thermally driven flows at
an asymmetric valley exit: Observations and model studies at the Lech Valley exit. Mon. Wea.
Rev., 137, 3437–3455.
Start, G. E., C. R. Dickson, and L. L. Wendell, 1975: Diffusion in a canyon within rough
mountainous terrain. J. Appl. Meteor., 14, 333–346.
Steinacker, R., 1984: Area-height distribution of a valley and its relation to the valley wind.
Contrib. Atmos. Phys., 57, 64–71.
Steinacker, R., and Coauthors, 2006: A mesoscale data analysis and downscaling method over
complex terrain. Mon. Wea. Rev., 134, 2758–2771.
Steinacker, R., and Coauthors, 2007: A sinkhole field experiment in the eastern Alps. Bull. Amer.
Meteor. Soc., 88, 701–716.
Stewart, J. Q., C. D. Whiteman, W. J. Steenburgh, and X. Bian, 2002: A climatological study of
thermally driven wind systems of the US Intermountain West. Bull. Amer. Meteor. Soc., 83,
699–708.
Stone, G. L., and D. E. Hoard, 1989: Low frequency velocity and temperature fluctuations in
katabatic valley flows. J. Appl. Meteor., 28, 477–488.
Stull, R. B., 1988: An Introduction to Boundary Layer Meteorology. Kluwer Academic, 670pp.
Sturman, A. P., and H. A. McGowan, 1995: An assessment of boundary-layer air mass character-
istics associated with topographically-induced local wind systems. Bound.-Layer Meteor., 74,
181–193.
Sturman, A. P., H. A. McGowan, and R. A. Spronken-Smith, 1999: Mesoscale and local climates.
Progress Physical Geography, 23, 611–635.
Sturman, A. P., and Coauthors, 2003a: The Lake Tekapo Experiment (LTEX): An investigation
of atmospheric boundary layer processes in complex terrain. Bull. Amer. Meteor. Soc., 84,
371–380.
Sturman, A. P., and Coauthors, 2003b: Supplement to the Lake Tekapo Experiment (LTEX): An
investigation of atmospheric boundary layer processes in complex terrain. Bull. Amer. Meteor.
Soc., 84, 381–383.
Sun, J., S. P. Burns, A. C. Delany, S. P. Oncley, T. W. Horst, and D. H. Lenschow, 2003: Heat
balance in the nocturnal boundary layer during CASES-99. J. Appl. Meteor., 42, 1649–1666.
Szintai, B., P. Kaufmann, and M. Rotach, 2010: Simulation of pollutant transport in complex
terrain with a numerical weather prediction–particle dispersion model combination. Bound.-
Layer Meteor., 137, 373–396.
Taylor, J. R., M. Kossmann, D. J. Low, and P. Zawar-Reza, 2005: Summertime easterly surges in
southeastern Australia: A case study of thermally forced flow. Austr. Met. Mag., 54, 213–223.
Thompson, B. W., 1986: Small-scale katabatics and cold hollows. Weather, 41, 146–153.
Trachte, K., T. Nauss, and J. Bendix, 2010: The impact of different terrain configurations on the
formation and dynamics of katabatic flows: Idealised case studies. Bound.-Layer Meteor., 134,
307–325.
Triantafyllou, A. G., C. G. Helmis, D. N. Asimakopoulos, and A. T. Soilemes, 1995: Boundary
layer evolution over a large and broad mountain basin. Theor. Appl. Climatol., 52, 19–25.
Urfer-Henneberger, C., 1970: Neure Beobachtungen über die Entwicklung des
Schönwetterwindsystems in einem V-förmigen Alpental (Dischmatal bei Davos). [New
observations of the development of a clear weather wind system in a v-shaped mountain valley
(Dischma Valley near Davos)]. Arch. Meteor. Geophys. Bioclimatol., Ser. B, 18, 21–42.
Van de Wiel, B. J. H., A. F. Moene, O. K. Hartogensis, H. A. R. De Bruin, and A. A. M. Holtslag,
2003: Intermittent turbulence and oscillations in the stable boundary layer over land. Part III.
A classification for observations during CASES-99. J. Atmos. Sci., 60, 2509–2522.
116 D. Zardi and C.D. Whiteman

Van den Broeke, M. R., 1997a: Momentum, heat and moisture budgets of the katabatic wind layer
over a midlatitude glacier in summer. J. Appl. Meteor., 36, 763–774.
Van den Broeke, M. R., 1997b: Structure and diurnal variation of the atmospheric boundary layer
over a mid-latitude glacier in summer. Bound.-Layer Meteor., 83, 183–205.
Van Gorsel, E., 2003: Aspects of flow characteristics and turbulence in complex terrain. Results
from the MAP-Riviera project. Dissertation. University of Basel, Geography Department.
ISBN 3-85977-247-1. 58pp.
Van Gorsel, E., A. Christen, C. Feigenwinter, E. Parlow, and R. Vogt, 2003a: Daytime turbulence
statistics above a steep forested slope. Bound.-Layer Meteor., 109, 311–329.
Van Gorsel, E., R. Vogt, A. Christen, and M. W. Rotach, 2003b: Low frequency temperature and
velocity oscillations in katabatic winds. Ext. Abstr., Vol A, Int. Conf. Alpine Meteor. and MAP
Meeting 2003. Publ. MeteoSwiss, No. 66, 251–254.
Varvayanni, M., J. G. Bartzis, N. Catsaros, P. Deligianni, and C. E. Elderkin, 1997: Simulation of
nocturnal drainage flows enhanced by deep canyons: The Rocky Flats case. J. Appl. Meteor.,
36, 775–791.
Vergeiner, I., 1983: Dynamik alpiner Windsysteme. [Dynamics of Alpine wind systems]. Bericht
zum forschungsvorhaben “3556” des Fonds zur Forderung der wissenschaftlichen Forschung,
Wien, Austria, 129pp.
Vergeiner, I., 1987: An elementary valley wind model. Meteor. Atmos. Phys., 36, 255–263.
Vergeiner, I., and E. Dreiseitl, 1987: Valley winds and slope winds – observations and elementary
thoughts. Meteor. Atmos. Phys., 36, 264–286.
Vertacnik, G, I. Sinjur, and M. Ogrin, 2007: Temperature comparison between some Alpine do-
lines in winter time. ICAM 2007. http://www.cnrm.meteo.fr/icam2007/html/PROCEEDINGS/
ICAM2007/extended/manuscript 95.pdf
Viana, S., E. Terradellas, and C. Yagüe, 2010: Analysis of gravity waves generated at the top of a
drainage flow. J. Atmos. Sci., 67, 3949–3966.
Vosper, S. B., and A. R. Brown, 2008: Numerical simulations of sheltering in valleys: The
formation of nighttime cold-air pools. Bound.-Layer Meteor., 127, 429–448.
Vrhovec, T., 1991: A cold air lake formation in a basin – a simulation with a mesoscale numerical
model. Meteor. Atmos. Phys., 46, 91–99.
Wagner, A., 1932: Neue Theorie der Berg- und Talwinde [New theory of mountain and valley
winds]. Meteor. Z., 49, 329–341.
Wagner, A., 1938: Theorie und Beobachtung der periodischen Gebirgswinde. [Theory and
observation of periodic mountain winds]. Gerlands Beitr. Geophys. (Leipzig), 52, 408–449.
[English translation: Whiteman, C.D., and E. Dreiseitl, 1984: Alpine Meteorology: Translations
of Classic Contributions by A. Wagner, E. Ekhart and F. Defant. PNL-5141/ASCOT-84-3.
Pacific Northwest Laboratory, Richland, Washington, 121pp].
Weber, R. O., and P. Kauffmann, 1995: Automated classification scheme for wind fields. J. Appl.
Meteor., 34, 1133–1141.
Weber, R. O., and P. Kaufmann, 1998: Relationship of synoptic winds and complex terrain flows
during the MISTRAL field experiment. J. Appl. Meteor., 37, 1486–1496.
Weigel, A. P., 2005: On the atmospheric boundary layer over highly complex topography.
Dissertation No. 15972, ETH, Zurich. [http://e-collection.ethbib.ethz.ch/view/eth:27927].
Weigel, A. P., and M. W. Rotach, 2004: Flow structure and turbulence characteristics of the daytime
atmosphere in a steep and narrow Alpine valley. Quart. J. Roy. Meteor. Soc., 130, 2605–2627.
Weigel, A. P., F. K. Chow, M. W. Rotach, R. L. Street, and M. Xue, 2006: High-resolution large-
eddy simulations of flow in a steep Alpine valley. Part II. Flow structure and heat budgets.
J. Appl. Meteor. Climatol., 45, 87–107.
Weigel, A. P., F. K. Chow, and M. W. Rotach, 2007: On the nature of turbulent kinetic energy in a
steep and narrow Alpine valley. Bound.-Layer Meteor., 123, 177–199.
Weissmann, M., F. J. Braun, L. Gantner, G. J. Mayr, S. Rahm, and O. Reitebuch, 2005: The Alpine
mountain-plain circulation: Airborne Doppler lidar measurements and numerical simulations.
Mon. Wea. Rev., 33, 3095–3109.
2 Diurnal Mountain Wind Systems 117

Whiteman, C. D., 1982: Breakup of temperature inversions in deep mountain valleys: Part I.
Observations. J. Appl. Meteor., 21, 270–289.
Whiteman, C. D., 1986: Temperature inversion buildup in Colorado’s Eagle valley. Meteor. Atmos.
Physics, 35, 220–226.
Whiteman, C. D., 1990: Observations of thermally developed wind systems in mountainous terrain.
In: W. Blumen, Editor, AMS Meteorological Monographs, 45 (23), 5–42.
Whiteman, C. D., 2000: Mountain Meteorology: Fundamentals and Applications. Oxford Univer-
sity Press, New York, 355pp.
Whiteman, C. D., and T. B. McKee, 1982: Breakup of temperature inversions in deep mountain
valleys: Part II. Thermodynamic model. J. Appl. Meteor., 21, 290–302.
Whiteman, C. D., and E. Dreiseitl, 1984: Alpine Meteorology. Translations of Classic Contri-
butions by A. Wagner, E. Ekhart and F. Defant. PNL-5141/ASCOT-84-3. Pacific Northwest
Laboratory, Richland, Washington, 121 pp.
Whiteman, C. D., and J. C. Doran, 1993: The relationship between overlying synoptic-scale flows
and winds within a valley. J. Appl. Meteor., 32, 1669–1682.
Whiteman, C. D., and X. Bian, 1998: Use of radar profiler data to investigate large-scale thermally
driven flows into the Rocky Mountains. Proc. Fourth Int. Symp. on Tropospheric Profiling:
Needs and Technologies, Snowmass, CO.
Whiteman, C. D., and S. Zhong, 2008: Downslope flows on a low-angle slope and their interactions
with valley inversions. I. Observations. J. Appl. Meteor. Climatol., 47, 2023–2038.
Whiteman, C. D., K. J. Allwine, L. J. Fritschen, M. M. Orgill, and J. R. Simpson, 1989a: Deep
valley radiation and surface energy budget microclimates. I. Radiation. J. Appl. Meteor., 28,
414–426.
Whiteman, C. D., K. J. Allwine, M. M. Orgill, L. J. Fritschen, and J. R. Simpson, 1989b: Deep
valley radiation and surface energy budget microclimates. II. Energy budget. J. Appl. Meteor.,
28, 427–437.
Whiteman, C. D., T. B. McKee, and J. C. Doran, 1996: Boundary layer evolution within a
canyonland basin. Part I. Mass, heat, and moisture budgets from observations. J. Appl. Meteor.,
35, 2145–2161.
Whiteman, C. D., X. Bian, and J. L. Sutherland, 1999a: Wintertime surface wind patterns in the
Colorado River valley. J. Appl. Meteor., 38, 1118–1130.
Whiteman, C. D., X. Bian, and S. Zhong, 1999b: Wintertime evolution of the temperature inversion
in the Colorado Plateau Basin. J. Appl. Meteor., 38, 1103–1117.
Whiteman, C. D., S. Zhong, and X. Bian, 1999c: Wintertime boundary-layer structure in the Grand
Canyon. J. Appl. Meteor., 38, 1084–1102.
Whiteman, C. D., S. Zhong, X. Bian, J. D. Fast, and J. C. Doran, 2000: Boundary layer evolution
and regional-scale diurnal circulations over the Mexico Basin and Mexican Plateau. J. Geophys.
Res., 105, 10081–10102.
Whiteman, C. D., S. Zhong, W. J. Shaw, J. M. Hubbe, X. Bian, and J. Mittelstadt, 2001: Cold pools
in the Columbia Basin. Wea. Forecasting, 16, 432–447.
Whiteman, C. D., S. Eisenbach, B. Pospichal, and R. Steinacker, 2004a: Comparison of vertical
soundings and sidewall air temperature measurements in a small Alpine basin. J. Appl. Meteor.,
43, 1635–1647.
Whiteman, C. D., B. Pospichal, S. Eisenbach P. Weihs, C. B. Clements, R. Steinacker, E. Mursch-
Radlgruber, and M. Dorninger, 2004b: Inversion breakup in small Rocky Mountain and Alpine
basins. J. Appl. Meteor., 43, 1069–1082.
Whiteman, C. D., T. Haiden, B. Pospichal, S. Eisenbach, and R. Steinacker, 2004c: Minimum
temperatures, diurnal temperature ranges and temperature inversions in limestone sinkholes of
different size and shape. J. Appl. Meteor., 43, 1224–1236.
Whiteman, C. D., S. F. J. De Wekker, and T. Haiden, 2007: Effect of dewfall and frostfall on
nighttime cooling in a small, closed basin. J. Appl. Meteor., 46, 3–13.
Whiteman, C. D., and Coauthors, 2008: METCRAX 2006 – Meteorological experiments in
Arizona’s Meteor Crater. Bull. Amer. Meteor. Soc., 89, 1665–1680.
118 D. Zardi and C.D. Whiteman

Whiteman, C.D., S.W. Hoch, and G.S. Poulos, 2009: Evening temperature rises on valley floors
and slopes: Their causes and their relationship to the thermally driven wind system. J. Appl.
Meteor. Climatol., 48, 776–788.
Whiteman, C. D., S. W. Hoch, M. Lehner, and T. Haiden, 2010: Nocturnal cold air intrusions
into Arizona’s Meteor Crater: Observational evidence and conceptual model. J. Appl. Meteor.
Climatol., 49, 1894–1905.
Wilczak, J. M., S. P. Oncley, and S. A. Stage, 2001: Sonic anemometer tilt correction algorithms.
Bound.-Layer Meteor., 99, 127–150.
Winkler, P., M. Lugauer, and O. Reitebuch, 2006: Alpines Pumpen [Alpine pumping]. Promet
(Deutscher Wetterdienst), 32, 34–42. [In German].
Wong, R. K. W., and K. D. Hage, 1995: Numerical simulation of pollutant dispersion in a small
valley under conditions with supercitical Richardson number. Bound.-Layer Meteor., 73, 15–33.
Wotawa, G., and H. Kromp-Kolb, 2000: The research project VOTALP – general objectives and
main results. Atmos. Environ., 34, 1319–1322.
Wu, P., M. Hara, J.I. Hamada, M.D. Yamanaka, and F. Kimura, 2009: Why a large amount of rain
falls over the sea in the vicinity of western Sumatra island during nighttime. J. Appl. Meteor.
Climatol., 48, 1345–1361.
Yang, K., T. Koike, H. Fujii, T. Tamura, X. Xu, L. Bian, and M Zhou, 2004: The daytime evolution
of the atmospheric boundary layer and convection over the Tibetan Plateau: Observations and
simulations. J. Meteor. Soc. Japan, 82, 1777–1792.
Yao, W. Q., and S. Zhong, 2009: Nocturnal temperature inversions in a small enclosed basin and
their relationship to ambient atmospheric conditions. Meteor. Atmos. Phys., 103, 195–210.
Ye, Z., M. Segal, J. R. Garratt, and R. A. Pielke, 1989: On the impact of cloudiness on the
characteristics of nocturnal downslope flows. Bound.-Layer Meteor., 49, 23–51.
Ye, Z. J., J. R. Garratt, M. Segal, and R. A. Pielke, 1990: On the impact of atmospheric thermal
stability on the characteristics of nocturnal downslope flows. Bound.-Layer Meteor., 51, 77–97.
Yoshino, M. M., 1975: Climate in a Small Area. Tokyo, University of Tokyo Press, 549pp.
Zammett, R. J., and A. C. Fowler, 2007: Katabatic winds on ice sheets: A refinement of the Prandtl
model. J. Atmos. Sci., 64, 2707–2716.
Zängl, G., 2003: The impact of upstream blocking, drainage flow and the geostrophic pressure
gradient on the persistence of cold-air pools. Quart. J. Roy. Meteor. Soc., 129, 117–137.
Zängl, G., 2004: A reexamination of the valley wind system in the Alpine Inn Valley with numerical
simulations. Meteor. Atmos. Phys., 87, 241–256.
Zängl, G., 2005: Wintertime cold-air pools in the Bavarian Danube Valley Basin: Data analysis
and idealized numerical simulations. J. Appl. Meteor., 44, 1950–1971.
Zängl, G., 2008: The impact of weak synoptic forcing on the valley-wind circulation in the Alpine
Inn Valley. Meteor. Atmos. Phys., 105, 37–53.
Zängl, G., and J. Egger, 2005: Diurnal circulation of the Bolivian Altiplano. Part II: Theoretical
and model investigations. Mon. Wea. Rev., 133, 3624–3643.
Zängl, G., and S. G. Chico, 2006: The thermal circulation of a grand plateau: Sensitivity to the
height, width, and shape of the plateau. Mon. Wea. Rev., 134, 2581–2600.
Zängl, G., and S. Vogt, 2006: Valley-wind characteristics in the Alpine Rhine Valley: Measure-
ments with a wind-temperature profiler in comparison with numerical simulations. Meteor. Z.,
15, 179–186.
Zängl, G., J. Egger, and V. Wirth, 2001: Diurnal winds in the Himalayan Kali Gandaki Valley. Part
II. Modeling. Mon. Wea. Rev., 129, 1062–1080.
Zardi, D., 2000: A model for the convective boundary layer development in an Alpine valley,
Proceedings, 26th Int. Conf. Alpine Meteor. ICAM2000, Innsbruck, Austria, 11–15 September
2000. Österreichische Beiträge zur Meteorologie und Geophysik. Heft Nr. 23/Publ. Nr. 362.
ISSN 1016–6254.
Zardi, D., R. Gerola, F. Tampieri, and M. Tubino, 1999: Measurement and modelling of a valley
wind system in the Alps. Proceedings of the 13th AMS Symposium on Boundary Layers and
Turbulence. 10–15 January 1999, Dallas, TX, 28–31.
2 Diurnal Mountain Wind Systems 119

Zaremba, L. L., and J. J. Carroll, 1999: Summer wind flow regimes over the Sacramento Valley.
J. Appl. Meteor., 38, 1463–1473.
Zawar-Reza, P., and A. P. Sturman, 2006a: Numerical analysis of a thermotopographically-induced
mesoscale circulation in a mountain basin using a non-hydrostatic model. Meteor. Atmos. Phys.,
93, 221–233.
Zawar-Reza, P., and A. P. Sturman, 2006b: Two-dimensional numerical analysis of a thermally
generated mesoscale wind system observed in the MacKenzie Basin, New Zealand. Aust.
Meteor. Mag., 55, 19–34.
Zawar-Reza, P., H. McGowan, A. Sturman, and M. Kossman, 2004: Numerical simulations of wind
and temperature structure within an Alpine lake basin, Lake Tekapo, New Zealand. Meteor.
Atmos. Phys., 86, 245–260.
Zhong, S., and J. Fast, 2003: An evaluation of the MM5, RAMS, and Meso-Eta models at
subkilometer resolution using VTMX field campaign data in the Salt Lake Valley. Mon. Wea.
Rev., 131, 1301–1322.
Zhong, S., and C. D. Whiteman, 2008: Downslope flows on a low-angle slope and their interactions
with valley inversions. II. Numerical modeling. J. Appl. Meteor. Climatol., 47, 2039–2057.
Zhong, S., C. D. Whiteman, X. Bian, W. J. Shaw, and J. M. Hubbe, 2001: Meteorological processes
affecting evolution of a wintertime cold air pool in the Columbia Basin. Mon. Wea. Rev., 129,
2600–2613.
Zhong, S., X. Bian, and C. D. Whiteman, 2003: Time scale for cold-air pool breakup by turbulent
erosion. Meteor. Z., 12, 229–233.
Zhong, S., C. D. Whiteman, and X. Bian, 2004: Diurnal evolution of three-dimensional wind and
temperature structure in California’s Central Valley. J. Appl. Meteor., 43, 1679–1699.
Zoumakis, N. M., and G. A. Efstathiou, 2006a: Parameterization of inversion breakup in idealized
valleys. Part I: The adjustable model parameters. J. Appl. Meteor. Climatol., 45, 600–608.
Zoumakis, N. M., and G. A. Efstathiou, 2006b: Parameterization of inversion breakup in idealized
valleys. Part II: Thermodynamic model. J. Appl. Meteor. Climatol., 45, 609–623.
Zoumakis, N. M., G. A. Efstathiou, A. G. Kelessis, J. Triandafyllis, D. Papas, M. Chasapis, M.
Petrakakis, P. Karavelis, 2006: A simple scheme for daytime estimates of surface energy budget
in complex terrain. Fresenius Environmental Bulletin, 15, 923–927.
Chapter 3
Dynamically-Driven Winds

Peter L. Jackson, Georg Mayr, and Simon Vosper

Abstract This chapter is concerned with dynamically-forced atmospheric flow


phenomena which occur when the wind encounters mountains. The range of effects
is wide and therefore attention is restricted to arguably the most important phenom-
ena in terms of weather forecasting. These are mountain waves, rotors, downslope
windstorms, gap winds and barrier jets. The essence of many of these phenomena is
described by mountain wave theory. Recent advances in observation technologies
and their application in field programs, as well as in numerical modeling, have
led to new understanding, including the incorporation of complicating factors
like boundary-layer processes. This chapter describes current theory for each of
these phenomena, along with recent observational studies and the latest forecast
techniques and models.

P.L. Jackson ()


Environmental Science & Engineering, Natural Resources and Environmental Studies Institute,
University of Northern British Columbia, 3333 University Way, Prince George, British Columbia,
Canada V2N 4Z9
e-mail: peterj@unbc.ca
G. Mayr
Institute of Meteorology and Geophysics, University of Innsbruck, Innrain 52,
6020 Innsbruck, Austria
e-mail: georg.mayr@uibk.ac.at
S. Vosper
Atmospheric Processes and Parametrizations, Met Office, FitzRoy Road, Exeter,
United Kingdom EX1 3PB
e-mail: simon.vosper@metoffice.gov.uk

F. Chow et al. (eds.), Mountain Weather Research and Forecasting, 121


Springer Atmospheric Sciences, DOI 10.1007/978-94-007-4098-3 3,
© Springer ScienceCBusiness Media B.V. 2013
122 P.L. Jackson et al.

3.1 Introduction: Flow Over, Around and Through


Mountain Ranges

This chapter deals with flow phenomena that are dynamically forced when wind
interacts with orography. As air approaches a mountain barrier, it can pass over
it, flow through gaps or valleys that dissect it, or be blocked by the mountain
and diverted horizontally around it. Each of these scenarios results in different
dynamically-driven wind phenomena. When air is able to pass over a mountain
barrier under specific conditions, mountain waves are generated that can result in lee
waves, rotors and downslope windstorms. Air flowing through channels that dissect
a mountain barrier can result in gap flow arising from along-channel acceleration.
If the air is blocked by a mountain barrier and diverted around it, then a barrier jet
may form. This chapter describes these various phenomena.
For air to be able to pass over a mountain, it must either have low stability
or large inertia (wind speed). Air that approaches a mountain barrier will be
forced to lift as it encounters higher terrain. As the air lifts, if the atmosphere is
stable, it will encounter negative buoyancy that will slow the wind in the across-
barrier direction. Unless the approaching air has sufficient inertia to cross the
barrier, it may be blocked entirely by the barrier and forced horizontally around
it. The ratio of energy needed to overcome the negative buoyancy imposed by
the mountain, to the horizontal inertia that can provide this energy, is given by a
fundamental dimensionless parameter, called the non-dimensional mountain height,
HO D NH=U , where N is the Brunt Väisälä, or buoyancy frequency, H is the
mountain barrier height, and U is the upstream wind speed component perpendicular
to the mountain barrier. The square of the Brunt Väisälä frequency is defined as

g d
N2 D ;
 dz

where z is height, g is the acceleration due to gravity and  is the potential


temperature. This will be discussed further in Sect. 3.2.2. If the air does not have
sufficient inertia to go over the mountain, then it will be diverted around the
mountain and may form an elevated barrier jet, as described in Sect. 3.6.
If the approaching air does have sufficient inertia to overcome stability and cross
a mountain barrier, the displacement caused by the barrier can have either horizontal
or vertical consequences depending on how long it takes for air to cross the barrier.
If the mountain barrier is so wide that it takes longer than f1 (f is the Coriolis
parameter) for air to cross it, then Coriolis effects and conservation of potential
vorticity will result in predominantly horizontal perturbations to the flow. This will
result in the generation of large-scale horizontal topographic Rossby waves with
ridges over the mountain barrier and troughs to the lee of the mountains. These
conditions depend on values of another important dimensionless number, the across-
barrier Rossby number, Ro D U/fL  1, where L is the mountain barrier width.
3 Dynamically-Driven Winds 123

On the other hand, when Ro  1, which would occur for mountain barriers less
than about 100 km in width for typical values of U (10 m s1 ) and f (104 s1 ),
the horizontal displacements are reduced, Coriolis effects can be largely ignored to
a linear approximation, and air parcels are displaced vertically. If the air is stably
stratified, but still has sufficient inertia to cross the barrier, the vertically displaced
air parcels will form gravity waves that oscillate about their equilibrium level. These
gravity waves, called mountain waves, may, depending on conditions, propagate
vertically or horizontally. Horizontally propagating gravity waves that are vertically
trapped can form lee waves and rotors (Sect. 3.3), and vertically propagating
gravity waves in some circumstances may amplify and cause downslope windstorms
(Sect. 3.4).
Dynamically-driven winds both affect, and are affected by, many other aspects
of mountain weather discussed in this book. For example, while early work
on dynamically-driven winds ignored “complicating” factors such as thermal,
boundary-layer, and moist processes (see discussions in Chaps. 2, 5, 6 and 7), it
is now recognized that considering these factors is often essential to go beyond a
first-order understanding of forced flows. In turn, strongly-forced dynamic flows
can overwhelm thermal and boundary-layer processes. The foehn, a well-known
and studied type of forced mountain flow is discussed in much more detail in
Chap. 4. Early progress in understanding dynamically-driven flows, was mainly
due to theoretical developments. While theory continues to be important, new
observations of the mountain atmosphere made possible by advances in technology
(Chap. 8) and high-resolution numerical simulation made possible by advances in
computing and models (Chaps. 9 and 10) have opened new avenues of inquiry
and understanding. As numerical weather prediction models (Chap. 11) begin to
achieve the resolution and model development necessary to adequately represent
dynamically-forced flows, they improve in their ability to forecast these flows
directly.
The chapter starts in Sect. 3.2 by introducing basic concepts: the equations used,
mountain waves, and hydraulic theory. The sections that follow this are organized
by flow phenomena. Sections 3.3 and 3.4 describe phenomena that are forced by
flow over mountains: Lee waves and rotors, and downslope windstorms. Section 3.5
describes gap flow that occurs when air is forced through valleys that dissect
mountain barriers. Section 3.6 describes barrier jets that arise when flow impinging
on a mountain barrier is blocked by it.

3.2 Fundamentals of Forced Flows in Complex Terrain

In order to understand the complex and diverse behavior of atmospheric flow in


mountainous terrain, knowledge of the most fundamental processes is first required.
Perhaps the two most important phenomena to consider are internal gravity waves
and the hydraulic behavior of shallow-layer flow across mountains and through
124 P.L. Jackson et al.

passes. Many aspects of dynamically driven flows in complex terrain are influenced
by these processes. In this section we shall briefly review the relevant equations
of motion and then separately consider internal gravity waves and hydraulic flow
theory. A number of simplifications and approximations can be applied to the
equations in order to bring out the most important aspects of the dynamics.

3.2.1 Flow Equations

The main aspects of the flows considered here are represented by the Euler equations
which govern the motion of inviscid, frictionless adiabatic flow of dry air. For flow
in a rotating frame of reference (the Earth), these equations can be written as

Du 1
C f k  u D  rp  gk (3.1)
Dt 
D
C r  .u/ D 0 (3.2)
Dt
D
D0 (3.3)
Dt
where u is velocity, ¡ is density, p is pressure and k is the unit vector in the vertical
(z) direction. Dissipative processes, such as the effects of viscosity or turbulent
mixing, and moist processes are not represented by these equations. Such processes
are clearly important for certain phenomena but our aim here is merely to illustrate
the primary aspects of dynamically driven winds. This can be accomplished without
the need to consider these additional effects.
Equations 3.1–3.3 describe the motion of a compressible gas and therefore permit
sound waves. Since these have limited dynamical significance and complicate
further analysis, it is useful to remove them from the system. This can be achieved
through the anelastic (e.g. Ogura and Phillips 1962) or Boussinesq approximations.
Subtly different versions of each exist and a comprehensive description is given
by Durran (1999) and references therein. It is useful to split the thermodynamic
variables, pressure, density and potential temperature into steady hydrostatically
balanced reference profiles (that are a function of height, z, only) and perturbations
which vary in space and time. In other words we define p.x; t/ D p.z/ N C p 0 .x; t/,
.x; t/ D .z/ 0 N 0
N C  .x; t/ and .x; t/ D .z/ C  .x; t/. It is also helpful to introduce
the Exner pressure,  D .p=p0 /R=cp and let .x; t/ D .z/ N C  0 .x; t/ where p0
is a reference pressure and R and cp are the gas constant and specific heat capacity
for dry air, respectively. The anelastic approximation consists of neglecting time
derivatives of the density term in the continuity equation and replacing (3.2) by

r  .u/
N D0 (3.4)
3 Dynamically-Driven Winds 125

The momentum equation, (3.1), can also be written in a more useful form as:

Du g 0
C f k  u C cp r 0 D k (3.5)
Dt N
Equations 3.5, 3.4 and 3.3 are the anelastic equations and their solutions do not
contain sound waves. The term on the right-hand side of (3.5) is the buoyancy term
and is of fundamental importance to stratified flows over mountains.
Perhaps the simplest example of a stratified flow is that of a shallow homoge-
neous layer of fluid with a free surface. The flow of such a fluid over an obstacle is
analogous to certain types of dynamically driven flows in mountainous terrain. For
example, the stationary patterns of ripples and the turbulent hydraulic jumps which
form in a stream as it flows over undulations in the river bed are analogous to some
of the features observed in atmospheric flows over mountains. The acceleration of
a channel flow through a narrow constriction also resembles that of an atmospheric
flow through a mountain pass. Consequently much progress can be made toward
understanding atmospheric flows by considering the equations which govern the
motion of these simpler single layers, the shallow water equations. For a free-surface
height, (x,y), the shallow water equations can be written as:

@u @u @u @
Cu Cv  f v D g (3.6)
@t @x @y @x
@v @v @v @
Cu Cv C fu D g (3.7)
@t @x @y @y
@ @.u/ @.v/
C C D0 (3.8)
@t @x @y

In the atmospheric context, (x,y) might describe the height of a discontinuity


in density or temperature, such as an inversion. The analogy between this simple
shallow water system, or its more complicated counterpart where the flow consists
of multiple homogeneous layers, provides a useful basis for interpretation of more
complex atmospheric flows.

3.2.2 Mountain Waves

Internal gravity waves, or buoyancy waves, are a fundamental feature of stably


stratified fluids and occur as a result of the buoyancy forces which act on fluid
parcels when they are vertically displaced from their equilibrium position. In the
atmosphere, internal gravity waves occur on the mesoscale, with typical horizontal
wavelengths lying in the range of a few hundred meters to several hundred
kilometers. Any process which involves lifting of air parcels or perturbations to
buoyancy through heating can result in gravity wave motion. Such processes include
126 P.L. Jackson et al.

convection, shear instability, convergence at frontal zones, drainage flows in stable


boundary layers (see Chap. 2) and, of course, the vertical displacements associated
with air flow over orography. The general term “mountain wave” is used to describe
this latter kind of internal gravity-wave motion.
Mountain waves are ubiquitous to flows over orography and are known to
have an important, if not controlling influence on the flow dynamics of a wide
range of physical processes. For example, the windstorms which form on the lee-
side of mountain ranges are intimately linked to the presence of mountain waves.
Mountain waves are also sometimes cited as the cause of clear-air turbulence, a
particular cause of concern for aviation safety. From the point of view of global
atmospheric circulation, mountain waves also play an important role. The waves
carry momentum and energy into the stratosphere, creating turbulence where they
break and imparting a net drag on the flow, influencing the planetary circulation and
thus the passage of synoptic-scale weather systems. Recently satellites have begun
to detect evidence for mountain waves in the middle and upper atmosphere (e.g.
Jiang et al. 2002). The use of microwave limb sounders to detect fluctuations in
temperature enables the distribution of gravity-wave activity to be examined on a
global scale. The most active areas appear to be the mountainous regions of the high
latitudes in wintertime, where the winds are strongest. It is likely that the observed
gravity waves are mountain waves which have propagated high into the stratosphere.
There is, of course, a vast body of literature on the subject of mountain waves
and there are several excellent and extensive reviews. Notably those by Smith
(1979, 1989), Baines (1995) and Nappo (2002) serve as very informative and useful
reviews of the subject and provide a thorough background to the theory of mountain
waves. In this section we aim only to present sufficient background to explain later
aspects of this chapter on downslope windstorms and lee waves and rotors. Over
the last two decades there have been several major field programs in which various
aspects of orographic flow, including mountain wave related phenomena, have been
examined in detail. Notable examples are the Pyrenees Experiment (PYREX, see
Bougeault et al. 1993), the Mesoscale Alpine Programme (MAP, see Bougeault
et al. 2001), and most recently the Terrain induced Rotor Experiment (T-REX, see
Grubišić et al. 2008). The observations made during these experiments have led
to advances in our understanding of various aspects of mountain-wave behaviour.
These include the generation, propagation and dissipation of mountain waves, the
drag associated with the waves, wave induced turbulence and the impacts of the
waves on the near-surface flow.
Two important non-dimensional parameters which determine the response of a
stratified fluid to orography are the parameters FL D U/(NL) and HO D NH=U ,
where L is the mountain width scale. As we shall see below, the quantity FL provides
information about the propagation characteristics of mountain waves. As described
in Sect. 3.1, the non-dimensional mountain height, HO , predicts the extent to which
the approaching flow will traverse over the mountain summit, or will be forced
around the mountain sides. In general when HO  1 the upstream flow contains
sufficient kinetic energy to overcome the potential energy barrier represented by the
mountain and the flow will therefore be largely over the mountain. When HO 1,
3 Dynamically-Driven Winds 127

however, the buoyancy forces associated with the stable stratification act to restrict
vertical motion and the flow is constrained to be largely horizontal. The flow will
then be blocked upwind and is forced to travel around the flanks of the mountain.
The parameter HO provides a measure of the importance of nonlinearity. For HO  1
the flow response to the mountains is generally well described by linear theory. For
HO > 1, however, nonlinear processes are important. Sometimes Ĥ1 D U/NH is
referred to as a Froude number, although this is different from the Froude number
used in hydraulic theory that will be discussed later in which the length scale is the
depth of the fluid, not the height of the mountain barrier.
In the interests of simplicity and because here our focus will be on waves
generated by individual peaks rather than very broad continental-scale mountain
ranges such as the Rockies or Andes, we will now neglect the effects of rotation
in our analysis. This amounts to neglecting the Coriolis term in (3.5). Neglecting
rotation is valid when the Rossby number, Ro D U/(fL) 1 and this is generally true
for typical mountain widths and values of U and f. For example, with a mountain
width of L D 10 km, a wind speed of U D 10 m s1 and a mid-latitude value
of f D 104 s1 , Ro D 10. Note, however, that while rotation may be relatively
unimportant for the dynamics of the mountain waves on these scales, it does affect
the dispersion properties of gravity waves with very long horizontal wavelengths
(e.g. several hundred kilometers). It is also known that in the nonlinear flow regime
(HO > 1), for moderate values of Ro (roughly 1  Ro  10), rotation can have
significant effects on the flow. These range from asymmetry in the flow as it splits
around the flanks of the mountain (e.g. Hunt et al. 2001), acceleration of the flow
through mountain passes (e.g. Sprenger and Schär 2001), asymmetry in mountain
wave breaking (e.g. Grisogono and Enger 2004) and modification of the drag
exerted by the mountain on the atmosphere (e.g. Ólafsson and Bougeault 1997).
Further progress can be made if we now assume that disturbances to the mean
flow are small. By splitting the velocity vector, u, into a mean height dependent
profile and a perturbation and then linearizing the equations (by neglecting products
of primed quantities) we obtain the linearized anelastic equations for a dry non-
rotating inviscid atmosphere. These can be written as

@u0 N  u0 C u0 r  U N
N C cp r 0 g 0
C Ur D k (3.9)
@t N

N 0/ D 0
r  .u (3.10)

@ 0 N
N  r 0 C w0 d  D 0
CU (3.11)
@t dz

where UN D .UN ; VN / is the horizontal mean flow vector and u0 .x; y; z; t/ D .u0 ; v0 ; w0 /
is the perturbation velocity vector. For mountain waves, this linearization of the
equations of motion is strictly only valid when both the mountain slope and the non-
dimensional mountain height are small (H/L  1 and HO  1). However the linear
128 P.L. Jackson et al.

equations prove to be useful both from the point of view of explaining many
of the dynamical properties of the flow, as well as providing useful quantitative
predictions well beyond their formal limits of applicability. We shall discuss the
role of nonlinearity, when HO is O(1) or larger, later in this chapter.
For the sake of simplicity we shall now further assume that the mean flow
contains no vertical wind shear (d U=d N z D 0) and that N is independent of
height. For now, we shall also make the Boussinesq approximation, which assumes
that the effect of changes in density can be neglected in the continuity equation
(3.10), allowing N to be replaced by a reference value 0 . Note that the Boussinesq
approximation is also commonly applied in studies of the atmospheric boundary
layer (see Chap. 2). Strictly, the Boussinesq approximation requires that the vertical
scale of motions is much less than the density scale height, H, defined as
 
1 d N 1
H D  (3.12)
N dz

This approximation simplifies the following analysis, but it also removes an im-
portant feature of mountain-wave dynamics, namely the growth of wave amplitude
with height which occurs as a result of the decrease in density, which can ultimately
lead to wave breaking in the stratosphere.
Without loss of generality we shall further assume that the flow direction is
aligned with the x-direction (this simplifies the problem to a two-dimensional one)
and assume wave-like solutions for all perturbation variables with wavevector
 D .k; m/ and frequency ! of the form
n o
O i.kxCmz!t / ;
0 D Re e (3.13)

where k and m are wavenumbers in the x and z directions, respectively, O is a


complex wave amplitude and Re denotes the real part. Substituting such expressions
into the Boussinesq versions of (3.9)–(3.11) gives the linear dispersion relation for
internal gravity wave motion, namely:

2 N 2k2
! D .!  UN k/ D 2
_ 2
(3.14)
k C m2

The quantity !O is the intrinsic frequency of the gravity wave motion and
describes the frequency of motion experienced in a frame of reference travelling
with the mean flow speed, UN . Equation 3.14 states that the intrinsic frequency of
gravity wave motion is always limited by the Brunt Väisälä frequency and there is
therefore a natural limit to the rapidity of oscillations. Note that !O D N can in
fact only be achieved when the wave motion has no phase variation with height (i.e.
when m D 0), the case where the lines of constant phase are vertical. For phase lines
more inclined towards the horizontal the intrinsic frequency decreases.
3 Dynamically-Driven Winds 129

Rearranging (3.14) we obtain


 
N2
m2 D k 2 1 (3.15)
!O 2

from which we can deduce that if !O 2 < N 2 then the vertical wavenumber is real
and the gravity wave motion is able to propagate in the vertical. If !O 2 > N 2 then
m is imaginary and the physically realistic solution is for the disturbances to decay
exponentially with height away from their source, in this case the orography. For
mountain waves, the source is of course stationary, implying that the frequency of
the waves (in a fixed frame of reference), ¨, is zero. Equation 3.15 then reduces to a
relationship between the vertical and horizontal wavenumbers and the background
flow properties, namely

N2
m2 D  k2: (3.16)
UN 2
According to (3.16) we should therefore expect mountain waves with long
horizontal wavelengths to have vertical wavelengths which are relatively short.
The vertical wavelength will increase with decreasing horizontal wavelength until,
for disturbances which are sufficiently short, m becomes imaginary and the wave
motion decays in the vertical. If we consider the case of monochromatic waves gen-
erated by a simple sinusoidal mountain range with wavelength,
, the significance
of the non-dimensional parameter FL becomes clear. Mountain waves will only exist
when N=UN > k, or alternatively when FL D UN =.NL/ < 1, where the scale L is
related to the wavelength by L D
/(2 ). For values of FL > 1 the waves generated
have a frequency greater than N, so that the flow cannot support mountain waves
and disturbances will decay exponentially with height. Such flows are illustrated
schematically by Fig. 3.1.
It is now useful to define the group velocity, cg D .@!=@k; @!=@m/. This
describes the direction and rate of propagation of information or energy by the
waves. The group velocity is the velocity with which a “packet” of waves travels
through the atmosphere. From (3.14) we can show that

@! Nm2
D UN ˙ 3 (3.17)
@k
and

@! Nmk
D
3 ; (3.18)
@m

where D jj. If the background wind was to vary sufficiently slowly with height
that the dispersion relations still hold locally, then from (3.17) we can see that as
UN ! 0, m ! 1 and from (3.18), @!=@m ! 0. This implies that as a wave
130 P.L. Jackson et al.

a b
5000 5000
4000 4000
3000 3000
z (m)

z (m)
2000 2000
1000 1000
0 0
-10000 -5000 0 5000 10000 -10000 -5000 0 5000 10000
x (m) x (m)

Fig. 3.1 Streamlines over sinusoidal mountains with wavelength 10 km. The wind speed and
Brunt Väisälä frequency are independent of height with N D 0.01 s1 and (a) UN D 10 ms1
and (b) UN D 20 ms1 . In (a) the value of FL is approximately 0.628 and therefore propagating
mountain waves occur. In (b) FL  1.26 and the wave motion decays exponentially with height

approaches a level where the wind (resolved in the direction of the horizontal
wavevector) falls to zero, the vertical wavelength tends to zero and the rate at which
wave energy propagates in the vertical also tends to zero. Therefore wave energy
does not pass through such a level. The wave energy is dissipated at such “critical
levels” and reflection may occur as a result of nonlinearity. The latter is thought to
be an important mechanism in the dynamics of downslope windstorms, where the
breaking of waves in the troposphere may result in a mountain wave induced critical
level above the mountain.
The angle at which a packet of waves will propagate away from the mountain
is determined by the ratio of expressions (3.17) and (3.18). Choosing the sign of
vertical propagation to be positive (so that wave energy radiates vertically away
from the mountain) it can be shown that the angle of wave energy propagation, ˛,
away from the mountain is given by
 
@! @! 1
tan ˛ D D m=k: (3.19)
@m @k

Equations 3.16 and 3.19 tell us that when the horizontal wavelength of mountain
waves is sufficiently long that k  N=UN , the vertical wavenumber m  N=UN
and the wave motion is non-dispersive; the angle of propagation is near vertical.
The motion associated with these long wavelengths is, to a good approximation,
in hydrostatic balance. For flows across broad smooth mountains (with FL  1)
for which the majority of the mountain spectrum consists of long wavelengths, the
wave motion is therefore contained in a narrow beam above the mountain. Wave
energy radiates vertically and little wave motion is encountered downwind. For
shorter horizontal wavelengths whose k is less than, but not significantly less than
N=UN , the waves are dispersive. In this case the angle ˛is reduced (and depends on
the wavenumber, k) and the wave motion is non-hydrostatic. Therefore for narrow
3 Dynamically-Driven Winds 131

Fig. 3.2 The mountain waves generated by flow over (a) broad and (b) narrow ridges when
the upwind profile consists of constant stability and wind speed (N D 0.02 s1 and U D 20 m
s1 ). Color shading denotes vertical velocity (m s1 ) and isentropes are denoted by line contours
(interval 5 K). In (a) the width-scale L is 10 km and FL D 0.1, resulting in a hydrostatic mountain
wave response in which the waves propagate vertically and the wave motion is restricted to a
vertical beam above the mountain. In (b) L is 1 km and FL D 1. In this case the wave motion is
non-hydrostatic. The waves propagate horizontally as well as vertically and wave motion is present
in a dispersive tail to the lee of the ridge

mountains, where FL < 1 and a significant portion of the mountain spectrum is made
up of short wavelengths, wave energy radiates in the downstream direction as well
as in the vertical, resulting in the appearance of “lee wave” motion downwind.
Examples of the hydrostatic and non-hydrostatic wave fields associated with
broad and narrow mountain ridges, respectively, are illustrated in Fig. 3.2. The
wave fields shown are those which occur in flow over two-dimensional ridges of the
 1
commonly used bell shaped form h.x/ D H 1 C x 2 =L2 . The upwind profiles
have constant stability and wind speed (N D 0.02 s1 and UN D 20 ms1 ) and the
wave fields shown are linear solutions, computed using (3.21) (see below). For the
broad mountain, where L D 10 km and FL D 0.1, wave motion is only observed
immediately above the mountain. These hydrostatic mountain waves propagate
vertically into the upper atmosphere. In contrast, for the narrower mountain, when
L D 1 km and FL D 1, the dispersive nature of the energy propagation (implied by
the k dependence in (3.19)) means that short wavelength waves have a horizontal
component to their propagation and hence wave motion appears downwind of the
ridge as well as above it.
In practice, the complexity of real mountain ranges means that they are made
up of a wide spectrum of horizontal scales and thus we should expect a range
of responses from the mountain wave field. The very shortest scales, perhaps
associated with individual peaks or narrow valleys, may in general be too short
to force wave motion and their associated disturbances may decay with height. For
longer scales, we can expect both a vertically propagating hydrostatic wave response
as well as non-hydrostatic waves which trail downwind of the mountain range. It is
132 P.L. Jackson et al.

worth bearing in mind that the increase of wind strength with height through the
troposphere, typical of mid-latitudes, implies that wave motion which is hydrostatic
(and hence vertically propagating) at low levels, may in fact respond in a more
dispersive non-hydrostatic manner aloft. The shorter, non-hydrostatic, wavelengths
may themselves be greatly affected by changes in wind and stability with height. As
we shall see in Sect. 3.3, this will often result in a trapped lee-wave response, where
the wave energy is contained within the troposphere.
In order to understand how mountain waves are affected by height variations in
wind and stability we shall return to (3.9–3.11) but this time drop the assumptions
of zero shear and constant stability. Note that we have dropped the Boussinesq
approximation for the time being, since density is a function of height in (3.10).
We seek solutions to the steady versions of the equations with the form

O
0 D .z/e i.kxCly/
: (3.20)

Substituting such expressions into (3.9–3.11) results, after some algebra, in the
vertical structure, or Taylor-Goldstein equation for mountain waves in a sheared
flow with non-uniform stability, namely
 
d 1 d  _
_
Nw C
2  h 2 w D 0; (3.21)
dz N dz

where h D .k; l/ is the horizontal wavevector, h D jh j, wO (h , z) is a complex


vertical velocity amplitude and the function
2 , often called the Scorer parameter,
is given by

N 2 h 2 1 d2 N

2 .z/ D  .U:h /: (3.22a)
N h/
.U:
2 N h dz2
U:

For flow over a two-dimensional ridge, the wave field is made up of horizontal
wavevectors with a single direction, perpendicular to the ridge. In this case the
expression for
2 simplifies further to

N2 1 d 2 Ur

2 .z/ D  ; (3.22b)
Ur 2 Ur d z2

where Ur is the wind speed resolved in the direction perpendicular to the ridge.
The presence of density in the first term of (3.21) results in the general growth
of mountain wave amplitude with height which can lead to wave breaking in the
upper atmosphere. This breakdown of wave motion can lead to the generation
of turbulence and a net drag force on the mean flow. The latter is commonly
parameterized in global weather prediction and climate models. Note that this wave
growth effect would not have been accounted for if we had retained the Boussinesq
approximation in the derivation of (3.21); the first term on the left-hand side would
O z2 .
then have been simply d 2 w=d
3 Dynamically-Driven Winds 133

The propagation of mountain waves is affected by the variations in


2 .z/ which
arise from changes in the background wind profile or static stability. As a rule,
the increase of wind speed with height through the troposphere means that
2 will
decrease with altitude, implying that short wavelength wave motion (for which
2 
h 2 > 0 at low levels) may decay with height at higher altitudes. In general the
first term on the right-hand sides of (3.22a and 3.22b) tend to dominate over the
second, although the latter term can be important across jets where there may be
rapid changes in vertical wind shear. The importance of variations in
2 will become
more apparent in Sect. 3.3 when we consider the phenomenon of wave reflection,
which can result in the channeling of wave energy downwind and the long trains of
waves, commonly known as trapped lee waves.
Traditionally most studies of mountain waves have assumed that the atmospheric
boundary layer is sufficiently shallow that it will have only a small impact on
their generation. Its role has been largely ignored and frictionless, free-slip lower
boundary conditions have been assumed. Recently, however, the importance of
boundary-layer processes has come to light in a number of studies. From the point
of view of wave generation, modeling studies have demonstrated how the boundary
layer will generally have a damping effect, reducing the amplitude of the waves, the
strength of downslope winds and the tendency for wave breaking (e.g. Richard et al.
1989; Peng and Thompson 2003; Vosper and Brown 2007). Notably Smith et al.
(2002) concluded that the presence of a stagnant boundary layer beneath the Alpine
mountain peaks had a significant attenuating effect on mountain waves observed
during MAP. Recently the influence of the boundary layer on wave generation has
been illuminated using linear theory by Smith et al. (2006), Smith (2007) and Jiang
et al. (2006) who showed that absorption of wave energy in the boundary layer can
result in a decay of trapped lee-wave motion downstream. We shall return to this
point in Sect. 3.3.
For situations where the combination of mountain height, upstream wind and
stability are such that the non-dimensional mountain height, HO , is of order unity
or greater, the linear equations of motion no longer provide an accurate description
of the flow. In such situations the flows are nonlinear and the deceleration of the
wind on the upwind side of the mountain can be sufficient to cause stagnation and
the appearance of a region of blocked flow upstream (e.g. Smith 1980). The flow
below mountain crest may then be diverted around the flanks rather than over the
summit, sometimes separating from sides and forming a wake region downwind.
In general the mountain waves will steepen as HO increases above unity, eventually
leading to wave breaking immediately above and downwind of the mountain (e.g.
Smith 1980; Miranda and James 1992). This low-level wave breaking enhances the
lee-slope acceleration and is thought to be an important process in the development
of downslope wind storms. The turbulent dissipation in the wave breaking region
also contributes to the development of the downstream wake (e.g. Schär and Durran
1997). The wake may take the form of a long uniform region of decelerated flow
extending from the mountain (e.g. Smith et al. 1997) or consist of an attached
vortex pair (Smolarkiewicz and Rotunno 1989) which under some circumstances
134 P.L. Jackson et al.

b(x)
F<1 F<1 F<1

(x) F=1
F<1

hydraulic
F>1
F>1

jump
F<1
h(x) F>1 F>1

Fig. 3.3 Schematics of hydraulic flow configuration through a channel (top, left) with a sill.
Flow is from left to right. Asymmetric flow configuration common for downslope windstorms
and gap flows with subcritical flow upstream and supercritical flow downstream adjusting to the
downstream conditions in a hydraulic jump (bottom, left). Subcritical flow everywhere (top, right);
supercritical flow everywhere (bottom, right)

will be unstable, resulting in the shedding of discrete and long-lived vertically


orientated vortices (Etling 1989; Schär and Durran 1997). Damping processes such
as boundary layer friction have been shown to limit the instability of the wake, and
thus prevent such vortex shedding (Grubišić et al. 1995).

3.2.3 Hydraulic Theory

Stratified air flowing over a mountain or through a valley is analogous to the flow
of water in a channel. This analogy can be exploited for a physical explanation of
dynamically forced flow in complex terrain as well as forecasting. Here, a short
introduction to the basic concepts of hydraulic theory will be given. The theory will
be developed further in the sections on downslope windstorms (Sect. 3.4) and gap
flow (Sect. 3.5). Baines (1995) describes hydraulic theory in even greater detail.
Starting with the one-dimensional version of the shallow water equations (3.6–
3.8), we consider flow in a channel of width b(x) over some topography h(x) as
shown in Fig. 3.3. The height of the free surface of the fluid is (x). As a further
simplification of (3.6), we are only interested in well-adjusted, steady-state flow
without friction, external pressure gradient, and Coriolis force. The advection in its
fully non-linear form, however, is kept. Then the momentum equation becomes

@u @ @h
u Cg D g (3.23)
@x @x @x
where the pressure gradient term has been split into a part resulting from the change
of the height of the free surface and a part due to changes in terrain elevation. The
3 Dynamically-Driven Winds 135

continuity equation requires the volume flux q through the cross-sectional area A to
be constant: @(A u)/@x D 0. If we take a channel with vertical sidewalls, the volume
flux q D u  b. The continuity equation becomes

@u @ u @b
 Cu D (3.24)
@x @x b @x
Note that when the free surface is an interface between two layers of comparable
densities, reduced gravity g0 D g (1  2 )/1 has to be used instead of g. The two
equations can be written as a quasi-linear differential equation relating derivatives of
the dependent flow variables (u, ) to the derivatives of the independent topographic
variables (h, b) (Armi 1986)
       
ug g 0 u h
C 'x D D‰x with C D ; DD ; 'D ; D
u 0 q  b 1
(3.25)

Solutions exist if det(C) ¤ 0 or if certain regularity conditions are fulfilled in the


neighborhood of locations where det(C) D 0. These locations are called “controls”
since they establish unique flow solutions. Regularity conditions ensure that the
magnitude of the slope of the free surface remains finite.
Characteristic velocities are given by det(C 
I) D 0 with I being the identity
matrix, and
p

Du˙ g (3.26)

The square-root term is the speed of (long) gravity waves at the fluid interface.
Its relation to the advective velocity u yields the Froude number

u2
F2 D (3.27)
g

The Froude number determines the characteristics of the flow: when F < 1 is
“subcritical” waves can travel up and down stream. When F > 1 is “supercritical”
waves can only travel downstream. The special situation of F D 1 is “critical” will
be discussed in more detail. In this case, det(C) D 0 and the characteristic velocity
in (3.26) vanishes.
The differential solutions to (3.25) are
   
1 @u 1 1 @b 1 1 @h
D C
u @x 1F 2 b @x 1F 2  @x
 2   
1 @ F 1 @b 1 1 @h
D  (3.28)
 @x 1F 2 b @x 1F 2  @x
136 P.L. Jackson et al.

The terms in square brackets are termed “influence coefficients”. Two special
cases occur for @b/@x D 0 and @h/@x D 0. For the special case of @b/@x D 0 (channel
of constant width or at the narrowest part), the change of height of the free surface
is directly but with opposite sign related to the change of elevation:
 
@ 1 @h
D (3.29)
@x 1F 2 @x

In the limit of F D 0, as the terrain rises the top of the fluid has to come down
equally. When the flow is subcritical (F < 1) everywhere, the top of the fluid still
changes in opposite direction to the terrain. However, it descends further than the
terrain rises on the upstream side and rises further than the terrain drops on the
downstream side (Fig. 3.3, top, right). The fluid depth is shallowest at the crest.
When the flow is supercritical (F > 1) everywhere, terrain and the height of the
interface move in tandem but the excursion of the interface height is less than the
excursion of the terrain (Fig. 3.3, bottom, right).
The most common situation for downslope windstorms, however, is an asym-
metric descent of the interface from upstream across the crest to some distance
downstream (Fig. 3.3, bottom, left). For that to happen, the influence coefficient
(i.e. the term in square parentheses) has to change sign as the flow crosses from
the upstream to the downstream side. A location has to exist where F2 D 1. For a
solution to exist, this location has to be at the crest where @h/@x D 0. Conditions of
asymmetry and regularity render the crest a so-called “control”. We are no longer
free to set an arbitrary combination of layer depth and volume flux rate q. If one is
set, e.g. the layer depth, the flow rate follows. And, since the interface has to descend
on both sides in the asymmetric case, the flow has to be subcritical (slow and thick)
upstream and supercritical (fast and shallow) downstream. The height descends with
subcritical flow upstream and supercritical flow downstream. At the control, i.e. the
crest, the Bernoulli equation (conservation of energy) is ½ u2 C g c D gH0 , where
H0 is the far-upstream depth of the fluid. Combining the Bernoulli equation with
(3.27) for F2 D 1, results in c D 2/3 H0 . At the control, the fluid depth has already
decreased by one third of its value far upstream.
Note that for the more general case when @b/@x ¤ 0 at @h/@x D 0, the control need
not be at the crest but where

1 @ 1 @h
 C D0 (3.30)
b @x  @x

This regularity condition includes one of the dependent variables, the height of
the interface .
Further downstream the flow will eventually return from a supercritical to a
subcritical state, slow down and thicken again. The turbulent transition zone is a
hydraulic jump, in which momentum is conserved but kinetic energy is dissipated.
For the asymmetric case of interest for gap flow and downslope windstorms, the
depth of the reservoirs on either side of the topographical constriction/crest will
differ with the larger depth on the upstream side.
3 Dynamically-Driven Winds 137

Using a single layer of constant density is the most frequent application of


hydraulic theory to the atmosphere. However, hydraulic theory can be formulated
for two layers (Armi 1986; Armi and Riemenschneider 2008) or more including
continuous stratification. For continuous stratification, an analytical solution exists,
which is described in Sect. 3.5 on gap flow.

3.3 Lee Waves and Rotors

A special kind of non-hydrostatic wave motion occurs when the variation of the
Scorer parameter,
2 , with height is such that vertically propagating wave motion
is partially reflected back down towards the ground. In the typical case where

2 > h 2 in the lower and mid-troposphere (due to relatively low wind speeds or
high stability, for example) and
2 < h 2 in the upper troposphere, the wave energy
will become trapped within the lower layer and ducted horizontally downstream.
Such trapping is generally only partial, due to the fact that
2  h 2 will often
be positive in the lower stratosphere and wave motion is able to “tunnel” through
layers of
2 < h 2 . In cases where the trapping is strong the waves may be ducted
horizontally for hundreds of kilometers downstream. Such waves are commonly
referred to as trapped lee waves and they are responsible for the long regular trains
of clouds which form in the wave crests downwind of mountain ranges and which
are readily seen in satellite imagery. A schematic diagram, illustrating the structure
of the trapped lee wave is provided in Fig. 3.4. Examples of satellite images of
lee-wave cloud formations are shown later in Figs. 3.6 and 3.9.
Typical horizontal wavelengths of trapped lee waves observed in the troposphere
lie in the range 5–30 km. Increased stability at low levels (for example due to
an inversion layer at the boundary layer top) or reduced low-level wind speeds
will favor the existence of shorter horizontal wavelengths, while reduced stability
and increased winds will increase the wavelength. As we shall see in Sect. 3.3.2,
according to linear theory at least, the horizontal wavelength of trapped lee waves
is a function of the upwind profile of wind and stability rather than the mountain
shape.
Mountain waves are often associated with turbulence, and one particularly severe
phenomenon which causes turbulence at low-levels is the mountain-wave rotor.
This term is used to refer to a region of recirculating airflow with horizontal
vorticity, orientated parallel to the ridge crest. Rotors are typically accompanied by
strong near-surface downslope winds which decelerate rapidly at the rotor’s leading
upwind edge, where boundary-layer flow separation takes place and air parcels
are lifted away from the surface. Downwind of the separation point, the mean
near-surface winds are generally slack or weakly reversed below the rotor, but are
usually highly turbulent, exhibiting a high degree of temporal and spatial variability.
Rotors are recognized as posing a significant hazard to aviation and have been
cited as contributing to numerous accidents and damage to aircraft. Additionally
138 P.L. Jackson et al.

2
kH
Height

Trapping layer

rgy

Wa
ene

ve
ve

ene
Wa

gyr
l2
Separation point Lee−wave (Type I) rotor

Fig. 3.4 Schematic diagram illustrating the structure of the trapped lee wave (flow is from left to
right). The Scorer parameter,
2 , profile is such that a trapping layer exists in which
2  h 2 > 0.
Upward propagating wave energy is reflected back down towards the ground at the top of this
layer, resulting in net downwind horizontal propagation of wave energy. In the situation shown
here the waves are of sufficient amplitude that flow separation occurs beneath the wave crests and
a turbulent (type I) rotor forms

rotor circulations may also have an important impact on the transport of pollutants
and aerosols in mountainous terrain, serving as a mechanism for transport from the
boundary layer into the free troposphere.
Based on observations and recent modeling (Hertenstein and Kuettner 2005),
rotors have been classified into two types: the rotors which form under the crests
of trapped lee waves (type I) and the jump-like rotor (type II), in which the wave
motion is replaced by a deep turbulent layer downwind of the mountain. The latter
is thought to contain the more severe turbulence, but is probably relatively rare. The
distinction between the type I and II rotor is clear, although as discussed later the
definition of the type II rotor itself is rather loose, and covers a variety of different
flow types including low-level wave breaking and hydraulic jumps.
Remarkable examples of rotor clouds and dust transport are provided by Fig. 3.5
which shows instances of rotor formation in the Owens Valley, to the east of the
Sierra Nevada mountain range in California, USA. The deep turbulent region and
lack of smooth wave clouds seen in Fig. 3.5a and b seems to fit the description of the
type II rotor. Figure 3.5c provides a clear example of rotor clouds associated with
a type I rotor beneath a train of lee waves which formed downwind of the Rocky
Mountains near Boulder, Colorado.
3 Dynamically-Driven Winds 139

Fig. 3.5 Rotor clouds and


blowing dust over the Owens
Valley during mountain wave
events in the lee of the Sierra
Nevada on (a) 5 March 1950
during the Sierra Wave
Program (SWP),
(photographed by Robert
Symons) and (b) 25 March
2006 during IOP-6 of the
Terrain induced Rotor
Experiment (T-REX)
(photographed by Barbara
Brooks from the FAAM
BAe-146 aircraft). In both
cases the flow is westerly
(from right to left) and the
Sierra Nevada Mountains are
on the right-hand side of the
photographs. (c) Rotor clouds
on 2 December 2003 in a
train of trapped lee waves in
the lee of the Rocky
Mountains, near Boulder,
Colorado (photographed by
Rolf Hertenstein) (From
Hertenstein and Kuettner
2005, reproduced by
permission of Tellus)

While mountain waves and trapped lee waves have been the subject of much
attention from the atmospheric science community over the last few decades, rotors
have received relatively little attention. However, improvements in observational
techniques and increases in computing power have prompted a surge in interest in
the rotor problem and there have been several recent field campaigns and numerical
modeling studies aimed at understanding the formation and characterizing the
properties of rotors.
140 P.L. Jackson et al.

a b
8

5
z/km

0
−4 −3 −2 −1 0 1 2 3 4
Vertical velocity /ms−1

Fig. 3.6 Measurements of lee-wave activity over Scotland on 11 March 2000 made by (a)
radiosonde measurements at 1012 UTC and (b) high-resolution visible satellite imagery at 1507
UTC. Plot (a) shows fluctuations in the ascent rate of a radiosonde launched from the Isle of
Arran after the mean ascent rate has been removed. Image (b) shows lee-wave cloud over much of
Scotland. The coastline and Isle of Arran are also shown (Figures are adapted from Vosper 2003.
Reproduced by permission of the Royal Meteorological Society)

3.3.1 Observations

3.3.1.1 Lee Waves

As discussed previously, lee-wave cloud patterns are frequently observed in high-


resolution satellite imagery. Their vertical motion is also often readily seen in
fluctuations in the ascent rates of radiosondes. Figure 3.6a shows an example of
a moderate amplitude lee wave as observed by a radiosonde launched on the Isle of
Arran, in south-west Scotland, UK. Fluctuations of amplitude around 2 m s1 are
evident in the ascent rate of the balloon between the ground and 6 km. As shown
by visible satellite imagery in Fig. 3.6b, the waves observed by the radiosonde form
part of an extensive train of lee waves in a north-westerly airstream. The origin of the
waves is clearly not the mountains and hills of Arran itself (though they undoubtedly
contribute towards the wave activity) since the wave motion extends over much of
Scotland and far upwind.
Research aircraft are commonly used to make measurements of lee waves
during field observation programs. Over the past few decades there have been
numerous observational research campaigns in which lee waves have been measured
by research aircraft above and downwind of mountain ranges. Recent examples
include the measurements made during PYREX, MAP and, most recently, T-REX.
Aircraft observations are particularly powerful when conducted with multiple
3 Dynamically-Driven Winds 141

Fig. 3.7 Vertical velocity (m s1 ) measured by three research aircraft over the Sierra Nevada
Mountains during a coordinated T-REX mission on 25 March 2006 (IOP-6). Measurements made
by the NCAR HIAPER are shown in blue, UK FAAM BAe146 measurements are shown in red
and University of Wyoming King Air (UWKA) measurements are shown in black. The UWKA
data above 7 km and just below 6 km share vertical velocity scales with the neighboring BAe146
legs. Data from the three aircraft were obtained over a period of just under 5 h. The HIAPER and
BAe146 tracks, which were almost perfectly aligned, were offset 3–4 km to the north from the
shown UWKA tracks (From Grubišić et al. 2008. © American Meteorological Society. Reprinted
with permission)

aircraft, coordinated in such a way to provide a comprehensive picture of the wave


field. Figure 3.7 provides such an example, with lee-wave measurements made
during T-REX over the Sierra Nevada Mountains. The measurements were made
using three research aircraft flying simultaneously at different altitudes. The utility
of aircraft measurements is enhanced further through the use of remote sensing
instruments, such as backscatter lidar, which can detect the presence of clouds. This
technique was used to good effect during MAP to obtain a detailed picture of the
wave and cloud field over the Alps (see Fig. 3.8).
142 P.L. Jackson et al.

Fig. 3.8 Vertical displacements (solid curves) inferred from aircraft vertical velocity measure-
ments and airborne lidar backscatter (colour, dBZ) on a north–south cross section over the Alps
on 20 September 1999 during MAP. The airflow is from the south (left-hand side) (From Smith
et al. 2007 reproduced from Doyle and Smith (2003). Reproduced by permission of the Royal
Meteorological Society)

3.3.1.2 Rotors

Compared with the waves themselves, observations of rotors are relatively scarce.
One of the earliest observations of rotors was that by Mohorovičić (1889), who
observed rotor clouds during bora events along the northern Adriatic coast in
Croatia. Grubišić and Orlić (2007) provide an interesting account of Mohorovičić’s
(largely forgotten) work, along with observations of rotor clouds elsewhere which
were published around the same time. Some of the best observations have come
from the pioneering work of the Sierra Wave Project (SWP) and the Jet Stream
Project (JSP) in the 1950s (Holmboe and Klieforth 1957; Grubišić and Lewis 2004)
of mountain waves and rotors over the Sierra Nevada. Rotors were also observed
3 Dynamically-Driven Winds 143

during the Colorado Lee Wave Program in the late 1960s and early 1970s (Lester
and Fingerhut 1974). Following this, there have been few organized campaigns
targeted at observing rotors. Recently, however, rotors have again become the focus
of observational programs. Mobbs et al. (2005) have obtained measurements of
the near-surface flow downwind of the Wickham range on the Falkland Islands
during episodes of rotor activity. Similar ground based measurements of lee-wave
induced flow separation were made by Sheridan et al. (2007) downwind of the
Pennines in the UK and Darby and Poulos (2006) observed rotor evolution in
the Rocky Mountains using several measurement techniques, including research
aircraft, Doppler lidar and wind profiler systems. Very recently, attention has
returned to the Sierra Nevada Mountains with a two-part observational campaign
focused on the rotor formation in the Owens Valley. The first, pilot-phase of
the campaign, the Sierra Rotors Project (SRP), took place during the spring
of 2004 and consisted primarily of ground based measurements in the Owens
Valley (Grubišić and Billings 2007). The second phase, the Terrain induced Rotor
Experiment (T-REX) took place during spring 2006 and involved a comprehensive
set of observing systems, including research aircraft and a wide range of ground
based and remote sensing instrumentation. T-REX has provided detailed systematic
measurements of mountain-wave rotors and the datasets will shed new light on the
dynamics of rotors.
For obvious safety and logistical reasons, in-situ aircraft measurements of rotors
are relatively rare, the exceptions being those made during the early SRP and JSP
and very recent T-REX campaigns. Most observations, by necessity, have been made
by ground-based instrumentation or use remote sensing instruments on airborne
platforms. One technique deployed in recent field campaigns (e.g. Mobbs et al.
2005; Sheridan et al. 2007; Grubišić and Billings 2007) has involved the deployment
of large numbers of automatic weather stations (AWSs) over a small area to provide
a well resolved picture of the surface “footprint” of rotor motion aloft. Figure 3.9
shows an example of such measurements obtained in the lee of the Pennine hills in
northern England, UK. The measurements show a region of relatively low surface
pressure, within which the near-surface winds undergo considerable acceleration.
Downwind of this region the flow is decelerated by an adverse (positive) pressure
gradient. A region of high surface pressure sits downwind of the low pressure and
the flow within this region is very slack, and at times is reversed. As we shall see
in Sect. 3.2, this type of pattern is consistent with the theoretical ideas of how lee
waves influence the near-surface flow.
Clearly AWS measurements offer some interesting insights into the impacts of
lee waves and rotors on the surface flow, but remote sensing techniques provide
the best chance of observing the vertical and three-dimensional flow structure.
Figure 3.10 shows Doppler lidar observations of the velocity field made during the
T-REX campaign. The lidar was situated in the Owens Valley and the measurements
shown are for IOP-6 which included periods of large amplitude mountain wave
144 P.L. Jackson et al.

Fig. 3.9 Lee-wave cloud over the UK and near-surface flow observations beneath the lee waves
in the Vale of York, northern England. Frame (a) shows high resolution visible satellite imagery
on 17 March 2005 at 1000 UTC. Frame (b) shows a snapshot of the 10 min average 2 m wind
vectors measured by AWSs in the Vale of York (in the lee of the Pennines hills) on the same day
at 2010 UTC. Colour contours indicate the surface pressure perturbations (units hPa) at each site
in the AWS array. Terrain contours are depicted by solid lines with an interval of 50 m. The inset
in (b) shows the flow measurements upwind of the Pennines. The boxed area in (a) indicates the
location of the northern Pennines and the Vale of York (Figures are adapted from Sheridan et al.
2007. Reproduced by permission of the Royal Meteorological Society)

and rotor activity. For the times shown in Fig. 3.10 strong westerly and south-
westerly winds were present on the west side of the valley (west of the lidar site)
with relatively weak northerly along-valley winds to the east. The scans show clear
evidence for a rapid reversal of the westerly flow approximately 1–2 km to the
west of the lidar, indicating the existence of a recirculating rotor within the valley.
The lidar scans showed a high degree of temporal variability at this time, indicating
that the flow within the rotor was highly turbulent and that the structure of the rotor
was very transient.
The few existing in-situ measurements made by aircraft in rotors suggest that the
presence of extreme turbulence is most likely connected with vortices whose scale is
perhaps only a few hundred meters, and therefore much smaller than that of the main
rotor. For example, several vertical gusts exceeding 10 m s1 (and as large as 20 m
s1 ) were experienced in a 50 s period during a JSP research flight when an aircraft
penetrated a rotor (Holmboe and Klieforth 1957; Doyle and Durran 2007). Lidar
measurements obtained near Boulder, Colorado (Banta et al. 1990) also suggest the
presence of small scale vortices within the main rotor recirculation. Similarly, the
recent T-REX lidar observations support the existence of these small-scale intense
subrotors.
3 Dynamically-Driven Winds 145

Fig. 3.10 Lidar scans of the low-level flow in the Owens Valley, east of the Sierra Nevada
mountains, during an episode of large amplitude mountain wave and rotor activity. The data
shown are from IOP-6 of the T-REX campaign. Frame (a) shows a PPI scan of Doppler lidar
radial velocity measurements obtained in the late afternoon on 25 March 2006. Green and blue
colors indicate flow toward the lidar while yellow and red indicate flow away from the lidar. The
dashed line indicates the orientation of the RHI scan shown in (b). Gray lines in (a) indicate terrain
contours, which are plotted every 200 m. The RHI scan was obtained about 4 min after the PPI
scan (From Grubišić et al. 2008. © American Meteorological Society. Reprinted with permission)

3.3.2 Theory

3.3.2.1 Lee Waves

As discussed earlier, we can expect trapped lee waves to form when the Scorer
parameter decreases with height in a suitable manner. The form of
2 .z/ determines
both the wavelength and the wave amplitude variation with height. At this stage it
is helpful to consider a very simple problem in order to illustrate the theory for lee
waves in more complex and realistic situations. Consider the case of a two-layer
stratified flow in which the stability takes constant (but different) values, Nl and Nu
146 P.L. Jackson et al.

in lower and upper layers, respectively. We shall assume that the wind speed, UN , is
constant across both layers and for convenience define the height of the interface
between the two layers to be at z D 0, with the ground located at z D d. We shall
also make the Boussinesq approximation, thus eliminating the density terms in
(3.21). We seek wave-like solutions to (3.21) of the form

O l D Al e iml z C Bl e iml z
w (3.31)

and

wO u D Au e imu z ; (3.32)

where wO l and wO u are the complex vertical velocity amplitudes in the lower and
upper layers, respectively, ml 2 D Nl 2 =UN 2  k 2 and mu 2 D Nu 2 =UN 2  k 2 are
the corresponding vertical wavenumbers (squared) and Al , Bl and Au are complex
constants. Note that (3.31) expresses the fact that we expect both upward and
downward propagation of wave energy in the lower layer, the latter being caused
by partial reflection at the layer interface, whereas (3.32) assumes that wave energy
in the upper layer will radiate freely in the vertical. We now impose matching
conditions at the interface. Firstly we require that an air parcel on the interface must
move with the vertical velocity of the interface. If the interface has shape z D &.x/
then this is described by

D&
D w on z D &: (3.33)
Dt
Letting & D &e O ikx and linearizing (3.33) then gives the matching condition
_
ikUN D w on z D 0 and this must be satisfied by air parcels both immediately above
_

and below the interface. Since the mean wind, UN , is the same in both layers this
becomes wO l D wO u at z D 0. The second matching condition applied at the interface
states simply that the pressure on both sides is the same, namely, POl D POu on z D 0.
This implies that

d wO l d wO u
D on z D 0: (3.34)
dz dz

Substituting these conditions into (3.31) and (3.32) and manipulating the result-
ing expressions for Al , Bl and Au then gives

2Al
wO l D Œml cos.ml z/ C i mu sin.ml z/ (3.35)
ml C mu

and
2Al ml imu z
wO u D e : (3.36)
ml C mu
3 Dynamically-Driven Winds 147

Fig. 3.11 The theoretical 12


resonant lee-wave horizontal
wavelength plotted as a d=3 km
function of wind speed U for d=4 km
the two-layer stability case 10 d=5 km
with Nu D 0.01 and
Nl D 0.02 s1 . Data are
plotted for lower layer depths,

Wavelength (km)
d, of 3, 4 and 5 km 8

2
10 12 14 16 18 20
U (ms-1)

The constant Al is obtained from a free-slip lower boundary condition which


simply states that air parcels at ground level follow the mountain profile, z D h(x).
Thus, on the mountain surface we have w D Dh/ Dt. Linearizing and writing h.x/ D
_
O ikx then gives w
he
_
N
l D i U kh on z D d. Substituting this into (3.35) then gives

_ m cos.m z/ C im sin.m z/
O l D UN ikh
l l u l
w : (3.37)
ml cos.ml d /  imu sin.ml d /

as the solution in the lower layer. Equation 3.37 contains singularities whenever ml
and mu are such as that the denominator is zero. This occurs when

tan.ml d / D i ml =mu (3.38)

and corresponds to the existence of resonant wavemodes. These resonant waves are
the naturally occurring trapped lee waves and as such require only an infinitesimal
amount of forcing by the orography. Their horizontal wavelength is determined
by the implicit relationship described by (3.38). Note that the occurrence of the
imaginary number, i, on the right-hand side of (3.38) means that solutions are only
possible when mu itself is imaginary, and hence the wave amplitude will decay
with height in the upper layer, consistent with the fact that the wave response is
trapped in the lower layer. Examples of solutions to (3.38) are presented in Fig. 3.11
which shows how the resonant horizontal wavelength varies with wind speed for a
case where the lower and upper-layer stabilities are Nl D 0.02 and Nu D 0.01 s1 ,
respectively.
148 P.L. Jackson et al.

According to (3.38) the theoretical horizontal wavelength of the resonant trapped


waves is purely a function of the background profile rather than the shape of the
orography. This is also true more generally, for any profile of wind and stability.
In general theoretical predictions for the resonant wavelength are only available
for a small class of idealized profiles such as the two-layer case considered above.
Vosper (2004) performed a similar analysis to that presented here and showed how
linear theory can be used to describe the behaviour of trapped lee waves which
form in a temperature inversion layer. More generally, however, for real profiles of
wind and stability, such resonant solutions cannot be determined analytically and
numerical approaches must be used. By discretizing (3.21) the resonant wavemodes
can be sought and the linear trapped lee wave solutions can then be computed
(e.g. Sawyer 1960; Vergeiner 1971). For three-dimensional flows the resonant
wavemodes will in general lie along lines in (k,l) space (Sawyer 1962) and the wave
field will be made up of contributions from wavemodes along such lines. Note that
although it is generally accepted that linear theory can provide a reliable prediction
of the horizontal wavelength of trapped lee waves (at least in cases where the non-
dimensional mountain height, HO  1) the same is not always true of the wave
amplitude. Although linear theory has been used widely for lee-wave prediction (and
as discussed in Sect. 3.3 certainly can prove useful from a forecasting perspective)
several studies (e.g. Smith 1976; Durran and Klemp 1982; Durran 1992; Vosper
2004) have shown how broad mountains, whose horizontal scale exceeds the lee-
wave wavelength, will provide little forcing at the resonant wavenumbers. The
amplitude of the waves will then be determined by nonlinear processes. For HO > 1
nonlinearity can also have an important influence on the horizontal wavelength,
when for instance, the mountain range consists of two (or more) ridges, separated by
a valley. Grubišić and Stiperski (2009) have shown how, in this case, when the two
ridges are of similar height, the ridge separation has a strong controlling influence
on the horizontal wavelength.
In common with most previous studies, the above discussion of trapped lee
wave behaviour makes no attempt to account for the influence of the atmospheric
boundary layer. As well as exhibiting a rich variety of phenomena itself in complex
terrain (see Chaps. 2 and 5), recent work has shown that the boundary layer can have
a significant impact on both the generation and then the attenuation of lee waves.
Notably Smith et al. (2002), motivated by observations of lee waves during MAP,
developed a simple theory for the absorption of wave energy by a stagnant layer
which filled the valleys between individual mountain peaks. Whereas a solid lower
boundary reflects downward propagating rays back upwards, the theory allows
for partial absorption of wave energy by the stagnant layer, which results in the
downwind decay of the lee wave motion. The process is illustrated by the schematic
diagram in Fig. 3.12. More recently, studies (Smith et al. 2006; Jiang et al. 2006)
have examined how the boundary layer acts to attenuate the trapped lee wave
motion and suggest a faster downwind decay in shallow nocturnal boundary layer
conditions, with slower decay when the boundary layer is deep and convective.
The boundary layer also plays an important role in mountain wave induced flow
separation, and thus has a crucial part to play in the dynamics of rotors.
3 Dynamically-Driven Winds 149

Fig. 3.12 A schematic of the propagation of wave energy for trapped lee waves when a stagnant
layer of air fills the valleys between mountain peaks. The triangles represent the lower mountains,
with the lee waves generated by the higher peak which penetrates into the ambient flow. The
dashed line labelled zref marks the top of the stagnant layer. The upper dashed line represents
the level of wave reflection due to a decreasing Scorer parameter. The slanting wave rays are
parallel to the local group velocity and indicate the direction of energy propagation. The down-
going wave amplitude, B, is reduced by partial absorption at zref (Figure taken from Smith et al.
2007. Reproduced by permission of the Royal Meteorological Society)

3.3.2.2 Rotors

The theoretical and numerical challenges posed by mountain wave rotors have
meant that, until recently, relatively little progress has been made toward under-
standing their dynamics or the conditions in which they form. Nevertheless, the
recent surge in interest in this area has resulted in some significant progress. By
necessity, most of the recent advances in understanding the dynamics of rotors
have come from numerical simulation rather than analytical theory, although some
limited progress towards understanding the process of flow separation, with which
rotors are intimately linked, has emerged from analytical treatments. The focus, to
date, has tended to be on the type I, lee-wave rotor.
The formation of rotors is clearly connected to the occurrence of boundary-layer
separation and the relationship between separation and the mountain-wave motion
aloft. Although theories have existed for several decades as to how the pressure
gradient due to lee waves might induce flow separation (e.g. Lyra 1943) these ideas
have only recently been tested using numerical simulations. Doyle and Durran
(2002) have shown, through numerical simulations of two-dimensional trapped
lee waves, how the adverse (positive) pressure gradient which occurs beneath the
leading upwind edge of a wave crest will act to decelerate the flow. Once this
gradient is sufficiently large the flow will separate away from the surface and a
(type I) rotor will form. Doyle and Durran (2002) and later Vosper (2004) have
shown how surface friction is vitally important to obtaining realistic simulations of
rotors. Its effect is to both reduce the wind speed towards zero at the surface, thereby
enhancing the influence of the pressure gradient, thus promoting separation, and to
generate a sheet of high horizontal vorticity along the lee slope of the mountain,
150 P.L. Jackson et al.

Fig. 3.13 Streamlines and horizontal vorticity (units s1 ) for a simulation of lee-wave (type 1)
rotors in flow over a two-dimensional ridge. Horizontal vorticities greater than 0.03 s1 are shaded
in color. Horizontal wind speeds less than or equal to zero are shown using blue isotachs (every
2 m s1 ) (From Doyle and Durran 2002. © American Meteorological Society. Reprinted with
permission)

which is lifted aloft by the waves when the flow separates. This is illustrated by
Fig. 3.13 which shows a two-dimensional simulation of a type I rotor (Doyle and
Durran 2002).
The occurrence of flow separation and its relationship with the trapped lee-wave
amplitude has been studied in further detail by Vosper et al. (2006) and Jiang et al.
(2007). In the former study a theory was developed which suggests that, for two-
dimensional waves, a critical lee-wave pressure amplitude exists, beyond which
flow separation will occur under the wave crests. This critical amplitude can be
expressed non-dimensionally as p=.u2 /, where p is the amplitude of the
pressure fluctuations exerted by the waves on the ground (assumed constant through
the boundary layer) and u is the surface friction velocity of the background flow.
Vosper et al. have shown how this critical non-dimensional amplitude increases as
a function of surface roughness and this can be explained by the fact that, for a
smoother surface, the surface stress will be reduced and the near-surface winds will
be stronger. A larger pressure gradient is therefore required to reverse the flow. Note
that this result is contrary to what occurs for a laminar flow where separation occurs
more readily for a smooth surface. Jiang et al. (2007) have also demonstrated the
link between boundary-layer separation and trapped lee wave amplitude and found
that separation depended largely on the ratio of the vertical velocity maximum to
the surface wind speed. Separation was far less sensitive to other factors such as the
3 Dynamically-Driven Winds 151

Fig. 3.14 The along ridge component of horizontal vorticity (colour shading, interval 0.02 s1 )
and flow vectors in a cross section through a type I rotor. The results shown are for a three-
dimensional simulation of flow over an infinitely long two-dimensional ridge. The vortex sheet is
lifted away from the surface and then broken up by shear instability, resulting in intense subrotors
which move along the interface between the main rotor and the wave field (From Fig. 7(c) of Doyle
and Durran 2007. © American Meteorological Society. Reprinted with permission)

mountain height and slope. Jiang et al. also found that the onset of separation was
sensitive to the surface sensible heat flux. A positive heat flux tends to inhibit the
occurrence of separation by increasing vertical turbulent mixing and thus enhancing
the surface wind speed. A negative (downward) heat flux appears to have the
opposite effect.
By performing very fine-resolution numerical simulations, Doyle and Durran
(2007) have begun to illuminate the internal dynamics of trapped lee-wave (type I)
rotors. Their simulations have shown how, once lifted away from the surface by flow
separation, shear instability along the upstream leading edge of the rotor can break
up the sheet of high horizontal vorticity into small intense vortices, or subrotors.
Doyle and Durran’s simulations illustrate the importance of three-dimensional
effects in the dynamics of these subrotors. Even for waves generated by a strictly
two-dimensional ridge, the tilting and in particular the stretching of vorticity which
is present in a three-dimensional turbulent flow, contributes to the vorticity of the
subrotors, meaning they can contain horizontal vorticity which far exceeds that of
the main rotor. The subrotors are relatively long lived, being swept downstream
along the interface between the main rotor and the lee wave and it seems possible
that their existence is connected to pulsations in the wind that have been observed
during downslope wind events (see Sect. 3.4.1). An example of this behaviour is
shown in Fig. 3.14 (Doyle and Durran 2007). In the less realistic situation where
the turbulence is restricted to be purely two-dimensional the subrotors are much
less intense and are short-lived, being eventually entrained into the main rotor
recirculation. Recently, new studies using eddy-resolving simulations have been
used to explore the role of turbulence in mountain-wave breaking (Epifanio and
152 P.L. Jackson et al.

Qian 2008) and interactions between the turbulent boundary layer and lee-wave
rotors (Smith and Skyllingstad 2009). This approach is clearly a powerful one and
the ability to explicitly resolve turbulent eddy motion will perhaps allow a full
understanding of the internal dynamics of rotors.
Compared with the type I rotors, the more severe jump-like rotors have received
much less attention. Vosper (2004), in a study of temperature inversion effects on
lee waves, found that a hydraulic-jump-like response occurred in place of trapped
lee waves if the potential temperature difference, , across the inversion was
sufficiently large. The results from a set of two-dimensional simulations were
used to construct a flow regime diagram which contained, in order of increasing
, trapped lee waves on the inversion layer, lee waves which were sufficiently
strong to induce flow separation and form (type I) rotors and, finally, regions of
overturning and severe turbulence in the lee of the mountain. These latter conditions
were somewhat loosely referred to as a “hydraulic jump” state. Hertenstein and
Kuettner (2005) also performed two-dimensional simulations for flow with a strong
temperature inversion and found that the solutions were sensitive to vertical wind
shear across the inversion in the upstream flow. For sufficiently strong forward
shear trapped lee waves formed with attendant type I rotors. Reduction in the shear,
however, resulted in a downstream response more akin to a hydraulic jump (see
Fig. 3.15). The relatively smooth lee-wave field is replaced by a deep turbulent layer
which extends far downwind. Hertenstein and Kuettner explained the sensitivity to
the wind shear in terms of the horizontal vorticity budget. The type of rotor that
forms depends on the dominant sign of the horizontal vorticity as the flow separates
from the surface. This balance is modified by the presence of vertical wind shear
across the upstream inversion.
Recently Jiang et al. (2007) have also examined the rotor types which can occur
in the presence of an elevated temperature inversion. They found that a variety of
different flow regimes could occur: (1) a steady undular jump regime, in which
a hydraulic jump forms in the lee of the ridge; (2) a regime in which trapped
lee-waves ride along the inversion layer behind the jump; (3) a propagating jump
regime in which the jump propagates away downwind and (4) a mixing jump regime
in which low-level wave breaking merges with a hydraulic jump below. These
flow regimes all involve the generation of turbulence and are capable of causing
boundary-layer separation, resulting in reversal in the lee of the mountain. In the
sense that these flows are all clearly different to that of the type I, trapped lee-wave
rotor, it is tempting to label the rotors associated with them all as type II, despite their
differences.

3.3.3 Models and Forecasting

The implications for aviation safety and risks to ground transport and property mean
that there is a clear need to forecast lee-wave motion and rotor activity. Glider pilots
can also benefit from wave forecasts since they can use the associated updrafts to
stay aloft for long periods of time and to achieve high altitudes and great distances.
3 Dynamically-Driven Winds 153

Fig. 3.15 Potential temperature (contour interval 0.5 K) downwind of a two-dimensional ridge in
a type II rotor simulation. The C signs mark the positions of three propagating eddies along the
top of the rotor (From Fig. 7(b) of Hertenstein and Kuettner 2005. Reproduced by permission of
Tellus)

The typical horizontal wavelengths of trapped lee waves mean that they are
generally poorly resolved by numerical weather prediction (NWP) models. Global
forecast models, whose horizontal resolutions are typically 25–50 km cannot hope
to explicitly represent lee wave motion at all. The same is true for most current
operational limited area models, whose resolutions are typically around 10 km. It is
only with grid spacings of 1–2 km that we can really expect to properly resolve
the waves. As for the prospects of explicitly resolving rotors, whose scales are
considerably smaller than those of the lee waves themselves, this is likely to remain
beyond the reach of operational models for some time.
The above limitations of NWP models mean that other techniques must be
used for forecasting lee-waves. Many of the methods used are based on obtaining
approximate solutions to (3.21) with profiles of wind and stability taken from
either NWP model or radiosonde data. Although it is possible to obtain numerical
solutions to (3.21) which utilize the full available resolution of the forecast or
radiosonde profiles, in the past the lack of computing resources available to bench
forecasters has necessitated far simpler approaches. For example, the widely used
Casswell (1966) calculation involves the assumption that
2 decays exponentially
with height and an estimate for this vertical decay scale is obtained using profile
data at only three heights. Because it utilizes such a small amount of information
about the atmospheric state, and incorporates no detail of the actual orography, the
Casswell method can at best provide only a very crude forecast of lee-wave activity.
154 P.L. Jackson et al.

Nowadays numerical solutions to (3.21) can be obtained (e.g. Lane et al. 2000;
Vosper and Mobbs 1996) with, by modern standards, only very modest computing
resources. Predictions can be made of the wavelength and vertical structure of
the resonant lee waves with very fine vertical resolution (e.g. tens of metres). An
example of such an approach, adopted for use in opertational forecasting is given by
Shutts (1997). A set of solutions to (3.21) is obtained over a range of horizontal
wavevector directions (centered on the low-level wind direction) and then one
particular solution is chosen. The actual vertical velocity associated with the chosen
wavemode can only be calculated if the spectral composition of the underlying
orography is known. Since the Fourier spectrum of real orography is generally
very irregular and for practical purposes can only be computed for a discrete set
of wavelengths, Shutts instead suggests assuming a power-law dependence (on the
horizontal wavenumber) for the vertical velocity at some height near the ground
(the wave “launching height”). A vertical profile of horizontally averaged vertical
velocity can then be estimated by applying suitable scaling to the solutions to (3.21).
Shutts (1997) has demonstrated the success of the technique by comparing forecasts
with observations.
Although the Shutts method can provide the forecaster with a prediction of wave
amplitude and horizontal wavelength, it does not provide a detailed picture of the
lee-wave motion. In terms of assessing the likelihood of rotors, since (3.21) is
based on inviscid theory which takes no account of boundary-layer processes, it
cannot be used to predict how the lee waves will affect the near-surface winds.
Detailed predictions such as these can only be made using a more complex three-
dimensional numerical model. While the computational requirements of using a
complex NWP model are, in practice, too restrictive, it is possible to obtain detailed
three-dimensional solutions using simplified models. One such example is provided
by Vosper (2003) who constructed a numerical finite-difference model to solve the
linearized equations of motion, the Boussinesq version of (3.9)–(3.11). The model
provides detailed predictions of the wave field over the real complex orography
during steady upwind conditions and the simplifications to the equations of motion
permit these solutions to be obtained with only a relatively modest computer, such
as a desktop PC. The neglect of nonlinear terms and moist and turbulent physical
processes undoubtedly compromises the accuracy of the predictions, however
Vosper (2003) demonstrated that for the Isle of Arran, which contains mountain
peaks as high as almost 900 m, the model provides accurate predictions of lee waves.
Comparisons with radiosonde observations show that, with sufficiently accurate
upwind profile data, both the amplitude and phase of the waves can be accurately
predicted.
Recent developments to the above linearized numerical model include the
incorporation of a linear turbulent boundary layer scheme (King et al. 2004) which
allows the representation of the influence of waves on the near-surface flow. The
model, 3DVOM (3-D Velocities Over Mountains), is currently used operationally
at the UK Met Office to produce 1 km-horizontal resolution predictions of the
lee-wave field every 6-h for separate hilly regions of the UK and Ireland, as well
3 Dynamically-Driven Winds 155

as the Falkland Islands. An example of such a forecast for the Pennines in northern
England, validated against observations made by a research aircraft, is shown in
Fig. 3.16. The model provides a good quantitative description of the wave field,
both in terms of wave amplitude and phase. Also shown are the predictions made
by a full NWP model (the Met Office Unified Model) run in a nested configuration
at 1 km resolution. The results are similar to those of the far simpler and cheaper
3DVOM calculation.
For larger mountains, where the linearized equations are no longer valid, we
would not expect a linear model to provide a quantitatively accurate forecast,
and a fully nonlinear model is required. Such an approach has been adopted as
a research activity in support of various recent field campaigns. For instance, the
NRL COAMPS model was used routinely to provide guidance for mission planning
during T-REX. Wave forecasts for the Sierra Nevada and Owens Valley area were
provided throughout the campaign on a series of nested grids, the finest of which
had a 2 km resolution (Grubišić et al. 2008). These forecasts proved to be a highly
useful and valuable part of the experiment planning process. An example of such
a forecast is shown in Fig. 3.17. The requirements of such models are discussed in
detail in Chap. 9, along with the issue of predictability of nonlinear mountain waves.
Compared with the techniques available for forecasting lee waves, the knowledge
available for forecasting rotors amounts to very little. Field campaigns have
provided some insight into the conditions under which rotors occur, but this tends to
be location-specific and the scientific knowledge is not sufficiently well developed
to be applied more generally. For example, the occurrence of a strong upwind
temperature inversion at, or around, the mountain height, along with unidirectional
flow across the mountain ridge, is known to be conducive to lee-wave motion
and it is generally thought that backward wind shear above the mountain top
will promote rotor formation. While these ideas are not inconsistent with current
scientific understanding, the theoretical and numerical studies which have addressed
the problem are so far too limited to provide any firm guidance as to the kind of rotor
motion which might occur, let alone the severity of the turbulence which might be
encountered. The ability of the very high-resolution numerical models to simulate
the dynamics of rotors (e.g. Doyle and Durran 2007) provides some hope that in
the future operational NWP models will have the ability to explicitly represent rotor
effects. Note, however, that much further research in this area is required, both in
terms of increasing our understanding, and also of the techniques needed to make
the best use of the high-resolution models.

3.4 Downslope Windstorms

Downslope windstorms have a long and rich history of scientific investigation and
have formed an important part of the meteorological literature for over 70 years.
They are a type of large-amplitude mountain wave that can occur downwind of
mountain barriers resulting in strong, often gusty low-level flow that accelerates
156 P.L. Jackson et al.

Fig. 3.16 3DVOM forecasts of lee-wave vertical velocity (blue line) over the Pennines, northern
England, UK, valid at 12 UTC on 17 November 2004. The forecasts are compared with
observations made by the UK FAAM research aircraft along east–west flight legs at 1,350 and
1,080 m above sea level (black line). Also shown are the predictions made by the Met Office
Unified Model run at 1 km resolution (red line) and the underlying terrain (shaded)

Fig. 3.17 Vertical velocity


(colour, interval 0.5 m s1 )
and potential temperature
(contour interval 6 K) from a
real-time 30-h COAMPS
forecast for the innermost
2 km resolution grid valid at
2100 UTC 25 March 2006
during T-REX IOP 6. The
cross section illustrates the
type of product that was
available for mission
planning purposes (From
Grubišić et al. 2008, Fig. 5.
© American Meteorological
Society. Reprinted with
permission)
3 Dynamically-Driven Winds 157

down the lee slopes of the mountain. Under the right conditions, wind speeds in
downslope windstorms can be two to three times the wind speed at the mountain top
level. Necessary ingredients for downslope windstorms include a sufficiently large
mountain barrier, as well as strong across-barrier winds and a stable atmosphere,
both near the mountaintop level. They are observed in many places downwind of
mountains and are often given local names.
Several mechanisms have been proposed to explain the strong winds associated
with these events. These include an analog to hydraulic flow (Long 1954; Smith
1985; Durran 1986; Klemp and Durran 1987), the amplification of vertically
propagating gravity waves by constructive interference as they reflect off a pre-
existing critical layer, usually near the tropopause (Klemp and Lilly 1975), and
a similar reflection of upwardly propagating waves off a self-induced critical
layer created by wave breaking (Peltier and Clark 1979). The theoretical and
practical understanding of downslope windstorms has also been informed by rich
datasets from several critically important field programs, most notably the Pyrenees
Experiment (PYREX Bougeault et al. 1993), the Alpine Experiment (ALPEX Smith
1987), the Mesoscale Alpine Programme (MAP Bougeault et al. 2001; Smith et al.
2007) and the Terrain induced Rotor Experiment (T-REX Grubišić et al. 2008).

3.4.1 Observations

Downslope windstorms are observed in many places downwind of mountains and


are often given names. Locally named downslope windstorms in specific areas in-
clude: Rocky Mountains (Boulder windstorm, chinook), Western Washington (East
winds), Alps (deep foehn winds – refer to Chap. 4 in this volume), Eastern Adriatic
(bora winds), Southern California (Santa Ana winds), Argentina (Zonda), Utah
(Wasatch winds), southeast Alaska (Taku wind) etc. These winds are sometimes
categorized depending upon whether they are a warm wind or a cold wind. Warm
downslope windstorms are often referred to as foehn or chinook-type winds, while
cold downslope windstorms are often called bora-type winds. Chinook or foehn
events normally occur when an existing cold pool is displaced by an air mass
descending a lee slope that is warmed adiabatically as it descends (see Chap. 4
for a detailed discussion). It is thought that latent heat release on the windward
side of the mountain barrier may contribute to the relative warmth of foehn winds.
Bora-type downslope windstorms typically occur (in Western Washington State, and
on the Adriatic coast) when cold continental air displaces a warmer maritime air
mass, so is manifested as a relatively cold wind. The circumstances forming a bora
may instead form a gap wind if there is a channel that cuts through the mountain
barrier. The classification of downslope windstorms as bora- (cold) type or as foehn-
(warm) type is not always helpful, as it is sometimes difficult to make a definitive
categorization and some phenomena like the Boulder downslope windstorm can
occur as both warm and cold events. Also, a foehn is a general warm downslope
wind that may or may not be classified as a windstorm.
158 P.L. Jackson et al.

One of the best known downslope windstorms is the storm that occurs period-
ically, usually during winter, in Boulder Colorado, that is characterized by very
strong winds with peak gusts as high as 60 m s1 . This downslope windstorm has
been studied extensively due to its intensity, but also due to the high concentration
of meteorologists who live nearby and directly experience it. Brinkmann (1974)
analysed 20 Boulder windstorms finding that they are characterized by strong
gusts, are non-stationary, are associated with a surface pressure minima under the
disturbance, followed by a pressure jump downstream, that results from a hydraulic
jump. Brinkmann found that common upstream conditions include an inversion or
stable layer just above the mountain top, in combination with strong winds at that
level. The most studied Boulder downslope windstorm event is that of 11 January
1972 (e.g. Klemp and Lilly 1975, 1978; Clark and Peltier 1977; Peltier and Clark
1979, 1983; Clark and Farley 1984; Durran 1986; Doyle et al. 2000), since it was
the subject of special aircraft observations that provided rare in-situ observations
of the mountain wave structure (Lilly and Zipser 1972; Lilly 1978) (Figs. 3.18
and 3.19). This event was characterized by a very large amplitude mountain wave
extending through the troposphere into the lower stratosphere. A stable layer below
the mountain wave is behaving like a hydraulic supercritical flow layer as the
isentropes lower in the lee of the mountain (Fig. 3.18) with associated strong winds
in excess of 60 m s1 (Fig. 3.19) and terminating in a hydraulic jump that is
indicated by sharply rising isentropes and decelerating wind speeds. The surface
winds during this event were very strong and gusty (Fig. 3.20). Observations of
severe gustiness are quite common in downslope windstorms and may be caused by
subrotors embedded in the flow. For the 11 January 1972 Boulder windstorm, gusts
appear to have a period of slightly more than a minute, however longer period wind
pulsations have also been observed for less severe events (Durran 1990). The gust
structure of a moderate Boulder windstorm was documented by Neiman et al. (1988)
using Doppler lidar. They found the gusts had a period of approximately 4 min,
extended several hundred meters above the surface and appeared as a coherent
structure that propagates downwind at approximately the mean wind speed.
The foehn-type wind (see Chap. 4) can take the form of a downslope windstorm.
An important similarity between the 11 January 1972 Boulder windstorm in
Figs. 3.18–3.20 and the deep Alpine foehn (look ahead at Fig. 3.30 (bottom)) is
the presence of a weakly stable, nearly mixed region above the descending flow.
The Alpine deep foehn case additionally had a nearly mixed region below the
flowing layer on the upstream side. An example of an Alpine foehn downslope
windstorm event from MAP is shown in Fig. 3.21 (Jiang and Doyle 2004). It has
many similarities with the 11 January 1972 Boulder downslope windstorm shown
in Figs. 3.18–3.20: specifically the lower stable layer that descends and accelerates
that is surmounted by a turbulent unstable region with sharply rising isentropes and
across-barrier flow decreasing or reversing.
The bora is a downslope windstorm that occurs along the northeastern Adriatic
coast from Trieste, Italy to southern Croatia (Dalmatia) and has been frequently
documented in the European literature since the mid-nineteenth century (Yoshino
1976; Jurčec 1981; Grisogono and Belušić 2009). Other bora-like winds that have
3 Dynamically-Driven Winds 159

Fig. 3.18 Cross section of potential temperature field observed in the 11 January 1972 downslope
wind storm illustrating the strong mountain wave extending through the troposphere and the lower-
level zone of supercritical flow and hydraulic jump between Boulder and Denver (From Lilly 1978.
© American Meteorological Society. Reprinted with permission)

Fig. 3.19 Cross section of u-component of the wind (m s1 ) from aircraft flight data and sondes
observed in the 11 January 1972 downslope wind storm (From Klemp and Lilly 1975. © American
Meteorological Society. Reprinted with permission)
160 P.L. Jackson et al.

Fig. 3.20 Anemometer trace from Southern Hills Junior High School in Boulder Colorado, on 11
January 1972 from 2,000 to 2,330 MST. Speed is in miles per hour (1 mph D 0.45 m s1 ) (From
Klemp and Lilly 1975. © American Meteorological Society. Reprinted with permission)

been studied occur near Novorossiysk on the Black Sea’s northeastern coast, at
Novaya Zemlya in the Russian arctic, and near Lake Baykal (Yoshino 1976). The
bora occurs under similar synoptic conditions as the gap flow through the fjords
and windstorms along the western coast of North America, described in Sect. 3.5.
A strong anticyclone over eastern or/and central Europe often in conjunction with
cyclonic conditions in the Mediterranean results in a strong pressure gradient
directed across the Dinaric Alps that separate the interior of the continent from
the Adriatic coast (Tutiš and Ivančan-Picek 1991; Ivančan-Picek and Tutiš 1995,
1996). Arctic air associated with the anticyclone deepens and eventually spills
through passes onto the coast (Jurčec 1981; Yoshino 1976; Smith 1987), usually
resulting in a decrease in air temperature. The strongest bora winds occur during
winter and at night, with generally ten “bora days” per month each winter season
(Yoshino 1972; Bajić 1989). A stable layer usually exists at mountain crest level
(Yoshino 1976) indicating the importance of hydraulic dynamics in describing
the flow.
Before ALPEX it was thought that the bora was a type of katabatic flow –
accelerating downslope due to its density. It is now understood that the bora is a
type of downslope windstorm, with hydraulic/mountain wave characteristics (Smith
1987). Klemp and Durran (1987) conducted two dimensional numerical simulations
of the bora and found that the flow resembled hydraulic flow over an obstacle,
and that the pressure gradients created by thermal differences on either side of the
Dinaric Alps were sufficient to generate the bora. The bora winds are strongest
downwind of gaps in the Dinaric Alps and can maintain their speed as localized jets
across the Adriatic sea for some distance (Zecchetto and Cappa 2001), presumably
decreasing speed when a transition from supercritical shooting flow to subcritical
flow occurs in a hydraulic jump. Gohm and Mayr (2005) simulated an Adriatic
deep bora case with a high resolution mesoscale model and found the strongest bora
3 Dynamically-Driven Winds 161

Fig. 3.21 Cross section through the central Alps on 21 October 1999 showing (a) isentropes
(every K), backscatter from SABL lidar (in grayscale) depicting clouds and regions of strong
turbulence (indicated by symbol ƒ), (b) has the along-flight wind component (every 2 m s1 )
added in black contours with the region of reversed flow hatched. The flow pattern resembles
classical downslope wind event with plunging isentropes indicating the shallow hydraulic flow
separated from the air aloft by a turbulent region with overturning isentropes and a wind reversal
(From Jiang and Doyle 2004. © American Meteorological Society. Reprinted with permission)
162 P.L. Jackson et al.

flow downwind of gaps. They found the simulated bora was diurnally modulated:
during the night the stable surface layer upstream increased the bora strength, while
the developing convective boundary layer during the day reduced the gravity wave
amplitude and strength of the downslope bora flow. Gohm et al. (2008) simulated a
bora downwind of a gap that formed a jet with hydraulic jump features in the middle
of the gap and lee wave/rotor features on the edges of the jet.
In western Washington, downslope windstorms have been documented to occur
when cold arctic air and associated high surface pressure moves southward over the
interior of the continent. The coast-parallel Cascade mountain chain separates the
interior arctic air from the maritime air on the coast resulting in a large pressure
gradient across the mountain barrier as the continental high moves south. In this
situation, strong downslope winds can form, especially downwind of passes or gaps
in the mountain barrier such as the Stampede gap (Mass and Albright 1985; Reed
1981) where winds during these events typically reach 15–25 m s1 (Mass and
Albright 1985). The air flows down the slope and over the inland waters of Puget
Sound where a readjustment in boundary layer depth takes place and wind speeds
diminish with the flow splitting around the Olympic Mountains resulting in strong
easterly flow through Juan de Fuca Strait.
A very similar downslope windstorm has been documented near Juneau in
Alaska, the so-called Taku windstorm. This windstorm also involves downslope
acceleration in the lee of a gap (Colman and Dierking 1992; Dierking 1998; Bond
et al. 2006) under very similar synoptic conditions as the Cascade windstorms. In the
case of the Taku wind, there is an interplay between downslope windstorm dynamics
as discussed below, and gap wind dynamics discussed in Sect. 3.5, since this area is
also comprised of many gaps dissecting the coastal mountain barrier, and both gap
winds through channels and the stronger downslope winds tend to occur at the same
time with the lee trough induced by the downslope windstorm enhancing gap flows
through local channels.
The Santa Ana is a type of downslope windstorm occurring along the coast of
southern California when a large area of high pressure forms over the Great Basin
between the Rocky Mountains and Sierra Nevada Mountains. When a thermal low
pressure area over coastal California is advected offshore, the pressure gradient and
wind speeds can be enhanced. Santa Anas tend to occur most frequently between
October and February with December being the month of peak occurrence (Raphael
2003). About 20 events occur per year with the average duration being 1.5 days
(Raphael 2003). Santa Anas are often enhanced during the night by a land breeze
with typical observed wind speeds of 15–20 m s1 , gusting to over 25 m s1 . The air
flowing out of the high turns anticyclonically, is channelled through the canyons and
warms adiabatically due to compression as it flows downslope towards the coast.
The hot, dry, strong wind is associated with increased wildfire risk in California
and can modify temperature and currents in the nearby coastal upper ocean. The
winds can persist offshore for a hundred or so kilometers and are associated with
high wave conditions on the northeast side of the Southern California Bight Channel
Islands.
3 Dynamically-Driven Winds 163

3.4.2 Theory

Downslope windstorms are large-amplitude mountain waves that undergo a hy-


draulic transition giving strong, supercritical flow on the lee slope. There is a
rich history of theoretical work on this phenomenon that has suggested several
mechanisms to account for downslope windstorms. These theories have been well
described and developed in a number of excellent reviews of this subject (e.g.
Smith 1979, 1989; Durran 1990; Baines 1995), and so will not be repeated in
detail here. The primary theories include an analog to hydraulic flow (e.g. Long
1954; Smith 1985; Durran 1986; Klemp and Durran 1987), the amplification of
vertically propagating gravity waves by constructive interference as they reflect off
a pre-existing critical layer, usually the tropopause (Klemp and Lilly 1975), and
a similar reflection and amplification of upwardly propagating waves off a self-
induced critical layer created by wave breaking (Peltier and Clark 1979). While
there has been debate over the relative importance and frequency of downslope
windstorms caused by these different mechanisms, it seems that the fundamental
dynamics are qualitatively explained as an analog to the hydraulic flow of water
over an obstacle resulting in rapid, supercritical flow along the obstacle’s lee slope,
which terminates in a turbulent hydraulic jump (see subsection 3.2.3). In the lower
left panel of Fig. 3.3, the supercritical flow of hydraulic theory over the lee slope,
corresponds to the zone of strong downslope winds near the surface.
The first mechanism, based on two-layer hydraulic theory with a deep upper
layer, was originally proposed by Long (1954), and used by Houghton and Kasahara
(1968), Houghton and Isaacson (1968) and Arakawa (1968, 1969). In the classical
hydraulic approach the flow of air over a mountain is modeled by fluid flowing over
a barrier using the shallow water equations (Eqs. 3.6–3.8, development described
in Sect. 3.2.3). Strong winds occur on the lee slope when the fluid (air) goes
from subcritical flow upstream, to supercritical flow over the mountain, becoming
subcritical again downstream as kinetic energy dissipates in a turbulent hydraulic
jump. In a classical hydraulic model, if the incoming subcritical flow is too rapid
to become critical at the mountain crest which acts as a control point, then a
wave can propagate upstream to “correct” the incoming flow. When applied to
the atmosphere, the main weaknesses of this theory are that it does not allow
continuous stratification, or vertical propagation of energy above either the free
surface or rigid lid upper boundary condition. However this theory is attractive
because of its simplicity and its apparent inclusion of important features observed in
downslope windstorms. Also, it is not uncommon for the atmosphere to have a sharp
change in stability, via elevated inversions and stable layers, which encourages the
atmosphere to behave as a simple hydraulic system. When a stable layer is present
near mountaintop level, hydraulic effects can occur.
A mechanism for downslope windstorms which would allow for the possibility
of vertical energy transport (Klemp and Lilly 1975) is the amplification of linear
vertically propagating gravity waves, by waves reflected off the tropopause (or some
other layer where the Scorer parameter, (3.22a), changes rapidly). The presence of
164 P.L. Jackson et al.

Fig. 3.22 Isentropes from a two-dimensional numerical simulation of a two-layer atmosphere over
a 600 m high mountain. Flow is from left to right. (a) Top of the lower stable layer is at 1,571 m
(one quarter of a vertical wavelength). (b) Top of the lower stable layer is at 3,142 m (one half of a
vertical wavelength) which corresponds to nonlinear resonance resulting in shooting supercritical
flow on the lee slope terminating in a hydraulic jump (Figure from Durran 1986. © American
Meteorological Society. Reprinted with permission)

vertically propagating mountain waves can be diagnosed from the isentropes. If


there is an upstream tilt of the isentrope phase, this indicates vertically propagating
wave energy that can be reflected downward by a critical layer (see Fig. 3.2 for
examples of this). The wave amplitude is determined by the superposition of these
upward and downward propagating modes and will depend upon whether or not
the atmosphere is “tuned” to constructive interference. Klemp and Lilly found that
the strongest downslope wind response occurred when the tropopause was located
half a vertical wavelength above the ground. With this linearized model of the
flow equations the vertically propagating and amplifying gravity waves cause lower
pressure and accelerated winds on the lee slopes with the wind strength proportional
to the mountain height.
As the theory is linear, its general applicability to large amplitude waves has
not been determined (Durran 1986). It seems that while the mechanism involving
linear theory proposed by Klemp and Lilly in which the predicted downslope
wind response scales with H2 can account for the strongest Boulder downslope
windstorms in the case where the profiles of U and N are constant (i.e. a single
layer structure), it is unsuccessful with nonlinear effects becoming important when
there are vertical variations in these profiles and the flow more closely resembles
classical hydraulic flow (Durran 1986, 1990) as shown in Fig. 3.22. Durran (1986)
examined the importance of nonlinear amplification for an atmosphere with a two-
layer stability structure, finding that in this case, nonlinear effects become important,
even with non-dimensional mountain heights as low as 0.3 – values which would
have insignificant nonlinear effects in an atmosphere with a single layer stability
structure.
3 Dynamically-Driven Winds 165

The third mechanism proposed (Clark and Peltier 1977, 1984; Peltier and Clark
1979, 1983; Clark and Farley 1984), suggests large amplitude waves and downslope
winds occur after a developing vertically propagating mountain wave breaks. Wave
breaking occurs when the isentropes become vertical and the air becomes unstable
and overturns. The area of wave breaking creates a self-induced critical layer with
strong mixing and wind reversal. In this layer, where the wind component in the
direction of the horizontal wavevector goes to zero, the vertical wavelength and the
rate of vertical energy propagation also tends to zero as discussed in Sect. 3.2.2.
In this situation, if the distance between the self-induced critical layer and the
mountain is suitably tuned, it may trap and reflect upward propagating waves
resulting in resonance and strong surface winds. Peltier and Clark’s (1983) theory of
wave amplification for the numerically simulated downslope windstorm uses linear
theory and the assumption that the wave breaking region acts like a critical layer
where the Richardson number, Ri D N2 /(dU/dz), is less than ¼. According to Peltier
and Clark’s results, the critical layer will produce amplification only when it is
located about ¼ C n/2 (n D 0,1,2 : : : ) vertical wavelengths over the terrain. The self-
induced critical layer produced by wave breaking is similar to a turbulent mean-state
critical layer that would be produced when the stability becomes too low to provide
a sufficient restoring force for gravity waves, which occurs when the Richardson
number is less than ¼. It is believed that nonlinear waves that encounter such a
critical layer are reflected by it without losing much amplitude (Durran 1990).
The importance of wave breaking for the 11 January 1972 downslope windstorm
as well as for other cases with constant upstream wind and stability (constant U
and N) has been verified by many numerical simulations (e.g. Doyle et al. 2000).
Whether or not wave breaking occurs will depend on the non-dimensional mountain
height exceeding a critical value. For flow across a two dimensional ridge, Huppert
and Miles (1969) found the critical non-dimensional mountain height for gravity
wave breaking was 0.85. Smith and Grønås (1993) found a critical non-dimensional
mountain height of 1.1 ˙ 0.1 for an isolated Gaussian hill in three dimensional
flow. In each simulation when the critical non-dimensional mountain height was
exceeded, the wind stagnated aloft over the lee slope, isentropes overturned in wave
breaking and there was a strong downslope wind response. A critical level will
explicitly exist when there is a reverse shear in the approaching wind – U decreases
with height and perhaps even reverses. If this occurs at the correct altitude for a
resonant response, then strong downslope winds can result.
The theory proposed by Smith (1985) and Smith and Sun (1987) extends classical
hydraulic theory, by using a concept of local hydraulics in which the atmosphere
can have constant stratification rather than layers separated by a density interface,
and the hydraulic nature of the flow only manifests itself when a transition to
supercritical flow occurs on the lee slope below a layer of stagnant fluid. In local
hydraulics, the flow approaching the mountain may “select” a critical streamline
which splits becoming the upper boundary of the accelerated flow (Fig. 3.23),
decoupling it from a turbulent layer of well-mixed air aloft. It is the presence of
this stagnant region that provides the decoupling allowing the flow to produce a
shallow layer-like flow that can spill over terrain and resemble classical hydraulic
166 P.L. Jackson et al.

Fig. 3.23 (a) Schematic diagram of a downslope windstorm based on aircraft observations and
numerical simulations. In this case a certain critical dividing streamline encompasses a turbulent
region of uniform density ¡c indicated by the inverted “v” shapes. In this local hydraulic approach,
continuously stratified air approaches the mountain barrier and acquires hydraulic characteristics
of supercritical flow after the critical streamline divides. (b) Flow over a mountain from the Smith
(1985) local hydraulics model. U0 D 20 m s1 , N0 D 0.01 s1 (From Smith 1985. © American
Meteorological Society. Reprinted with permission)

supercritical flow and even produce a hydraulic jump. In this theory, when the non-
dimensional mountain height, Ĥ D NH/U is small, or if a wind reversal exists at the
wrong altitude the fluid may not form a neutral decoupled region with hydraulic
flow, but instead form continuous vertically propagating gravity waves (Smith and
Sun 1987). According to this theory, amplification of a mountain wave is possible
across a range of critical layer heights from 1/4 C n to 3/4 C n, where (n D 0,1,2 : : : )
vertical wavelengths above the terrain, in contrast to the predictions of Peltier
and Clark.
Durran and Klemp (1987) and Bacmeister and Pierrehumbert (1988) used
numerical experiments to compare the theories of Smith (1985) and Peltier and
Clark (1979, 1983) and found that when there is a mean state critical layer, the
mountain waves interact with this layer to form a local region of turbulent flow, and
that the essential elements of hydraulic theory explain the acceleration on the lee
slopes of mountains. This occurred in Durran and Klemp’s experiments whenever
Ĥ exceeded 0.2 and there was a mean state critical layer between ¼ and ½ vertical
wavelengths above the terrain (Durran 1990). In their experiments a deep layer with
Ri < ¼ developed above the topography and a strong downslope wind response was
found with this configuration when Ĥ exceeded 0.4.
Reed (1981) and Mass and Albright (1985) were able to use a simple Bernoulli
equation and gap wind dynamics based on a balance between the pressure gradient
force and inertia to account for the observed wind speeds in the cases they studied
in US Pacific Northwest. Colle and Mass (1998) set out to examine the relative
importance of mountain wave/hydraulic dynamics versus simple gap flow dynamics
by analysing a number of past events in western Washington State, as well as
using idealized mesoscale model simulations. They found that a combination of the
3 Dynamically-Driven Winds 167

across-barrier pressure gradient and across-barrier flow at crest level explained 82%
of the variance in windstorm severity there. The presence of a critical layer (region
of flow reversal needed for reflection of vertically propagating waves) appeared to
be of less importance in storm evolution. However in cases with strong cross-barrier
flow at mountain crest level, the lack of a critical layer below 400 hPa resulted in
weaker surface winds unless a region of strong stability existed near crest level, in
which case the flow behaved hydraulically. For cases with weak cross barrier flow
at crest level, the presence of a critical layer was less important, as they postulated
a self-induced critical layer could be created and strong surface winds could
ensue.
One remarkable feature of many severe downslope windstorm events is the severe
gustiness of the wind. The observed gustiness of the 11 January 1972 Boulder
downslope windstorm (Fig. 3.20) was investigated by Clark and Farley (1984) using
two and three-dimensional numerical models. They found gust structures were only
apparent in the three-dimensional simulations and hypothesized that the gustiness
was produced by an intrinsically three-dimensional competition between wave
enhancement by gravity wave dynamics and wave destruction by convective insta-
bility. The apparent requirement of three-dimensions for gustiness was contradicted
by Scinocca and Peltier (1989) however who were able to produce realistic gusts
(which they call pulsations) in their two-dimensional simulations. Peltier and
Scinocca (1990) demonstrate that the simulated pulsations are due to Kelvin-
Helmholtz instability at the wind shear interface between the low level shooting
flow and the wave-induced stagnant layer. The Kelvin-Helmholtz instability is a
secondary instability that occurs after the downslope windstorm has formed when
Ri < ¼ in the flow and appears to initiate just upstream of the mountain crest.
Afanasyev and Peltier (1998) demonstrate the importance of the Kelvin-Helmholtz
mechanism and further demonstrate how in a fully three-dimensional simulation,
the Kelvin-Helmholtz billows are eroded by a convective instability and dissolve
into a turbulent background flow as they move downstream. The Kelvin-Helmholtz
instability as a mechanism to explain the pulsations has been suggested and modeled
for the bora as well (Belušić et al. 2007).
Most of the earlier theoretical work using linear theory, as well as numerical mod-
eling studies in the past has assumed that boundary-layer effects are not important
for downslope windstorms. However numerical simulations of the 1972 Boulder
windstorm by Richard et al. (1989) as well as work by Georgelin et al. (1994),
Ólafsson and Bougeault (1997), Peng and Thompson (2003), Smith et al. (2002,
2007) and Smith et al. (2006) has illustrated the importance of the boundary layer
on absorbing mountain wave energy and both attenuating downslope windstorm
strength and limiting its extent downwind. Georgelin et al. (1994) studied the
effect of subgrid scale orographic roughness on orographic flows from PYREX.
They found that correctly accounting for subgrid scale orography improved model
predictions. It reduced mountain wave amplitude, but increased blocking as well as
leeside low-level turbulence. Ólafsson and Bougeault (1997) conducted idealized
simulations to investigate the effects of rotation (f ) and surface friction on mountain
wave drag. They found that the effect of friction was to suppress wave breaking,
168 P.L. Jackson et al.

especially for NH/U less than 3, and that the combined effects of rotation and drag
were to constrain the drag and flow patterns to resemble those predicted by linear
theory. Smith et al. (2002) discussed the effects of a low-level stagnant layer on
the mountain wave, finding that it reduced the wave amplitude and was able to
absorb downward reflected gravity waves, preventing lee wave generation. Jiang and
Doyle (2008) modeled a downslope wind event in Owen’s Valley during the Terrain-
Induced Rotor Experiment, and found in this narrow valley that the downslope
winds varied diurnally through an interaction with the thermal structure in the
valley. Overnight, when the valley filled with cool air, strong downslope winds
were restricted to the upper lee slope. After sunrise, surface heating in the valley
weakened the downslope winds, eventually decoupling the valley air from the strong
westerlies above the mountain top. As the valley air continued to warm during the
afternoon, the mixed layer extended beyond the top of the valley which caused a
transition as a shallow layer of cooler air from the westerlies above the mountain top
descended and flowed across the whole valley. These simulations serve to illustrate
the important connections with diurnal mountain circulations (Chap. 2).
Until recently, “dry” dynamics were used to understand downslope windstorm
formation. Doyle and Smith (2003) used a numerical model to isolate the influence
of latent heat in mountain wave generation by running the COAMPS model with and
without latent heating and examining the differences (Fig. 3.24). They found that if
the latent heating in a shallow layer upstream was strong, the lee downslope wind
response increased significantly. This occurred because the latent heating resulted in
a decrease in stability aloft creating a critical layer there that decoupled the low level
flow allowing it to behave as a hydraulic layer. Such a mechanism – the interaction
between low level latent heating and the establishment of critical layer aloft resulting
in a downslope windstorm – suggests a mechanism explaining how warm foehn
winds are able to descend on the lee side of the Alps.

3.4.3 Models and Forecasting

The previous section has outlined the theoretical basis for understanding downslope
windstorm formation. This understanding has developed over time through a fusion
of analytical approaches involving careful analysis and judicious simplifications to
the governing equations, as well as through idealized and realistic experiments with
numerical models of increasing complexity and accuracy. This section will discuss
how this theoretical understanding can be applied to model and forecast downslope
windstorms using synoptic interpretation and NWP as well as statistical methods.
More information on the application of mesoscale models and NWP models is
provided in Chaps. 9, 10, and 11.
While NWP models are essential to understand and forecast most meteorological
phenomena, there are limitations in the application of NWP to downslope wind-
storms. These limitations, such as model resolution and the vertical coordinate
system, as well as predictability will be discussed later in this section as well as
3 Dynamically-Driven Winds 169

Fig. 3.24 Cross section of potential temperature (contours every 3 K) and vertical velocity
(contours every 2 m s1 with upward velocities greater than 2 m s1 shaded) for a COAMPS
simulation of flow from south (left) to north (right) of a mountain wave event across the central
Alps on 20 September 1999, (a) includes latent heating and (b) has latent heating turned off.
Note the more strongly plunging flow and downslope wind response in (a). The shallow upslope
latent heating has tuned the atmosphere for nonlinear resonance (From Doyle and Smith 2003.
Reproduced by permission of the Royal Meteorological Society)

in Chaps. 9 and 10. However, even if NWP models do not accurately represent
mountain waves directly, they can still provide important information on upstream
vertical profiles of wind and stability. This information may be interpreted using
knowledge of downslope windstorm mechanisms, to infer whether or not a strong
downslope wind response is likely. Let us summarize some of the key findings from
observations and theoretical work that relate conditions in the upstream environment
to downslope windstorm occurrence. The type of mountain wave, and whether or
not any waves will occur depends on topography, wind and stability profiles. The
following conditions are necessary, although not necessarily sufficient for a strong
downslope wind response:
• A significant mountain barrier. Long mountain barriers that are asymmetrical
with a shallow windward slope and a steep lee slope are particularly conducive
to downslope windstorm development (Smith 1977; Lilly and Klemp 1979;
Hoinka 1985). This limits the occurrence of downslope windstorms to specific
geographic regions.
• The presence of a moderate to strong wind component in the across-barrier
direction (7–15 m s1 or more, depending on the mountain) at an elevation near
the mountaintop level.
• Stable stratification. This is necessary to provide a restoring force for a gravity
wave response.
• A non-dimensional mountain height, Ĥ D NH/U, somewhat close to 1 (i.e. Ĥ not
1 or 1). If Ĥ is too small, the flow will easily go over the mountain without
a mountain wave response. If Ĥ is too large, the flow will be completely blocked
by the mountain barrier and not able to flow over the mountain.
170 P.L. Jackson et al.

A strong downslope wind response closely resembles supercritical hydraulic


flow. The circumstances that encourage this type of flow can arise in several ways.
One way is if the background atmospheric state has a layered stratification. The
presence of an inversion or isothermal layer at or just above the mountain top
level enhances hydraulic effects and the possibility of a strong supercritical flow
response on the lee slope. The importance of such a stable layer is verified in
the Boulder downslope windstorm climatologies of Brinkmann (1974) and Colson
(1954). Reduced stability or cross barrier winds further aloft can also decouple
the low-level flow and allow it to behave like a shallow hydraulic layer (Smith
1985; Baines 1995). If the cross-barrier winds increase aloft in the presence of an
inversion near mountain top level (i.e. there is forward shear), there is a possibility
of either lee waves or enhanced downslope winds. A rule of thumb that can be used
to differentiate the two mountain wave forms, is to compare the cross-barrier wind
component at mountain top level with the winds 2,000 m above the mountain top. If
the winds aloft increase to over 1.6 times the winds at mountain top level, then lee
waves rather than a downslope windstorm are more likely (COMET 2008).
Since gravity waves require a restoring force (buoyancy) and therefore a stable
atmosphere – they will not propagate through a neutral or near neutral turbulent
layer, which can therefore act as a mean-state critical layer (Ri < ¼) resulting in the
reflection of vertically propagating mountain waves. Therefore the presence of such
a mean-state critical layer above a stable layer can be diagnostic of a downslope
windstorm event. Vertically propagating mountain waves can also be reflected by a
self-induced critical layer (a layer where the cross-barrier wind component goes to
zero) that can form when vertically propagating mountain waves break, forming
a turbulent “dead” zone of air that effectively decouples the lower level flow,
allowing it to behave hydraulically. Such a wave breaking situation can occur in an
atmosphere with constant U and N where the mountain is large enough (Clark and
Peltier 1977). In addition to wave breaking resulting in a self-induced critical layer
and the existence of a mean-state critical layer, a critical layer can also be formed
by wind shear, when the wind component in the cross-barrier direction goes to zero,
either by the wind speed decreasing or by the direction shifting to a barrier-parallel
direction.
It is thought that jet streak dynamics could also play a role in downslope
windstorm development. Specifically the right-exit region of a jet streak is a zone
of upper level convergence and downward vertical velocity. If this zone occurs near
a ridgeline of sufficient height in a jet that is oriented perpendicular to the ridge,
then it can contribute to a subsidence inversion or stable layer near the mountain
top, which can lead to a downslope windstorm. Many of these factors for practical
forecasting of downslope windstorms are well illustrated in a COMET MetEd
interactive module on forecasting downslope windstorms (register for MetEd and
then go to http://www.meted.ucar.edu/mesoprim/mtnwave/) (COMET 2008).
The previous section has discussed numerical modeling of downslope wind-
storms from a research and process perspective – high resolution models have
been used extensively in idealized and realistic simulations to shed light on the
3 Dynamically-Driven Winds 171

mechanisms responsible for downslope windstorm events. There are also numerous
examples where numerical models have successfully simulated many aspects of
downslope windstorms, in hindcast mode (e.g. Doyle and Shapiro 2000; Colle and
Mass 1998). NWP models are of course a key tool for operational forecasting. Are
operational NWP models of use in downslope windstorm predictions? One major
issue is whether the models have sufficient vertical and horizontal resolution and
accuracy to capture mountain wave activity.
It is generally accepted that to properly resolve a wave-like feature in a
numerical model, there must be six to eight grid points per wave, so that the model
resolution must be finer than 1/6 of a wavelength. To properly resolve mountain
waves, model horizontal resolution should be on the order of 1 km, and certainly less
than 10 km. Current operational NWP models (as of 2011) are close to achieving
10 km or finer horizontal resolution, so in the future will be able to provide realistic
simulations of mountain wave activity and downslope windstorms, although there
are still questions concerning the predictability of these phenomena which will be
discussed below.
Other issues related to downslope wind forecasting by numerical models that
are summarized by Smith et al. (2007) include inaccuracies associated with the
vertical coordinate system over mountains. The commonly used terrain following
coordinate system has difficulty in accurately representing the horizontal pressure
gradient force. This led Schär et al. (2002) to propose a new coordinate system
that has smoother and more horizontal levels at mid and upper levels and thereby
reduce the truncation errors and high frequency noise in the model (see Chap. 9).
Klemp et al. 2003 found that errors and spurious gravity waves could be generated
in numerical models in the conversion between terrain-following and Cartesian
coordinates if the transformation of the advection terms and pressure gradient
terms was not done consistently. Another issue related to the sigma coordinate
system concerns the application of horizontal diffusion on sloping sigma surfaces.
Horizontal diffusion is commonly applied to control model errors. If it is applied
along the terrain following coordinates which slope with the terrain, this can conflate
vertical gradients with horizontal gradients leading to errors especially when there
are strong vertical gradients such as inversions (Smith et al. 2007). Zängl (2002b)
developed a true horizontal diffusion scheme for the MM5 model that was shown to
improve its ability to simulate topographic flows (see further discussion in Chap. 9).
Although increasing computer power and improving NWP models with ever-
increasing resolution holds promise for the prediction of downslope windstorms, the
predictability of these phenomena remains a major unresolved issue. Lorenz (1969)
hypothesized that the growth rate of perturbations increases as the scale of a phe-
nomenon decreases, implying that the predictability of mesoscale phenomena whose
scales are on the order of 10 km, may only be a few hours in advance. However, the
fact that downslope windstorms are forced by topography implies that they ought
to be inherently more predictable than mesoscale phenomena that do not have such
a strong organizing influence (Anthes et al. 1985). The enhanced predictability of
downslope windstorms was first demonstrated by Klemp and Lilly (1975), who used
a linear two-dimensional model to predict the occurrence of downslope windstorms
172 P.L. Jackson et al.

in Boulder, Colorado. They initialized their model using soundings from NWP
model forecasts, and made predictions valid approximately 5 h in advance of the
windstorm, successfully predicting events where observed gusts were in excess of
25 m s1 . More recent studies have not been as encouraging. Nance and Colman
(2000) used a high-resolution two-dimensional nonlinear mesoscale model that was
initialized with upstream soundings from 12 to 18 h NWP model forecasts in order
to support operational forecasting of downslope wind events in seven regions in the
United States between 1993 and 1997. They found that the model improved upon
existing forecasting methods which were based on a decision tree (Brown 1986), but
also had a high “false alarm” rate, implying difficulty in distinguishing windstorm
events from null events.
During MAP, many studies were able to successfully simulate mountain waves of
various types (e.g. Smith et al. 2002; Smith and Broad 2003; Doyle and Smith 2003;
Volkert et al. 2003). As pointed out by Smith et al. (2007) these successful cases
were characterized by stationary non-breaking waves, and results from other cases
that were characterized by non-stationary waves were not as successful (e.g. Jiang
and Doyle 2004; Smith 2004; Jiang et al. 2005; Doyle and Jiang 2006). Doyle et al.
(2000) compared two-dimensional simulations by 11 different numerical models
initialized with an identical sounding representing the 11 January 1972 Boulder
downslope windstorm event. While all models predicted downslope winds and wave
breaking in the lower stratosphere, the models had different temporal evolutions of
the wave breaking. Doyle and Reynolds (2008) tested the sensitivity of the same
downslope windstorm event to small variations in initial conditions by performing
ensemble simulations in a high-resolution two-dimensional numerical model. Each
ensemble member was constructed by perturbing the upstream sounding by small
amounts corresponding to typical radiosonde measurement errors. They found that
for this case there was a large spread in the ensembles with a 25 m s1 range in the
modeled downslope wind speed. Half the ensemble members simulated a trapped
lee wave response and the other half simulated large-amplitude wave breaking in the
lower stratosphere with a hydraulic jump and a strong downslope wind response.
The bimodal distribution of the ensemble members led them to conclude that when
there is such a transition across a regime boundary (between non-breaking and wave
breaking responses), the use of ensemble statistics such as the ensemble mean may
not be appropriate. Reinecke and Durran (2009a) used a three-dimensional NWP
model to conduct ensemble simulations of two downslope windstorms observed
during T-REX. One case was a wave breaking downslope windstorm, while the
second case was a layered case characterized by strong low-level stability in which
wave breaking was not important. For each case they conducted 70 ensembles
by perturbing the initial conditions across a range representing uncertainty in the
initial conditions. They found that the forecasts were very sensitive to the initial
conditions, especially for the wave breaking case where the difference in the
downslope wind speeds predicted by the strong ensemble members minus the weak
ensemble members grew to 28 m s1 after only 6 h of forecast. The layered case
was found to be somewhat more predictable with the difference in downslope wind
speeds growing to 22 m s1 after 12 forecast hours. The results of these recent
3 Dynamically-Driven Winds 173

downslope wind predictability studies indicating sensitivity to initial conditions, do


not suggest much confidence should be placed in the accurate and precise strength,
timing, or location of downslope winds predicted beyond 12 h in advance. Despite
this, operational NWP models are capable of predicting the overall synoptic setting
for these storms up to several days in advance. This currently allows a forecast
of downslope windstorm occurrence without necessarily being able to precisely
predict the location, timing or strength. As computer power continues to increase,
high resolution NWP models could be implemented in operational forecast settings
as a nowcasting tool to make short-range precise and accurate forecasts. Further
discussion on predictability issues can be found in Chap. 9 (Sect. 9.9).
Žagar and Rakovec (1999) discuss a dynamic downscaling approach to obtain
a high resolution wind field that they call “dynamic adaptation”. The approach
interpolates a coarse resolution NWP forecast onto a fine-resolution mesoscale
model grid, and then runs a simplified version of the mesoscale model for between
30 and 45 min to obtain a high-resolution wind field. The approach was successfully
applied to a strong wind event: a foehn in the Ljubljana basin in Slovenia, in
which diurnal boundary layer variations which cannot be properly represented
by the approach, were not important. The advantage of the approach is its low
computational cost, and it seems to hold promise as a forecast tool in an operational
setting for strong dynamically forced flows.
As a less computationally and technically intensive forecast tool, it is possible
to develop statistical models to predict downslope windstorms at specific locations.
Mercer et al. (2008) used a 10 year Boulder windstorm dataset together with a set
of 18 predictors to train and test statistical models to predict the downslope wind
speed. They tested linear regression and neural network models but found that a
support vector regression model performed best and was able to predict downslope
wind speed with an RMSE of 4–6 m s1 .
Given the likelihood of limited predictability (i.e. short prediction lead times), at
least for strong, highly nonlinear downslope windstorms, the ability to observe their
occurrence is very important. Mountain waves, including downslope windstorms
can be observed in a variety of ways. Of course, direct observation of the downslope
wind response at surface observing stations is the best indication of occurrence;
however stations do not exist in every location. Satellite imagery can also be useful.
If there is sufficient moisture for clouds to be present, a zone of descending air is
often indicated in satellite imagery by a gap in the cloudiness that may indicate
mountain wave formation.

3.5 Gap Winds

Gaps criss-cross mountain ranges all over the world. Gaps also exist between
islands, or islands and the mainland. They can be level or elevated with an elevation
maximum in the gap whence they are commonly called passes or cols. Gaps provide
a means for an air mass to advance to the other side of the barrier formed by
174 P.L. Jackson et al.

Fig. 3.25 Cold continental air crossing the Coast Mountains through gaps and emanating as
accelerating flows onto the Pacific Ocean. Flow is from right to left. Wind speed (see color
bar) over water is derived from Synthetic Aperture Radar (SAR) data. Image date: 0224 UTC
29 January 2008 (Data courtesy of Alaska SAR Demonstration [AKDEMO] of NOAA’s STAR-
Center for Satellite Applications and Research)

mountain ranges. A distinctive feature of most gap flows is their asymmetry between
the region upstream of the gap entrance and downstream of the gap. Upstream, the
layer which flows through the gap is thick and relatively slow moving (subcritical
in hydraulic terminology). As it approaches the gap, it thins and accelerates and
becomes supercritical downstream of the gap.

3.5.1 Observations

Regions in the world where gaps join two persistently different air masses see
frequent gap flows. In this subsection, we will first look at observations from the
earth’s atmosphere and the laboratory of flows through level gaps. Next we will
look at elevated gaps, which have a vertical constriction added to the lateral one to
form a pass or a sill.
The northern part of the North American Pacific coast where long, coast-parallel
mountain ranges separate cold wintertime continental air masses from warmer
maritime ones is a preferred region for gap flows. Figure 3.25 shows a striking
visualization of the surface wind field with jets emanating from various gaps in
the Coast Mountains of British Columbia and Alaska. High wind speeds are only
3 Dynamically-Driven Winds 175

visible downstream of gaps, which provide a connection to the continental air mass
east of the Coast Mountains. Highest wind speeds occur near the exit or even further
downwind as seen in the southern part of the figure close to the coast. There, several
gap jets merge to form one wide high wind speed streak. Even further away from
the gap exits, the flow slows down but has to pass through another gap formed by
the islands off the coast, where it reaches its overall maximum wind speeds near the
exit of that last gap.
Further visualizations of gap flow wind fields over water from synthetic aperture
radar (SAR) measurements can be found in Pan and Smith (1999), Winstead et al.
(2006), Liu et al. (2008), and Alpers et al. (2009), for example. Alpers et al. (2009)
provide highly resolved wind fields over the Adriatic Sea and the Black Sea during
six bora – gap flow events.
Shelikof Strait, formed by Kodiak Island and the Alaska Peninsula (Fig. 3.26),
is an example of a level gap. Colder air frequently lies northeastward on the other
side of the mountain range barely visible at the horizon of the figure. Research
aircraft flights (Lackmann and Overland 1989) provided the horizontal wind and
temperature field at 80 m AGL shown in Fig. 3.26. Wind speeds between the
entrance and exit regions nearly triple. Like in the previous example, the maximum
is close to the exit of the gap and not at its narrowest part as simple continuity
reasoning (“Venturi flow”) might suggest. Another important characteristic feature
is that the upstream air is colder than the dowstream air, here by about 4 K.
The Strait of Juan de Fuca (Fig. 3.27a) forms another level gap along the Pacific
North American coast. In wintertime it is fed by cold continental air from east of
the Cascade Range, which has passed through elevated gaps in the Cascade Range,
i.e. the Strait of Juan de Fuca is the second gap through which the air streams.
In the gap flow event shown in Fig. 3.27, the presence of hydrometeors provided
backscatter for a Doppler radar on board of the NOAA-P3 aircraft, from which the
three-dimensional wind field could be computed. The vertical structure of the flow
along the strait is shown in Fig. 3.27. Wind speeds between entrance and exit of
the gap double. It is remarkable that the abrupt thinning of the gap flow layer is
coincident with a further speed-up. The gap flow layer is isolated from the flow
aloft through a layer with southerly flow in almost perpendicular direction to the
gap flow.
Observations of gap flows in the laboratory (Armi and Williams 1993) repro-
duced in Fig. 3.28 show similar structures to the ones observed in the atmosphere.
Similar to its cold and stably stratified counterpart in the atmosphere, the fluid
in the laboratory experiment is continuously stratified, which is visualized by
dyeing selected isopycnals to give an equivalent to isentropic cross sections in the
atmosphere. An arrow indicates the narrowest part. When fluid is withdrawn at
the bottom of the channel, the flow is similar to the one observed in the Strait of
Juan de Fuca, shown in Fig. 3.27b. The flow accelerates through the channel as
the depth of the gap flow decreases. Near the exit, the flow accelerates even more
and its depth drops. A homogeneous stagnant layer forms as the streamlines split
at the gap exit. This layer isolates the gap flow from the fluid aloft. A similar,
nearly stagnant layer was observed in the Strait of Juan de Fuca example as well
176 P.L. Jackson et al.

Fig. 3.26 Horizontal wind vectors (see reference vector for scale), isotachs (m s1 , dotted), and
temperature (ı C, colored) of the gap flow through Shelikof Strait measured with the NOAA-P3
aircraft 80 m above the ocean on 28 March 1985. Inset shows a pseudo-3D view of Shelikof Strait
between the Alaskan Peninsula and Kodiak Island viewed from SSE (Adapted from Figs. 6 and 8
of Lackmann and Overland 1989. © American Meteorological Society. Reprinted with permission.
Inset figure © Google and NOAA)

as the location of the speed maximum closest to the ground (Fig. 3.27b). The
equivalent to withdrawing fluid near the bottom of the laboratory channel is having
the strongest pressure gradient in the flowing parts of the atmospheric gap flow close
to the surface. Since the temperature differences of the air masses on either side of
the gap are hydrostatically responsible for the bulk of the pressure difference, this
translates into having the strongest temperature differences between the upstream
and downstream side of the gap close to the surface.
3 Dynamically-Driven Winds 177

Fig. 3.27 (a) A pseudo-3D


view from the W of the Strait
of Juan de Fuca between
Vancouver Island and the
Olympic Mountains. (b)
Wind vectors in the along-gap
section and isotachs of
along-section speed (contours
in 3 m s1 steps and shading)
computed from Doppler radar
measurements onboard of the
NOAA P3 aircraft. (c) Wind
vectors and isotachs of the
along-section speed (dashed
contours in 3 m s1 steps
andshading) and isentropes
(solid, 3 K intervals) from
numerical simulations. The
gap flow in (b) and (c) is in
the offshore direction from
right to left, and is capped by
southerly flow around 2 km
MSL and onshore flow
further aloft. (Part (a) ©
Google and NOAA. Parts (b)
and (c) are from Colle and
Mass 2000. © American
Meteorological Society.
Reprinted with permission)

When fluid is withdrawn from mid-levels of the channel (Fig. 3.28b), the flow
response changes. The fluid accelerates through the gap and the flowing layer thins
and accelerates with a parabolic wind profile and the jet maximum at the center of
the gap flow layer at the exit of the gap. In addition to the isolating, stagnant and
178 P.L. Jackson et al.

Fig. 3.28 Steady continuously stratified flow through a contraction. The linear continuous stratifi-
cation is made visible by dyeing selected isopycnals in the upstream reservoir. The fluid accelerates
through the narrowest section of the contraction (marked by the arrow) as the flow moves from left
to right. (a) Level gap with fluid withdrawn through a slit at the bottom of the channel downstream
of the gap. (b) Level gap with fluid withdrawn at the center height of the channel downstream of the
gap. Vertical dye streaks are distorted by the flow and show the self-similar velocity profile of the
Wood (1968) solution (cf. subsection 3.5.2). (c) Elevated gap, i.e. a sill was added at the narrowest
section of the gap. Note the upstream flow in the contraction is the self-similar solution but that
this solution changes in the neighborhood of the sill and accelerates down the lee slope (Adapted
from Armi and Williams 1993. Reproduced by permission of Cambridge University Press)

homogeneous layer aloft, another such layer forms at the bottom. Just like for the
case of bottom withdrawal, the flow is self-similar in the sense that at any position,
the density and velocity profiles can be scaled with the height of the moving
stratified layer. This self-similar structure upstream is a necessary component of
any steady stratified flow from a reservoir and for it to exist, at least an isolating
3 Dynamically-Driven Winds 179

Fig. 3.29 A pseudo-3D from the N of the Wipp Valley in the central Alps at the border between
Austria and Italy. The main gap with an altitude of about 2.1 km MSL with a small deep incision
of the Brenner Pass at only 1.4 km MSL is embedded in the main Alpine crest (approximately
3 km MSL). The Wipp Valley merges with the Inn Valley via a nearly 200 m terrain drop at
approximately right angles (© Google and NOAA)

stagnant intermediate layer needs to be formed above (cf. Armi and Williams 1993).
The equivalent to a withdrawal at mid-levels of the gap flow is to have the strongest
temperature differences between the upstream and downstream side and thus the
largest hydrostatically produced pressure gradient at mid-levels.
Elevated gaps, which are more commonly called “passes”, are constricted both
laterally and vertically. One extensively studied one, Brenner Pass in the Alps,
is shown in Fig. 3.29, a pseudo-3D view looking at the gap from the north.
Topography is much more complicated than in the level gaps shown so far. The
main gap, approximately 20 km wide, at 2–2.2 km MSL (“channel crest”) is cut
into the main crest line of 3 km MSL. The channel formed by the valley north of
the pass is no longer nearly straight as for the two previous strait examples but
contains tributaries and a bend and its exit drops 200 m into another valley, which
runs nearly perpendicular. Gap flows are possible in both directions, northwards and
southwards, but only northward gap flows have been studied extensively.
Three prototypical gap flows were found (Armi and Mayr 2007). The first one,
shown in Fig. 3.30, has a (nearly) mixed flowing layer on the upstream side, which
accelerates across the gap, before descending and thinning. Upstream, the flow is
isolated by a blocked layer underneath and a cap with strong stability seen in the
180 P.L. Jackson et al.

abrupt increase of potential temperature aloft. Downstream, the well-mixed gap


flow layer has sped up. The speed maximum lies a few hundred meters above
ground. The flow is no longer isolated underneath by a stagnant layer. Aloft,
however, a wedge of (nearly) neutral and stagnant air has formed. Its lower boundary
follows the descending gap flow whereas its upper boundary remains approximately
horizontal. The flow above the gap flow is therefore unaffected by the underlying
gap and channel topography. Since the gap flow layer is well mixed on both sides of
the pass, this case is the “single-layer prototype”.
The second prototype (second row in Fig. 3.30) differs from the first one by
having a stably stratified gap flow layer upstream, typically with a linear increase
of potential temperature with height. Upstream, the gap flow is isolated from the
surface and from the flow aloft, respectively, by neutrally stratified, (nearly) stagnant
layers.1 Downstream of the narrowest part, however, turbulent mixing takes place,
both in an internal hydraulic jump and over the rough surface, quickly turning
the stable stratification into a neutral one. The downstream part of the second
prototype is thus similar to the first one. Also, a wedge of (nearly) stagnant and
neutrally stratified air forms on top of the accelerating and descending gap flow layer
downstream of the pass. Since a thin isolating layer is already present upstream of
the pass, the thickness of the wedge is larger than for the single-layer prototype.
Another common feature of both prototypes is that the flow aloft is unaffected by
the underlying channel and pass topography and follows the horizontal isentropes.
The upstream part of the second prototype corresponds to the analytical solution
of Wood (1968), the downstream part is a single layer, making this case the
“Wood – single-layer prototype”. Wood’s solution describes a self-similar flow that
accelerates from an upstream reservoir to the control. It is the only known analytical
solution in such a setting for a continuously stratified fluid and is described and
derived in more detail in subsection 3.5.2.
The second prototype is also observed in the laboratory (Fig. 3.28c) when a sill
is added to the lateral constriction to form a pass or col. Upstream of the pass, the
stably stratified gap flow is isolated from the channel bottom and the fluid aloft by
neutrally stratified layers. The flow response downstream of the pass changes: the
gap flow layer accelerates even more, thins and rebounds in an internal hydraulic
jump at the downstream end of the sill. Further downstream, the gap flow layer is
well mixed. The depth of the upstream isolating layer on top of the gap flow layer
increases as the gap flow layer plunges downstream of the pass to form a thick
(nearly) stagnant layer. This layer both isolates the gap flow from the fluid aloft and
the fluid aloft from the underlying topography. The isopycnals remain horizontal,
like the isentropes in the atmospheric case.
The third, “hybrid” prototype (Fig. 3.30, bottom row) is so deep that the flow no
longer only crosses through the gap but also over the crest. A downslope windstorm
(or “deep foehn” as it is called in the Alps) is superimposed on the gap flow. The

1
Below the lower neutral isolating layer, the air next to the surface might be stably stratified (e.g.
nocturnally) and blocked
3 Dynamically-Driven Winds 181

6 6

5 5
altitude (km MSL)

4 4

3 isolating layer 3

2 2

1 1

290 300 310 320 -10 0 10 20 30 40 290 300 310 320

6 6

5 5
altitude (km MSL)

4 4
isolating layer
3 isolating layer 3

2 2
isolating layer

1 1

290 300 310 320 -10 0 10 20 30 40 290 300 310 320

6 6
isolating layer
isolating layer
5 5
altitude (km MSL)

4 4

3 3

2 isolating layer 2

1 1

290 300 310 320 -10 0 10 20 30 40 290 300 310 320


pot. temperature (K) distance pot. temperature (K)
(km)

Fig. 3.30 Prototypes of gap flow across Brenner Pass in the central Alps. Single layer hydraulics:
shallow gap flow (first row). Shallow, stably stratified gap flow with Wood’s (1968) self similar
solution upstream turning into a single layer flow downstream (second row). Deep, stably stratified
flow through the gap and across the crest (often called “deep foehn” or “downslope windstorm”)
(third row). Soundings of potential temperature from upstream and downstream are shown to the
left and right, respectively. Arrows in the vertical cross section show their respective locations.
Topography in the sections is shown shaded, bounded by a dashed line at the valley floor.
The solid bold line above the valley floor shows the average altitude across the gap “felt” by
the flow. Isentropes are at 2 K intervals and arrows indicate along-valley speeds upstream and
downstream of the gap (Adapted from Armi and Mayr 2007. Reproduced by permission of the
Royal Meteorological Society)
182 P.L. Jackson et al.

flowing layer on the upstream side is stably stratified as in the “Wood – single-layer
prototype” and is isolated from the surface and from the flow aloft by a neutrally
stratified (nearly) stagnant layer. The upper isolating layer is no longer at or below
crest but clearly above it. The gap-crest flow layer plunges, accelerates and thins.
However, the internal jump occurs only within the channel so that only the lowest
portion of the deep flow can be mixed into a neutrally stratified part; the largest part
of the flowing layer remains stably stratified downstream of the pass. Again, the
upper isolating layer thickens into a deep wedge keeping the flow aloft from being
affected by the underlying topography.
Before the gap flow becomes established, all three prototypes have lower
potential temperatures in the gap flow layer on the upstream side of the pass
than in the air mass on the downstream side (Mayr and Armi 2008). Figure 3.31
shows stages in the evolution of a particular gap flow case using a sounding on
the far upstream side of the pass and at the end of the channel. Prior to the onset
of northward gap flow (i.e. towards the viewer in the accompanying topography
Fig. 3.29), potential temperatures are higher on the southern side of the Alps than on
the northern side to above the main Alpine crest as seen in the difference between the
southern and northern radiosounding (center of Fig. 3.31). No northward gap flow
occurred.2 Changes on the synoptic scale brought a warmer air mass to the north of
the Alps and a colder one to the south so that potential temperature differences
(center of Fig. 3.31) are colder south of the pass to an altitude just above the
channel crest. Higher up in the channel to about 1 km above the main Alpine
crest, however, potential temperatures are still higher on the southern side, Gap
flow had commenced a few hours prior to the soundings. With the continuation
of the synoptic development, the air mass to the south kept getting colder between
2 and 4 km MSL, i.e. channel to 1 km above the main Alpine crest, whereas the air
mass to the north remained unchanged. Consequently, the gap flow layer deepened
and extended even above the main crest, causing hybrid gap flow with superimposed
cross-crest foehn flow.
It must be emphasized that the driving mechanism behind this gap flow event was
the air mass difference on both sides of the crest. During pure gap flow at times 3 and
4, the layer on top of the descending gap flow layer moved at speeds comparable to
and even higher than the gap flow layer on the upstream side. Yet it did not descend
but remained nearly horizontal since its potential temperatures were not low enough
to descend.3
A colder air mass upstream of the gap was also noted in other observations
of gap flow around the world. On the North American Pacific coast it is mostly
cold wintertime continental air contrasting with warmer air over the Pacific Ocean
(e.g. Cameron 1931; Overland 1984; Lackmann and Overland 1989; Colman and
Dierking 1992; Bond and Macklin 1993; Finnigan et al. 1994; Jackson and Steyn
1994a; Pan and Smith 1999; Colle and Mass 2000; Collier 2002; Sharp and Mass

2
Actually a gap flow in the opposite direction – southward – happened.
3
Spatial details were observed with two aircraft flights.
3 Dynamically-Driven Winds 183

4
km MSL

Alpine
3 crest
5 4 3 2 1

6 5 43 2
1 channel 5
crest 6
2 1 2 43

6
deepest
cut

1
295 300 305 310 -10 -5 0 5 10 295 300 305 310
θ Milano Δθ = θ Milano - θ Innsbruck θ Innsbruck

Fig. 3.31 Changes in the air masses upstream and downstream of the Brenner Pass and their
effects on the flow through the pass. At time 3 the potential temperature was for the first time
lower to the south (center column) and very shallow gap flow occurred. The northward gap flow
became deeper (time 4) and changed into a deep hybrid flow composed of deep foehn (“downslope
windstorm”) and gap flow at time 5. At the end (curve 6), gap flow seized as the potential
temperature difference vanished (From Mayr and Armi 2008. Reproduced by permission of the
Royal Meteorological Society)

2004; Bond et al. 2006; Neiman et al.2006). Similarly, gap flows along the Adriatic
and Black Sea coasts are driven by a colder continental air mass compared to the
warmer air mass over the Mediterranean (e.g. Bajić 1989; Grubišić 1989; Glasnovic
and Jurčec 1990; Gohm and Mayr 2005; Gohm et al. 2008; Grisogono and Belušić
2009; Alpers et al. 2009).
Sometimes cold air over the North American continent surges southwards from
where it was formed and reaches as far south as the Central American Pacific Coast,
causing gap flows in Mexico (e.g. Hurd 1929; Schultz et al. 1997; Steenburgh et al.
1998). Over Japan, where gap flows are called “dashikaze”, temperature differences
might be caused by typhoons (e.g. Arakawa 1969; Ishii et al. 2007; Mashiko 2008).
Colder air advected across the ocean can cause gap flows, e.g. on the Falkland
Islands (Mobbs et al. 2005), through the Strait of Gibraltar (Scorer 1952; Dorman
et al.1995), and through the Rhone Valley (e.g. Pettre 1982; Jiang et al. 2003).
184 P.L. Jackson et al.

Different mechanisms lead to colder air upstream of gap flows in the interior of
a continent (cf. Sect. 3.5.3). Examples of such observations are in Seibert (1990),
Doran and Zhong (2000), and Mayr and Armi (2008).
In many of the gap wind cases cited in the preceding paragraphs, cold continental
air of the gap flow extends over the oceans. In this case, significant air-sea
interactions can result as sensible and latent heat fluxes from the warm water surface
alter the overlying air, and in turn the cold wind affects the oceans. Examples of the
impact of strong, cold low level flow on the ocean, in this case of the bora, which
often emerges as a gap-like flow onto the Adriatic, are discussed by Orlić et al.
(1994, 2006) who describe the complex response of the sea to strong bora winds.
Enger and Grisogono (1998) use two-dimensional simulations to demonstrate that
bora flow is extended over the warm sea due to the effect of the sea surface
temperature on reducing N which allows the flow to remain supercritical for a
greater distance by postponing the hydraulic jump. Gap flow over the Gulf of
Tehuantepec has been found to induce coastal upwelling, as well as a lowering of
sea surface temperature by as much as 8ı C (e.g. Schultz et al. 1997).

3.5.2 Theory

Linearizing the governing equations about a basic state has been successfully used
to (semi)analytically describe orographic flows. However, this approach fails to
describe gap flows as shown by Pan and Smith (1999). In linear theory (e.g. Queney
1948; Klemp and Lilly 1975; Smith 1979, 1980), vertically propagating gravity
waves cause lower pressure and accelerating winds on the lee slopes of a mountain
ridge; their strength is proportional to the height of the mountain. Accordingly, the
strongest pressure anomaly occurs along the lee slopes of the mountains on either
side of the gap and not in the gap itself even when confluence at the gap entrance
is included (Pan and Smith 1999; Zängl 2002c). Linearized gap flows are always
weaker than the flow down the adjacent crests, contrary to what is often observed.
Hydraulic theory, on the other hand, can describe gap flows. Its application
to mountain flows was pioneered by Prandtl (1942) and Long (1953, 1954). It
is fully nonlinear, which is an essential characteristic of asymmetric gap flows.
Steady state is assumed in hydraulic theory. Further assumptions are made about the
vertical structure of the air/fluid either by vertically averaging layers or by requiring
self-similarity. Self-similarity means that, at any position, the density and velocity
profiles can be scaled with the height of the moving stratified layer.
Hydraulics in the atmospheric science community is often taken synonymously
with its simplest, “single layer” case of one layer of constant vertical profile of
potential temperature (or density in case of fluids). The theory of the single layer
case is described in Sect. 3.2.3.
Gap flows frequently have stably and continuously stratified rather than ho-
mogeneous, neutrally stratified air masses on the upstream side, as was shown in
3 Dynamically-Driven Winds 185

Fig. 3.32 Wood’s (1968) self-similar solution for stratified flow from a reservoir through a
contraction, i.e. a level gap. Upper panel: vertical section with isopycnals and the vertical density
profiles upstream and downstream of the contraction; lower panel: plan view. Flow from left to
right

the previous section. The single layer theory must therefore be expanded. Wood
(1968) could show that an analytical hydraulic solution exists for withdrawal from
a reservoir of fluid with a known, arbitrary stratification through a level confluent-
diffluent gap as shown in Fig. 3.32. Upon withdrawal, the uppermost and lowermost
streamlines will respectively drop down and rise up as indicated in the figure. If
the volume in these wedge-shaped regions is relatively small then the fluid filling
them will originate from a relatively small range of depth in the reservoir and
be thus of approximately constant density. This implies that the uppermost and
lowermost streamlines are (approximately) also lines of constant pressure. The
analytical solution is for the general case of arbitrary density distribution. Let 0
be the density at the uppermost streamline. The density increases by a total 
to the lowermost streamline over the depth  of the flowing layer. At any height
z, the density is thus 0 C ı. An analytical solution exists when the density and
velocity distributions are self-similar, i.e. when they scale with the depth of the
flowing layer
 at each section. Then the density profile at all sections has the form
ı
 F  D F .ƒ/ where F() is determined by the far-upstream density
z
186 P.L. Jackson et al.

distribution. Exploiting the fact of constant pressure along the streamline and using
Bernoulli’s equation for a streamline in the flow, yields the velocity, v, at an arbitrary
height
"Z Z #
.0 C ı/ v2 1 1
D2 F .ƒ/ dƒ C F .ƒ/  F .ƒ/ F .ƒ/ dƒ 2G .ƒ/
g Œzt .1/  zt   0
(3.39)

where zt is the height of the uppermost streamline. For example, with a linear
stratification, F and G become F(ƒ) D 1  ƒ and 2G ./ D 3ƒ  4ƒ2 C ƒ3 ,
respectively, and the velocity distribution is parabolic-like. The analytical solution
could be reproduced in the laboratory (Armi and Williams 1993, their Fig. 5) and
is shown here in Fig. 3.28b. Note the continuous stratification far upstream in the
reservoir. As it is withdrawn through the gap, well-mixed neutrally stratified and
stagnant layers form above and below the flowing layer, which isolate it from the
remaining fluid. Whether both isolating layers form depends on the height range
over which fluid is withdrawn downstream of the gap. When it is withdrawn at
the bottom, an isolating layer forms only at the top. The atmospheric equivalent is
shown in observations and numerical simulations of a gap flow event through Juan
de Fuca Strait in Figs. 3.27b and c. Similarly, when the fluid is withdrawn from the
top of the channel, an isolating layer forms underneath only. And when the fluid
is withdrawn from the center, both upper and lower isolating layers form as shown
in Fig. 3.28b.
For a particular flow rate, a range of densities 0 to 0 C  flows from
the reservoir. As the flow rate changes, the range of densities drawn from the
reservoir changes so that the flow at the narrowest section of the gap remains critical
with respect to the longest wave mode. Note that a continuously varying density
stratification results in an infinite number of wave modes. Thus there is an infinite
number of sections which act as virtual controls in addition to the physical control of
the narrowest cross section. Benjamin (1981), in a generalization of Wood (1968),
showed that self-similar flows will tend to be realized whenever fluid is withdrawn
from a continuously stratified reservoir. With a sufficiently high extraction rate
downstream, the resulting flow is critical at the narrowest section and supercritical
downstream of it. Even the longest (and fastest) wave mode is swept downstream in
supercritical regions of the flow.
The equivalent to a reservoir of continuously stratified fluid in the atmosphere
is a different air mass on the upstream side of the mountain range. The withdrawal
of that air through a gap is started when a relatively warmer air mass is on the
downstream side. The pressure gradient force between the different air masses
substitutes for the pump, which withdraws fluid in the laboratory experiments. At
what vertical range that pressure gradient and air mass difference exist, determines
where the isolating neutral layers will form. In the case of the Alps (Wipp Valley), it
was mid-level since both upper and lower isolating layers were observed (Fig. 3.30
middle and bottom).
3 Dynamically-Driven Winds 187

Upstream b0 F 22 = 2 Upstream b0 F12 = 2


2 1

layer 1
or
layer 2 or

Lower layer dominant Upper layer dominant


downstream downstream

Upstream bv Crest F 22 = 2 Upstream bv Crest F12 = 2

Lower layer dominant Upper layer dominant


downstream downstream

Fig. 3.33 Two-layer hydraulic flow across a pass (lateral and vertical contraction). Upper panel:
low to moderate flow rates with flow controlled at the narrowest part of the gap. Lower panel: high
flow rates with supercritical flow at the narrowest section. When the lower layer dominates (left
column) it remains supercritical downstream of the gap center and vice versa for an dominating
upper layer (right column). In the figures Fi is the Froude number in layer i, b0 is the location of the
primary control for the flow at the top of the hill, bv is the virtual control where the Froude number
becomes unity (Adapted from Armi and Riemenschneider 2008. Reproduced by permission of
Cambridge University Press)

Atmospheric gap flows can consist of more than one dynamically active layer.
For two layers of comparable depths and flow rates, two unique controlled asym-
metric flows are often possible (e.g. Armi and Riemenschneider 2008). They are
determined by the difference in the fluid reservoirs (or air masses in case of the
atmosphere) upstream and downstream of the gap. For relatively low flow rates,
either the lower or upper layer can be active layers. As flow rates increase, there
may no longer be the possibility of the two asymmetric solutions satisfying critical
conditions at the topographic control, the narrowest section of the (level) gap.
Instead, they merge upstream and the control moves upstream of the narrowest
section as a “virtual” control. An additional possibility of two-layer flows are
“exchange flows” with flow in opposite direction in each layer (cf. Baines 1995,
Sect. 3.11).
As was seen in Sect. 3.5.1 on observations of gap flows, the contraction in
a gap can be only lateral (“level gap”), only vertical (constant width but bottom
rises; “sill”), or a combination of lateral and vertical (“pass”). For two layer flow,
a level gap directly contacts both layers. A sill, on the other hand, only directly
contacts the bottom layer, and the response of the upper layer is indirect. Armi
and Riemenschneider (2008) found analytical solutions for two-layer flows over
a pass. Four different flow configurations are summarized in Fig. 3.33. When the
lower layer dominates (upper left), it accelerates as the interface between the two
188 P.L. Jackson et al.

layers first rises slightly but then descends even upstream of the crest. Compared to
a single-layer flow, the presence of the upper layer keeps the interface higher at the
center of the gap and also increases the flow rate, but both only by less than 5%. Such
a small change is currently not detectable observationally in the atmosphere. The
flow in the upper layer can even be in the opposite direction. When the upper layer
dominates (upper right part of Fig. 3.33), the vertical contraction part of the pass is
only communicated indirectly via the dynamics of the lower layer. The upper layer
depth decreases as it accelerates towards the pass and becomes supercritical. As
flow rates increase, only supercritical conditions can exist at the gap center and the
transition from the subcritical conditions in the reservoir occurs further upstream at
the virtual control (lower panel in Fig. 3.33). On its way towards the pass, the upper
layer accelerates more than the lower layer (Armi and Riemenschneider 2008).
Whether upper or lower layer dominate and remain supercritical downstream of
the gap center, depends on the matching downstream conditions.

3.5.3 Models and Forecasting

Modeling of gap flows has spanned the whole spectrum from idealized to realistic
4D-simulations. Shallow water models are used with surprising success along with
non-hydrostatic numerical weather prediction models. Forecasting gap flows has
not attracted much research interest yet but well-established practices exist in
forecasting offices which have gap flows in their forecast region. An ingredients-
based method for forecasting gap flows is presented.

3.5.3.1 Models

The most basic models to simulate gap flows are shallow water models, which
numerically solve the single-layer hydraulic flow equations for one or two well-
mixed layers. Pan and Smith (1999) used such a model developed by Schär and
Smith (1993) to study gap flows at Unimak Island southwest of the Alaskan
Peninsula but had limited data with which to initialize and compare their model.
Jackson and Steyn (1994b) used a different hydraulic model, which also included
friction. Flow through Howe Sound could be reproduced. Horizontal pressure
gradients and turbulent friction were dominant in a force-balance analysis.
The same model as in Pan and Smith (1999) was used for more intensively
observed gap flows in the Wipp Valley with data from the Mesoscale Alpine
Programme (Gohm and Mayr 2004) and for a case study of bora flow at the Adriatic
coast (Gohm et al. 2008). Systematic parameter studies with the single-layer shallow
water model for the Brenner Pass – Wipp Valley by Gohm and Mayr (2004)
emphasized the dominant role of small-scale topographical features for details of
the gap flow. Ridges protruding into the valley downstream of the pass, a widening
or a turning of the valley can all be locations where supercritical flow returns
3 Dynamically-Driven Winds 189

to a subcritical state in a dissipative, turbulent hydraulic jump. For the Brenner


Pass geometry, the vertical constriction rather than the lateral constriction was
found to be largely controlling the flow. A comparison with detailed observations
showed that the shallow water simulations included all basic features of the
gap flows.
A similarly successful replication was achieved for a bora flow at the Adriatic
coast (Gohm et al. 2008), where the gap flow jets and hydraulic jumps were
captured. At first, it is surprising that a single layer model can reproduce the essence
of gap flows in realistic settings and continuous upstream stratification – a deviation
from the neutral stratification in the numerical model. As Wood (1968) showed in
his analytical solution discussed in Sect. 3.5.2 and as Armi and Williams (1993)
showed in the laboratory, well-mixed nearly stagnant layers isolate the gap flow
from the flow aloft so that disregarding the effects of the flow aloft as is done in the
shallow water model has only small adverse affects on the simulations. Furthermore,
Durran and Klemp (1987, cf. their Fig. 8) and Farmer and Armi (1999, their Fig. 3)
showed that a continuously stratified flow over a ridge (vertical constriction), despite
its continuous stratification is qualitatively similar to a single-layer hydraulic flow,
especially when the depth of the continuously stratified system is less than one half
of its vertical wavelength (Durran and Klemp 1987). The third reason for the success
of the simple model is that turbulence near the surface, in hydraulic jump(s) and at
the upper interface of the gap flow mixes shallower gap flows from a continuous to
a neutral stratification.
Numerical simulations solving the complete set of nonlinear governing equations
have been used in varying degrees of idealization for the simulations of gap
flows. Idealization of both topography and flow/air masses will be considered first,
followed by idealization of only flow/air mass, and finally fully realistic flows.
These fully non-linear simulations with mesocale numerical models confirmed
turbulent friction identified in the hydraulic modeling of Jackson and Steyn (1994b)
as an additional key parameter for gap flows, besides the shape of the underlying
topography. Gap flow depths are of comparable magnitude to the depth of the
planetary boundary layer. Highly idealized simulations with a geostrophically
balanced frictionless background with uniform static stability leave the gap flow
direction dependent on the shape of the topography, which is not supported by
observations. For the balanced background flow parallel to an infinite ridge Sprenger
and Schär (2001) and Zängl (2002a) found flow towards lower pressure. An
isolated ridge, however, reverses the direction (Zängl 2002a). It splits the flow
on its upwind edge, leading to anticyclonic flow on the left side (looking in flow
direction) of the isolated ridge. This flow has an ageostrophic component directed
towards the mountain, thus counteracting and slightly overpowering the larger-scale
geostrophic pressure gradient. The addition of friction turns the low-level wind
counterclockwise compared to the flow aloft, i.e. towards the region of lower large-
scale pressure (Zängl 2002a; Gaberšek and Durran 2006). Gaberšek and Durran
(2006) quantified the importance of friction with volume-averaged momentum-flux
budgets.
190 P.L. Jackson et al.

Turbulent friction in the boundary layer produces filaments of anomalous


potential vorticity, called potential vorticity banners, at the side walls of the gap
(Ross and Vosper 2003), which are then advected further downstream. They can
be found at large distances away from the gap and can also occur in radiatively
driven downvalley flows with a lateral constriction at the terminus of the valley
(Zängl 2004).
Including the realistic topography of the Alps but keeping the geostrophically
(except in the boundary layer) balanced initial flow fields with constant static
stability throughout the troposphere, Zängl (2003) found 3D-propagation of gravity
waves excited over adjacent ridges to be crucial for the acceleration of a gap flow in
the Wipp Valley. The initial conditions, however, strongly deviate from the typical
situation of two different air masses of limited vertical extent on either side of the
gap. The limited depth leads to the self-isolation of the gap flow from the flow aloft
(cf. Sect. 3.5.2).
Including realistic initial conditions is one of the main difficulties for the
numerical modeling of realistic gap flows. Detailed observations, especially vertical
profiles in the proximity of gaps are typically sparse or none-existent. Even if
they exist, making full use of them is difficult since more sophisticated methods
than nudging for assimilating such data into mesoscale models (e.g. 4D-VAR,
ensemble Kalman filters) are just emerging. As a result, the initial state of the
mesoscale models is still dominated by the fields from global models, into which
the mesoscale models are nested. The horizontal resolution of global forecasting
models is currently (2010) at 16 km or coarser, so that model topography differs
significantly from the real topography (e.g. the wider part of the valley south
of Brenner Pass is about 0.7 km below the ECMWF T1279 topography). Initial
conditions of the models therefore lack the sharp potential temperature steps on top
of the gap flow layer, which occur often within only dozens or hundreds of meters.
The smoothing introduced by the numerics during model integration aggravates the
situation. In contrast, the sharp separation can easily be introduced to the much
simpler hydraulic models.
Gohm et al. (2004) compared model simulations of a gap flow and foehn event
in the Wipp Valley with the rich data set of MAP. They attributed considerable
discrepancies at the beginning of the event to deficiencies in the model profile
upstream of the pass, where the flowing layer (and consequently the potential
temperature step at its top) was too shallow. Given that the height of the gap flow
layer on the upstream side, the mass flux through the gap, and the “withdrawal” on
the downstream side are all related (cf. Sect. 3.5.2), a combination of deviations of
the model from the real atmosphere both upstream and downstream might have been
the cause. Later in the event, too much mass flowed across the pass and the gap flow
layer did not descend as much as observed downstream of the pass. However, most
of the prominent observable gap flow features like hydraulic jumps and cross-valley
asymmetries of the flow are so much dominated by the underlying topography, that
they were still found in the simulations. The simulation of a level gap flow event
through the Strait of Juan de Fuca (cf. Fig. 3.27) also found cross-gap asymmetries
3 Dynamically-Driven Winds 191

caused by flow into the channel over one of its sides. One of the main advantages
of the fully realistic simulations is the ability to obtain a complete (time-dependent)
three-dimensional depiction of gap flows (Jackson and Steyn 1994a).
Topography must be sufficiently resolved in the model for the correct amount
of air to be withdrawn from the upstream side. Horizontal mesh widths for Wipp
Valley simulations were as fine as 0.27 km in the innermost model domain and
0.8 km in the domain that spans both sides of the gap in Gohm et al. (2004) and
Zängl et al. (2004). Reducing the 0.8 km grid spacing to 0.4 km further improved the
results (Zängl 2006). A mesh of at least approximately 1 km was also needed for gap
flow through the Columbia Gorge (Sharp and Mass 2002), with 0.44 km horizontal
spacing also improving the results.
When the flow (and the different air masses) are deep enough to “flood” the
main crest into which a gap or pass is cut, downslope windstorms (or deep foehns
as they are called in the Alps) are superimposed on the gap flows. Gap flows play
then a relatively minor role in the total mass transport but otherwise their main
characteristics remain similar (e.g. Colle and Mass 1998; Gohm et al. 2004; Zängl
et al. 2004).
Numerics and turbulence schemes are further challenges in correctly simulating
gap flows with numerical models. Reinecke and Durran (2009b) in simulating
mountain waves showed that second-order finite differencing schemes overam-
plified a standing mountain wave by about a third compared to fourth-order
schemes within the nonhydrostatic regime with typically used horizontal resolutions
of the mountain. Applied to the American Sierra Nevada, this translated to an
overestimation of downslope speeds at the surface of 20 m s1 and of vertical
velocity maxima by 10 m s1 (as seen in the study of Reinecke and Durran 2009b,
see Fig. 9.18 in Chap. 9).
Most numerical simulations of flows over mountains until recently were per-
formed using second-order or third-order schemes. The large mountain waves and
strong rebounds (often termed “wave breaking”) seen in Fig. 9.18 were common
features of these simulations and also part of a scientific dispute over flow over an
oceanic sill, where observations (Farmer and Armi 1999, 2001; Armi and Farmer
2002) could not be reconciled with numerical simulations (Afanasyev and Peltier
2001a, b). A large part of the discrepancy occurred at the top of the flowing layer,
where small scale entrainment was observed but wave breaking on a much larger
scale was modeled. Small scale entrainment was also found to be the important
mechanism at the top of a stratified flow down a slope in the laboratory (Baines
2001, 2005). Gohm et al. (2008) found a strong sensitivity to the turbulence scheme
of the top of the gap flow and the layer immediately above. An improvement of these
schemes seems to be needed. At the ground, the height of the lowest model surface
above the ground had a surprisingly large impact in sensitivity studies of gap flow
reported in Zängl et al. (2008), even more so than the type of PBL scheme.
Essential for the correct simulation of gap flows and the interactions between
the different air masses involved is that horizontal diffusion of the mass field is
truly computed horizontally, not along the terrain following coordinate surfaces
(Zängl 2002b).
192 P.L. Jackson et al.

3.5.4 Forecasting

Forecasting methods for gap flows have been developed in local forecasting offices
but have not yet progressed into easily accessible peer-reviewed literature. They
are typically empirical and statistically based without establishing a direct physical
connection between predictor and predictand. Some combine statistics and the
fundamental characteristics of gap flows. An example is Drechsel and Mayr (2008),
who use the frequency distribution of potential air mass difference on a model
surface and reduced pressure difference across the barrier to probabilistically
forecast the binary event of gap flow – yes or no.
Following Doswell et al. (1996) an “ingredients-based” method is proposed here,
which uses the physical understanding of gap flows. It is not regionally dependent
(gap flows occur all over the globe) but flexible enough to be easily adaptable
to local forecasting demands and to incorporate new observational or modeling
advances as they become available.
Gap flows occur due to different air masses on either side of a gap. The
resulting pressure gradient is the main driving force. The altitude range of the
air mass difference will determine the depth of the gap flow. Most forecasting
applications will be interested in gap flows reaching the surface downstream of
a gap. Unless the potential temperature of the air mass upstream of the gap is
lower or at most the same as the potential temperature of the air mass down-
stream, an asymmetric gap flow with higher wind speeds downstream will not
happen.
The ingredients to create different air masses on either side of the gap will
be examined progressing from large to small scales. On the large scale is the
quasi-horizontal differential advection of air masses. Synoptic-scale systems may
transport different air masses to opposite sides of the gap. Examples are cold fronts
and subsequent mid-level troughs typically co-located with the coldest air passing
over only one side of the obstacle (Fig. 3.34a). A complementary situation is that of
warm fronts and mid-level ridges being confined to only one side. A combination of
these two is also possible.
Another mechanism on the large scale is air mass formation under anticyclones.
Different surface characteristics on either side of the gap lead (diabatically) to
different air masses (Fig. 3.34b). An example is the wintertime formation of cold
air over the continent, which is separated by mountain ranges from relatively warm
air over the ocean. Well-known regions are the Pacific Northwest and the Adriatic
coast. Ocean-land contrasts also lead to large differences over the Falkland Islands
and create a particularly large potential temperature step on top of the flowing layer
(Mobbs et al. 2005).
Advective and diabatic processes are also at work on the mesoscale to create
sufficiently lower potential temperatures at one side of the gap to cause gap flows.
Lower pressure further downstream of the gap can lead to ageostrophic flow of
air towards the low pressure. Compensating subsidence with adiabatic temperature
increase might be sufficient to raise potential temperatures higher than the ones
3 Dynamically-Driven Winds 193

a b

H
F 500

c d
q

Fig. 3.34 Forecasting gap flows: processes creating lower potential temperatures in the flowing
layer upstream than near the surface downstream. (a) Large-scale differential advection from a
cold front and upper level trough followed by a warm front. (b) Formation of air masses under
a high pressure system (cold wintertime continent, warmer ocean). (c) Ageostrophic drainage of
cold air towards low pressure caused by flow over the larger mountain range into which the gap
is embedded. (d) Differential heating caused by upstream cloud-cover and (nearly) clear skies
downstream

upstream of the gap. Causes for lower pressure further away from the gap can be an
approaching cyclone or the topographic low pressure caused by flow over the whole
mountain range, into which the gap is embedded (Fig. 3.34c).
Low and/or mid-level cloud cover on one side of the gap with nearly clear skies
on the other side will lead to much stronger diurnal warming near the surface on
the clear side of the gap, raising potential temperatures above the ones on the
cloudy side (Fig. 3.34d). Consequently, many gap flow regions have a climatological
maximum in the afternoon. At nighttime, however, the radiative cooling can lower
potential temperatures sufficiently to prevent the gap flow from reaching the ground.
Differential heating between a valley atmosphere and the adjacent (flat) forelands
will create valley wind systems (see Chap. 2). Given sufficient heating difference
and a pronounced lateral constriction, e.g. at the exit of the valley, an asymmetric
gap flow with strong acceleration downstream of the constriction can develop. It is
most pronounced for nocturnal downvalley flow and has been observed and modeled
for several Alpine valleys draining into the forelands (Müller et al.1984; Zängl 2004,
and references therein, and Chap. 2 in this volume).
Finally, at the meso-gamma scale, small-scale topographical features might
exclude gap flow from small regions. A kink in the terrain will lead to flow
separation when the potential temperature is similar to the one further underneath
(i.e. nearly neutrally stratified). Lateral flow separation will keep some regions in
the wake of the flow.
194 P.L. Jackson et al.

Shear-induced turbulence from gap flow shooting over a cold pool downstream
of the gap acts to increase the potential temperature step across the interface and
the thickness of the interface by mixing in air of higher potential temperature
from aloft and lower potential temperature from the cold pool, respectively. The
thickness of the induced turbulent eddies depends on the thickness of the shear
layer and the stratification of the flow, as expressed by a (gradient) Richardson
number. These eddies pair with neighboring eddies to form larger vortices (Koop
and Browand 1979), which increase the vertical extent of the mixing region at
the interface several-fold (Koop and Browand 1979, Fig. 8). When the cold pool
is shallow enough to be comparable to the size of the shear-induced eddies, these
eddies and thus the gap flow can penetrate down to the surface and cause a sudden
temperature increase there. The details are sensitive to the value of the Richardson
number. Currently no routine observational methods exist to be able to exploit this
“turbulence ingredient” for forecasting the break-in of gap flow through a cold pool.
Observations and NWP forecasts still need to be improved to make the best
use of such a conceptual, ingredients-based forecasting method. Vertical profiles
of air masses on either side of the gap are observed in very few locations of the
world only. The increase in the size of automatic weather station networks has
put more stations into exposed locations like mountain tops, which provide coarse
information about the vertical structure of the air mass. Vertical structures in NWP
analysis and forecast products are strongly smoothed versions of the actual profile.
However, sharp potential temperature steps over a few decameters only and well-
mixed layers of a few hundred meters and their exact location are crucial for gap
flow dynamics.
An upper limit to the increase of speed from a station in the bottom part of the
flowing layer upstream of the gap, vu , and a downstream station, vd can be computed
from the potential temperature difference across the top of the flowing layer, ™,
and how far the top of the layer has descended between upstream and downstream
locations (e.g. Armi and Mayr 2007), h, which yields a pressure difference p:


p D g h (3.40)
0

Alternatively the pressure difference can be taken from weather stations (as long
as they are in the flowing layers, see below) but have to be reduced to a common
altitude (cf. Mayr et al. 2002, for a discussion of reduction methods). With the
Bernoulli equation along a streamsurface:

p
v2d D v2u  2 (3.41)


This is an upper limit since it does not include the effects of surface friction, hy-
draulic jumps and the fact that the stations might not lie on the same streamsurface.
In many locations, the only diagnostic tool available is pressure observations
from valley/plain stations. Since pressure is an integral measure of the temperature
3 Dynamically-Driven Winds 195

profile aloft, the pressure difference across the gap can suffice for diagnosing (and
forecasting) air mass differences and gap flows. Caution must be taken to account
for blocked layers underneath gap flows. Especially upstream of passes they can be
of substantial depth. As they do not flow across the pass, their contribution to the
pressure difference needs to be discounted. A decrease of temperature by 3 K over
a depth of 1 km increases pressure by 1 hPa.

3.6 Barrier Jets

A barrier jet is an elevated wind maxima on the windward side of a mountain


barrier, blowing parallel to the barrier. Barrier jets can occur when stably stratified
flow approaches an extra-tropical mountain barrier and is blocked by the barrier
for hours to days or longer. The blocked flow is forced to move up the barrier, and
as the low level cool air is forced against buoyancy to rise, it cools adiabatically
and creates higher pressure along the slope. The higher pressure against the barrier
acts to decelerate and block the flow. With enough time, geostrophic adjustment
can occur between the flow and the high pressure along the slope. This adjustment
causes the flow to turn left (right) in the northern (southern) hemisphere, so that the
high pressure is to the right (left) of the flow. Turbulent friction reduces the wind
near the surface, and thermal wind considerations imply the strongest jet will lie
above the largest horizontal temperature gradient. The general structure of a barrier
jet is shown in Fig. 3.35.

Fig. 3.35 Schematic diagram showing a barrier jet as a low-level flow of strong winds (under
the dashed line) parallel to the mountain barrier (flow is into the page). The flow is formed
by the upslope movement of stratified air, increasing the pressure along the slope so that the
pressure gradient force is balanced by the Coriolis force in the cross-barrier direction. The flow
is ageostrophic in the along-barrier direction, where an antitriptic balance forms with the along-
barrier pressure gradient balanced by friction and inertia. H is the mountain height, L is the
mountain half-width, and Lr is the internal Rossby radius of deformation (Lr D NH/f)
196 P.L. Jackson et al.

3.6.1 Observations

The across-barrier blocking, turning and acceleration of the wind has been doc-
umented in many coastal and inland mountain regions such as: north of the
Brooks range in Alaska (Schwerdtfeger 1974) and southeast Alaska (Lackmann and
Overland 1989; Overland and Bond 1993; Loescher et al. 2006; Olson et al. 2007),
southwest British Columbia (Overland and Bond 1995; Doyle and Bond 2001),
Washington State (Mass and Ferber 1990), Oregon (Braun et al. 1997), California
(Doyle 1997; Yu and Smull 2000), east of the Appalachian and Rocky Mountains
(Bell and Bosart 1988; Colle and Mass 1995), Taiwan (Li and Chen 1998; Yeh
and Chen 2003), the Colorado Front Range (Marwitz and Toth 1993), windward
of the Sierra Nevada (Parish 1982), southern Norway (Økland 1990), Antarctica
(Schwerdtfeger 1975), and as a related phenomenon, the Southerly Buster in
southern Australia (Baines 1980). A recent field experiment – the Southeast Alaska
Regional Jets (SARJET) experiment (Winstead et al. 2006) – completed over
the Gulf of Alaska in September and October of 2004 has provided new data
(aircraft and Synthetic Aperture Radar) and modeling results that have contributed
significantly to better understanding Barrier Jets.
Barrier Jets can be categorized into a few broad groups:
Classic barrier jets. In this case the onshore flow of stratified marine air onto a
western coast is blocked by coastal mountains, resulting in wind turning into an
along-barrier direction as a barrier jet near and upwind of the coastal mountains.
These conditions: stratified air impinging on a flow-perpendicular mountain barrier,
are common along mountainous coastlines on the western edges of continents –
especially of North and South America. During winter, west coasts at mid-latitudes
can be characterized by the regular passage of frontal systems across the coastal
mountains often resulting in the onshore flow of stable air ahead of a surface
warm front. The onshore (across-barrier) flow of cool marine air below the frontal
inversion can be blocked by the coastal mountains, decelerate in the across-barrier
direction and deepen, thereby creating a high pressure ridge along the mountain
barrier. In the northern hemisphere, the onshore flow turns to the left along the
barrier, both in geostrophic adjustment to this pressure ridge, and in response to
decreased along-barrier Coriolis force as the onshore wind speed drops (Smith
1979). The enhanced across-barrier pressure gradient can result in a barrier jet in
the coastal zone (e.g. Schwerdtfeger 1979; Parish 1982, 1983; Li and Chen 1998;
Overland and Bond 1993, 1995).
In a numerical simulation of a coastal jet that formed ahead of a land-falling cold
front approaching the northern California coast, Colle et al. (2002) found the block-
ing of the onshore flow resulted in coastal pressure ridging and terrain-enhanced
southerly winds that exceeded 25 m s1 (90 km h1 ). Considerable along-coast
variability in the jet was noted, with reductions in speed downwind of Cape Blanco
and Mendocino and accelerations adjacent to higher coastal topography. Numerical
model diagnostics suggest that the low level blocking of flow impinging on coastal
3 Dynamically-Driven Winds 197

mountain ranges is often important for generating ageostrophic coastal jets in


California and elsewhere (e.g., Doyle 1997). Essentially, this blocking contributes
to pressure ridging upwind of the mountain barrier which then adjusts to form
an along-barrier jet. The winds in the model jet were enhanced by 45% due to
topographic blocking and peak winds of 22 m s1 were simulated (Doyle 1997).
Olson et al. (2007) discuss the structure of various barrier jets observed in southeast
Alaska during the SARJET experiment (Winstead et al. 2006), and found that
classical jets had maximum winds over 30 m s1 at the coast between 600 and
800 m above sea level, extending approximately 60 km offshore. An illustration of
a Classic Barrier Jet, over the waters of Southwest Alaska using Synthetic Aperture
Radar (SAR) is shown in Fig. 3.36a (Loescher et al. 2006). Aircraft observations
and MM5 simulations from SARJET of a classic barrier jet from Olson et al. (2007)
are shown in Fig. 3.37a and b
Hybrid barrier jets. In regions of complex terrain, such as the Pacific Northwest
region of North America, interactions between different orographically forced
mesoscale flows often occur. For example, the source of cold air along the windward
slopes of North American west coast mountains can be gap flow exiting coastal
valleys which is turned poleward by the Coriolis force as the flow undergoes
geostrophic adjustment upon exiting the gap. This is the so-called “hybrid” barrier
jet that is reported by Doyle and Bond (2001) and Loescher et al. (2006) who
discuss the similar cases of high-latitude (strong f) gap flows in British Columbia
and southeast Alaska where gap flows exiting onto a mountainous coastline develop
characteristics of terrain-parallel barrier jets. This is illustrated over southeast
Alaskan waters using SAR imagery in Fig. 3.36c (Loescher et al. 2006).
The geostrophic adjustment of a low-latitude gap flow was modeled by
Steenburgh et al. (1998). They found that the adjustment curvature was determined
by an inertial circle – in their case due to weak Coriolis forcing, the gap flow
never became a jet, however at higher latitudes since the inertial circle radius is
inversely proportional to the Coriolis parameter, the gap flow can turn rapidly
poleward, merging with the synoptic flow becoming a barrier jet. Overland and
Bond (1995) suggest that cold air supplied by gap flow from the Strait of Juan de
Fuca that separates Vancouver Island from the Olympic Peninsula on the west coast
of North America, enhanced the barrier jet observed during the Coastal Observation
and Simulation with Topography Experiment (COAST). Numerical modeling
experiments by Doyle and Bond (2001) indicate that the cold air supplied by
the gap flow was as important as the topography of Vancouver Island in enhancing
the barrier wind. Olson et al. (2007) in an analysis of barrier jets over southeast
Alaska found that hybrid jets with a maximum speed of approximately 30 m s1 at
500 m above sea level were displaced 30–40 km offshore (in contrast to classical
jets which had maximum winds over the coastline). Aircraft observations and MM5
simulations from SARJET of a hybrid barrier jet from Olson et al. (2007) are shown
in Fig. 3.37c and d. A conceptual model of the hybrid barrier jet over southeastern
Alaskan waters is shown in Fig. 3.38d (Olson et al. 2007).
198 P.L. Jackson et al.

Fig. 3.36 Synthetic Aperture Radar imagery of Barrier Jet surface winds over the coastal waters
of southwest Alaska (Source: Loescher et al. 2006. © American Meteorological Society. Reprinted
with permission)
3 Dynamically-Driven Winds 199

Fig. 3.37 Vertical cross sections across the width of barrier jets from IOP1 – classical barrier jet
(panels a, b) and IOP7 – hybrid barrier jet (panels c, d) of the SARJET experiment in southeastern
Alaska. Panels (a) and (c) are from aircraft observations and panels (b) and (d) are from MM5
simulations (Figures from Olson et al. 2007, Figs. 6c, d and 12c, d. © American Meteorological
Society. Reprinted with permission)

Cold Air Damming. In situations where flow impinges on an eastward-facing


mountain barrier, cold air damming can occur if cold air is present to the north
(in the northern hemisphere). In this situation, the mountain may block the low-
level flow, creating higher pressure along the mountain slope and northerly flow
in which the across barrier pressure gradient is balanced by the Coriolis force and
there is an antitriptic balance between friction, inertia and the pressure gradient
in the along barrier direction so as cold air is advected to the south (northern
hemisphere). The jet is effectively trapped against the topography by the Coriolis
force and high static stability at the top of the cold air. These events have been
documented east of the Appalachian Mountains (e.g. Bell and Bosart 1988), east of
the Rocky Mountains (Colle and Mass 1995), among other locations. Marwitz and
Toth (1993) in a Cold Air Damming event along the Colorado Front Range, found
a northerly barrier jet formed by upslope flow with adiabatic cooling enhanced by
cooling from evaporation and melting of falling precipitation. The jets formed by
cold air damming on the eastern slopes of mountain barriers which advect cold air
200 P.L. Jackson et al.

Fig. 3.38 Conceptual model of a hybrid barrier jet over the coast of southeastern Alaska. Thick
arrows represent the gap flow, medium arrows the 1,000–1,500 m flow and thin arrows the
low-level flow. T represents temperature anomalies (Source: Olson et al. 2007. © American
Meteorological Society. Reprinted with permission)

southward, are somewhat analogous to the hybrid jets along the western slopes of
mountains discussed above. In the case of hybrid jets, the cold air is supplied by gap
flow of cold continental air through mainland inlets and the barrier jet advects the
cold air poleward, while in cold air damming events the cold air originates on the
barrier-jet side of the mountains and the cold air is advected equatorward.
The “Southerly Buster” of southeastern Australia is an along-barrier jet resulting
from the interaction of a midlatitude frontal system with the coastal orography. In
this case, a relatively shallow cold front approaches southeastern Australia from the
ocean and transitions into a barrier jet-type flow as the front is partially blocked
on the inland side of the mountains but accelerates along the coast of New South
Wales as the upper disturbance moves eastward across the region. A rapidly moving
along-coastal barrier jet may form, whose leading edge (the front, which resembles
a topographically trapped gravity current) is often marked by a roll cloud over the
coastal Tasman Sea (Baines 1980). Since this leading edge separates the preceding
warm offshore flow by the cool along-barrier jet, the local term “Southerly Buster”
has arisen. The situation is in some ways analogous to the Cold Air Damming
situation in North America discussed above. However there are also significant
differences since the Australian landmass ends at mid-latitudes, the cold air source
region is the cool ocean south of Australia rather than the continent, the westerly
jet in the region contributes to the Southerly Buster, and the topography of the
region is low compared to that of North America. Studies with mesoscale models
3 Dynamically-Driven Winds 201

(e.g., Howells and Kuo 1988; McInnes and McBride 1993; Reid and Leslie 1999)
have attempted to isolate the various roles of season, time of day, topography,
surface friction and heating in the evolution of the Buster. The Southerly Buster is
similar to the orographic coastal jets observed and modeled over central California
by Doyle (1997).
Loescher et al. (2006) used Synthetic Aperture Radar (SAR) imagery from the
Gulf of Alaska to document the spatial and temporal variability of barrier jets
over the ocean in that region. SAR detects centimeter-scale capillary waves on
the water surface that, with some knowledge of wind direction, can be related to
wind speed. Therefore SAR imagery gives high resolution spatial images of wind
speeds over water. Figure 3.36 documents the different categories of barrier jets that
they found and suggested. Figure 3.36a shows a classic barrier jet – the flow has
an onshore component with wind speeds higher than 20 m s1 near the coastline.
Figure 3.36c depicts a hybrid barrier jet – the flow has an offshore component, and
offshore gap flow can be observed exiting the mainland fjords, which undergoes
rapid geostrophic adjustment into a coast-parallel direction upon exiting the gap.
In Fig. 3.36d, a pure gap flow case, there is no turning of the wind upon exiting
the gaps, therefore no barrier jet forms. They found that about 1/3 of all barrier
jets were observed to have a sharp transition between the barrier jet flow and the
weaker ambient flow, and called these “shock jets” – Fig. 3.36e. This form was
associated with anomalously cold, dry air over the continent (Colle et al. 2006).
Variable jets were observed 23% of the time (Fig. 3.36f) in which the along-barrier
flow was found to vary considerably along the coast – this was associated with more
convective conditions in which it was hypothesized that higher-momentum air from
aloft was transferred locally to the surface via convection. Loescher et al. (2006)
observed more barrier jets in the cool season than in the spring and summer. This
was true of all barrier jets, but especially so for hybrid jets. This is expected since
the gap flow component of hybrid barrier jets occurs mainly during the cold season
when conditions exist for a large offshore-directed low-level pressure gradient:
there is a strong contrast between the cold continental air favoring high surface
pressure and the warmer coastal air favoring lower surface pressure. They also
found a close association between percent occurrence of barrier jets and the coastal
orography, with the most frequent occurrence of barrier jets associated with the
highest topography within 100 km of the coast – this is illustrated in Fig. 3.39.
The median speed was found to be 20 m s1 for both classic and hybrid jets. The
median width, measured from the base of the terrain to the outer edge of the barrier
jet was found to be 50 km for classic jets and 60 km for hybrid jets. They found
that some hybrid jets detached from the coastline over part of their length, often by
about 10 km. Olson et al. (2007) note that barrier jets over southeast Alaska reach a
maximum strength when the onshore flow is oriented approximately 40ı from coast
parallel suggesting the influence of ambient wind direction on barrier jet intensity
and evolution.
At this point, it is useful to point out that not all jets blowing parallel to mountain
barriers are barrier jets. Coastally Trapped Wind Reversals (CTWRs) (also called
Coastally Trapped Disturbances) are a type of coastal jet that has been extensively
202 P.L. Jackson et al.

Fig. 3.39 Maximum terrain height found within 100 km of the Gulf of Alaska coastline compared
with the frequency of barrier jets (From Loescher et al. 2006. © American Meteorological Society.
Reprinted with permission)

studied along the West Coast of North America from California to British Columbia
(e.g. Reason and Steyn 1992; Bond et al. 1996; Rogers et al. 1998; Jackson et al.
1999; Nuss et al. 2000; Reason et al. 2001; and others). While there are some
apparent similarities – both phenomena result in southerly flow within a Rossby
radius of the coastal mountains – the initiation mechanism is quite different. Rather
than being forced by the onshore flow of stratified air, CTWRs are initiated by
offshore (downslope) flow at mountain crest level to the north, and the creation of
lower surface pressure offshore to the north of the disturbance, creating an along
barrier pressure gradient (Mass and Bond 1996). CTWRs also propagate from
south to north along the coast as surface ridges and are primarily a warm-season
phenomenon.

3.6.2 Theory

In its simplest form, at steady-state, a barrier jet is semi-geostrophic: it is


in geostrophic balance with the across-barrier pressure gradient, while in the
along-barrier direction, an antitriptic balance exists, in which the pressure gradient
force is balanced by inertia and friction. However this simple force balance does not
adequately address the normal situation when the atmosphere is not in steady state,
3 Dynamically-Driven Winds 203

nor does it tell us much about the mechanisms or conditions under which barrier jets
form. Scale analysis of the governing equations (e.g. Smith 1979; Pierrehumbert
and Wyman 1985; Overland and Bond 1993, 1995) has characterized the general
dynamical nature of stratified flow impinging on a mountain barrier, and can help
us understand the conditions under which barrier jets form even though these
studies neglect complicating factors such as time variation, mesoscale topographic
variation, friction, and diabatic effects.
Important atmospheric properties that determine the amount of blocking and
the distance away from the barrier that blocking will occur are determined by
several dimensionless numbers: the non-dimensional mountain height (HO ), Rossby
number (Ro), Burger number (B), and the Rossby radius of deformation (Lr). The
non-dimensional mountain height, HO D NH=U , as discussed in Sect. 3.2.2, is
a measure of the importance of vertical displacement by an obstacle and tells us
whether or not the cross-barrier flow has sufficient kinetic energy to overcome the
potential energy needed to cross a mountain barrier. If HO < 1 the flow has sufficient
kinetic energy to go over the mountain, if HO > 1 the flow will tend to be blocked by
the mountain leading to upstream deceleration (Pierrehumbert and Wyman 1985).
The distance of upstream deceleration, and therefore the extent of the barrier jet, is
approximately a Rossby radius of deformation from the mountain crest (Pierrehum-
bert and Wyman 1985). The across-barrier Rossby number, Ro D U/fL is the ratio of
inertial to Coriolis forces and is a measure of the importance of the Earth’s rotation
on the flow. It is 1 for geostrophic flow, and 1 when the effect of the Earth’s
rotation is unimportant. The Burger number, B D NH/fL D RoHO , a dynamically
scaled mountain slope, defines the general hydrodynamic regime for the influence of
the mountain barrier on impinging flow. Finally, the Rossby radius of deformation,
Lr D NH/f D BL is the length scale at which rotation effects are as important as
stratification, and represents the maximum distance upstream of blocking by the
mountains. In the expressions above, U is the across-barrier (onshore) velocity up-
stream, f is the Coriolis parameter, L is the mountain half-width, H is the mountain
height, and N is the Brunt-Väisälä frequency. If the Burger number, B  1, the
mountain is effectively gently sloped and the flow is over the mountain barrier and
quasi-geostrophic. If B 1, the mountain is “hydrodynamically steep” and block-
ing is complete and does not depend on the mountain half width (Overland and Bond
1993, 1995). If 0.1 < B < 1, the flow is semigeostrophic and is modified by the slope
of the mountain, primarily over the mountain (Pierrehumbert and Wyman 1985).
For HO > 1 (i.e. when the flow is shallow compared to the mountain height),
Overland and Bond (1995) suggest that the appropriate offshore distance scale for
deceleration and the barrier jet should be Lr/Ĥ D U/f (instead of Lr), because the
vertical scale in this case is not the mountain height, H, but an inertial or gravity
height U/N found by setting HO D 1. In other words, the flow does not “feel” the
full height of the mountain when HO > 1. Overland and Bond (1995) found that
scaling arguments suggest the barrier jet strength enhancement is on the order of
the incident across-barrier speed for HO > 1 and on the order of NH for HO < 1.
The largest barrier jet response should occur when B > 1 and HO  1, since when
HO > 1 that usually means U is small, and when HO < 1, N is likely to be small
204 P.L. Jackson et al.

(Overland and Bond 1995). Smith (1979) discusses the importance of the width of
an elevated plateau downwind of the mountain barrier, on upstream and downstream
flow. He demonstrates that far upstream, the influence of a broad plateau is to induce
cyclonic turning of the wind due to stretching of a column of air as isentropes aloft
are displaced upwards. This causes the flow impinging on a mountain barrier to shift
to more of a barrier-parallel orientation and enhances the formation of a barrier jet
near the mountain.
The scale analysis discussed above gives general guidance on barrier jet for-
mation, but is probably not applicable to many aspects of jets created by land-
falling storms. These storms are not uniform and in steady state, are subject to
mesoscale topographic variability, diabatic and frictional effects, and hence the
scaling assumptions are violated. However idealized 2D (Braun et al. 1999) and
3D numerical simulations (Olson and Colle 2009) can provide further insight.
Idealized 2D numerical simulations conducted by Braun et al. (1999) explored the
importance of mountain half width, height, as well as the width of an elevated
plateau downstream of the mountain barrier, on barrier jet characteristics. They
found that while the upstream deceleration is determined by the mountain half width
and height, it is the downstream plateau size that determines the amount of upstream
cyclonic turning and hence the strength of the barrier jet. The barrier jet strength
also depends on the duration of onshore flow: reaching a steady value only after the
time required for the flow to cross the mountain barrier. This finding is consistent
with the theoretical work of Smith (1979) and was confirmed in the idealized 3D
simulations with the realistic terrain of the Gulf of Alaska conducted by Olson and
Colle (2009). Therefore it is expected that wide mountain barriers or mountains with
wide plateaus would have the strongest barrier jet formation.

3.6.3 Models and Forecasting

Mesoscale numerical models have shown good skill at capturing the essential
features of barrier jets in a number of both idealized (e.g. Cui et al. 1998; Yeh and
Chen 2003; Barstad and Grønås 2005; Olson and Colle 2009) and realistic numerical
studies (e.g. Olson et al. 2007; Doyle 1997; Colle and Mass 1995). This suggests
that NWP models with sufficient spatial resolution should be able to adequately
forecast these events. With a width in the across-barrier direction of approximately
50 km, it is likely that at least 10 km horizontal resolution would be required for
accurate representation of classical barrier jets and cold air damming barrier jets.
In cases of hybrid barrier jets, the model must have adequate resolution to simulate
both the gap flow that delivers the cold air onto the coast, as well as the barrier jet
itself. Since topographic gaps in mountain barriers are typically much narrower than
the across-barrier jet scale, it seems likely that models with horizontal resolutions
of 1–2 km would be required to adequately simulate hybrid barrier jets when gap
widths are a few km. In southeast Alaska, Olson et al. (2007) successfully simulated
hybrid jets with an inner grid of 4 km horizontal resolution which was sufficient
3 Dynamically-Driven Winds 205

to resolve the Cross Sound gap which is approximately 50 km wide. Routinely


operated NWP models will achieve these resolutions as computing power continues
to increase.
While NWP models will continue to improve their ability to forecast barrier
jets, much has been learned from studies that can aid in both the interpretation of
NWP, as well as the creation of rules of thumb for diagnosis and forecasting of
barrier jet events. The basic ingredients of course are that there must be a flow of
stable air toward a mountain barrier for a sufficiently long period of time before
a barrier jet can form. However this leaves several unanswered questions such as:
How high must the barrier be? Are there critical values of Ĥ or B before barrier
jets can form? What factors will determine the barrier jet width and wind speed
enhancement? How long must the across-barrier flow exist before barrier jets can
form?
A climatology of barrier jets by Loescher et al. (2006) using SAR imagery over
the Gulf of Alaska found that terrain height plays a key role in discriminating
locations where barrier jets may form. They found that barrier jets occurred most
frequently in locations where the maximum terrain heights within 100 km of the
coast exceeded 2 km (Fig. 3.38), and that the maximum topography within 100 km
of the coast was more important than the terrain height at the coast in determining
barrier jet frequency. Hybrid barrier jets in this location were associated with major
gaps in the mountain barrier as the gap flow turned to the right due to geostrophic
adjustment upon exiting the gap. The synoptic conditions conducive to barrier jet
formation in this region were investigated by Colle et al. (2006) who found that
cool season jets (without shock or variable features) were associated with a deeper
than normal upper-level trough approaching the Gulf of Alaska, and an anomalously
high ridge over the northwestern North American continent, resulting in low-level
southerlies and warm advection. They found that shock barrier jets had significant
cold anomalies at low levels over the interior, and that variable barrier jets had
weaker low-level stability that favored mixing of higher momentum air to the
surface in localized areas.
Idealized MM5 simulations using the realistic topography of southeast Alaska
with an inner grid of 6 km resolution conducted by Olson and Colle (2009) provide
insight into both the structure and synoptic characteristics useful for forecasting
both classical and hybrid barrier jets in this location and possibly elsewhere too.
They found that the broad inland plateau is essential to the strength of the barrier
jet and results in an anticyclonic pressure perturbation along the windward slopes of
mountains, especially for mountain perpendicular flow. This acts to turn the winds
into a more barrier-parallel direction as much as 500–1,000 km upstream of the
mountains. The timescale for this to develop was found to be the time required
for the flow to cross the full mountain barrier. They found that barrier jet width
tends to be narrower when the impinging flow is more perpendicular to the barrier.
Gap flow from an inland cold pool could contribute to hybrid jets when the low
level ambient flow was nearly barrier-parallel – this results in a hybrid barrier jet
of greater width, with a jet maximum shifted further offshore. When the ambient
flow was nearly barrier-perpendicular, the gap flows were reduced and only classical
206 P.L. Jackson et al.

barrier jets formed. The greatest wind speed enhancements (nearly two times the
component of ambient flow that is normal to the barrier) were for
• High Ĥ (2.5–3.3) consistent with other studies (Ólafsson and Bougeault 1996;
Braun et al. 1999; Petersen et al. 2003; and also with Colle and Mass 1995 for
cold air damming east of the Rocky Mountains)
• Lower wind speeds, and
• Ambient wind directions that are 30–45ı from barrier-parallel.
Olson and Colle (2009) also found that the barrier jet height was most dependent
on wind speed with strongest winds (25 m s1 ) resulting in the deepest jets (900 m).
Scaling arguments can provide guidance on the across-barrier scale, with the Rossby
radius of deformation, Lr D NH/f defining the maximum barrier jet width, and
this value being reduced by a factor of 1/HO to Lr=HO when the non-dimensional
mountain height is greater than one.

3.7 Summary

This chapter detailed flow phenomena that are dynamically forced when wind
interacts with orography. Flow approaching a mountain barrier, will pass over it,
flow through gaps or valleys that dissect it, or be blocked by the mountain and
diverted horizontally around it. As we have seen, each of these scenarios results
in different dynamically-driven wind phenomena. When air is able to pass over a
mountain barrier under specific conditions, mountain waves are generated that can
result in lee waves, rotors and downslope windstorms. Air flowing through channels
that dissect a mountain barrier can result in gap flow arising from along-channel
acceleration. If the air is blocked by a mountain barrier and diverted around it, then
a barrier jet may form. This chapter described these various phenomena.
Mountain waves are a common and important feature of atmospheric flow in
mountainous regions. Long wavelength mountain waves have a tendency to prop-
agate vertically into the middle and upper atmosphere where the waves eventually
break, causing turbulence and a drag force on the flow. Shorter wavelength waves
tend to propagate horizontally as well as vertically and wave motion appears
downwind of mountain ranges. Wave reflection, caused by height variations in
wind and stability, can enhance the downwind propagation through the formation
of trapped lee waves in some circumstances and downslope windstorms in others.
Mountain waves of sufficiently large amplitude can give rise to severe turbulence
in the troposphere. The term mountain-wave rotor is used to describe horizontally
oriented vortices which form under the wave crests. Rotors can be broadly classified
into two types. The type I rotor is probably the most common and forms beneath
the crests of trapped lee waves. Its formation is linked to boundary layer separation,
which is itself caused by the strong pressure gradient exerted on the near-surface
flow by the lee waves aloft. The rarer type II rotor resembles a hydraulic jump
and is a deep layer of severe turbulence which forms downwind of the mountain,
3 Dynamically-Driven Winds 207

associated with low-level wave breaking and downslope windstorms. Its structure is
not as well defined as the type I rotor and it is less well understood. Both types of
rotor pose a significant risk to aviation safety and present a difficult challenge for
forecasting.
Downslope windstorms are a consequence of large-amplitude mountain waves
that can form downwind of a mountain barrier in which the low level winds are
accelerated on the lee slope, sometimes to two to three times the wind speed at
mountain top level. They resemble the hydraulic flow of fluid over a rock with
shallow supercritical flow corresponding to the strong downslope winds on the
lee slope which terminate in a hydraulic jump. The circumstances resulting in
this flow can arise in a few ways. If the atmosphere has an upstream layered
structure that already resembles the structure of classical hydraulic flow, then the
transition to a lower shooting flow is facilitated. If the atmosphere above the
ridge has reverse shear (cross-barrier winds decreasing with height either through
changing wind speed or direction) or decreasing stability above a lower stable layer
then this may result in a mean-state critical layer which can act to reflect upward
propagating mountain waves and decouple the lower layer, creating the conditions
for shallow supercritical flow on the lee slope. If the atmosphere has constant
stratification and the non-dimensional mountain height is sufficiently large, then
vertically propagating mountain waves may break in the upper troposphere, creating
a stagnant turbulent zone in which the cross barrier wind goes to zero resulting in a
self-induced critical layer that acts like a mean-state critical layer.
Gaps or valleys that dissect a mountain provide a means for an air mass to cross a
barrier. Potential temperatures on the upstream side have to be lower than in the air
mass downstream for asymmetric gap flows to develop where the layer, which flows
through the gap, thins and accelerates with the highest speeds occurring downstream
of the narrowest part of the gap. Gaps can be level or elevated and asymmetric
gap flows are found all over the world. Hydraulic, non-linear theory best describes
gap flows. An analytical solution exists even for continuously stratified flow, where
nearly neutrally stratified and stagnant layers form above and/or below the gap flow
layer to isolate it from the flow aloft and/or the underlying terrain (on the upstream
side). Usually it is the narrowest section of the gap, which forms a “control” for
the flow. Given a particular value of the upstream depth of the gap flow layer, only
one value of the flow rate through the control is possible when flow transits from
subcritical on the upstream side to supercritical on the downstream side, where it
will eventually adjust to the downstream air mass in a hydraulic jump. Shallow
water models are successful in numerically simulating gap flows. Full NWP models
currently still face challenges to (1) properly resolve the underlying topography, to
(2) handle turbulence near the ground and in the shear layer at the top of the gap flow
and in the hydraulic jump, and to (3) contain the details of the air mass differences
across the gap and the processes producing the hydraulic flow. An ingredients-based
forecasting method identifies processes that lead to the formation of air masses with
different characteristics on both sides of the gap.
Barrier jets are elevated wind maxima that blow parallel to a mountain barrier
on its windward side. When flow approaching a mountain barrier is blocked and
208 P.L. Jackson et al.

diverted around the mountain then barrier jets may occur if the flow is blocked
by the barrier for hours to days or longer. The blocked flow is forced to move up
the barrier, and cools adiabatically creating higher pressure along the slope which
acts to decelerate and block the flow. Geostrophic adjustment turns the flow left
(right) in the northern (southern) hemisphere and a low-level barrier-parallel jet is
formed between the surface and the mountain crest. Barrier jets have been observed
in many places – and can occur in the extra-tropics (i.e. where rotational effects are
important) wherever stratified flow approaches a significant mountain barrier. They
have been described as classic barrier jets along the western slopes of mountains,
as hybrids where gap flow of cold air from the east side of mountains contributes
the low-level cold air needed for stratification, as cold-air damming events along the
eastern slopes of mountains, and as resulting from the interaction of fronts as they
approach mountains.
Understanding of dynamically forced flows has progressed considerably thanks
to observational, theoretical, numerical modeling (Chaps. 9 and 10), and com-
putational advances over the past three decades. We have progressed from a
quite limited linear, two-dimensional, “dry”, constant stability/constant wind un-
derstanding of these phenomena, to the consideration of more realistic conditions.
Specifically, an understanding of the importance of effects such as non-linearity,
three-dimensionality, moisture, varying wind and stability profiles, and boundary
layer (Chaps. 2 and 5) effects, are now beginning to be more fully understood.
As knowledge has developed, the consideration of these other factors has become
increasingly important.

Acknowledgements The authors thank the volume editors (Fotini Kapodes Chow, Stephen De
Wekker and Bradley J. Snyder) for keeping us on track with their encouragement and organization,
and for all of their efforts in bringing this book to completion. We also thank the three anonymous
reviewers who provided many helpful comments that improved the chapter. We thank the authors
and organizations who allowed us to use many of the figures in this chapter. The first author would
like to acknowledge partial funding support of his research program from an NSERC Discovery
Grant.

References

Afanasyev, Y. D., and W.R. Peltier, 1998: The three-dimensionalization of stratified flow over two-
dimensional topography. J. Atmos. Sci., 55, 19–39.
Afanasyev Y.D., and W.R. Peltier, 2001a: On breaking internal waves over the sill in Knight Inlet.
Proc. R. Soc. Lond. A, 457, 2799–2825.
Afanasyev Y.D., and W.R. Peltier, 2001b: Reply to comment on the paper ‘On breaking internal
waves over the sill in Knight Inlet’. Proc. Roy. Soc. A, 457, 2831–2834.
Alpers, W., A. Ivanov, and J. Horstmann, 2009: Observations of bora events over the adriatic sea
and black sea by spaceborne synthetic aperture radar. Mon. Wea. Rev.,137, 1150–1161.
Anthes, R.A., Y.H. Kuo, D.P. Baumhefner, R.P. Errico, and T.W. Bettge, 1985: Predictability of
mesoscale atmospheric motions. Advances in Geophysics, 28B, Academic Press, 159–202.
Arakawa, S. 1968: A proposed mechanism of fall winds and dashikaze. Met. Geophys., 19,
69–99.
3 Dynamically-Driven Winds 209

Arakawa, S. 1969: Climatological and dynamical studies on the local strong winds, mainly in
Hokkaidō, Japan. Geophys. Mag., 34, 359–425.
Armi, L. 1986: The hydraulics of two flowing layers with different densities. J. Fluid Mech., 163,
27–58.
Armi, L., and D. Farmer, 2002: Stratified flow over topography: bifurcation fronts and transition to
the uncontrolled state. Proc. R. Soc. London, 458, 513–538.
Armi, L., and G.J. Mayr, 2007: Stratified flow across an alpine crest with a pass: Shallow and deep
flows. Quar. J. Roy. Meteorol. Soc., 133, 459–477.
Armi, L., and U. Riemenschneider, 2008: Two-layer hydraulics for a co-located crest and narrows.
J. Fluid Mech., 615, 169–184.
Armi, L., and R. Williams, 1993: The hydraulics of a stratified fluid flowing through a contraction.
J. Fluid Mech., 251, 355–375.
Bacmeister, J.T., and R. T. Pierrehumbert, 1988: On high-drag states of nonlinear flow over an
obstacle. J. Atmos. Sci., 45, 63–80.
Baines, P.G., 1980: The dynamics of the Southerly Buster. Australian Meteorological Magazine,
28, 175–200.
Baines, P.G., 1995: Topographic effects in stratified flow. Cambridge University Press, 482 pp.
Baines, P.G., 2001: Mixing in flows down gentle slopes into stratified environments. J. Fluid Mech.,
443, 237–270.
Baines, P.G., 2005: Mixing regimes for the flow of dense fluid down slopes into stratified
environments. J. Fluid Mech., 538, 247–267.
Bajić A., 1989: Severe bora on the northern Adriatic Part 1: Statistical analysis. RASPRAVE,
24, 1–9.
Banta, R.M., L.D. Olivier, and J.M. Intrieri, 1990: Doppler lidar observations of the 9 January
1989 severe downslope windstorm in Boulder, Colorado. Preprints, Fifth Conf. on Mountain
Meteorology, Boulder, CO, Amer. Meteor. Soc., 68–89.
Barstad, I., and S. Grønås, 2005: Southwesterly flows over southern Norway – Mesoscale
sensitivity to large-scale wind direction and speed. Tellus, 57A, 136–152.
Bell, G.D., and L.F. Bosart, 1988: Appalachian cold-air damming. Mon. Wea. Rev., 116, 137–161.
Belušić, D., M. Žagar, and B. Grisogono, 2007: Numerical simulation of pulsations in the bora
wind. Quar. J. Roy. Meteorol. Soc., 133, 1371–1388.
Benjamin TB. 1981. Steady flows drawn from a stably stratified reservoir. J. Fluid Mech.,106,
245–260.
Bond, N.A., C.F. Dierking, and J.D. Doyle, 2006: Research aircraft and wind profiler observations
in Gastineau Channel during a Taku wind event. Wea. Forecasting, 21, 489–501.
Bond, N.A. and S.A. Macklin, 1993: Aircraft observations of offshore-directed flow near wide bay,
Alaska. Mon. Wea. Rev.,121, 150–161.
Bond, N.A., C. F. Mass, and J.E. Overland, 1996: Coastally trapped wind reversals along the United
States west coast during the warm season, Part I: Climatology and temporal evolution. Mon.
Wea. Rev.,124, 430–445.
Bougeault, P., P. Binder, A. Buzzi, R. Dirks, R. Houze, J.P. Kuettner, R.B. Smith, R. Steinacker, and
H. Volkert, 2001: The MAP Special Observing Period. Bull. Amer. Meteor. Soc., 82, 433–462.
Bougeault, P., and coauthors, 1993: The atmospheric momentum budget over a major mountain
range: First results of the PYREX field program. Ann. Geophys., 11, 395–418.
Braun, S.A., R.A. Houze Jr., B.F. Smull, 1997: Airborne dual-Doppler observations of an intense
frontal system approaching the Pacific Northwest coast. Mon. Wea. Rev.,125, 3131–3156.
Braun, S.A., R. Rotunno, J.B. Klemp, 1999: Effects of coastal orography on landfalling cold fronts.
Part I: Dry, inviscid dynamics. J. Atmos. Sci., 56, 517–533.
Brinkmann, W.A.R., 1974: Strong downslope winds at Boulder, Colorado. Mon. Wea. Rev., 102,
592–602.
Brown, J.M., 1986: A decision tree for forecasting downslope windstorms in Colorado. Preprints,
11th Conf. on Weather Forecasting and Analysis, Kansas City, MO, Amer. Meteor. Soc., 83–88.
Cameron, D.C., 1931: Easterly gales in the Columbia River Gorge during the winter of 1930–
1931 – Some of their causes and effects. Mon. Wea. Rev., 59, 411–413.
210 P.L. Jackson et al.

Casswell, S.A., 1966: A simplified calculation of maximum vertical velocities in mountain lee
waves. Meteorol. Mag, 95, 68–80.
Clark. T.L., and R.D. Farley, 1984: Severe downslope windstorm calculations in two and three
spatial dimensions using anelastic interactive grid nesting: A possible mechanism for gustiness.
J. Atmos. Sci., 41, 329–350.
Clark, T.L., and W.R. Peltier, 1977: On the evolution and stability of finite-amplitude mountain
waves. J. Atmos. Sci., 34, 1715–1730.
Colle, B.A., and C.F. Mass, 1995: The structure and evolution of cold surges east of the Rocky
Mountains. Mon. Wea. Rev., 123, 2577–2610.
Colle, B.A., and C.F. Mass, 1998: Windstorms along the Western Side of the Washington Cascade
Mountains. Part I: A High-Resolution Observational and Modeling Study of the 12 February
1995 Event. Mon. Wea. Rev., 126, 28–52.
Colle, B.A., and C.F. Mass, 2000: High-resolution observations and numerical simulations of
easterly gap flow through the Strait of Juan de Fuca on 9–10 December 1995. Mon. Wea.
Rev., 128, 2398–2422.
Colle, B.A., B.F. Smull, and M.-J. Yang, 2002: Numerical simulations of a landfalling cold front
observed during COAST: Rapid evolution and responsible mechanisms. Mon. Wea. Rev., 130,
1945–1966.
Colle, B.A., K.A. Loescher, G.S. Young, and N.S. Winstead, 2006: Climatology of barrier jets
along the Alaskan coast. Part II: Large-scale and sounding composites. Mon. Wea. Rev., 134,
454–477.
Collier, C., 2002: Developments in radar and remote-sensing methods for measuring and forecast-
ing rainfall. Phil. Trans. Roy. Soc., series A, 360, 1345–1361
Colman, B.R., and C. F. Dierking, 1992: The Taku Wind of Southeast Alaska: Its Identification
and Prediction. Wea. Forecasting, 7, 49–64.
Colson, D., 1954: Meteorological problems in forecasting mountain waves. Bull. Amer. Meteor.
Soc., 35, 363–371.
COMET, 2008: MetEd: Mountain Meteorology training module. Available online at: http://www.
meted.ucar.edu/mesoprim/mtnwave/, University Corporation for Atmospheric Research.
Cui, Z., M. Tjernström, and B. Grisogono, 1998: Idealized simulations of atmospheric coastal flow
along the central coast of California, J. Appl. Meteor., 37, 1332–1363.
Darby, L.S. and G.S. Poulos, 2006: The evolution of lee-wave rotor activity in the lee of Pike’s
Peak under the influence of a cold frontal passage: Implications for aircraft safety. Mon. Wea.
Rev., 134, 2857–2876.
Dierking, C.F., 1998: Effects of a mountain wave windstorm at the surface. Wea. Forecasting, 13,
606–613.
Doran, J.C., and S. Zhong, 2000: Thermally driven gap winds into the Mexico City basin. J. Appl.
Meteorol., 39, 1330–1340.
Dorman, C.E., R.C. Beardsley, and R. Limeburner, 1995: Winds in the Strait of Gibraltar. Quart.
J. Roy. Meteorol. Soc.,121, 1903–1921.
Doswell, C.A., H.E. Brooks, and R.A. Maddox, 1996: Flash flood forecasting: An ingredients-
based methodology. Wea. Forecasting,11, 560–581.
Doyle, J.D., 1997: The influence of mesoscale orography on a coastal jet and rainband. Mon. Wea.
Rev., 125, 1465–1488.
Doyle, J.D., D.R. Durran, C. Chen, B.A. Colle, M. Georgelin, V. Grubišić, W.R. Hsu, C.Y.
Huang, D. Landau, Y.L. Lin, G.S. Poulos, W.Y. Sun, D.B. Weber, M.G. Wurtele, and M. Xue,
2000: An intercomparison of model-predicted wave breaking for the 11 January 1972 Boulder
windstorm. Mon. Wea. Rev., 128, 901–914.
Doyle, J.D., and M.A. Shapiro, 2000: A multi-scale simulation of an extreme downslope
windstorm over complex topography. Meteor. Atmos. Phys., 74, 83–101.
Doyle, J.D., and N.A. Bond, 2001: Research aircraft observations and numerical simulations of a
warm front approaching Vancouver Island, Mon. Wea. Rev., 129, 978–998.
Doyle, J.D., and Durran, D.R. 2002. The dynamics of mountain-wave induced rotors. J. Atmos.
Sci., 59, 186–201.
3 Dynamically-Driven Winds 211

Doyle, J.D., and R.B. Smith, 2003: Mountain waves over the Hohe Tauern. Quart. J. Roy. Meteor.
Soc.,129, 799–823.
Doyle, J.D., and Durran, D.R. 2007. Rotor and sub-rotor dynamics in the lee of three-dimensional
terrain. J. Atmos. Sci., 64, 4202–4221.
Doyle, J.D., and Q. Jiang, 2006: Observations and numerical simulations of mountain waves in the
presence of directional wind shear. Q. J. R. Meteorol. Soc., 132, 1877–1905.
Doyle, J.D., and C.A. Reynolds, 2008: Implications of regime transitions for mountain-wave-
breaking predictability. Mon. Wea. Rev., 136, 5211–5223.
Drechsel, S., and G.J. Mayr, 2008: Objective forecasting of foehn winds for a subgrid-scale alpine
valley. Wea. Forecasting, 23, 205–218.
Durran, D.R., and Klemp, J.B. 1982. The effects of moisture on trapped lee waves. J. Atmos. Sci.,
39 2490–2506.
Durran, D.R., 1986: Another look at downslope windstorms. Part I: The development of analogs
to supercritical flow in an infinitely deep, continuously stratified fluid. J. Atmos. Sci., 43,
2527–2543.
Durran, D.R., and J.B. Klemp, 1987: Another look at downslope winds. Part II: Nonlinear
amplification beneath wave-overturning layers. J. Atmos. Sci., 44, 3402–3412.
Durran, D.R., 1990: Mountain waves and downslope winds. Atmospheric Processes over Complex
Terrain, Meteor. Monogr., No. 45, Amer. Meteor. Soc., 59–81.
Durran, D.R., 1992: Two-layer solutions to Long’s equation for vertically propagating mountain
waves: How good is linear theory? Quart. J. Roy. Meteor. Soc., 118, 415–433.
Durran, D.R., 1999: Numerical methods for wave equations in geophysical fluid dynamics.
Springer, 465 pp.
Enger, L. and B. Grisogono, 1998: The response of bora-type flow to sea surface temperature.
Quart. J. Roy. Meteorol. Soc., 124, 1227–1244.
Epifanio, C.C., and T. Qian, 2008: Wave–turbulence interactions in a breaking mountain wave.
J. Atmos. Sci., 65, 3139–3158.
Etling, D. 1989: On atmospheric vortex sheets in the wake of large islands. Meteor. Atmos. Phys.,
41, 157–164.
Farmer, D.M., and L. Armi, 1999: Stratified flow over topography: the role of small-scale
entrainment and mixing in flow establishment. Proc. Roy. Soc. London A, 455, 3221–3258.
Farmer, D.M., and L. Armi, 2001: Stratified flow over topography: models versus observations.
Proc. Roy. Soc. A, 457, 2827–2830.
Finnigan, T.D., J.A.Vine, P.L. Jackson, S.E. Allen, G.A. Lawrence, and D.G. Steyn, 1994:
Hydraulic physical modeling and observations of a severe gap wind. Mon. Wea. Rev., 122,
2677–2687.
Gaberšek, S., and D.R. Durran, 2006: Gap Flows through Idealized Topography. Part II: Effects of
rotation and surface friction. J. Atmos. Sci., 63, 2720–2739.
Georgelin, M., E. Richard, M. Petitdidier, A. Druilhet, 1994: Impact of subgrid-scale orography
parameterization on the simulation of orographic flows. Mon. Wea. Rev.,122, 1509–1522.
Glasnovic, D., and V. Jurčec, 1990: Determination of upstream bora layer depth. Meteor. Atmos.
Phys., 43, 137–144.
Gohm, A., and G.J. Mayr, 2004: Hydraulic aspects of föhn winds in an Alpine valley. Quart. J. Roy.
Meteorol. Soc., 130, 449–480.
Gohm, A., and G.J. Mayr, 2005: Numerical and observational case study of a deep Adriatic bora.
Quart. J. Roy. Meteorol. Soc., 131, 1363–1392.
Gohm, A., G.J. Mayr, A. Fix, and A. Giez, 2008: On the onset of bora and the formation of rotors
and jumps near a mountain gap. Quart. J. Roy. Meteorol. Soc., 134, 21–46.
Gohm, A., G. Zängl, and G.J. Mayr, 2004: South foehn in the Wipp Valley on 24 October 1999
(MAP IOP 10): Verification of high-resolution numerical simulations with observations. Mon.
Weather Rev., 132, 78–102.
Grisogono, B., and D. Belušić, 2009: A review of recent advances in understanding the meso- and
microscale properties of the severe bora wind. Tellus A, 61, 1–16.
212 P.L. Jackson et al.

Grisogono, B. and L. Enger, 2004: Boundary-layer variations due to orographic-wave breaking in


the presence of rotation. Quart. J. Roy. Meteorol. Soc., 130, 2991–3014.
Grubišić, V., 1989: Application of the hydraulic theory in cases of bora with strong upstream flow.
RASPRAVE, 24, 21–27.
Grubišić, V. and B.J. Billings, 2007: The intense lee-wave rotor event of Sierra Rotors IOP8.
J. Atmos. Sci., 64, 4178–4201.
Grubišić, V., J.D. Doyle, J. Kuettner, S. Mobbs, R.B. Smith, C.D. Whiteman, R. Dirks, S. Czyzyk,
S.A. Chon, S. Vosper, M. Weissmann, S. Haimov, S.F.J. De Wekker, L.L. Pan, and F.K.
Chow, 2008: The Terrain-induced Rotor Experiment. A field campaign overview including
observational highlights. Bull. Amer. Meteor. Soc., 89, 1513–1533.
Grubišić, V. and Lewis, J.M. 2004. Sierra Wave Project revisited: 50 years later. Bull. Amer. Meteor.
Soc., 85, 1127–1142.
Grubišić, V. and M. Orlic 2007: Early observations of rotor clouds by Andrija Mohorovičić. Bull.
Amer. Meteor. Soc., 88, 683–700.
Grubišić, V., R.B. Smith, and C. Schär, 1995: The effect of bottom friction on shallow-water flow
past an isolated obstacle. J. Atmos. Sci.,52, 1985–2005.
Grubišić, V. and I. Stiperski, 2009: Lee-wave resonances over double bell-shaped obstacles. J.
Atmos. Sci., 66, 1205-1228.
Hertenstein, R.F. and J.P. Kuettner, 2005. Rotor types associated with steep lee topography:
influence of the wind profile. Tellus A, 57, 117–135.
Hoinka, K.P., 1985: A comparison of numerical simulations of hydrostatic flow over mountains
and observations. Mon. Wea. Rev., 113, 719–735.
Howells, P.A.C, and Y.H. Kuo, 1988: A numerical study of the mesoscale environment of a
southerly buster event. Mon. Wea. Rev., 116, 1771–1788.
Holmboe, J. and H. Klieforth, 1957: Investigation of mountain lee waves and the air flow over
the Sierra Nevada. Final Report. Contract No. AF19(604)-728, University of California ADNo.
133606, Dept. of Meteorology, University of California, Los Angeles, 290 pp.
Houghton, D.C., and E. Isaacson, 1968: Mountain winds. Stud. Numer. Anal., 2, 21–52.
Houghton, D.D., and A. Kasahara, 1968: Nonlinear shallow fluid flow over an isolated ridge.
Commun. Pure and Appl. Math., 21, 1–23.
Huppert, H.E., and J.W. Miles, 1969: Lee waves in a stratified flow Part 3: Semi-elliptical obstacles.
J. Fluid Mech., 35, 481–496.
Hunt, J.C.R., H. Olafsson and P. Bougeault, 2001: Coriolis effects on orographic and mesoscale
flows. Quart. J. Roy. Meteorol. Soc.,127, 601–633.
Hurd, W.E., 1929: Northers of the Gulf of Tehuantepec. Mon. Wea. Rev., 57, 192–194.
Ishii, S., K. Sasaki, K. Mizutani, T. Aoki, T. Itabe, H. Kanno, D. Matsushima, W. Sha, A. Noda,
M. Sawada, M. Ujiie, Y. Matsuura, and T. Iwasaki, 2007: Temporal evolution and spatial
structure of the local easterly wind “kiyokawa-dashi” in Japan part I: Coherent Doppler lidar
observations. J. Meteorol. Soc. Japan, 85, 797–813.
Ivančan-Picek,B. and V.Tutiš, 1995: Mesoscale bora flow and mountain pressure drag. Meteorol.Z.,
4, 119–128
Ivančan-Picek,B. and V.Tutiš, 1996: A case study of a severe Adriatic bora on 28 December 1992.
Tellus A, 48, 357–367
Jackson, P.L., and D.G. Steyn, 1994a: Gap winds in a fjord. Part I: Observations and numerical
simulation. Mon. Wea. Rev., 122, 2645–2665.
Jackson, P.L., and D.G. Steyn, 1994b: Gap winds in a fjord. Part II: Hydraulic analog. Mon. Wea.
Rev., 122, 2666–2676.
Jackson, P.L., C.J.C. Reason, S. Guan, 1999: Numerical modeling of a coastal trapped disturbance.
Part II: Structure and dynamics, Mon.Wea. Rev., 127, 535–550.
Jurčec, V., 1981: On mesoscale characteristics of bora conditions in Yugoslavia, Pure Appl.
Geophys.,119, 640–657.
Jiang, J.H., D.L. Wu, and S.D. Eckermann, 2002: Upper Atmosphere Research Satellite
(UARS) MLS observation of mountain waves over the Andes. J. Geophys. Res., 107:
10.1029/2002JD002091.
3 Dynamically-Driven Winds 213

Jiang, Q., R.B. Smith, and J.D. Doyle, 2003: The nature of the mistral: Observations and modelling
of two MAP events. Q. J. R. Meteorol. Soc. 129, 857–875.
Jiang, Q., and J.D. Doyle, 2004: Gravity wave breaking over the central Alps: Role of complex
terrain, J. Atmos. Sci., 61, 2249–2266.
Jiang, Q., J.D. Doyle, and R.B. Smith, 2005: Blocking, descent and gravity waves: Observations
and modeling of a MAP northerly föhn event. Quart. J. Roy. Meteorol. Soc., 131, 675–701.
Jiang, Q., J.D. Doyle, and R.B. Smith, 2006: Interactions between lee waves and boundary layers.
J. Atmos. Sci., 63, 617–633.
Jiang, Q., J.D. Doyle, S. Wang, and R.B. Smith, 2007: On boundary layer separation in the lee of
mesoscale topography. J. Atmos. Sci., 64, 401–420.
Jiang, Q., and J.D. Doyle, 2008: Diurnal variation of downslope winds in Owens Valley during the
Sierra Rotor Experiment. Mon. Wea. Rev., 136, 3760–3780.
King, J.C., P.S. Anderson, D.G. Vaughan, G.W. Mann, S.D. Mobbs, and S.B. Vosper, 2004: Wind-
borne redistribution of snow across an Antarctic ice rise. J. Geoph. Res., 109, No. D11, D11104.
Klemp J.B., and D.K. Lilly, 1975. The dynamics of wave-induced downslope winds. J. Atmos. Sci.,
32, 320–339.
Klemp J.B., and D.K. Lilly, 1978. Numerical simulation of hydrostatic mountain waves. J. Atmos.
Sci., 35, 78–107.
Klemp, J.B., and D.R. Durran, 1987: Numerical modeling of bora winds. Meteor. Atmos. Phys.,
36, 215–227.
Klemp, J.B., W.C., Skamarock, and O. Fuhrer, 2003: Numerical consistency of metric terms in
terrain-following coordinates. Mon. Wea. Rev., 131, 1229–1239.
Koop, C.G., and F.K. Browand, 1979: Instability and turbulence in a stratified fluid with shear.
J. Fluid Mech., 93, 135–159.
Lackmann, G.M., and J. Overland, 1989: Atmospheric structure and momentum balance during a
gap-wind event in Shelikof Strait, Alaska. Mon. Wea. Rev., 117, 1817–1833.
Lane, T.P., M.J. Reeder, B.R. Morton, and T.L. Clark, 2000: Observations and numerical modelling
of mountain waves over the Southern Alps of New Zealand. Quart. J. Roy. Meteorol. Soc., 126,
2765–2788.
Lester, P.F., and W.A. Fingerhut, 1974: Lower turbulent zones associated with mountain lee waves.
J. Appl. Meteor., 13, 54–61.
Li, J., and Y.-L. Chen, 1998: Barrier jets during TAMEX. Mon. Wea. Rev., 126, 959–971.
Lilly, D.K., 1978: A severe downslope windstorm and aircraft turbulence event induced by a
mountain wave. J. Atmos. Sci., 35, 59–77.
Lilly, D.K., and J.B. Klemp, 1979: Effects of terrain shape on non-linear hydrostatic mountain
waves. J. Fluid Mech., 95, 241–261.
Lilly, D.K., and E.J. Zipser, 1972: The Front Range windstorm of January 11 1972. Weatherwise,
25, 56–63.
Liu, H., P. Olsson, and K. Volz, 2008: SAR observation and modeling of gap winds in the Prince
William Sound of Alaska. Sensors, 8, 4894–4914.
Loescher, K.A., G.S. Young, B.A. Colle, and N.S. Winstead, 2006: Climatology of barrier jets
along the Alaskan Coast. Part I: Spatial and temporal distributions. Mon. Wea. Rev., 134,
437–453.
Long, R.R., 1953: Some aspects of the flow of stratified fluids. I. A theoretical investigation. Tellus,
5, 42–57.
Long, R.R., 1954: Some aspects of the flow of stratified fluids. II. Experiments with a two-fluid
system. Tellus, 6, 97–115.
Lorenz, E.M., 1969: The predictability of a flow which possesses many scales of motion. Tellus,
21, 289–307.
Lyra, G. 1943. Theorie der stationären Leewellenstromung in freier Atmosphäre. Z. Angew. Math.
Mech., 23, 1–128.
Marwitz, J., and J. Toth, 1993: The Front range blizzard of 1990. Part I: Synoptic and mesoscale
structure. Mon. Wea. Rev., 121, 402–415.
214 P.L. Jackson et al.

Mashiko, W., 2008: Formation mechanism of a low-level jet during the passage of typhoon Ma-on
(2004) over the Southern Kanto District. J. Meteorol. Soc. Japan. Ser. II, 86, 183–202, URL
http://ci.nii.ac.jp/naid/110006633437/en/.
Mass, C.F., and M.D. Albright, 1985: A severe windstorm in the lee of the Cascade mountains of
Washington State, Mon. Wea. Rev., 113, 1261–1281.
Mass, C.F., and G.K. Ferber, 1990: Surface pressure perturbations produced by an isolated
mesoscale topographic barrier. Part I: General characteristics and dynamics. Mon. Wea. Rev.,
118, 2579–2596.
Mass, C.F., and N.A. Bond, 1996: Coastally trapped wind reversals along the United States west
coast during the warm season, Part II: Synoptic evolution. Mon. Wea. Rev., 124, 446–461.
Mayr, G.J., and L. Armi, 2008: Föhn as a response to changing upstream and downstream air
masses. Quart. J. Roy. Meteorol. Soc., 134, 1357–1369.
Mayr, G.J., J. Vergeiner, and A. Gohm, 2002: An Automobile Platform for the Measurement of
Foehn and Gap Flows. J. Atmos. Oceanic Tech., 19, 1545–1556.
McInnes, K.L., and J.L. McBride, 1993: Australian southerly busters. Part I: Analysis of a
numerically simulated case-study. Mon. Wea. Rev., 121, 1904–1920.
Mercer, A.E., M.B. Richman, and H.B. Bluestein, 2008: Statistical modeling of downslope
windstorms in Boulder, Colorado. Wea. Forecasting, 23, 1176–1194.
Miranda, P.M.A. and I.N. James, 1992: Nonlinear three-dimensional effects on gravity wave drag:
Splitting flow and breaking waves. Quart. J. Roy. Meteor. Soc., 118, 1057–1081.
Mobbs, S., S. Vosper, P. Sheridan, R. Cardoso, R. Burton, S. Arnold, M. Hill, V. Horlacher, and A.
Gadian, 2005: Observations of downslope winds and rotors in the Falkland Islands. Quart. J.
Roy. Meteorol. Soc., 131, 329–351.
Mohorovičić, A. 1889: Interessante Wolkenbildung über der Bucht von Buccari (with a comment
from the editor J. Hann). Meteor. Z., 24, 56–58.
Müller, H., R. Reiter, and R. Sládkovič, 1984: Die vertikale windstruktur beim merkur-
schwerpunkt ’tagesperiodische windsysteme’ aufgrund von aerologischen messungen im inntal
und im rosenheimer becken (engl.: Vertical structure of the diurnal wind system within the inn
valley and the adjacent plain: results from aerological soundings during the merkur campaign).
Arch. Met. Geophys. Biokl. B, 33, 359–372.
Nance, L. B., and B. R. Colman, 2000: Evaluating the use of a nonlinear two-dimensional model
in downslope windstorm forecasts. Wea. Forecasting, 15, 715–729.
Nappo, C. J. 2002: An introduction to atmospheric gravity waves. Academic Press, 279 pp.
Neiman, P.J., R.M. Hardesty, M.A. Shapiro and R.E. Cupp, 1988: Doppler lidar observations of a
downslope windstorm. Mon. Wea. Rev., 116, 2265–2275.
Neiman, P.J., F.M. Ralph, A.B. White, D.D. Parrish, J.S. Holloway, and D.L. Bartels, 2006: A
multiwinter analysis of channeled flow through a prominent gap along the northern California
coast during CALJET and PACJET. Mon. Wea. Rev.,134, 1815–1841.
Nuss, W.A., J. Bane, W. Thompson, C. Dorman, M. Ralph, R. Rotunno, J. Klemp, W. Skamarock,
R. Samelson, A.Rogerson, C.J.C. Reason, P.L. Jackson, 2000: Coastally trapped wind reversals:
A new level of understanding from the experiment on coastally trapped disturbances, Bull.
Amer. Meteor. Soc., 81, 719–743.
Ogura, Y., and N. Phillips, 1962: Scale analysis for deep and shallow convection in the atmosphere.
J. Atmos. Sci., 19, 173–179.
Økland, H., 1990: The dynamics of coastal troughs and coastal fronts. Tellus, 42A, 444–462.
Ólafsson, H., and P. Bougeault, 1996: Nonlinear flow past an elliptic mountain ridge. J. Atmos.
Sci., 53, 2465–2489.
Ólafsson, H., and P. Bougeault, 1997: The effect of rotation and surface friction on orographic
drag. J. Atmos. Sci., 54, 193–210.
Olson, J.B., B. Colle, N. Bond, and N. Winstead, 2007: A comparison of two coastal battier jet
events along the southeast Alaskan Coast during the SARJET field experiment. Mon. Wea.
Rev., 135, 3642–3663.
Olson, J.B., and B. A. Colle, 2009: Three-dimensional idealized simulations of barrier jets along
the southeast coast of Alaska, Mon. Wea. Rev., 137, 391–413.
3 Dynamically-Driven Winds 215

Orlić, M., M. Kuzmić, and Z. Pasarić, 1994: Response of the Adriatic Sea to the bora and sirocco
forcing. Continent. Shelf Res., 14, 91–116.
Orlić, M., V. Dadić, B. Grbec, N. Leder, A. Marki, F. Matić, H. Mihanović, G. Beg Paklar, M.
Pasarić, Z. Pasarić, and I. Vilibić, 2006: Wintertime buoyancy forcing, circulation systems
produced in the Adriatic. J. Geophys. Res. Oceans, 111, C03S07.
Overland, J.E., 1984: Scale analysis of marine winds in straits and along mountainous coasts. Mon.
Wea. Rev., 112, 2530–2534.
Overland, J.E., and N.A. Bond, 1993: The influence of coastal orography: The Yakutat storm. Mon.
Wea. Rev., 121, 1388–1397.
Overland, J.E., N.A. Bond, 1995: Observations and scale analysis of coastal wind jets. Mon. Wea.
Rev., 123, 2934–2941.
Pan, F., and R.B. Smith, 1999: Gap winds and wakes: SAR Observations and numerical
simulations. J. Atmos. Sci., 56, 905–923.
Parish, T.R., 1982: Barrier winds along the Sierra Nevada Mountains. J. Appl. Meteor., 21,
925–930.
Parish, T.R., 1983: The influence of the Antarctic peninsula on the wind-field over the western
Weddell Sea. J. Geophys. Res., 88, 2684–2692.
Peltier, W.R., and T.L. Clark, 1979: the evolution and stability of finite-amplitude mountain waves.
Part II: Surface wave drag and severe downslope windstorms. J. Atmos. Sci., 36, 1498–1529.
Peltier, W.R., and T.L. Clark, 1983: Nonlinear mountain waves in two and three spatial dimensions.
Quart. J. Roy. Meteor. Soc.,109, 527–548.
Peltier, W.R., and J.F. Scinocca, 1990: The origin of severe downslope windstorm pulsations.
J. Atmos. Sci., 47, 2853–2870.
Peng, M.S., and W.T. Thompson, 2003: Some aspects of the effect of surface friction on flows over
mountains. Quart. J. Roy. Meteorol. Soc., 129, 2527–2557.
Petersen, G.N., J.E. Kristjánsson, and H. Ólafsson, 2003: The effect of upstream wind direction
on atmospheric flow in the vicinity of a large mountain. Quart. J. Roy. Meteorol. Soc., 131,
1113–1128.
Pettre, P., 1982: On the problem of violent valley winds. J. Atmos. Sci., 39: 542–554.
Pierrehumbert, R.T., and B. Wyman, 1985: Upstream effects of mesoscale mountains. J. Atmos.
Sci., 42, 977–1003.
Prandtl, L. 1942: Führer durch die Strömungslehre. Friedr. Vieweg & Sohn, Braunschweig, 3rd
edn.
Queney, P. 1948: The problem of airflow over mountains: A summary of theoretical studies. Bull.
Amer. Meteor. Soc., 29, 16–26.
Raphael, M.N., 2003: The Santa Ana winds of California. Earth Interactions, 7(8), 13 pp.
Reason, C.J.C., D.G. Steyn, 1992: The dynamics of coastally trapped mesoscale ridges in the lower
atmosphere, J. Atmos. Sci., 49, 1677–1692.
Reason, C.J.C., K. Tory, P.L. Jackson, 2001: A model investigation of the dynamics of a Coastally
Trapped Disturbance. J. Atmos. Sci., 58(14), 1892–1906.
Reed, R.J., 1981: A case of a bora-like windstorm in western Washington. Mon. Wea. Rev., 109,
2383–2393.
Reinecke, P. A., and D. R. Durran, 2009a: Initial-condition sensitivities and the predictability of
downslope winds. J. Atmos. Sci., 66, 3401–3418.
Reinecke, P. A., Durran, D., 2009b: The Overamplification of Gravity Waves in Numerical
Solutions to Flow over Topography. Mon. Wea. Rev., 137, 1533–1549.
Reid, J.J., L.M. Leslie, 1999: Modeling coastally trapped wind surges over southeastern Australia.
Part I: Timing and speed of propagation. Wea. Forecasting, 14, 53–66.
Richard, E., P. Mascart, and E.C. Nickerson, 1989: On the role of surface friction in downslope
windstorms. J Appl. Meteor., 28, 241–251.
Rogers, D.P., C.E. Dorman, K. A. Edwards, I.M. Brooks, W.K Melville, S.D. Burk, W.T.
Thompson, T. Holt, L. M Ström, M. Tjernström, B. Grisogono, J.M. Bane, W.A. Nuss, B.
M. Morley and A.J. Schanot, 1998: Highlights of Coastal Waves, 1996, Bull. Amer. Meteor.
Soc., 79, 1307–1326.
216 P.L. Jackson et al.

Ross, A.N., and S.B. Vosper, 2003: Numerical simulations of stably stratified flow through a
mountain pass. Quart. J. Roy. Meteor. Soc., 129, 97–115.
Sawyer, J.S., 1960: Numerical calculation of the displacements of a stratified airstream crossing a
ridge of small height. Quart. J. Roy. Meteor. Soc., 86, 326–345.
Sawyer, J.S., 1962: Gravity waves in the atmosphere as a three-dimensional problem. Quart. J. Roy.
Meteor. Soc., 88, 412–425.
Schär, C., and R.B. Smith, 1993: Shallow-water flow past isolated topography. Part I: Vorticity
production and wake formation. J. Atmos. Sci., 50, 1373–1400.
Schär, C., and D.R. Durran, 1997: Vortex formation and vortex shedding in continuously stratified
flows past isolated topography. J. Atmos. Sci., 54, 534–554.
Schär, C., D. Leuenberger, O. Fuhrer, D. Lüthi, and C. Girard, 2002: A new terrain-following vical
coordinate formulation for atmospheric prediction models. Mon. Wea. Rev., 130, 2459–2480.
Schultz, D.M., W.E. Bracken, L.F. Bosart, G.J. Hakim, M.A. Bedrick, M.J. Dickinson, K.R. Tyle,
1997: The 1993 superstorm cold surge: Frontal structure, gap flow, and tropical impact. Mon.
Wea. Rev.,125, 5–39.
Schwerdtfeger, W., 1974: Mountain barrier effect on the flow of stable air north of the Brooks
Range. Proc. 24th Alaskan Science Conf., Fairbanks, AL, Geophysical Institute, University of
Alaska Fairbanks, 204–208.
Schwerdtfeger, W., 1975: The effect of the Antarctic Peninsula on the temperature regime of the
Weddell Sea. Mon. Wea. Rev., 103, 45–51.
Schwerdtfeger, W., 1979: Meteorological aspects of the drift of ice from the Weddell Sea towards
the mid-latitude westerlies. J. Geophys. Res., 84, 6321–6327.
Scinocca, J.F., and W.R. Peltier, 1989: Pulsating downslope windstorms. J. Atmos. Sci., 46,
2885–2914.
Scorer, R., 1952: Mountain-gap winds: A study of surface wind at Gibraltar. Quart. J. Roy. Meteor.
Soc., 78, 53–61.
Seibert, P., 1990: South foehn studies since the ALPEX experiment. Meteor. Atmos. Phys.,43,
91–103.
Sharp, J., C.F. Mass, 2002: Columbia Gorge gap flow. Bull. Amer. Meteor. Soc., 83, 1757–1762.
Sharp, J., C.F. Mass, 2004: Columbia Gorge Gap Winds: Their Climatological Influence and
Synoptic Evolution. Wea. Forecasting, 19, 970–992.
Sheridan, P.F., V. Horlacher, G.G. Rooney, P. Hignett, S.D. Mobbs, and S.B. Vosper, 2007:
Lee waves and flow separation downwind of the Pennines. Quart. J. Roy. Meteor. Soc.,
133,1353–1369.
Shutts, G.J., 1997: Operational lee wave forecasting. Meteorol. Appl. 4, 23–35.
Smith, C.M. and E.D. Skyllingstad 2009: Investigation of upstream boundary layer influence on
mountain wave breaking and lee wave rotors using a large-eddy simulation. J. Atmos. Sci., 66,
3147–3164.
Smith, R.B., 1976: The generation of lee waves by the Blue Ridge. J. Atmos. Sci., 33, 507–519.
Smith, R.B., 1977: The steepening of hydrostatic mountain waves. J. Atmos. Sci., 34, 1634–1654.
Smith, R.B., 1979: The influence of mountains on the atmosphere. Adv. Geophys., Vol. 21,
Academic Press, 87–230.
Smith, R.B., 1980: Linear theory of stratified hydrostatic flow past an isolated mountain.
Tellus,32A, 348–364.
Smith, R.B., 1985: On severe downslope winds. J. Atmos. Sci., 42, 2597–2603.
Smith, R.B., 1987: Aerial observations of the Yugoslavian bora. J. Atmos. Sci., 44, 269–297.
Smith, R.B., 1989: Hydrostatic airflow over mountains. Adv. Geophys., 31, 1–41.
Smith, S.A., 2004: Observations and simulations of the 8 November 1999 MAP mountain wave
case. Quart. J. Roy. Meteor. Soc., 130, 1305–1326.
Smith, S.A., and A.S. Broad, 2003: Horizontal and temporal variability of mountain waves over
Mont Blanc. Quart. J. Roy. Meteor. Soc., 129, 2195–2216.
Smith, R.B., A.C. Gleason, and P.A. Gluhosky, 1997: The wake of St Vincent. J. Atmos. Sci., 54,
606–623.
3 Dynamically-Driven Winds 217

Smith, R.B., and S. Grønås, 1993: Stagnation points and bifurcation in 3-D mountain air-flow.
Tellus, 45A, 28–43.
Smith, R.B., S. Skubis, J.D. Doyle, A.S. Broad, C. Kiemle, and H. Volkert, 2002: Mountain waves
over Mont Blanc: Influence of a stagnant boundary layer, J. Atmos. Sci.,59, 2073–2092.
Smith, R.B., and J. Sun, 1987: Generalized hydraulic solutions pertaining to severe downslope
winds. J. Atmos. Sci., 44, 2934–2939.
Smith, R.B., Jiang, Q. and Doyle, J.D. 2006. A theory of gravity wave absorption by boundary
layers, J. Atmos. Sci., 63, 774–781.
Smith, R.B., 2007: Interacting mountain waves and boundary layers. J. Atmos. Sci., 64, 594–607.
Smith, R.B., J.D. Doyle, Q. Jiang, and S.A. Smith, 2007: Alpine gravity waves: Lessons from MAP
regarding mountain wave generation and breaking. Quart. J. Roy. Meteor. Soc., 133, 917–936.
Smolarkiewicz, P.K. and R. Rotunno, 1989: Low Froude number flow past three-dimensional
obstacles. Part I: Baroclinically generated lee vortices. J. Atmos. Sci., 46, 1154–1164.
Sprenger, M., and C. Schär, 2001: Rotational aspects of stratified gap flows and shallow foehn.
Q. J. R. Meteorol. Soc., 127, 161–187.
Steenburgh, W.J., D.M. Schultz, and B.A. Colle, 1998: The structure and evolution of gap outflow
over the gulf of Tehuantepec, Mexico. Mon. Wea. Rev., 126, 2673–2691.
Tutiš,V. and B. Ivančan-Picek, 1991: Pressure drag on the Dinaric Alps during the ALPEX SOP,
Meteor. Atmos. Phys, 47, 73–81.
Vergeiner, I., 1971: An operational linear lee wave model for arbitrary basic flow and two-
dimensional topography. Quart. J. Roy. Meteor. Soc., 97, 30–60.
Volkert, H., C. Keil, C. Kiemle, G. Poberaj, J-P. Chaboureau, and E. Richard, 2003: Gravity waves
over the eastern Alps: A synopsis of the 25 October 1999 event (IOP10) combining in situ and
remote-sensing measurements with a high-resolution simulation. Quart. J. Roy. Meteor. Soc.,
129, 777–797.
Vosper, S.B., 2003: Development and testing of a high resolution mountain-wave forecasting
system. Meteor. Appl.,10, 75–86.
Vosper, S.B., 2004: Inversion effects on mountain lee waves. Quart. J. Roy. Meteor. Soc., 130,
1723–1748.
Vosper, S.B., and A.R. Brown, 2007: The effect of small-scale hills on orographic drag. Quart.
J. Roy. Meteor. Soc., 133, 1345–1352.
Vosper, S.B. and S.D. Mobbs, 1996: Lee waves over the English Lake District, Quart. J. Roy.
Meteor. Soc., 122, 1283–1305.
Vosper, S.B., P.F. Sheridan, and A.R. Brown, 2006: Flow separation and rotor formation beneath
two-dimensional trapped lee waves. Quart. J. Roy. Meteor. Soc., 132, 2415–2438.
Winstead, N.S., B. Colle, N. Bond, G. Young, J. Olson, K. Loescher, F. Monaldo, D. Thompson,
and W. Pichel, 2006: Using SAR remote sensing, field observations, and models to better
understand coastal flows in the Gulf of Alaska, Bull. Amer. Meteor. Soc., 87, 787–800.
Wood I.R., 1968: Selective withdrawal from a stably stratified fluid. J. Fluid Mech.32: 209–223.
Yeh, H.-C., and Y.-L. Chen, 2003: Numerical simulations of the barrier jet over northwestern
Taiwan during the mei-yu season. Mon. Wea. Rev., 131, 1396–1407.
Yoshino, M.M., 1972: An annotated bibliography on bora. Climatol. Notes, 10, 1–22.
Yoshino, M.M., 1976: Bora studies: A historical review and problems today. In Local Wind Bora,
M.M. Yoshino Ed., 3–18.
Yu, C.-K., and B.F. Smull, 2000: Airborne Doppler observations of a landfalling cold front
upstream of a steep coastal orography. Mon. Wea. Rev., 128, 1577–1603.
Žagar, M. and J. Rakovec, 1999: Small scale surface wind prediction using dynamical adaptation.
Tellus, 51A, 489–504.
Zängl, G., 2002a: Idealized numerical simulations of shallow föhn. Quart. J. Roy. Meteor.
Soc., 128, 431–450.
Zängl, G., 2002b: An Improved Method for Computing Horizontal Diffusion in a Sigma-
Coordinate Model and Its Application to Simulations over Mountainous Topography. Mon.
Wea. Rev.,130, 1423–1432.
218 P.L. Jackson et al.

Zängl, G., 2002c: Stratified flow over a mountain with a gap: Linear theory and numerical
simulations. Q. J. R. Meteor. Soc.,128, 927–949.
Zängl, G., 2003: Deep and shallow south foehn in the region of Innsbruck: Typical features and
semi-idealized numerical simulations. Meteor. Atmos. Phys., 83, 237–262.
Zängl, G., 2004: A reexamination of the valley wind system in the Alpine Inn Valley with numerical
simulations. Meteor. Atmos. Phys., 87, 241–256.
Zängl, G., 2006: North foehn in the Austrian Inn valley: A case study and idealized numerical
simulations. Meteor. Atmos. Phys., 91, 85–105.
Zängl, G., A. Gohm, and G. Geier, 2004: South foehn in the Wipp Valley – Innsbruck region:
Numerical simulations of the 24 October 1999 case (MAP-IOP 10). Meteor. Atmos. Phys., 86,
213–243.
Zängl, G., A. Gohm, and F. Obleitner, 2008: The impact of the PBL scheme and the vertical
distribution of model layers on simulations of Alpine foehn. Meteor. Atmos. Phys., 99, 105–128.
Zecchetto, S., and C. Cappa, 2001: The spatial structure of the Mediterranean Sea winds as revealed
by the ERS-1 scatterometer. Int. J. Remote Sensing, 22, 45–70.
Chapter 4
Understanding and Forecasting Alpine Foehn

Hans Richner and Patrick Hächler

Abstract This chapter focuses on the history, physics, climatology, forecasting


and the broad effects of Alpine foehn on human populations. In the European
Alps, foehn winds have been studied since the mid-1800s. The main focus of
the investigations was the question of why foehn winds are so warm. While it
soon became clear that adiabatic processes provide an explanation, the role of wet
adiabatic rising on the upwind side of the Alps continued to be strongly debated.
The so-called textbook theory for foehn – heat gain by wet adiabatic, forced lifting
on the upwind side followed by dry adiabatic descent in the lee – represents only
an extreme situation. Foehn occurs also with partial or complete blocking of the
upwind air mass, i.e., with limited or no heat gain by wet adiabatic expansion. The
second focus is on processes which lead to descending air masses after passing
the mountain ridge. A discussion of the most important processes shows that there
seems to be no theory which is applicable in all situations. Forecasting foehn is still
a challenge to meteorologists. While the general foehn situation can be predicted
reliably, today’s numerical models still often poorly simulate the sudden break in
of potentially devastating foehn air in the lee. Efforts to improve this must continue
because foehn storms have a significant societal impact (threat to transportation
systems and massive increase of fire danger) as several recent incidents show.

H. Richner ()
Institute for Atmospheric and Climate Science (IACETH), ETH Zurich, Universitätsstrasse 16,
Zurich CH-8092, Switzerland
e-mail: hans.richner@ethz.ch
P. Hächler
Federal Office of Meteorology and Climatology MeteoSwiss, Krähbühlstrasse 58, Zurich
CH-8044, Switzerland
e-mail: patrick.haechler@meteoswiss.ch

F. Chow et al. (eds.), Mountain Weather Research and Forecasting, 219


Springer Atmospheric Sciences, DOI 10.1007/978-94-007-4098-3 4,
© Springer ScienceCBusiness Media B.V. 2013
220 H. Richner and P. Hächler

4.1 Introduction

Foehn is a prominent meteorological phenomenon occurring in practically all


extended mountain ranges. “Foehn” is a generic term for a downslope wind that
is strong, warm, and dry; WMO (1992) defines foehn wind as a “wind [which is]
warmed and dried by descent, in general on the lee side of a mountain.” Its societal
impact is considerable, be it in a favorable manner with regard to climatology (mild
climate) or in a negative way (windstorms) with respect to aviation and traffic.
Outside the Alps, foehn storms are often called downslope windstorms.
While the term “foehn” originated in the European Alpine area, foehn winds
occur all over the world where there are extended mountain ranges. Foehn winds are
even observed in places where the mountain ridges are not that high such as in the
United Kingdom. Depending on the place, foehn winds might have a different name
such as, e.g., Chinook (North America) or Helm wind (UK). The rapid temperature
rise and the dry air of foehn type winds have also led to numerous local, descriptive
names such as “snow eater,” “grape cooker,” etc. Examples of some of these flows
are given in Chap. 3, Sect. 3.4.
This chapter continues the discussion of dynamically-driven winds from Chap. 3,
but specifically focusing on the combined history, physics, climatology, forecasting
and the broad effects of Alpine foehn on human populations. Alpine foehn presents
a significant forecasting and modeling challenge, and has been the topic of study
for close to two centuries. This chapter highlights this particular flow phenomenon
because of the breadth of its impact and the peculiar challenges associated with
forecasting. The chapter centers on recent results of the Mesoscale Alpine Program
(MAP) and on ongoing research activities mainly in Switzerland, but touches also
on activities in Austria and France. For research on foehn-type winds in other parts
of the world, see e.g. Brinkmann (1974) (Rocky Mountains), Raphael (2003) (Santa
Ana Winds), or McGowan et al. (2002) (New Zealand Alps).
The overview encompasses statistical analyses of foehn occurrence in different
Alpine regions, on the interaction of foehn flow and the cold pool, and on
current techniques for forecasting foehn-related windstorms. Figure 4.1 shows a
topographical map of the Alpine massif. A detailed list of terms and definitions is
provided in the appendices to help clarify the varying terminology used to describe
foehn events.

4.2 History of Foehn Research

For many decades, foehn was the outstanding example to explain thermodynamic
processes and the role of latent heat in the atmosphere. Hence, nearly every textbook
contains a graph similar to Fig. 4.2. Driven by the synoptic scale pressure field,
humid air is forced towards a mountain range. As it ascends by forced convection,
it cools dry-adiabatically until reaching saturation. From then on, the rise is
4 Understanding and Forecasting Alpine Foehn 221

Fig. 4.1 Topographical map of the Alpine massif. The two dashed lines show the Gotthard cross
section (western) and the Brenner cross section (eastern). Only those Alpine regions are labelled
which are referred to in the text

Fig. 4.2 Schematic description of the “textbook theory” of foehn. Diagrams similar to this one are
found in many textbooks on meteorology

wet-adiabatically until the air reaches the crest of the mountain range; as a
consequence, clouds are formed and precipitation occurs. As the air descends in the
lee of the mountain, it is heated dry-adiabatically; as consequence, the air becomes
dry and reaches temperatures that – at equal elevation – are higher than the original
222 H. Richner and P. Hächler

Fig. 4.3 The “proof” that Hann was aware of the two types of foehn: Original illustration depicting
Hann’s (1866) foehn theory in Ficker (1920). “Nordseite”: north side; “Südseite”: south side;
“Föhn, Erwärmung”: foehn, warming; “Luft in Ruhe”: air at rest. Note that north–south is reversed
with respect to Fig. 4.2

temperature on the windward side. Because of the very clean and dry air on the lee
side, visibility is outstanding (D>foehn window1 ), and over the crest, the piled-up
clouds can be seen as the D> foehn wall (Fig. 4.2).
As beautiful as this example is, in reality foehn winds often do not follow
this classical textbook theory that is attributed to Hann (nineteenth century). In
his “Lehrbuch der Meteorologie”, Hann (1901) describes both types of foehn (see
Figs. 4.2 and 4.3) and mentions that many forms in between the two have been
observed. Seibert (2005) concludes that Hann’s explanation was seriously distorted
in the first half of the twentieth century. Particularly Austrian researchers repeatedly
questioned the validity of the textbook approach. In his book “Environmental
Aerodynamics”, Scorer (1978) wrote a chapter entitled “Foehn Fallacy” which gives
at least two explanations for the warming of air masses which do not require the heat
of condensation on the windward slope of the mountain ridge: mechanical stirring
of a stably stratified air mass, and – more important – blocking. In the first case, the
lower part becomes warmer and the upper part cools because the mixing produces a
constant potential temperature, in the second case, potentially warmer air subsides
in the lee.
Analyses of data collected during and after the Alpine Experiment ALPEX in
1982 clearly confirmed that, at least in the Brenner cross-section, there was blocked
air during most foehn cases. This is particularly true for the “century foehn” of
November 8, 1982. In the Ticino region, i.e., south of the Gotthard, however, there
was some light precipitation towards the end of this extreme foehn event (hourly
mean winds up to 35 m s1 and gusts over 50 m s1 at the station Gütsch, see
Fig. 4.7).
It is a little disturbing that not only popular publications but also modern
textbooks often present only the theory depicted in Fig. 4.2 without discussing
alternative foehn schemes.
Austria and Switzerland were clearly the hotspots for foehn research. Innsbruck
(Austria) is the place where scientists like Hann, Ficker, Hoinkes, Kuhn, Dreiseitl,

1
Terms preceded by “D>” are further explained in Appendix A.
4 Understanding and Forecasting Alpine Foehn 223

Steinacker, Seibert, Hoinka, Mayr, and Gohm have carried out their research; in the
Swiss Alps it is mainly Billwiller, Wild, Streiff-Becker, Schüepp, Frey, Gutermann
and others who have made observations and developed their theories since the
nineteenth century. In Switzerland, research was not quite as active as in Austria.
However, in Switzerland, monks and pastors meticulously observed and recorded
foehn events, this being particularly true for the Reuss Valley. Thanks to them, a
quite homogenous time series for foehn events in Altdorf is available since 1864.

4.3 Characteristics of Foehn

4.3.1 Textbook Theory and Real Dynamics

South foehn occurs in the Alpine region when a synoptic pressure field according
to Fig. 4.4 exists. (For simplicity the following discussions is restricted to south
foehn, however, some remarks related to north foehn and west foehn are given
in Sect. 4.3.4.) Nowadays it is widely accepted that foehn winds can and do
occur without precipitation. The lower the crest height, the more likely it is that
advected air simply crosses the mountain ridge and, subsequently, descends, as
already debated by Hann, Ficker and others and as depicted in Fig. 4.3. Trajectory
analyses as well as tracer studies with ozone confirm this mechanism (Baumann
et al. 2001).

Fig. 4.4 Synoptic pressure field producing foehn in the Alps


224 H. Richner and P. Hächler

Fig. 4.5 Suggested parameters for characterizing foehn types. A: depth of blocked air, C: crest
height,  : difference in potential temperature

The fact that much foehn research was and still is being done along the rather
flat Brenner cross-section, explains why in Austria the textbook theory was more
fiercely queried than, e.g., in Switzerland. As a matter of fact, Hann (1866) already
distinguished between foehn type I or “Swiss foehn type” (for what is here called
the textbook theory) and foehn type II or “Austrian foehn type” (for the airflow
without gain of latent heat and with reduced or no precipitation on the windward
side). These two foehn types must be regarded as extremes, as anything in between
can be observed.
It must be stressed that even heavy precipitation on the upwind side is no proof
for gain of latent heat; precipitation can stem from the lower air mass that is
not part of the foehn flow. For describing a foehn case, it would be desirable to
use as additional characterizing parameters (a) the height of the internal boundary
between the lower, blocked air and the pressure driven foehn flow aloft, and (b) the
difference ™ between the potential temperature of the two air masses (see Fig. 4.5).
If sufficient information on the vertical structure is available, this parameter can be
determined by trajectory calculations. In a textbook foehn case (i.e., type I), the
value for A would be zero; in a foehn situation with total blocking on the upwind
side (i.e., type II), the value for A would correspond to the ridge height C. The
value for A could be expressed either in absolute units or in percent of the ridge
height.
There are at least two reported cases when foehn winds were observed simulta-
neously on both sides of the same mountain range (Frey 1986). This can only be
explained by assuming subsidence of significant air masses over the ridge. In this
1986 event, south winds were in excess of 40 m s1 north of the Alps!
On the other hand, it would be wrong to dismiss textbook theory completely.
Observations clearly show that heavy precipitation can and does occur on the
windward side of high mountain complexes during foehn. There are indications
that this is more often the case south of the (higher) Gotthard than south of the
(lower) Brenner massif. Although this suspicion has been brought up repeatedly
after ALPEX, there are no statistical analyses that would prove this, and no firm
statement can be made.
4 Understanding and Forecasting Alpine Foehn 225

4.3.2 The Effect of Topography

Foehn areas are primarily defined by topography. Basically, airflow penetrates a


given valley more easily the better the valley axis is aligned with the flow, and hence
the fewer ridges further downwind hinder the flow. This situation can be described
by saying that the flow is diffluent, the degree of the diffluence being controlled
by topography. Based on these principles, the topography defines the foehn areas.
In general, major valleys perpendicular to the ridge represent the areas with the
highest foehn frequency (see Sect. 4.5).
Valleys have a strong channeling effect on the flow. Because the cross section
decreases as the flow further penetrates into the valley, the wind speeds can be higher
near the ground than they are above. As observations show, the channeling effect (in
the sense that the valley axis determines the direction of the flow) can reach up to
heights well above the crest of the mountains bordering the valley in question.
A forecaster’s rule says that foehn winds rarely descend more that 2,000 m
behind the crest over which they streamed in. For this reason, foehn is seldom
experienced in the region of Lake Thun, a tourist area just north of the Jungfrau
region in Switzerland. A sound explanation in terms of dynamics for this rule is not
available.
When talking about foehn frequency at a given location, it is important to exactly
define what is meant by saying, “There is foehn”. For some people it is sufficient
to have a foehn airflow above the location which does not reach the ground, for
others, the winds must actually touch down. During the winter period September
to March, foehn reaches the downtown area of Zurich, Switzerland, on average
only on 2–3 days, however, on 50–70 days, there is a foehn situation where the
downtown remains in the calm cold pool while a possibly strong, certainly warmer
foehn airflow exists a few hundred meters above.
In some cases it is very clear whether foehn is present or not (see next section).
However, for a foehn climatology, the definition of foehn can become crucial
and problematic. Barry (2008, p. 173) discusses this issue and gives examples
for definitions used in different foehn climatologies. The “Alpine Research Group
Foehn Rhine Valley/Lake Constance” (AGF) has adopted objective definitions for
both manual (AGF 2007) and automatic (see Sect. 4.6.1) foehn identification.

4.3.3 A Real Example of Foehn Characteristics

Figure 4.6 shows how foehn manifests itself in one of Switzerland’s important
foehn valleys, the Reuss Valley which is part of the so-called Gotthard cross-
section (Fig. 4.7). The rise of temperature, the drop in relative humidity, the onset of
high winds, and the constant wind direction occur simultaneously within minutes.
Consequently, the onset and the breakdown of foehn can easily be determined. This
is true for almost any foehn station that is located in the center of a not too broad
foehn valley.
226 H. Richner and P. Hächler

Fig. 4.6 Example of foehn in Altdorf on January 2–7, 2008 as observed by the surface station.
The horizontal axes are time axes, date and time is indicated below the first frame. TT temperature,
UU humidity, ff wind speed, dd wind direction. The two red vertical lines mark the simultaneous
abrupt changes in the variables at the onset of the main and of a secondary short foehn phase. Note
how clearly foehn can be identified (Data kindly provided by MeteoSwiss)

Figure 4.6 also demonstrates that the breakdown of a foehn occurs usually in
several phases. Or, as one of our colleagues once put it, foehn breakdown is a
gigantic battle between two air masses, namely the foehn and the approaching cold
front. As the foehn weakens, the cold air moves in, only to be pushed back for a
limited period, a back-and-forth process which is repeated several times before the
foehn flow breaks down for good.
4 Understanding and Forecasting Alpine Foehn 227

Fig. 4.7 The so-called Gotthard cross-section, the only south–north profile with only one ridge,
and valleys practically perpendicular to it. In the south is the Leventina, in the north the Reuss
Valley. The distance between Lugano and Altdorf is almost exactly 100 km

Table 4.1 Characteristic parameters for stations


on the upstream side, the ridge, and the lee side
for March 10, 2008, 1200 UTC (Meteorological
data kindly provided by MeteoSwiss)
Param Lugano Piotta Gütsch Altdorf
z (m) 273 1,007 2,287 449
p (hPa) 970.0 887.1 752.8 943.1
T (ı C) 6.5 2.3 4.5 13.1
™ (ı C) 8.9 11.9 18.1 17.9

Figure 4.8 shows a foehn occurrence, which proves that there are cases which
indeed follow the textbook theory. There is significant precipitation and south wind
on the upwind side. Table 4.1 shows that on the top of the ridge the potential
temperature is markedly higher than on the upwind side (for the location of the
stations see Fig. 4.7). Precipitation on the windward slope alone is not sufficient
proof for the textbook theory, it merely indicates blocking. An increase of the
potential temperature along the upwind slope, however, implies forced advection
and, consequently, a type I foehn.
For the Mesoscale Alpine Program, MAP, special efforts were made to densely
instrument an area for foehn studies in the Rhine Valley (Richner et al. 2005). For
some foehn cases, detailed investigations of the three-dimensional structure of foehn
were made (Drobinski et al. 2003). A general summary of the main results related to
airflow over and in mountains can be found in Drobinski et al. (2007). In these, the
distinction between D> “shallow foehn” and D> “deep foehn” that had evolved
during the last decades was clearly found in the life cycle of the foehn episode.
As expected, compared to simple mountain cross-sections (like the aforementioned
Gotthard cross-section), complex topographical features cause significantly higher
complexity of foehn flow; a glance at the topography shown in Fig. 4.9 makes this
statement understandable.
228 H. Richner and P. Hächler

Fig. 4.8 A foehn case on March 10/11, 2008 with precipitation on the upwind side. The horizontal
axes are time axes, date and time is indicated below the first frame. Top frame: precipitation (RR) in
Lugano (blue) and Magadino (red); next frames show temperature (TT), humidity (UU) and wind
(ff ) for Altdorf (red), Lugano (blue), and Gütsch (green); all data are surface data. The station
Magadino lies 18 km north of Lugano, it is included to give a more representative measure of the
precipitation intensity (Data kindly provided by MeteoSwiss)

Model runs for a MAP foehn case concluded that the simulation outputs
essentially confirm the observed foehn dynamics (Jaubert and Stein 2003). However,
it is also realized that the role of passes and valleys is not fully understood yet.
4 Understanding and Forecasting Alpine Foehn 229

Fig. 4.9 Topography around the Rhine Valley where foehn studies during MAP were conducted
(Drobinski et al. 2003, reproduced by permission of Wiley and Sons)

Fig. 4.10 TKE values in the


lee of the Alps during foehn.
Yellow: values measured by
aircraft, blue: values
computed by the French
“meso-NH” model (Lothon
2002, reproduced by
permission of Marie Lothon)

At any rate, for a reliable foehn forecast, today’s mesoscale models are not yet
sufficiently accurate, this despite the fact that they do produce realistic flow and
temperature fields. Turbulence, on the other hand, is very poorly reproduced in
model runs. Computed turbulent kinetic energy (TKE) values were orders of
magnitudes lower than those observed by aircraft (Fig. 4.10). A detailed study of
the observed turbulent properties of Alpine foehn events is found in Lothon et al.
(2003).
230 H. Richner and P. Hächler

4.3.4 Foehn from Directions Other than South

As mentioned, Alpine foehn also occurs in the reverse direction, i.e., with a north-
to-south airflow. This north foehn is almost always connected with precipitation
in the north, i.e., north foehn does often follow the textbook theory. In addition
to the difference of the dominating mechanisms, there are additional differences
between south and north foehn. North foehn blows in the direction of the positive
north–south (i.e., meridional) temperature gradient, hence, the warming effect of the
foehn is less pronounced because the ambient temperature in the lee is also higher.
Nevertheless, because of its low humidity, north foehn regularly dries the extended
chestnut forests in southern Switzerland, and forest fires occur recurrently during
north foehn season. Strong north foehn occurs also regularly in the Trentino, in
South Tyrol, and in the Lombardy region of Italy; particularly strong events such as
the one on January 27/28, 2008, are called “foehn bollente” (bollente [Italian] D hot,
boiling).
In the Piedmont region (Northern Italy, west of Milan), foehn occurs with
westerly flow because the Alps bend here towards the south. West winds coming
from France over the Ligurian Alps and the Maritime Alps (i.e., the part of the Alps
between France and Italy, see Fig. 4.1) generate foehn with all its attributes in the
region of Turin (Musso and Cassardo 2004; Di Napoli and Mercalli 2008).
In France, very little research on foehn has been published. Buchot (1977)
compared the foehn winds blowing from northeast to west in the Tarantaise
(basically the region of the Isère Valley in Savoy) with the south foehn in Austria
and Switzerland. From a statistical analysis covering 15 years, he concluded that
there are no significant differences in the characteristics.

4.4 Foehn Dynamics

4.4.1 Leeside Motion

Probably the least understood mechanism in the flow dynamics is the behavior of
the air masses after passing the mountain ridge. Why does the air descend and not
simply continue at the same height level? This question is even more justified by
the fact that its potential temperature is in most cases higher, i.e., that a stable
stratification is present! Steinacker (2006) has compiled the different theories that
have emerged over the last century; the six theories summarized here follow largely
his work; Figs. 4.11–4.16 are slight modifications of those found in Steinacker
(2006).
Probably the first person to write about the problem of the descending air was
Wild (1901), a Swiss physicist who became interested in meteorological problems
because he realized that weather influenced his astronomical observations. Quite
some time later, Streiff-Becker (1933) came up with the so-called vertical aspiration
4 Understanding and Forecasting Alpine Foehn 231

Fig. 4.11 Schematic representation of the vertical aspiration theory after Streiff-Becker (1933)

Fig. 4.12 Schematic representation of horizontal aspiration theory after Ficker (1931). Solid lines
represent isentropes before, dashed lines isentropes after the aspiration by the low

Fig. 4.13 Schematic representation of lee waves. Solid lines represent isentropes
232 H. Richner and P. Hächler

Fig. 4.14 Solenoid theory according to Frey (1944): effect of the rotational term due to the non-
coincidence of temperature and pressure surfaces

Fig. 4.15 Schematic representation of the waterfall theory according to Rossmann (1950)

Fig. 4.16 Schematic depiction of the hydraulic jump theory according to Schweitzer (1953)
4 Understanding and Forecasting Alpine Foehn 233

theory (Fig. 4.11), an explanation which builds on the ideas of Wild, that potentially
warmer air moves over colder air. After passing the ridge, it successively removes
the stagnant, potentially colder air by turbulent erosion. By this process, the
potentially warmer air from above sinks gradually into the valleys, replacing the
colder air masses.
In 1912, Ficker published the so-called horizontal aspiration theory or passive
replacement flow theory (Fig. 4.12), a concept that he slightly modified again in
1931 and 1943. It was based on an earlier hypothesis by Billwiller (1899): An
approaching low over the Atlantic sucks the near-ground air masses away, causing
the potentially warmer air above to descend. The topography prevents the advection
of air from the sides. Ficker assumed that air would flow out of the valleys just
like water would flow out of a basin. A south–north oriented pressure gradient
would cause an (ageostrophic!) near-surface flow. Hence, foehn is nothing but
a replacement flow for the air lost in the valleys due to the depression. As a
consequence, all places at a given elevation would simultaneously be subjected to
the adiabatically warmed foehn air that represents this southerly replacement flow.
Observations showed that flow over mountains caused the formation of waves
that extend well beyond the obstacle (see Sect. 3.2.2 by Jackson et al.). Lyra (1940)
and Queney (1948) were the first to present a theoretical analysis of waves in the lee
of a mountain massif. Given the streamlines that emerged from their computations,
the sinking of air masses in the lee could be regarded as a forced downslope motion
(Fig. 4.13). In their often-referenced paper, Klemp and Lilly (1975) provide a
detailed analysis of a strong downslope wind induced by lee waves such as can
occur with foehn. They state that the observed amplification can occur only “if the
upstream wind and stability profiles lie within sharply limited but plausible ranges.”
Hence, depending on temperature, humidity, and wind profile, such a mechanism
may be responsible for the downslope flow in certain cases, but will not in
general.
Based on observations, Frey (1944) developed the so-called solenoid theory
(Fig. 4.14). Measurements showed that temperature in the lee of the Alps increases
towards the north while, at the same time, the pressure falls. Hence, isothermal and
isobaric surfaces are not at the same angle, resulting in a rotational acceleration
(or solenoid). According to Frey’s theory, the kinetic energy of the moving foehn
is not only determined by the pressure gradient, but also to a substantial degree to
the strength of the rotational term; the kinetic energy is sufficient to remove the cold
pool by a plain mechanically forced replacement.
Rossmann (1950) and Schüepp (1952) saw the cause for the descent of the flow
in the D> foehn wall, i.e., in the clouds formed above the crest of the mountain
ridge (Fig. 4.15). They assumed that the air in the cloud is colder and, therefore,
denser than the air outside the cloud. This would result in a downward acceleration,
the kinetic energy to be sufficient to penetrate the lee side valleys. This explanation
became known under the descriptive name waterfall theory.
Using fluid dynamics theory, Schweitzer (1953) proposed the phenomenon of the
“stationary hydraulic jump” as an explanation for the descent of the foehn air after
234 H. Richner and P. Hächler

passing the ridge (Fig. 4.16; see also Sect. 3.2.3 by Jackson et al.). This phenomenon
is observed (and theoretically well understood) on rivers and dams when the water
is discharged at high speed into a region of significantly lower flow speed. The result
of this transition is an abrupt rise of the slower moving fluid.
Obviously, there is no generally applicable theory. Meteorologists have to live
with the fact that various physical mechanisms are needed to explain the same
phenomenon, but occurring under different conditions. It seems that the lee wave
theory is rarely a reasonable explanation, while that hydraulic jump theory is often
a quite plausible description of the observations. Brinkmann (1971) and Hoinka
(1985a) discuss in detail how large amplitude lee waves and/or hydraulic jumps
accelerate foehn flow in valleys perpendicular to the ridge. The solenoid theory
explains quite well thermally driven wind fields such as mountain-valley wind
systems. Whether it is an explanation in the dynamically driven foehn flow is not
clear, here the rotational term could just as well be the result rather than the cause.
The waterfall theory might be useful to explain local effects, however, it can hardly
be regarded as general mesoscale phenomenon controlling the lee side flow. (For
additional theoretical discussions related to these issues, see Chap. 3 Sect. 3.2 in
this book.)

4.4.2 Dimmerfoehn

Under certain conditions, precipitating clouds may be drawn over the ridge, far
into the lee-side valleys. Nevertheless, foehn is present here. As a consequence
of the significant cooling due to evaporation, the dewpoint depression is rather
small, typically 1–7 K instead of the 10–20 K observed normally. An extreme case
happened on November 16, 2002, when there was foehn in Altdorf with 95 mm of
precipitation.
This type of foehn is called D> dimmerfoehn (“dimmerig” or “dimmrig” [Swiss
German] meaning dim, obscure). The airstream aloft shows typically cyclonic shear,
which transports strong condensation from S to N. It was first described in the first
half of last century by Swiss researchers and prompted a fierce and not always
very scientific dispute between Austrian and Swiss meteorologists. The reservation
Austrians had against this foehn type may be explained by the fact that dimmerfoehn
is very rare; there are years without any such case, and it seems that this foehn type
occurs even more rarely in Austria.
Sometimes, the expression “dimmerfoehn” is also used for situations with
southerly flow and strong haze, often caused by advected Saharan dust. Such a case
occurred on April 27, 1993; it is well documented in Burri et al. (1999). (It must be
noted that the definitions found for dimmerfoehn in different publications are not
entirely consistent! The explanation given above is based on the original description
by the Swiss meteorologist Streiff-Becker (1947), it was somewhat modified by
subsequent, more accurate observations.)
4 Understanding and Forecasting Alpine Foehn 235

4.4.3 Interaction with Cold Pool

When foehn winds descend in the lee of a mountain ridge, they clash with much
cooler, stagnant air. If topography allows, this cold pool is usually flushed away,
typically within hours. However, there are many cases where topography prevents
the outflow of the cold pool; the Wipp and Inn Valley in Austria (with the Nordkette
as downstream obstacle), and the lower Reuss Valley in Switzerland (with the Jura
Mountains as obstacles) are prominent examples for this constellation. In addition,
mere bends in the main valley axis can prevent the cold pool from being flushed out
rapidly.
On the Swiss Plateau, i.e., the area between the Alps in the south and the Jura
Mountains in the north, often a cold pool of a few hundred meters depth remains
while foehn flows over it. In winter, such a situation can persist for up to several
days. The cold pool has a wedge-like form, the angle is typically 2ı . Of course,
air in the cold pool remains calm, while the line where the foehn flow leaves the
ground to flow over the cold air is very gusty. Over water, this border is easily seen
by a spray line and by the haze usually present in the cold air.
As the strength of foehn flow increases, it works its way downstream, gradually
forcing the cold pool back. This can occur by three possible mechanisms or any
combination thereof: (a) heating by convection, (b) erosion of the top by mixing
and entrainment, and (c) static and dynamic displacement.
During MAP, an attempt was made to directly measure the heat flux near the
internal boundary between foehn and cold air by a small aircraft. The daily mean
found was about 15 W m2 , almost exactly the same value that was measured at the
ground. Hence, heating by convection and mixing at the top of the cold pool seem
to be of similar importance for cold pool removal.

4.4.4 Waves

Figure 4.17 depicts the two types of waves which commonly occur with foehn,
namely lee waves at about crest height and above, and shear-induced gravity waves
on top of the cold pool.
Lee waves or simply mountain waves are stationary, hence they belong to the
group of standing waves. Their wavelength is – depending on the width and shape of
the mountain massif – anywhere between 5 and 10 km. Also the amplitudes depend
on the geometry; they are typically of the order of 100–200 m. The amplitudes
are largest at some intermediate level just above ridge height. (See also Chap. 3,
Sect. 3.3.)
In Alpine D> south foehn, the air has passed over the Mediterranean Sea and
picked up substantial moisture. Consequently, when the air reaches saturation by
adiabatic expansion at the wave crests, lens-shaped clouds (cumulus lenticularis or
altocumulus lenticularis, see Fig. 4.18, D> foehn cloud) may be formed. Naturally,
236 H. Richner and P. Hächler

Fig. 4.17 The two wave types occurring with foehn: (a) above crest height the mountain lee waves
(sometimes also called D> foehn waves) and (b) on top of the cold pool shear-induced gravity
waves

Fig. 4.18 Lenticularis clouds over the Monte Rosa Massif near Zermatt in the Swiss Alps
(Reproduced by permission of Andreas Fuchs)

these clouds are also quasi-stationary. Glider pilots were the first to describe lee
waves; this phenomenon enables them to reach very high altitudes by cleverly
utilizing the strong updrafts. However, lee waves can also cause severe turbulence,
particularly when they break or when rotors are formed. Numerous accidents with
small aircraft, commercial aircraft, balloons, paragliders, etc. were caused by lee
waves or by turbulence associated with them.
For a more thorough discussion of lee waves and rotors, see the reports of a recent
study in the Sierra Nevada (Grubišić et al. 2008; Grubišić and Billings 2008), or the
Pennines (Sheridan et al. 2007); the interaction between foehn-type wind and large
amplitude lee waves is treated in detail by Beran (1967). Finally, a broad discussion
is found in Chap. 3, Sects. 3.3 and 3.4.
4 Understanding and Forecasting Alpine Foehn 237

Fig. 4.19 Waves on cold pool in the Rhine Valley, looking SE. Foehn is from right to left, notice
the weak counter flow in the cold air mass (Reproduced by permission of Andreas Walker)

Fig. 4.20 Waves on the cold pool seen by a sodar. In this time-height diagram, the height covered
is 900 m, the total recording time about 2 h. Observations were made on October 1, 1976,
about 45 km north of Altdorf in the Reuss Valley (at the north end of Fig. 4.7 in the cold pool)
(Reproduced by permission of Werner Nater)

At the internal boundary between foehn flow and cold pool, there is a strong
shear as well as a large temperature gradient. Wind speed in the foehn air might
be typically 20 m s1 while it is virtually zero in the cold pool, the temperature
differences between the two air masses is typically 5–15 K. Shear and different
densities (due to temperature difference) cause gravity waves on top of the cold
pool by the same mechanism by which waves appear on a lake’s surface when the
wind blows over it. Waves on the cold pool can be seen when strong haze or fog is
found in the cold air mass (Fig. 4.19), however, they can be made visible indirectly,
e.g., with sodar, RASS or lidar (Fig. 4.20). These waves produce small fluctuations
in surface pressure (0.1–1 hPa). Depending on the vertical profile of temperature
and wind, their period is somewhat longer than the Brunt-Vaisala period (of the
238 H. Richner and P. Hächler

order of 10–20 min). Unlike lee waves, gravity waves propagate with a phase speed
of roughly half the shear. Obviously, period (measured at a fixed location), phase
velocity and wavelength are interdependent as described by the wave equation.

4.4.5 The Three-Dimensionality of Foehn

While it is the aim of any science to find common mechanisms, to formulate general
laws, describe systematic behavior etc., it must not be forgotten that in reality no two
foehn cases are exactly alike. The case studies presented here were chosen more
or less arbitrarily as “typical” and recent examples. Many disputes happened only
because scientist X adjusted the theory to a specific foehn case while scientist Y
did the same, but (a) in another topography und (b) for another foehn case. Hence,
generalization must always be made with great caution. Insight into processes can
be gained both by statistical analyses of many cases and by detailed case studies.
While systematic statistical analyses of a sufficiently large number of foehn cases is
lacking, there are numerous case studies, many of them very detailed.
In the above discussions, foehn mechanisms are treated in two dimensions only.
For a more precise analysis, the third dimension must, of course, be taken into
consideration, too. However, if this is done, generally valid statements can only be
made for the particular topography and the particular meteorological situation. It is
obvious that tributary valleys do play an important role, and it is to be expected that a
kink in the valley’s axis changes the flow pattern. But each synoptic weather pattern,
and each vertical profile has its distinct effect on the dynamics and thermodynamics
on the air masses involved.
Because foehn represents an airflow more or less perpendicular to a mountain
ridge, and because there are often gaps or passes in the ridges, there is a close
connection between foehn and D> gap flow. The dynamics of foehn-related gap
flow manifests itself in a particularly clear manner along the Brenner cross section,
this because (a) the topography in the ridge area is well suited and (b) because foehn
occurs here sometimes without clouds over the ridge and rarely with heavy upwind
precipitation. Further details on D> gap flow and foehn can be found in Gohm and
Mayr (2004), in Gohm et al. (2004), and in Chap. 3, Sect. 3.5 in this book.
There are several detailed case studies on the three-dimensionality of foehn,
each giving insight into characteristic properties of the case in question. Seibert
(1990) collected south foehn studies that were made since the Alpine Experiment
ALPEX in 1982 in the eastern part of the Alpine massif. Based on these studies
she concluded that the textbook theory is not correct. As discussed above, such a
statement goes too far.
Sprenger and Schär (2001) show how – under D> shallow foehn conditions –
the complex topography of the Alpine ridge can cause flow splitting. Similar studies
were carried out by Zängl (2002). Using observations made during the Mesoscale
Alpine Program MAP, Jaubert and Stein (2003) analyzed and modeled a foehn case
that showed wave breaking and hydraulic jumps. Drobinski et al. (2003) looked at
4 Understanding and Forecasting Alpine Foehn 239

the scale interactions by which smaller scale topography influences the development
and the characteristics of foehn. Across mountain passes, so-called gap winds occur
in foehn situations. Examples of observations and a discussion of their theory are
given in Chap. 3, Sect. 3.5.

4.5 Foehn Climatology

In Austria, a map showing foehn regions was produced. The intention was to advise
people suffering from discomfort supposedly caused by foehn (see Sect. 4.7.4).
This map (Fig. 4.21) is primarily based on the experience of forecasters and only
minimally on quantitative observations. As mentioned in Sect. 4.3.2, universal
criteria for foehn in a given area are almost impossible to define. For a case with
strong foehn, Gutermann (private communication) has produced a foehn map based
on surface temperature (Fig. 4.22). While this map has the limitation of being a case
study, it does show clearly where the foehn regions are.
Statistical analyses of data from dedicated foehn stations show that foehn
occurrence is highly variable, and that no trend is discernible in the last 140 years
(Fig. 4.23; Richner and Gutermann 2007). In addition, the regional and seasonal
variability is considerable. Figures 4.24 and 4.25 give some ideas about variability.
Observations are made three times a day, morning, noon, and evening; the numbers
refer to the number of such observations.

Fig. 4.21 Foehn regions in Austria and Bavaria. The different shadings represent the frequency
of foehn (arbitrary scale) (Reproduced by permission of the Central Institute for Meteorology and
Geodynamics [ZAMG], Vienna, Austria)
240 H. Richner and P. Hächler

Fig. 4.22 Case study showing the main areas in Switzerland where foehn reaches the surface
(Reproduced by permission of Thomas Gutermann, data kindly provided by MeteoSwiss)

Fig. 4.23 Year-to-year variation of foehn occurrence at the station Altdorf. The vertical axis
represents the sum of daily observations in the morning, at noon, and in the evening, the heavy line
represents the 20-year running mean (Richner and Gutermann 2007, basic data kindly provided by
MeteoSwiss)

Seasonal variations in foehn frequency (upper frame of Fig. 4.25) are caused by
the changing general circulation pattern and, subsequently, by the resulting synoptic
situations. The changes in the diurnal distribution of foehn activity (lower frame of
Fig. 4.25) are the result of interactions of foehn flow with thermally driven local
wind systems, i.e., mountain-valley winds. Note that the relative frequency of foehn
at noon remains the same throughout the year.
4 Understanding and Forecasting Alpine Foehn 241

Fig. 4.24 Comparison of the yearly foehn occurrence at different Swiss stations. Values represent
the mean yearly sum for the period 1973–1982. The value for Altdorf (ALT) for this period is 62.2
(Richner and Gutermann 2007, basic data kindly provided by MeteoSwiss)

4.6 Forecasting Problems

Foehn forecasting rests on three pillars: (a) probabilistic methods based on a


few observed or forecasted parameters, (b) on model output, and (c) (still very
important!) on the skill of experienced forecasters.

4.6.1 Probabilistic Methods

Forecasting foehn means predicting a mesoscale phenomenon based on a synoptic


situation. Such forecasts are primarily based on the pressure field at different
altitudes. Depending on the orientation of the isobars, foehn might develop in
one valley and not in another, hence, different locations would require specific
forecasting procedures. In the 1960s, Widmer developed for the foehn station
Altdorf, Switzerland, a “foehn test” that was refined by Courvoisier and Gutermann
(1971); it remains the operational tool until today. Two pressure gradients across
the Alps plus the trend of one of these are used to compute an index. If the index
is below a certain, season-dependent threshold value, the probability of foehn at
242 H. Richner and P. Hächler

Fig. 4.25 Seasonal variation of total foehn occurrences per month (upper frame) and relative
frequencies of occurrence at the three observing times (lower frame) for the foehn station Altdorf.
Yellow: morning, red: noon, blue: evening observations (Richner and Gutermann 2007, basic data
kindly provided by MeteoSwiss)

Altdorf within the next 36 h is over 70%. The index allows also predicting the
breakdown of foehn with even a slightly higher success rate. Figures 4.26 and 4.27
give an example of the Widmer Index and its potential for forecasting a foehn case
in January 2008. Note that foehn breakdown is predicted when the Widmer index
has passed though its maximum and starts to decrease again, hence, the prediction
of the breakdown was not very good in this example.
Dürr (2008) developed an automated method for identifying, i.e., nowcasting
foehn. His procedure is based on 10-min real-time data from the automated Swiss
surface network. The most important predictors are the differences in potential
temperature between the reference station Gütsch (2,282 mASL, close to the Alpine
ridge) and the locations for which foehn should be nowcasted. For the time being,
4 Understanding and Forecasting Alpine Foehn 243

Fig. 4.26 Example of Widmer Index for January 2008. The indices are computed from pressure
differences across the Alps as forecasted for defined gridpoints. Different lines relate to different
forecasts and pressure levels, or combinations thereof. The horizontal line represents the winter
threshold; a value above this indicates foehn at Altdorf (note inverted vertical scale!). Data used
here are the ECMWF and COSMO-7 forecasts of January 1 and 2; ECMWF forecasts for 10
days (dashed and dotted lines), COSMO-7 for 3 days (black and orange lines). COSMO-7 is
the MeteoSwiss mesoscale forecasting model for the Alpine region with 6.6 km resolution (Data
kindly provided by MeteoSwiss)

the application of this technique is restricted to locations on valley floors or near


valley exits; since July 2008, the procedure is implemented as an automated routine
diagnostic tool at MeteoSwiss.
Quite recently, Drechsel and Mayr (2008) developed an objective, probabilistic
forecasting method for foehn in the Wipp Valley (Innsbruck) based on ECMWF
model output. As predictors, they use cross-barrier pressure differences and,
additionally, the isentropic descent. A test over 3 years proved that – based on the
two variables – reliable forecasts of up to 3 days for foehn and the associated gust
winds can be made.

4.6.2 Model Forecasts

As the resolution of models is improved, the topography of the complex Alpine ter-
rain (and with this the foehn valleys) is more accurately represented. Consequently,
244 H. Richner and P. Hächler

Fig. 4.27 Comparison of observed (via humidity) and forecast (using the Widmer index) onset of
foehn; zooming in on Fig. 4.26; January 2–5, 2008 (Data kindly provided by MeteoSwiss)

Fig. 4.28 Wind speed and temperature for the foehn stations Altenrhein (ARH, red) and Vaduz
(VAD, blue). The time series with 10-min resolution are observations (fine lines), the series with
1-h resolution (heavier lines) are model data (Hächler et al. 2011, reproduced by permission of
Klaus Burri)

there are well-founded hopes that models will provide sufficiently accurate foehn
predictions.
At MeteoSwiss, COSMO-2, a 2.2 km grid size model, has been run operationally
since February 2008. Figure 4.28 assesses its capability to analyze a foehn case
4 Understanding and Forecasting Alpine Foehn 245

(a test case from the pre-operational phase, Burri et al. 2007). For both stations,
the analysis represents the onset of foehn too early. The increase in wind speed
is significantly below the observations, the same is even more pronounced for
temperature.
In summary it can be said that COSMO-2 forecast fields and analysis fields
overestimate the spatial foehn extension and mostly underestimate temperature and
wind speed not only at the two stations investigated here, but at most locations.
The modeled temperature gradient between valley stations and the Alpine ridge
site Gütsch (not shown here) never reached dry adiabatic conditions, this in
contradiction to the observations. In an attempt to improve the situation, not only
deterministic, but also statistical methods will be used the future. In particular,
ongoing work shows that model output statistics (MOS) are a promising tool for
improving foehn forecasting.

4.6.3 Open Problems

In practice, the skills of experienced forecasters who are familiar with the local
situations are still an indispensable prerequisite for a successful forecast. They
know from experience how a somewhat different wind direction might influence
the onset or breakdown of foehn in a given valley, how observed wind data must
be interpreted to arrive at a correct prediction. On the other hand, any tool, be it
based on probability or on model output, is a welcome and appreciated support
giving a first approximation which is subsequently modulated with the forecaster’s
experience and skill.
The synoptic chart and particularly the surface pressure pattern is one of the
most useful tools for predicting the onset of foehn. Here, the forecaster will focus
primarily on the pressure pattern and the formation of the so-called foehn knee
or foehn nose, the characteristic deformation of the isobars on the upwind side
of the Alps (Fig. 4.4). It must be stressed, however, that the formation of the
foehn knee alone is an insufficient indicator for foehn since this is purely a surface
characteristic and does not say anything about the situation aloft. In this respect, the
pressure gradient at ridge height (e.g. from 850 hPa maps) is a much more useful
parameter.
The breaking down of foehn is in most cases associated with an incoming
coldfront from the west. In these cases, forecasting the breakdown can be achieved
quite reliably by observing the progressive increase in surface pressure. If, however,
a high moves in from the east, predicting the breakdown is rather difficult, there
are no clear rules for this case. Nevertheless, it seems that further improvements in
models – both in resolution and parameterization – will improve future forecasting
of foehn.
After the Mesoscale Alpine Program MAP, a few open problems became
obvious. Among these is, as indicated, the role of the tributary valleys which is
246 H. Richner and P. Hächler

still not well understood. This is partly due to a lack of sufficient high-density
observations, but also to a poor understanding of the interaction of air masses
coming from different valleys. There are many locations where two or even more
foehn valleys merge. Observations indicate that the flow does not necessarily merge,
but that one foehn “stream” might cross over the other. Which flow will go over the
other depends on the synoptic fields and the orography. The higher the elevation
at which the winds cross the Alps, the higher are the resulting temperatures in the
valleys due to dry stable structure of the atmosphere in the south. Thus, the warmer
foehn streak (which crossed the ridge at greater height) flows over the colder one.
This effect can be observed, e.g., in the area of the lake Walensee where airflow
down the Rhine Valley (directly towards the lake of Zurich) interacts with airflow
from the Linth Valley in the Glarus Alps. Also here, refined and higher resolving
models might soon provide better solutions.

4.7 Societal Impact of Foehn

4.7.1 Climate Impact

Foehn has serious impacts on the local climate that, in turn, influence agricultural
possibilities in foehn valleys. Thanks to foehn winds, wine can be grown in areas
where it otherwise would be impossible. Foehn storms (also called downslope
windstorms) have also a major effect on the snowmelt, an effect, which is not
particularly liked in skiing resorts.

4.7.2 Air Quality

Foehn situations provide the most spectacular views of the mountains. As a


consequence of the precipitation on the upwind side, the foehn air is usually very
clean, there is no haze, and distant objects seem to be much closer (see Burri et
al. 1999). Likewise, the air mass originating at high levels in foehn type II has
much lower aerosol concentrations than the air it replaces in the valleys. The entire
fantastic Alpine panorama can be seen from places where one normally does not
see any mountains at all. On the negative side of these stunning postcard-views,
however, are the increases in ozone concentration. Although the values seldom
reach alarm levels, foehn air easily triples existing ozone concentrations (Baumann
et al. 2001). Trajectory calculations and aircraft measurements prove that the higher
concentrations are simply due to the descending of ozone-richer air from about
4,000 m over the Mediterranean region. Figure 4.29 depicts the situation for Altdorf
for the foehn case discussed above.
4 Understanding and Forecasting Alpine Foehn 247

Fig. 4.29 Ozone during foehn at Altdorf in January 2008 (All data taken from an air quality station
that is not collocated with the foehn station. Hence, temperature data is not exactly the same as in
Fig. 4.5.) (Data kindly provided by “inLuft”, Central Switzerland [www.in-luft.ch])

4.7.3 Fires and Traffic Accidents

The most striking danger from foehn events, however, was and still is the spreading
of fires. The warm and very dry air combined with high wind speed supports and
proliferates fires very efficiently. In the course of time, numerous towns have burned
down completely. In 1861, 600 houses of the canton capital Glarus, Switzerland,
were completely devastated during a foehn storm, and only recently, in 2001, a fire
maintained by foehn winds in excess of 15 m s1 destroyed 15 houses in Balzers,
Principality of Liechtenstein. During foehn situations, a few towns still activate a
fire watch during nights, and smoking outside houses is strictly forbidden. In some
mountain regions, it is – as a matter of principle – illegal to start fires outside
specially designated fire areas.
Foehn winds can be dangerous to flying. Professional pilots and local airports
are well aware of the problems and issue the necessary warnings. At Innsbruck
airport, e.g., special foehn procedures ensure that the most turbulence prone parts
of the valley are avoided during climb out and approach in order to enhance safety
and passenger comfort. However, when hot-air ballooning, paragliding, and similar
sports became popular during the last third of last century, the number of severe
accidents due to high winds and large shears increased significantly. Improved
training, special safety courses, and specific information and forecasts have reduced,
but not eliminated, this problem.
248 H. Richner and P. Hächler

Cable car accidents and even train accidents can be caused by strong and gusty
foehn winds. Although all cable transportation systems are required to monitor wind
speed and to have an alarm system, gusts and shears that slip between the different
anemometers can surprise operators. Sadly, there was such an accident in January
2008 in the Jungfrau region. High winds during a foehn storm “derailed” the cable
of a double chair cable lift. First, the lift stopped, and the cable was caught in the
cable catcher, but a successive gust threw the cable out of the catcher, and the chairs
dropped to the ground. One person died and several were severely injured; mean
wind speed was about 25 m s1 which is below the alarm level of 28 m s1 . (An
almost identical incident happened 2003 in Wangs-Pizol, Switzerland. That time the
directly affected gondolas were empty and nobody was injured. The cause was again
gust winds, this time associated with a thunderstorm front.)
On February 15, 1925, a train was thrown from its track in a foehn storm
in Strobl near Salzburg, Austria. Another spectacular accident happened during a
severe foehn storm in the Jungfrau region on November 11, 1996: a double motor
coach of a narrow gauge railway was blown from its track. Its mass of 43 t could
not withstand a foehn gust of 52 m s1 ; four persons were injured, fortunately none
seriously.
On lakes, foehn storms hamper scheduled boat traffic; in extreme cases, opera-
tions have to be suspended. The Swiss town of Brunnen (which is directly in the
main axis of the Reuss Valley) has built a special “foehn harbor.” It serves two
purposes: (a) it is used by boats in storm situations as shelter, and (b) if wind and
waves still allow the scheduled ships to navigate, they can dock in this harbor, which
is better protected against the waves and gusts than the wharf in the center of the
town.
The most significant recurring damage caused by foehn winds is most probably
that to boats, piers, and shores. After severe foehn storms, pictures of loose-torn
boats lying on shores or damaged piers appear regularly in the news. The previously
mentioned “century foehn” on November 8, 1982, destroyed not only numerous
boats and yachts, but also the newly built pier of the town of Sisikon (Lake Uri).

4.7.4 Biometeorological Effects

Still much debated are biometeorological effects of foehn winds. Interestingly, it is


primarily in the Alpine area that people blame foehn winds for almost any ailment,
accident, crime, and in particular for headaches. Sferics (electromagnetic radiation
originating in the atmosphere), ion concentrations, and pressure fluctuations were
considered as possible causes of foehn-related ailments. While recent measurements
prove that neither sferics nor ion concentrations are correlated with foehn events,
pressure fluctuations induced by gravity waves on the cold pool remain a possible
link between foehn and man.
4 Understanding and Forecasting Alpine Foehn 249

Statistical analyses of pressure fluctuations and subjective well-being show that


there is indeed a statistically significant correlation between the two. However, there
is no proof of any cause-and-effect mechanism. The study of short-term pressure
fluctuations did not take into account the actual weather situation. Since fronts also
cause high amplitude pressure fluctuations, the positive correlation might simply
reflect that people feel subjectively better when the weather is good. A direct
analysis of the frequency of headaches and the occurrence of foehn (as defined in
the Alpine weather statistics) did not produce any result.
Interestingly, on the American continent the interest in biometeorological re-
search seems to be picking up after a long phase of disinterest. There have
been several research projects dealing with Chinook winds and headache, strokes,
etc. However, so far no significant progress was made in relating ailments to
foehn-type winds or weather in general (see, e.g., Cooke et al. 2000; Field and
Hill 2002).

Acknowldgments We thank our colleagues in the “Alpine Research Group Foehn Rhine Val-
ley/Lake Constance” for their contributions and support. In addition, we acknowledge valuable
suggestions from three anonymous reviewers for improvements of our text.

Appendix A: Explanation and Definition of Foehn-Related


Terms

Alpine Foothills Foehn


A warm but humid wind on the Alpine foothills, in the lee the air is mostly calm.
This is not a foehn in the meteorological sense.
Anticyclonic Foehn
Foehn in an anticyclonic situation with a high potential height of the 500 hPa
level. Only partly cloudy in the south, and usually dry. Temperature gain in
the foehn valleys due to adiabatic sinking, flow forced by southerly pressure
gradient.
Cyclonic Foehn
Foehn in connection with a cyclonic regime, i.e. low potential height of the
500 hPa level. Causes cloudiness even in the classical foehn areas, but criteria for
foehn are well pronounced. Develops in extreme cases to dimmerfoehn.
Deep Foehn, High-reaching Foehn
The classical foehn situation with a high-reaching SW flow driven by a high over
northern Italy and a low over southern Germany; isobars show the typical foehn
knee; lee waves, lenticularis clouds, foehn wall, often also lee side rotors are present;
warm, dry winds with high speed reach the valley floors (Hoinka 1990).
Dimmerfoehn
A south foehn which does not immediately follow the topography in the lee, but
touches the surface further downwind. The mountain ridge is in clouds that extend
250 H. Richner and P. Hächler

downwind. The comparatively calm area right downwind of the ridge is dark due to
heavy clouds, hence the name (“dimmerig” or “dimmrig” [Swiss German] means
dim, obscure). In rare cases, there is no precipitation but it is very hazy due to
Saharan dust.
Double Foehn
The simultaneous existence of north and south foehn. Double foehn is a short-
lived, very rare phenomenon which occurs when a cold high-pressure system in the
lower troposphere moves quickly over the Alpine ridge while a stormy westerly flow
is present a greater height (Frey 1986).
Flat Foehn D> Shallow Foehn
Foehn Air
The warm and dry air produced by foehn.
Foehn Bank D> Foehn Wall
Foehn Break D> Foehn Window
Foehn Clearance
At distances larger than about 50 km downwind from the Alpine ridge, foehn
winds become light, while they remain warm and dry. This leads to a significant
reduction of cloudiness, i.e., to a clearing (Hoinka 1985b).
Foehn Cloud
Lenticularis clouds in the lee of the mountain barrier. They are caused by the
lee waves associated with the foehn flow. Their orientation is parallel to the ridge,
and because the lee waves are stationary, also the foehn clouds are quasi-stationary.
Note: Lenticularis clouds are neither a prerequisite for, nor proof of foehn, they can
occur with any airflow over a barrier. The term “foehn cloud” is not be confused
with the D> foehn wall.
Foehn Cyclone
A cyclone in the lee of a mountain massif which is formed or enhanced by a
foehn process.
Foehn Gap D> Foehn Window
Foehn Island
The surface area where foehn touched down while the surrounding area is still
covered with cold air.
Foehn Knee, Foehn Nose, Foehn Wedge
The typical deformation of the isobars (or isohypses) in a foehn situation. A high-
pressure wedge lies on the windward side of the mountain range, while in the lee
a trough is formed. As a consequence, the isobars (or isohypses) bulge and take a
nose-like or knee-like form. In the German literature, “foehn knee” (“Föhnknie”) is
more commonly used, in the English literature the term “foehn nose” (Föhnnase) is
normally found.
4 Understanding and Forecasting Alpine Foehn 251

Foehn Nose D> Foehn Knee


Foehn Pause
Foehn winds do not blow constantly, but there is varying intensity and sometimes
cessations. Such an interruption is called foehn pause (the German “Pause” means
break, pause, intermission). Foehn pauses occur predominantly during the early
morning hours before sunrise when cold air masses intrude.
Rarely, the term foehn pause is used for the region delimiting the foehn air from
the ambient air, this analogous to “tropopause”, “stratopause”, etc.
Foehn Period
This term deals with the duration or simply with the occurrence of foehn at one
or several stations. The use of the term by different research groups is very diverse,
hence, there is no generally accepted definition.
Foehn Phase
A rather general term used by Alpine foehn researchers for classifying the
different development stages of a foehn situation; the definition of the phases might
differ among different authors. The systems used by Ficker and by Streiff-Becker
(1933) differentiated three phase, the recent system introduced by Seibert (1990)
has four:
Phase I: cold air masses on both sides of the Alps
Phase II: warm advection from W or SW, development of temperature gradient and,
as consequence, pressure gradient across Alpine ridge, D> shallow foehn develops
Phase III: approaching trough changes flow to SW or S, speed and humidity in-
crease, pressure gradient reaches maximum, D> deep foehn develops, precipitation
in the upwind region
Phase IV: passing of cold front, breakdown of foehn, often formation of a cyclone
in the N.
Foehn Storm
A potentially destructive storm as result of a foehn situation. A foehn storm
is characterized by its sudden occurrence in a practically calm situation and its
intermittent nature during its onset. After the initial phase, wind speed remains high
but steady. Additionally, the fast moving air is, as a result of the foehn mechanism,
significantly warmer than the displaced air. Foehn storms belong into the category
of downslope windstorms where they are often referred to as warm downslope
windstorms.
Foehn Tongue
The area covered by foehn air flowing out of foehn valleys.
Foehn Trough
The trough in the lee of the mountain range formed dynamically by the flow
across the barrier.
Foehn Wall, foehn bank
In those situations where foehn occurs with precipitation on the windward side
of the mountain range, the clouds can be seen from the lee side as a “wall” topping
252 H. Richner and P. Hächler

the mountain ridge. Depending on the characteristics of the foehn flow, the clouds
can be dragged over the ridge where they are dissolved while sinking. Under these
conditions, the foehn wall has the appearance of a cap.
Foehn Waves
The lee waves or mountain waves associated with a foehn situation. Sometimes
foehn waves become visible because they produce “foehn clouds”, i.e., cumulus
lenticularis. Foehn waves must not be confused with the gravity waves occurring on
the top of the cold pool.
Foehn Wedge D> Foehn Knee
Foehn Window, Foehn Gap, Foehn Break
A cloud-free gap or clearance in the lee of the mountain range over which a foehn
flow occurs. This gap is caused by the descending foehn air that is dry-adiabatically
heated and, consequently, becomes dry.
Foehnic Situation, Foehn-Like Situation
Southerly flow which does not penetrate into the valleys. Foehn wall, foehn
window, and lenticularis clouds may be present, visibility is very high.
Free Foehn
The sinking of air masses in the free atmosphere on a synoptic scale, particularly
in high-pressure areas. The vertical motion is often brought to a standstill by an
inversion that sometimes tops a fog layer. The increase in temperature and decrease
in humidity caused by the dry-adiabatic sinking leads to dissipation of clouds. This
process has nothing to do with mountain winds, and the use of the term “foehn” in
this context is strongly discouraged.
Gap Foehn (Gap Flow) D> Pass Foehn
High-Reaching Foehn D> Deep Foehn
Light Foehn
Corresponds to phase II of the Innsbruck foehn strength scale (D> foehn phase).
Pass Foehn, Gap Foehn
A pass foehn exists when the foehn criteria are met at least at one mountain or
pass station, but not at a valley station. The distinction between pass foehn and D>
valley foehn was introduced to characterize also foehn situations in which the flow
does not penetrate into the valleys. Another name for a D> foehnic situation.
Sandwich Foehn
A complex and short-lived three-layer flow situation along the Brenner cross
section, first described during the Mesoscale Alpine Program MAP (Vergeiner and
Mayr 2000): a shallow pressure-driven south foehn event below a decoupled flow
from W or even NW, below the foehn flow a very stable cold pool which cannot be
removed by the southerly flow. (In other areas, e.g., over the Swiss Plateau, such a
situation can persist for up to several days; here the term is not used.)
4 Understanding and Forecasting Alpine Foehn 253

Shallow Foehn, Flat Foehn


A foehn with all its characteristics in the downwind side valley but without the
dominant southerly flow at high altitude, here flow can be from W or even NW.
Shallow foehn can be regarded as compensation flow between different air masses
on both sides of the mountain massif; it is often found in the lee of passes which
form gaps in the mountain barrier (Seibert 1990).
Strong Foehn
Corresponds to phase III of the Innsbruck foehn strength scale (D> foehn phase).
Talfoehn D> Valley Foehn
Valley Foehn, Talfoehn
A valley foehn is present when the criteria for foehn are fulfilled at least for one
mountain station and at least for one valley station. Used in this sense, it is not a
foehn type but a foehn phase. See also pass foehn.
Also, the term “valley foehn” is a somewhat obsolete name for deep foehn used
by Streiff-Becker. He used it to distinguish between deep foehn and Alpine foothills
foehn (the latter not being a foehn in a meteorological sense).

Appendix B: Names of Foehn-Type Winds in Different Regions

The German term Föhn originated most probably during Roman times. When the
Romans conquered the territories north of the Alps in the first century BC, they
found this stormy, warm, and dry mountain wind. From home, they knew a similar
wind, the “favonius” (meaning “the favorable”), a warm and dry wind originating in
northern Africa. Albeit the two winds are meteorologically very different, they do
have similar characteristics; hence, it was logical that the Romans gave this name
also to the warm wind they found north of the Alps. In the Raeto-Roman language
(still spoken today as Rumansch by a minority in the Swiss Alps) “favonius” became
“favuogn”, in its dialect “fuogn”. In the Old High German language “fuogn” became
“phōnno” which over time gradually developed to the German name “Föhn” (the öh
spoken as in b-i-rd or l-ea-rn).
As mentioned in Sect. 4.1, foehn winds in the meteorological sense occur
wherever large mountain massifs exist. In meteorology, the general term “foehn”
for the warm dry wind was chosen, because research started in the Alps, hence, the
local name became also the scientific label. The following list is based on Schamp
(1964), on information collected through personal contacts by Gubser (2006), and
on internet searches.
Aperwind (Alps)
Foehn in Spring which melts the snow (apertus [lat.]: open; apern [ger.]: melting
of snow).
254 H. Richner and P. Hächler

Aspre (Western Slope of Central Massif, France)


Dry, warm easterly winds blowing from the Central Massif over the Garonne
plain, also called “Lou Cantalié”.
Appenzeller Föhn (Eastern Switzerland)
Regional foehn in the foothills of the Alps around Appenzell, St. Gallen, and
Lake Constance. No southerly wind component at the Alpine ridge, but locally
favorable foehn conditions. Occurs usually with westerly winds and falling pressure
north of the Alps.
Austru (Walachia, Romania)
Foehn downwind of the Balkan Massif and the Transylvanian Alps from W
to SW; occurs predominantly in winter, the associated clearing leading to severe
radiation frost.
Autan (Southern France)
The Foehn in the lee of Cevenne Mountains and the Pyrenees from SE to E.
Autan Blanc (Southern France)
Particularly strong D> Autan with cloudless sky; can reach the Atlantic coast
south of the Gironde.
Autan Noir (Southern France)
An D> Autan followed by rain steered by a low over the Bay of Biscay.
Berg Wind (South Africa)
Hot and dry foehn-like wind that blows from the highlands down to the coastal
plateau; particularly Southwest Africa.
Bohorok, Also Bokorot (Eastern Sumatra)
Dry wind blowing from the Karo Plateau down to the plains of northeast Sumatra,
a passat wind warmed by the foehn mechanism.
Boulder Windstorm (Boulder, U.S.A.)
A windstorm occurring basically with D> Chinook in Boulder, CO.
Bregenzer Fallwind (Bay of Bregenz, Austria)
An easterly to northeasterly foehn-type wind blowing form the Gebhardsberg and
Pfänder, also called “Ostföhn” (east foehn), “Falscher Föhn” (wrong foehn, because
it comes from the unusual direction), or “Pfänderwind”.
Broebroe, Brubru (Sulwesi, Indonesia)
A gusty foehn-type easterly wind, part of the east monsoon near Makassar on the
west coast of the island.
Canterbury Northwester D> Northwester
Cape Doctor (Capetown, South Africa)
Maritime air masses, heated by a foehn-type process but still cold, flowing from
False Bay over Capetown towards Table Bay.
4 Understanding and Forecasting Alpine Foehn 255

Chanduy (Mexico)
Foehn-type wind in the area of Guayaquil, Mexico; occurs primarily during the
dry season in the afternoon.
Chinook (Rocky Mountains, U.S.A.)
Warm and dry foehn-type west wind on the eastern slope of the Rocky Moun-
tains. (In coastal regions of Oregon and Washington also a warm, humid sea breeze
from SW passing over coastal areas, has nothing to do with foehn.) The American
analogue to Alpine foehn, for details of the mechanism see Brinkmann (1973).
Chinook Arch
A striking feature of the D> Chinook, a band of stationary stratus clouds caused
by air undulations over the mountains due to orographic lifting.
Falscher Föhn D> Bregenzer Fallwind
Favogn, Favuogn (Grisons, Switzerland)
Romansh name for foehn.
Gending (Java)
Foehn winds over the northerly plains of Java during the SE monsoons.
Ghilbi (Libya)
Basically a hot desert wind often connected with sand storms. However, when
passing over coastal mountain ranges, there is an additional foehn effect.
Great Wind (Inner Asia)
Foehn-type NE wind on the SW slopes of the Alatau.
Guggifoehn (Bernese Oberland, Switzerland)
The Alpine foehn in the region of the Lauberhorn mountain.
Halny Wiatr (Poland)
A south foehn in the High Tatra.
Helm Wind (Northern England)
A strong foehn-type wind from the Cross Fell, characterized by cloud caps on
the mountain peaks, hence the name.
Himmelsbesen D> Sky Sweeper
Ibe (Western China)
Strong wind through a gap, the Dzungarian Gate (separating the depressions of
Lake Balkhash [Kazakhstan] and Lake Ala-Kul [Kyrgyzstan]). The wind is a gap
wind similar to foehn.
Jauk (Klagenfurt Basin, Austria)
South foehn over the Karawanken mountains.
Kâchan (Sri Lanka)
Foehn-type wind connected with the SW monsoon in the easterly parts of Sri
Lanka. The Tamil speaking population calls it “solaha-kâchan”, i.e., dry, burning
monsoon.
256 H. Richner and P. Hächler

Kumbang, Koembang (Java)


SE wind near Tijriban and Tegal on the north coast, the foehn effect is due to the
Pembarisan mountains.
Lenzbote (Alpine countries)
Name of the spring foehn melting the snow (“Lenzbote” [ger.]: spring messen-
ger).
Levanto (Canary Islands)
Foehn-type E wind particularly in the Orotava valley on Teneriffa.
Livas (Greece)
Any foehn-type wind. The name “Livas” (plural: “Lives”) is derived from “Lips”,
a warm SW wind from the direction of Libya (which is a warm wind but not a foehn-
type wind).
Ljuka (Carinthia, Austria)
Local name of foehn, most likely derived from the Slovenian word “jug” meaning
south wind.
Llebetg, Llebejado (Roussillon and Eastern Pyrenees, France)
Arabic or Catalan name of a foehn-type SW wind on the northern slopes of the
eastern Pyrenees.
Lou Cantalié D> Aspre
Megas (Greece)
A foehn-type SW wind blowing from Parnassus towards Boeotia, particularly in
the plain of the dried-up Kopais lake.
Mikuni-Oroshi (Japan)
The “downwind of Mikunitoge” is a foehn-type W to NW wind in the Tone valley
near Maebashi (eastern coast).
Moazagotl (Riesengebirge, Germany)
The name of a peculiar cloud showing the lee waves. It is formed in the prefrontal
stage of a foehn cyclone.
Canterbury Northwester D> Northwester
Montana Monsoon (Montana, U.S.A.)
Popular name for the D> Chinook in the prairies of Montana.
Northwester, Canterbury Northwester (New Zeeland)
A warm and dry northwestly wind that reaches Canterbury Plain on South Island.
Norder, Norther or Nortes (Gulf of Mexico, Central America)
Actually a cold stormy wind, however, when blowing from the southeast it
passes over the Isthmus of Tehuantepec, the blocking on the upwind side causes
precipitation while in the lee, on the side of the Pacific foehnic clearing occurs.
4 Understanding and Forecasting Alpine Foehn 257

Ostföhn D> Bregenzer Fallwind


Pacific Wind (Colorado, U.S.A.)
A name sometimes used for the Chinook.
Pfänderwind D> Bregenzer Fallwind
Puelche (Southern Chili)
Foehn wind in the Andes, in Argentina the same wind is called Zonda.
Rotenturm Wind, Talmac Wind (Transylvania, Romania)
A foehn-type south wind near Sibiu (Hermanstadt) blowing from the southern
Carpatians over the Rotenturm Pass into the Transylvanian Basin; sometimes also
called Talmac Wind after the town Nagy-Talmacs near Sibiu.
Santa Ana (California, U.S.A.)
Northeastern foehn wind in the basin of Los Angeles, named after the river and
the pass.
Sky Sweeper (Majorca, Spain)
Foehn-type NW wind over the coastal mountain ranges of Majorca, Spain.
Note that among seamen the term “Sky Sweeper” (or its German translation
“Himmelsbesen”) signifies any northeasterly wind in northerly latitudes. It causes a
very clear view after “sweeping the clouds from the sky.”
Solano (Spain)
Foehn-type wind from south to southeast. (The name is also used for easterly
winds from the Mediterranean Sea bringing monsoon-like precipitation.)
Talmac, Talmesch Wind D> Rotenturm Wind
Tenggara (Speimonde Archipel near Sulawesi, Celebes)
Föhn-type winds during easterly monsoons in the wake of the southern peninsula
of Sulawesi.
Toureillo (Arriège, Southern France)
South foehn from the Pyrenees in the Arriège Valley.
Tsinias (Aegean Sea)
Foehn wind from the southern cliffs of the islands in the Aegean Sea, name is
mainly used on the island of Tinos.
Türkenwind (Northern Tyrol, Austria; Rhine Valley, Switzerland)
Foehn accelerates the ripening of corn (and other crop). In this region, corn is
also called Türkenkorn (Turkish grain, instead of the German “Mais”, hence the
name Türkenwind [Turkish wind]).
Vent d’Espagne (Southern France)
West to southwest foehn-type wind from the Pyrenees when a depression moves
in north of them.
258 H. Richner and P. Hächler

Wambraw, Wambru (New Guinea)


Southwesterly foehn-type wind over the Geelvink Bay in northwestern New
Guinea during summer monsoon.
Wasatch (Utah, U.S.A.)
Easterly Chinook in the valleys of the Wasatch mountains.
Zephyr (Colorado, U.S.A.)
Another name for Chinook.
Zonda (Argentina)
Foehn wind in the Andes, in Chili the same wind is called Puelche.

References

AGF, 2007: http://www.agf.ch


Barry, R.G., 2008: Mountain Weather and Climate. 3 rd edition, Cambridge University Press.
506 pp.
Baumann, K., H. Maurer, G. Rau, M. Piringer, U. Pechinger, A. Prévôt, M. Furger, B. Neininger,
and U. Pellegrini, 2001: The influence of south Foehn on the ozone distribution in the Alpine
Rhine valley – results from the MAP field phase. Atmos. Environ., 35, 6379–6390.
Beran, D.W., 1967: Large amplitude lee waves and chinook winds. J. Appl. Meteorol., 6,
865–877.
Billwiller, R., 1899: Über verschiedene Entstehungsarten und Erscheinungsformen des Föhns.
Meteorol. Z., 16, 204–215.
Brinkmann W.A.R., 1971: What is Foehn? Weather, 26, 230–239.
Brinkmann, W.A.R., 1973: A climatological study of strong downslope winds in the Boulder area.
NCAR, Coop. Thes. 27, Univ. of Colorado, INSTAAR Occasional Pap. No. 6, 229 pp.
Brinkmann, W.A.R., 1974: Strong downslope winds at Boulder. Mon. Weather Rev., 102, 592–602.
Buchot, C., 1977: Le fœhn en Haute-Tarentaise. Rev. géographie alpine, 65, 257-276 l.
Burri K., P. Hächler, M. Schüepp, and R. Werner, 1999: Der Föhnfall vom April 1993. Arbeits-
bericht 196, MeteoSchweiz, 193 pp.
Burri, K., B. Dürr, Th. Gutermann, A. Neururer, R. Werner, and E. Zala, 2007: Foehn Verification
with the COSMO Model. Poster, Int. Conf. Alpine Meteorol., June 4 to 8, 2007, Chambéry,
France.
Cooke, L.J., M.S. Rose, and W.J. Becker, 2000: Chinook winds and migraine headache. Neurology,
54, 302.
Courvoisier, H.W. und Th. Gutermann, 1971: Zur praktischen Anwendung des Föhntests von
Widmer. Arbeitsbericht MeteoSchweiz, 21. (in German, out of print).
Di Napoli, G. and L. Mercalli, 2008: Il vento (capitolo 27). In: Il Clima di Torino. Società
Meteorologica Subalpina, Bussoleno, 629–660. ISBN 978-88-903023-4-3.
Drechsel S. and G. Mayr, 2008: Objective Forecasting of Foehn Winds for a Subgrid-Scale Alpine
Valley. Wea. Forecasting, 23, 205–218.
Drobinski, P., C. Haeberli, E. Richard, M. Lothon, A.M. Dabas, P.H. Flamant, M. Furger, and R.
Steinacker, 2003: Scale interaction processes during the MAOP IOP 12 south foehn event in
the Rhine Valley. Q.J.R. Meteorol. Soc., 129, 729–753.
Drobinski P, R. Steinacker, H. Richner, K. Baumann-Stanzer, G. Beffrey, B. Benech, H. Berger,
B. Chimani, A. Dabas, M. Dorninger, B. Dürr, C. Flamant, M. Frioud, M. Furger, I. Gröhn,
S. Gubser, Th. Gutermann, C. Häberli, E. Häller-Scharnhost, G. Jaubert, M. Lothon, V. Mitev,
4 Understanding and Forecasting Alpine Foehn 259

U. Pechinger, M. Piringer, M. Ratheiser, D. Ruffieux, G. Seiz, M. Spatzierer, S. Tschannett, S.


Vogt, R. Werner, and G. Zängl, 2007: Foehn in the Rhine Valley during MAP: A review of its
multiscale dynamics in complex valley geometry. Q.J.R. Meteorol. Soc., 133, 897–916.
Dürr, B., 2008: Automatisiertes Verfahren zur Bestimmung von Föhn in Alpentälern. Arbeitsbericht
223, MeteoSchweiz, 22 pp.
Ficker, H., 1920: Der Einfluss der Alpen auf die Fallgebiete des Luftdruckes und die Entstehung
von Depressionen über dem Mittelmeer. Meterol. Z., 55, 350–363.
Ficker, H., 1931: Warum steigt der Föhn in die Täler herab? Meteorol. Z., 48, 227–229.
Field, T.S. and M.D. Hill, 2002: Weather, Chinook, and Stroke Occurrence. Stroke, 33, 1751–1758.
Frey, K., 1944: Zur Entwicklung des Föhns. Verh. Schweiz. Naturforsch. Ges., 124. Jahresvers.
Sils. 90–93.
Frey, K., 1986: The simultaneous occurrence of North and South Foehn during a westerly flow
over Central Europe. Meteor. Atmos. Phys., 34, 349–366.
Gohm, A. and G. Mayr, 2004: Hydraulic Aspects of foehn winds in an Alpine Valley, Quart. J.R.
Meteorol. Soc., 130, 449–480.
Gohm, A., G. Zängl, and G.J. Mayr, 2004: South Foehn in the Wipp Valley on 24 October 1999
(MAP IOP 10): Verification of High-Resolution Numerical Simulations with Observations.
Mon. Wea. Rev., 132, 78–102.
Grubišić, V., J.D. Doyle, J. Kuettner, S. Mobbs, R.B. Smith, C.D. Whiteman, R. Dirks, S. Czyzyk,
S.A. Cohn, S. Vosper, M. Weissmann, S. Haimov, S.F.J. De Wekker, L.L. Pan, and F.K. Chow,
2008: The Terrain-Induced Rotor Experiment. Bull. Amer. Meteor. Soc., 89, 1513–1533.
Grubišić, V. and B.J. Billings, 2008: Climatology of the Sierra Nevada Mountain-Wave Events.
Mon. Wea. Rev., 136, 757–768.
Gubser, S., 2006: Wechselwirkung zwischen Föhn und planetarer Grenzschicht. Dissertation Nr.
16286, ETH Zürich, 121 pp.
Hächler, P., K. Burri, B. Dürr, T. Gutermann, A. Neururer, H. Richner und R. Werner, 2011: Der
Föhnfall vom 8. Dezember 2006 – Eine Fallstudie. Arbeitsbericht 234, MeteoSchweiz, 52 pp.
Hann, J., 1866: Zur Frage über den Ursprung des Föhns. Z. Österr. Ges. Meteorol., 1, 257–263.
Hann, J., 1901: Lehrbuch der Meteorologie. Tauchnitz, Leipzig, 805 pp.
Hoinka, K.P., 1985a: Observations of the airflow over the Alps during foehn. Quart. J.R. Meteorol.
Soc.,111, 199–224.
Hoinka, K.P., 1985b: What is a Foehn Clearance. Bull. Am. Meteor. Soc., 66, 1123–1132.
Hoinka, K.P., 1990: Untersuchungen der alpinen Gebirgsüberströmung bei Südföhn. Forschungs-
bericht der DLR, FB 90–30, 186 pp.
Jaubert, G. and J. Stein, 2003: Multiscale and unsteady aspects of a deep foehn event during MAP.
Q.J.R. Meteorol. Soc.,129, 755–776.
Klemp, J.B. and D.R. Lilly, 1975: The Dynamics of Wave-Induced Downslope Winds. J. Atmos.
Sci., 32, 320–339.
Lothon, M., 2002: Etude phénoménologique du foehn dans la vallée du Rhin dans le cadre de
l’expérience MAP (Mesoscale Alpine Programme). Thèse, Université Paul Sabatier, Toulouse,
260 pp.
Lothon, M., A. Druilhet, B. Bénech, B. Campistron, S. Bernard, F. Saı̈d, 2003: Experimental
study of five foehn events during the Mesoscale Alpine Programme: from synoptic scale to
turbulence. Q.J.R. Meteorol. Soc.,129, 2171–2193.
Lyra, G., 1940: Über den Einfluss von Bodenerhebungen auf die Strömung einer stabil
geschichteten Atmospäre. Beitr. Phys. frei. Atmos., 26, 197–206.
McGowan H.A., A.P. Sturman, M. Kossmann, and P. Zawar-Reza, 2002, Observations of foehn
onset in Southern Alps, New Zealand, Meteorol. Atmos. Phys., 79, 215–230.
Musso A. and C. Cassardo, 2004: Climatologia del foehn in Piemonte. Nimbus, 31–32, 40–45.
Queney, M. P., 1948: The problem of airflow over mountains: A summary of theoretical studies.
Bull. Amer. Meteor. Soc., 29, 16–26.
Raphael, M.N., 2003: The Santa Ana Winds of California. Earth Interact., 7, 1–13.
260 H. Richner and P. Hächler

Richner, H. and Th. Gutermann, 2007: Statistical analysis of foehn in Altdorf, Switzerland.
Extended Abstracts Volume 2, Int. Conf. Alpine Meteorol., June 4 to 8, 2007, Chambéry, France,
457–460.
Richner, H., K. Baumann-Stanzer, B. Benech, H. Berger, B. Chimani, M. Dorninger, P. Drobinski,
M. Furger, S. Gubser, Th. Gutermann, C. Häberli, E. Häller, M. Lothon, V. Mitev, D. Ruffieux,
G. Seiz, R. Steinacker, S. Tschannett, S. Vogt, and R. Werner, 2005: Unstationary aspects of
foehn in a large valley, part I: operational setup, scientific objectives and analysis of the cases
during the special observing period of the MAP subprogramme FORM., Meteorol. Atmos.
Phys., 92, 255–284.
Rossmann, F., 1950: Über das Absteigen des Föhns in die Täler. Ber. deutsch. Wetterd. US-Zone,
12, 94–98.
Schamp, H., 1964: Die Winde der Erde und ihre Namen. Franz Steiner Verlag, Wiesbaden, 94 pp.
Schüepp, W., 1952: Die qualitative und quantitative Bedeutung der Föhnmauer. Meteorol. Rdsch.,
5, 136–138.
Schweitzer, H., 1953: Versuch einer Erklärung des Föhns als Luftströmung mit überkritischer
Geschwindigkeit. Arch. Meteor. Geophys. Bioklimatol., A5, 350–371.
Scorer, R.S., 1978: Environmental Aerodynamics. Wiley & Sons, 488 pp.
Seibert, P., 1990: South foehn studies since the ALPEX experiment. Meteorol. Atmos. Phys., 43,
91–103.
Seibert, P., 2005: Hann’s Thermodynamic Foehn Theory and its Presentation in Meteorological
Textbooks in the Course of Time. From Beaufort to Bjerknes and Beyond, Algorismus, Issue
52, 169–180; ISBN 978-3936905-13-7.
Sheridan, P.F., V. Horlacher, G.G. Rooney, P. Hignett, S.D. Mobbs, and S.B. Vosper, 2007:
Influence of lee waves on the near-surface flow downwind of the Pennines. Q.J.R. Meteorol.
Soc., 133, 1353–1369.
Sprenger, M., and C. Schär, 2001: Rotational aspects of stratified gap flows and shallow foehn.
Quart. J.R. Meteorol. Soc., 127, 161–187.
Steinacker, R., 2006: Alpine foehn – a new verse to an old song. Promet, 32, 3–10 (in German).
Streiff-Becker, R., 1933: Die Föhnwinde. Viertelj.schr. Naturforsch. Ges. Zürich, 78, 66.
Streiff-Becker, R., 1947: Der Dimmerföhn. Viertelj.schr. Naturforsch. Ges. Zürich, 92, 195–198.
Vergeiner J. and G. Mayr, 2000: Case study of the MAP-IOP “Sandwich” foehn on 18th October
1999. MAP Newsletter, 13, 36–37.
Wild, H., 1901: Über den Föhn und Vorschlag zur Beschränkung seines Begriffs. Denkschr.
Schweiz. Naturf. Ges., 38, 99 p. plus appendix.
WMO (ed.), 1992: International Meteorological Vocabulary. WMO No. 182, World Meteorologi-
cal Organization, Geneva, Switzerland, 784 pp.
Zängl, G., 2002: Idealized numerical simulations of shallow föhn. Quart. J.R. Meteorol. Soc., 128,
431–450.
Chapter 5
Boundary Layers and Air Quality
in Mountainous Terrain

Douw G. Steyn, Stephan F.J. De Wekker, Meinolf Kossmann,


and Alberto Martilli

Abstract In this chapter, the general problem of air pollution in mountainous


terrain is discussed. Terrain effects on the spatial distribution and temporal vari-
ability of air pollutants are described including specific effects encountered in
stable- and convective boundary layers and under dynamically driven conditions.
The uses of numerical models, scale models, and observational studies are described
as tools for understanding air pollution in mountainous terrain. Specific integrated
regional studies of air pollution conducted in the Lower Fraser Valley, Canada and
in the Upper Rhine Valley, Germany are used to illustrate the terrain effects. The
chapter concludes with a discussion of approaches to forecasting air pollution in
mountainous terrain.

5.1 Introduction

In very simple terms, once emitted, the concentration of air pollutants in the
atmospheric boundary layer is determined by a combination of chemical (production
and destruction) and dispersion processes (the combined effects of dilution by mean

D.G. Steyn ()


Department of Earth and Ocean Sciences, The University of British Columbia,
Vancouver, BC, Canada
e-mail: dsteyn@eos.ubc.ca
S.F.J. De Wekker
Department of Environmental Sciences, University of Virginia, Charlottesville, VA, USA
e-mail: dewekker@virginia.edu
M. Kossmann
Climate and Environment Consultancy, Deutscher Wetterdienst, Offenbach am Main, Germany
e-mail: Meinolf.Kossmann@dwd.de
A. Martilli
CIEMAT Unidad de Contaminacion Atmosferice, Madrid, Spain
e-mail: alberto.martilli@ciemat.es

F. Chow et al. (eds.), Mountain Weather Research and Forecasting, 261


Springer Atmospheric Sciences, DOI 10.1007/978-94-007-4098-3 5,
© Springer ScienceCBusiness Media B.V. 2013
262 D.G. Steyn et al.

Fig. 5.1 Schematic of air pollution in a valley, illustrating elevated temperature inversion, multiple
pollution sources, pollution plumes and cities with industrial, commercial and residential buildings.
The country breeze is a thermally driven wind resulting from the horizontal temperature gradient
between an urban area and the surrounding country (Modified from Elsom 1987)

wind advection and cross-wind diffusion). In flat terrain, or over the oceans or lakes,
the dispersion processes are limited in the vertical by the depth of the atmospheric
boundary layer, and in the horizontal only by wind speed, wind trajectory sinuosity,
and the intensity of crosswind turbulence. In the near field (generally within 10 km
from a pollution source), diffusion effects dominate while the pollution is still in
a defined plume, while further downwind, boundary layer depth and mean wind
speed become dominant. In mountainous terrain the matter becomes much more
complicated. The atmospheric boundary layer responds to topographic forcing, and
most of these meteorological variables are strongly altered, generally in ways which
result in more severe pollution than would be expected from the same emissions
in flat locations. These conditions are particularly severe in valley bottoms, along
mountain slopes and in mountain basins, as indicated in Fig. 5.1. Many of the
mechanisms which affect air quality in mountainous terrain have been described
in detail in Chaps. 2 and 3, so references are provided here as appropriate.
Humans settle in mountains for economic, strategic, historical or political
reasons. In the year 2000, 720 million people (12% of world population) lived in
mountainous areas, and 663 million of these people lived in developing or transition
countries (Huddlestone et al. 2003), where restrictions on emissions of pollutants
are generally less stringent than elsewhere. It is thus important to understand
the fate and distribution of pollutants in mountainous environments. The history
of pollution episodes shows that many notably polluted locations are in valleys
or mountain basins, or topographically constrained coastal plains. These include
examples like the Plateau of Switzerland, the Po Plain in Italy, The Los Angeles
Basin in USA, Mexico City (also see Chap. 2), the Lower Fraser Valley in Canada
(also see Chap. 8), and the Rhine Valley in Germany. The latter two examples will be
examined as interesting case studies in this chapter, which will touch on particular
atmospheric meteorological phenomena that are important in limiting atmospheric
dispersion in mountainous regions.
5 Boundary Layers and Air Quality in Mountainous Terrain 263

5.2 Boundary Layer Processes Relevant for Air Quality


in Mountainous Terrain

In locations of flat, homogeneous terrain, the boundary layer depth is constant in


space, and varies temporally due to mechanically or thermally induced mixing or
due to stabilization by radiative loss from the surface. In convective conditions,
the boundary layer depth undergoes strong diurnal variation, as shown in Fig. 5.2.
The overall result is that the boundary layer depth is identical to the mixed layer
depth – the depth to which surface emitted substances (including air pollutants) will
be mixed. In regions of complex, mountainous terrain, boundary layer depth varies
strongly in both time and space. In general, boundary layer height will mirror terrain,
but with suppressed amplitude. Local, topographic scale, up- and down-slope flows
(as discussed in Chap. 2, Sect. 2.3) will enhance turbulent transport by adding small-
scale advection to thermally driven convective motions. Most notably, thermally
driven upslope flows will add substantially to the vertical transport of pollutants.
In many cases the net effect will be a transport of pollutants well above the locally-
determined convective boundary layer (CBL) depth to form a regional pollution
layer, often visible as a haze layer blanketing the mountains. This will contrast
strongly with times under which strong surface cooling results in a surface based
temperature inversion and very little turbulent mixing, and with katabatic winds
carrying pollutants or clean air down into valleys and basins. Clearly, different large-
scale weather conditions produce strongly varying pollution potential everywhere,
but particularly in the mountains. This section will elaborate on pollution in stable,
convective and dynamically driven boundary layers in mountainous terrain.

Fig. 5.2 Conceptual model of the diurnal evolution of the convective boundary diurnal in flat
terrain, showing various sub-layers. NBL is nocturnal boundary layer
264 D.G. Steyn et al.

Fig. 5.3 Schematic representation of processes affecting the dispersion of air pollutants in a stably
stratified boundary layer over complex terrain. (1) drainage flow, (2) slope flow detrainment,
(3) urban heat island circulation, (4) terrain induced flow splitting and convergence, (5) mesoscale
subsidence, (6) regional wind systems, (7) gravity waves, (8) intermittent turbulence (Jerome Fast,
Pacific Northwest National Laboratory, operated by Battelle for the U.S. Department of Energy.
Reprinted with permission)

5.2.1 Processes in Stable Boundary Layers

Episodes of severe pollution are frequently associated with poor dispersion


conditions occurring in stable boundary layers. While this association is found
everywhere, mountain induced processes can lead to particularly poor air quality.
A schematic illustration of nighttime air pollution dispersion processes over
mountainous terrain is presented in Fig. 5.3. Most prominent is the negative effect
of cold air pools forming in basin and many valleys during nighttime, as described
in detail in Chap. 2, Sect. 2.4. Such cold air pools are characterized by very stable
stratification, very low wind speeds, and weak or intermittent turbulent mixing.
With human settlements and hence anthropogenic emission sources being mostly
located at low terrain elevations, e.g. at the bottom of basins or along valley floors,
pollutants emitted into cold air pools tend to accumulate to high concentrations.
Particularly hazardous air quality conditions are likely to occur in persistent (multi-
day) cold air pools (Whiteman et al. 2001).
Under synoptically quiescent conditions, thermally driven slope- and valley
winds (Chap. 2) are very important as they maintain some degree of ventilation
in valleys and basins. However, flow convergence associated with cold air draining
from surrounding terrain elevations towards cold pools may favor the occurrence
of stagnant conditions with weak winds (Kossmann and Sturman 2004; Corsmeier
et al. 2006).
Down-slope and down-valley wind systems typically advect relatively unpolluted
air from rural surroundings at higher terrain elevations towards the densely settled
and therefore more polluted valley and basin sites (Banta et al. 1997; Raga et al.
1999; Baumbach and Vogt 1999; Poulos and Zhong 2008). This phenomenon
often results in the increase of concentrations of secondary pollutants such as
ozone because of an influx of precursor substances stored in the residual layer
5 Boundary Layers and Air Quality in Mountainous Terrain 265

(Löffler-Mang et al. 1997; Raga et al. 1999; Salmond and McKendry 2002).
Furthermore, down-slope or down-valley winds might re-circulate pollutants that
have been carried away from the emission sources by daytime up-slope or up-valley
winds. An example of mountain and coastal effects on pollutant re-circulation is
described in Sect. 5.4.1 for the Lower Fraser Valley, BC (also see Chap. 8).
In many cases the application of the topographical amplification factor concept
(see Chap. 2, Sect. 2.4) is useful in distinguishing between pooling and draining
valleys, and for a better understanding of the occurrence of valley/basin inversion
break-up (McKee and O’Neal 1989; Sakiyama 1990; Whiteman et al. 1996).
However, forecasting stable boundary-layer processes for air quality applications
still represents a major challenge. Reasons for this include the high spatial model
resolution required to resolve the most important processes and the difficulties with
turbulence parameterization under very stable conditions. Stable boundary-layer
processes in mountainous terrain have also been recognized as important in pro-
viding an understanding of carbon dioxide budgets, which are essential for climate
change monitoring and projections (Sun et al. 1998; Eugster and Siegrist 2000).

5.2.2 Processes in Convective Boundary Layers

Figure 5.4 presents a conceptual model of the diurnal development of the CBL over
mountainous terrain (Fiedler et al. 1987; Whiteman et al. 2000). During the night
and in the early morning hours, cold air accumulates in the valleys, resulting in
greater atmospheric stability in valleys than over the mountain ridges (Fig. 5.4a).
After sunrise, this leads to a slower development of the CBL in the valleys than
over the mountain ridges (Fig. 5.4b). Development of the CBL in valleys is further
suppressed by compensatory sinking motions as a result of mass removal by upslope
flows (e.g. Whiteman 1982; De Wekker 2008; see also Chap. 2). After the inversion
in the valleys has broken up the CBL approximately follows the terrain (Fig. 5.4c).
Further development of the CBL depends on factors such as the total sensible heat
input, the ambient stability and the strength of the synoptic scale sinking motions.
This often leads to parallel CBL growth in the valley and surrounding mountain,
and maintains the terrain-following CBL structure up to early afternoon. By late
afternoon, due to faster CBL development over the valley, the CBL top may become
a roughly horizontal surface (Fig. 5.4d).
This general conceptual model has important implications for air pollution in
mountain regions. Air pollutants that have accumulated in the valleys at night dis-
perse upward during the morning hours. In deep valleys, breakup of the temperature
inversion can take many hours and on some days does not occur, especially in
the winter when incoming radiation is small. Once the surface-based temperature
inversion breaks up, dispersion can occur in several ways, transport by upslope flows
and convective mixing in the growing CBL being the most important of these. Due
to these transport and mixing processes, mountain top locations often experience
increased levels of air pollution after the break-up of surface based temperature
inversions in adjacent valleys (e.g. Seibert et al. 1998).
266 D.G. Steyn et al.

Fig. 5.4 Conceptual model of the daytime boundary layer development over mountainous terrain.
(a) before sunrise, (b) early morning, (c) at noon, (d) late afternoon. Light grey shading represents
topography. Dark grey shading illustrates strongly stable stratification in nocturnal cold air pools.
Arrows denote horizontal and vertical flows. The wavy lines depict the upper limit of the surface
generated convection. Vertical profiles of potential temperature are shown at several locations along
the cross section (After Fiedler et al. 1987. Reprinted with permission)

The presence of mesoscale circulations and their interactions with the ambient
flow and with boundary layer convection complicates boundary layer structure in
mountainous terrain relative to that found over flat terrain. Very often, multiple
elevated temperature inversions are found in the lower atmosphere over mountains
leading to an ambiguous mixing height. In addition, aerosol layers are often found
above these temperature inversions, making the common definition of mixing
heights in flat terrain inoperable in complex terrain (De Wekker et al. 2004).
Mechanisms that can explain the presence of aerosol (and other pollutants) above
the CBL top are summarized in Fig. 5.5 (De Wekker et al. 2004). In this figure,
aerosol layer (AL) heights are relatively uniform and reach a maximum altitude
of about 4 km in the afternoon. The CBL heights are considerably lower than AL
heights and show more spatial variability. CBL heights tend to follow the terrain
more closely than AL heights, although the extent to which the CBL height follows
5 Boundary Layers and Air Quality in Mountainous Terrain 267

5
ha

4
height (km asl)

zi 3 4
1 + 2
3
1
2 5
2

SOUTH NORTH
0
0 10 20 30 40 50
horizontal distance (km)

Fig. 5.5 Conceptual diagram of mid-afternoon CBL in mountainous terrain (indicated by the
shaded grey area). The curve labelled zi is the CBL height and the curve labelled ha is the
AL height. The depicted mechanisms are (1) mountain venting, (2) cloud venting, (3) advective
venting, (4) advection of aerosols from airmasses elsewhere and (5) plume convection that does
not overshoot the CBL top (After De Wekker 2002. Reprinted with permission)

the terrain decreases during the day. The mechanisms contributing to deeper AL
than CBL heights and indicated in Fig. 5.5 are the following: (1) mountain venting,
(2) cloud venting, (3) advective venting, and (4) horizontal advection of aerosols
from airmasses upwind of the area of interest. Mountain venting and cloud venting
can occur simultaneously when an orographically induced thermal reaches lifting
condensation level. All of these mechanisms have been observed to occur in a
variety of field studies. For an overview, see Kossmann et al. (1999).
As noted, observations and model results show that the extent to which the CBL
height follows the terrain typically decreases during the day (Lenschow et al. 1979;
Fiedler 1983; De Wekker et al. 2004). Stull (1998) derived an equation (applicable
under free convection conditions) for the tendency of the CBL top to become more
horizontal in the course of the day by applying the mass conservation equation for a
CBL over hilly terrain. The results showed that advection, entrainment and friction
contribute to making the CBL follow the terrain, while gravitational forces tend
to level the top of the CBL. If there is more than just a light synoptic wind, the
advection term was found to be dominant in making the top of the CBL follow the
terrain. A major shortcoming of Stull’s theory is that it does not take into account
the effect of thermally-driven flows.

5.2.3 Dynamical Processes

The dynamic modification of a larger scale airflow passing over mountainous


terrain plays an important role in the transport, accumulation, and diffusion of air
pollutants. In addition to mean wind speed and direction, turbulence intensities
268 D.G. Steyn et al.

and structure are also subject to dynamical modification by the terrain. Dynamical
dispersion processes in mountainous terrain include flow over and around hills
and mountains, airflow channeling in valleys, barrier winds, gravity waves, flow
separation, and the formation of traveling or standing eddies with horizontal or
vertical axes (see Chap. 3). Turbulence intensities are often enhanced by flow
separation because of mechanically generated turbulence in wind shear at separation
zones, while highest wind speeds generally occur over exposed convex terrain
formations such as mountain ridges or due to airflow funneling through terrain
constrictions. Lowest mean wind speeds are usually observed at sheltered valley
or leeside locations.
Chapter 3 provides a detailed description of dynamical airflow processes over
mountainous terrain. Furthermore, the current understanding of the well-known
foehn winds (a strong downslope wind) is described in detail in Chap. 4. Foehn
winds are important for the turbulent erosion of cold air pools in valleys and basins
(Petkovsek 1992; Zhong et al. 2003). Hence, the onset of foehn winds often ends
periods of poor air quality. A useful tool to determine whether air pollutants emitted
at a certain height on the windward side of a mountain are likely to impinge on
the mountain slope or pass over a mountain is the dividing streamline concept
(see Chap. 3, Sect. 3.4.2). This concept is based on considerations of the kinetic
and potential energy of the approaching flow via the Froude number (Chap. 3,
Sect. 3.2.3). The Froude number can also be used to give a good indication of
the characteristics of gravity waves possibly developing on the leeward side of a
mountain.
The wind climate in mountain valleys is often dominated by airflow channeling
along the valley axis (Kalthoff and Vogel 1992; Smedman et al. 1996; Eckman
1998; Cogliati and Mazzeo 2006). A detailed discussion of mechanisms generating
airflow channeling is given in Chaps. 2 and 3. The consequences of such channeling
on plume dispersion are demonstrated in Fig. 5.6, which illustrates the dispersion
of plumes in a valley and over the adjacent plain. Over the plain, plumes from
neighboring sources only have intersecting centerlines if the regional wind blows
parallel to the connecting line between the sources. Plumes from sources located
along a valley, however, can be superimposed onto each other for almost all regional
wind directions, because airflow channeling leads to a bimodal wind direction
climatology in valleys. A consequence is that poor air quality situations resulting
from plume interference are far more likely to occur in valleys than over a plain.
Even more complex dispersion conditions occur in curved valleys, or at valley
bifurcations, where airflow channeling is likely to generate zones of converging or
diverging along-valley winds (Enger et al. 1993; Kossmann et al. 2002a; Kossmann
and Sturman 2003; Dobrinski et al. 2006).
Interaction of channeled along-valley winds and airflow across a mountain
barrier potentially causes strong wind shear and often results in eddy formation.
An example of such a dynamically-induced eddy with a horizontal axis is the
rotor which is associated with the existence of lee waves. A favorable location for
dynamical rotor formation, airflow channeling, and dust storm development is the
Owens Valley in California (Grubišić et al. 2008; Zhong et al. 2008). A lidar scan of
5 Boundary Layers and Air Quality in Mountainous Terrain 269

Fig. 5.6 Idealised plumes


from two sources within and
two sources outside an
north–south oriented valley.
Circles in red show the
location of sources. Dashed
lines indicate the valley
sidewalls. VG is the
geostrophic wind vector and
V0 is the channeled surface
wind vector in the valley. The
top panel shows terrain
elevations along a vertical
cross section, ‚(z) is the
potential temperature profile

aerosol backscatter intensity across the Owens Valley during strong westerly flows
across the Sierra Nevada Mountains is depicted in Fig. 5.7. The convergence of the
aerosol-free westerly surface winds and of channeled aerosol-filled southerly winds
along the Owens Valley leads to a lifting of aerosols up to 3 km above valley floor
in this situation.
The formation of eddies with a vertical axis and re-circulation of air pollutants
within these eddies is quite common over mountainous terrain. For example, the
Fresno Eddy shown schematically in Fig. 5.8 is generated by the interaction of local
mountain winds in the San Joaquin Valley and a regional low-level jet, which is
triggered by marine air intrusions in the San Francisco Bay area. Re-circulation
over strong emission sources and hence accumulation of air pollutants within the
Fresno Eddy is a very important mechanism for the build-up of high photochemical
pollution levels in the area and in adjacent National Parks (Lin and Jao 1995; Bao
et al. 2008). Such topographically induced eddies with vertical axes can occur under
a wide range of atmospheric stratifications. Observations and modeling studies
270 D.G. Steyn et al.

Fig. 5.7 Range Height Indicator (RHI) (vertical) lidar scan of aerosol backscatter intensity
captured during the T-REX campaign in the Owens Valley around 0015 UTC 3 March 2006
(After De Wekker and Mayor 2009). The azimuth angle of the scan is 282ı . The white arrow
directed toward the right indicates westerly flow down the east side of the Sierra Nevada slopes.
This relatively clean air undercuts the aerosol-laden southerly flow in the Owens Valley, which has
an easterly component, indicated by the white arrow directed toward the left. The lifting generated
by the near surface flow convergence and the resulting recirculation is illustrated by the black
arrows. The recirculation of the aerosol-laden air lasted for several minutes, as inferred from the
animation of RHI scans. Range rings are drawn at every 2 km distance from the lidar (© American
Meteorological Society. Reprinted with permission)

Fig. 5.8 Formation of eddies with vertical axes in the Central Valley of California (Bao et al.
2008). Recirculation and trapping of polluted air causes poor air quality in the area and often in
surrounding National Parks (© American Meteorological Society. Reprinted with permission)
5 Boundary Layers and Air Quality in Mountainous Terrain 271

indicate that conditions with a stably stratified atmosphere and moderate ambient
winds are particularly favorable for the formation of these eddies, which have been
discovered at many different locations around the world. Examples include the
Melbourne Eddy or Spillane Eddy (Spillane 1978; McGregor and Kimura 1989),
the Kanto Plain Eddy (Harada 1981; Kimura 1986), the Graz City Eddy (Oettl
et al. 2001), the Schultz Eddy (Zaremba and Carroll 1999; Bao et al. 2008), and
the Denver Cyclone (Wilczak and Christian 1990; Levinson and Banta 1995).

5.2.4 Interactions and Other Processes

Under most synoptic conditions air pollution dispersion over mountainous terrain
is complex because multiple thermally and dynamically generated processes occur
simultaneously. For example, nighttime down-slope winds may interact with gravity
waves on the lee side of mountains (Poulos et al. 2000), or airflow along valleys
may be characterized by layers of opposing wind directions with either dynamically
channeled winds or thermally driven mountain winds dominating in each layer
(Schmidli et al. 2009). During daytime, atmospheric stability at valley and basin
sites can be strongly modified by the advection of mixed layers formed over
neighboring mountain ranges (Arritt et al. 1992; Stensrud 1993). The advection of
elevated mixed layers and anabatic winds favor the formation of elevated pollution
layers (Kossmann et al. 1999; Frioud et al. 2003; Henne et al. 2004) that may
become subject to long-range transport. Down-mixing of these pollution layers due
to growing CBLs can significantly enhance surface concentrations of air pollutants
(Neu et al. 1994; McKendry et al. 1997).
In most mountain areas surface cover is not uniform. Water surfaces and urban
areas are often present in valley bottoms, for example. These surface heterogeneities
generate thermally driven wind systems and roughness effects which are superim-
posed on mountain induced processes, often resulting in enhanced dispersion (Lu
and Turco 1994; McGowan and Sturman 1996; Bischoff-Gauß et al. 1998; Millan
et al. 1997; McKendry and Lundgren 2000; Ohashi and Kida 2002; Troude et al.
2002; Kalthoff et al. 2004; Bastin et al. 2004; de Foy et al. 2005).
Figure 5.9 shows a situation from central Japan where sea/land breezes and multi-
scale orographically induced circulation systems cause a complex dispersion pattern
of air pollutants emitted at the coast in Tokyo and in neighboring cities on the
Kanto Plain (Kurita et al. 1990). The pollutants are carried inland during daytime
by the sea breeze and anabatic valley and slope winds. Particularly strong heating
in the central mountain basins causes the formation of a thermal low (Kuwagata
and Sumioka 1991) and triggers the onset of plain-to-basin winds (Kimura and
Kuwagata 1993). By the evening, the air pollutants emitted at the coast arrive in
the central mountain basins of Japan where local emissions are low. Air quality
in these remote regions therefore degrades significantly in the evening. During the
night some of the polluted air is transported back towards the coast and possibly
offshore by katabatic winds and land breezes.
272 D.G. Steyn et al.

Fig. 5.9 Vertical cross section of pollutant transport processes across Central Japan, from the
coastal regions to the mountainous inland region. Broken line denotes the average altitude of the
mountainous central region. Vertical hatching a depression in sea level pressure, PA polluted air,
TL thermal low, CL convergence line, UW upper level wind (only shown for 0600 Japan Standard
Time (JST)), LB land breeze, MW mountain wind, VW valley wind, ESB extended sea breeze, LSW
large scale wind toward the thermal low (Kurita et al. 1990) (© American Meteorological Society.
Reprinted with permission)

In densely populated mountain basins with higher emissions than in the sur-
rounding region, plain-to-basin winds are important because they advect relatively
unpolluted air into the basin. However, the advection of cold air into the basin
associated with the plain-to-basin wind might also cause strong reductions in
mixing depth inside basins (Regmi et al. 2003). These effects of plain-to-basin
winds on atmospheric stability and on the transport of clean or polluted air across
mountain ranges or mountain passes have made these winds the subject of several
investigations (De Wekker et al. 1998; Fast and Zhong 1998; Whiteman et al. 2000;
Kossmann et al. 2002b; Zawar-Reza and Sturman 2006).

5.3 Research Tools

Research tools typically employed in the study of air pollution in complex,


mountainous terrain include observations, numerical models and physical (scale)
models. While these tools are distinct, they are generally used in concert to provide
maximum information from a combination of tools and approaches.
5 Boundary Layers and Air Quality in Mountainous Terrain 273

5.3.1 Observations

Many technical and logistical challenges exist when attempting to make meteo-
rological measurements in mountainous terrain. Sites are often difficult to access
and/or electrical power is not available. All those challenges recur when adding
measurements of pollutants to meteorological measurements. In addition, the
representativeness of single measurements in mountainous terrain is often for much
smaller areas than corresponding flat terrain measurements because of the large
spatial variability in pollution fields. Mountain top measurements are an exception,
as under certain conditions they are representative of a regional or even larger scale.
The solutions to the limited representativeness of air pollution measurements in
mountainous terrain are exactly those for meteorological measurements – remote
sensing, very dense networks and aircraft borne sensors. Aircraft measurements in
mountainous terrain are made extraordinarily difficult by restricted and crowded
airspace in mountains. Ultimately, the solutions to these difficulties involve a mix
of fixed sensors monitoring the temporal changes (at many locations to also sample
the spatial variability), with mobile platforms to fill in the spatial gaps, and remote
sensing systems to provide broad coverage and data in inaccessible places. Chapter 8
provides a thorough discussion of these observational approaches. Examples of
these solutions are provided throughout this chapter in descriptions of various field
studies.

5.3.2 Numerical Modeling

Difficulties involved in making atmospheric measurements in complex terrain


would imply that much reliance would be placed on modeling as an adjunct to
measurement. In many ways this has happened, but experience has shown that mod-
eling in extremely steep, mountainous terrain has its own set of difficulties. Many of
these difficulties are outlined in Chaps. 9, 10, and 11. Undoubtedly, the overriding
advantage of numerical modeling of air pollution in complex terrain is that model
output is available at locations where measurement would be impractical or even
impossible. This allows investigation of the spatial structure of pollution fields,
rather than the apparent structure, represented by the distribution of monitoring
sites, which are only in locations that allow instrument operation. In many cases,
this is driven by the availability of electrical power to drive the instruments. The
major weaknesses of numerical models of air quality in complex terrain are similar
to those of meteorological models in complex terrain that are used to drive them
(see Chap. 10). Of paramount concern must be a consideration of mesoscale model
grid resolution in relation to the scale of terrain-induced motions. In an ideal world,
the model grid must be designed so that resolved scales of motion are smaller than
dominant topographically induced scales of pollution advection and dispersion. In
far too many cases, this ideal has to be compromised because of the computational
274 D.G. Steyn et al.

cost of fine resolution grids, or because of the scales of motion being smaller than
the finest resolvable scale of the mesoscale model being used. The technique of
large-eddy simulation has been developed in recent years to allow application to real
complex topography. However, successful application of this technique in weather
and air quality forecasting is far from being realized. Furthermore, an improvement
in spatial resolution in numerical models does not always imply improvements in
simulations of atmospheric processes (see Chap. 10, Sect. 10.2.1). More research is
certainly needed in this area.

5.3.3 Physical Modeling

Physical scale models (either wind tunnels or water tanks) have been used to
study a wide variety of atmospheric phenomena, including many in the field of
air pollution meteorology. Most commonly, wind tunnel studies focus on plume
dispersion, often in complex terrain, or around buildings (Plate 1982). Water tank
studies are primarily used to study density driven phenomena (primarily anabatic or
katabatic winds) in complex terrain (Chen et al. 1996). In many cases these studies
are motivated by an interest in the topographic-scale (therefore smaller mesoscale)
advection of pollutants in regions of complex terrain. Common target phenomena
are the venting of pollutants out of valleys in anabatic flows, or conversely, the
pooling of pollutants in low-lying areas because of katabatic flows.
In common with all physical scale modeling of flow phenomena, the major
challenge is ensuring that dimensionless governing parameters in the scale model
match those in the full-scale world. This generally means that it is not possible to re-
alistically model both turbulent diffusion and mesoscale advection simultaneously.
These difficulties are compensated for by the fact that tank experiments are highly
repeatable and permit measurements that are impossible in the field, and allow easy
control and manipulation of external governing variables.
An example of a tank study of an air pollution phenomenon in complex terrain
is provided by Reuten et al. (2007) who examine pollutant trapping in a growing
morning boundary layer along a steep sided valley wall. The technique has large
untapped potential for studies that are locally specific, as well as generic. Some
more examples of tank studies are provided in Chap. 2, Sect. 2.1.4.

5.4 Examples from Integrated Studies of Air Pollution


in Complex Terrain

The range and complexity of processes discussed so far indicate that local and
regional air quality in mountainous areas could be severely compromised because
of boundary layer meteorological processes found there. The need for scientifically
5 Boundary Layers and Air Quality in Mountainous Terrain 275

supported air quality management strategies, and the development of air quality
forecast tools has resulted in integrated, comprehensive air quality studies in many
mountainous areas. These studies all capture, to varying degrees the processes
discussed in isolation in this chapter.
Understanding the effects of these processes on air quality is an important part
of the air quality forecasts that have become increasingly popular and available
in the last few years (Eder et al. 2010). Even though the level of sophistication
of air quality models has increased, their performance in mountainous regions
such as the Appalachian Mountains is poor (Eder et al. 2006). The incorporation
of sophisticated air quality models in daily air quality forecasting follows the
same process as the incorporation of numerical weather prediction models in daily
weather forecasting. This process includes detailed evaluations of the performance
of these forecast models. Model evaluations in mountainous terrain are particularly
challenging for reasons mentioned earlier but they are extremely important to
provide guidance to air quality forecasters. There are different ways in which
such guidance for forecasters could be envisioned. One example would be to
quantitatively assess the effects of terrain on atmospheric processes in case studies
where these effects are shown to contribute to poor forecasts. These effects,
such as modified boundary layer heights, wind channeling, thermally-driven flows,
stagnation, etc., could then be incorporated into the daily forecasts. The detailed
model evaluation that is at the core of improved forecast guidance requires: (1)
significant interactions between operational forecasters and researchers, and (2)
integrated, comprehensive air quality studies in mountainous areas. Such studies
have been performed in the past, primarily driven by the need for scientifically
supported air quality management strategies.
In order to provide a glimpse into the fascinating complexity that is encountered
when some of these processes occur simultaneously, or in sequence, we provide
brief summaries of two examples of mountainous regions whose air quality has
been studied in some detail. In undertaking this exercise, it must be recognized that
human activities of notable size (and therefore resulting in substantial emissions),
are seldom found on mountains, but rather in valleys, basins or plateaus surrounded
by mountains. Our two examples will typify this.

5.4.1 Lower Fraser Valley, BC

The Lower Fraser Valley (LFV) in British Columbia, Canada is home to the city of
Vancouver and its surrounding municipalities, with a total population of roughly
2.2 million people. The valley is roughly triangular, with all three sides being
approximately 100 km in length. The western margin of the valley is the inland
waterway of the Strait of Georgia (Fig. 5.10). The LFV is flat bottomed, bordered to
the north by the Coast Mountains, which rise steeply to an elevation of 2,200 m ASL,
and to the south, by the slightly lower Cascade Mountains of Washington State,
USA. The valley is subject to mild, wet winters with long periods under the influence
276 D.G. Steyn et al.

Fig. 5.10 Map of the Lower Fraser Valley showing typical ozone plumes. (a) 1988 (pre controls),
(b) 2003 (post controls)

of weak, occluded frontal systems, interspersed by often deep, cold, dry anticyclonic
conditions (Oke and Hay 1994). In summertime, the LFV comes under the influence
of an eastward extension of the Pacific High, resulting in warm, dry conditions,
and very light synoptic winds. Emissions sources in the LFV are primarily due
to the transportation sector, concentrated in the northwest portion of the valley,
there being very little primary industry in the region. Wintertime air pollution is
largely oxides of nitrogen (NOx) and particulate matter (PM), and is closely linked
to transportation corridors. The most important summertime air pollution concern in
the LFV is ozone and secondary PM (Steyn et al. 1997). The occurrence and spatio-
temporal variability of summertime ozone and PM episodes is strongly influenced
by the presence of interacting land/sea breeze circulations, and valley- and slope-
flow systems. The strong recirculation of pollutants in these topographically driven
flows and the presence of very shallow convective boundary layers (Steyn and Oke
1982; Hägeli et al. 2000) lead to far more severe pollution episodes than might
be expected, given the relatively small population and absence of heavy industry.
Contributing to the severity of the episodes is vertical, horizontal and day-to-day
5 Boundary Layers and Air Quality in Mountainous Terrain 277

Fig. 5.10 (continued)

recirculation of pollutants as indicated in Fig. 5.11. These three recirculation modes


are driven by a complex interplay of slope flows, land/sea breezes and boundary
layer mixing. These quasi-closed recirculation pathways are characteristic of the
LFV summertime ozone episodes which occur under very weak synoptic forcing,
and contribute to day-to-day buildup of pollutants within an episode. Evidence for
the formation of the elevated polluted layer above the CBL by topographically
induced advection of polluted air is given in Fig. 5.12. Chapter 8, Sect. 8.2 shows
some of the meteorological, and therefore pollutant distribution complexities that
arise from exchange of air between side valleys and the major LFV. Undoubtedly,
pollutants stored in these side valleys by day are recirculated by down-valley flows
at night.
Figure 5.10 indicates typical ozone plumes before and after a program of NOx
and VOC emission controls. The overall shape of both plumes is strongly affected
by topographic barriers, and sea breeze channeling. The eastward (downwind)
displacement of the more recent plume is hypothesized to be a result of changing
NOx /VOC ratios resulting in a less reactive chemical mixture under essentially
unchanged wind speeds (Ainslie and Steyn 2007).
278 D.G. Steyn et al.

Fig. 5.11 Three modes of pollutant recirculation in the Lower Fraser Valley. (a) Pollutants emitted
into the morning sea breeze are advected inland during the day and return by the landward arm of
the land breeze at night. (b) Pollutants emitted into the evening land breeze area advected seaward
during the night and return by the seaward arm of the sea breeze by day. (c) Pollutants emitted into
the sea breeze by day are advected landward (solid heavy arrow) and carried aloft by upslope flows
into the return flow, and are subsequently carried downward by subsidence (open arrow), entrained
into the convective boundary layer and mixed downwards (small curved arrows)

5.4.2 Upper Rhine Valley

The Upper Rhine Valley located in south western Germany, eastern France, and
northern Switzerland is a densely populated and industrialised area where complex
terrain meteorology and air quality issues have been of great interest for many years
(e.g. Fiedler and Prenosil 1980; Fiedler et al. 1987; Fiedler 1992a, b; Fiedler and
Zimmermann 1993; Fiedler and Borell 2000).
In September 1992, the TRACT (Transport of Air Pollutants over Complex
Terrain) field campaign was conducted in the Upper Rhine Valley providing a
dataset with a high temporal and spatial resolution of the CBL over mountainous
terrain (Fiedler 1992a; Fiedler and Borell 2000). Measurements from rawinsonde,
tethersonde, research aircraft and sodar (see Fig. 5.13) on a cross section through the
Black Forest Mountains were used to determine the spatial and temporal behavior of
the CBL (De Wekker 1995; Kossmann et al. 1998; Kalthoff et al. 1998). The terrain
is characterized by a sharp transition from the Rhine Valley to the Black Forest
region, with a height difference of roughly 1,000 m over a horizontal distance of
5 Boundary Layers and Air Quality in Mountainous Terrain 279

Fig. 5.12 Polluted boundary layer in LFV indicated by lidar backscatter superposed on profiles
of potential temperature (white solid line) and specific humidity (white dashed line) and inferred
AL height (dark line) measured on July 26, 2001 from 1314 to 1355 Pacific Standard Time. The
mountain slope can just be seen rising to the right (north) of the figure. Sodar data indicate winds
with a slight northerly component above the CBL top (Reuten et al. 2005) (© Springer. Reprinted
with permission)

about 10 km. The analyses showed that during the morning hours, the mixed layer
top followed the underlying terrain but generally tended to become more level,
i.e., less terrain-following, over the course of the day. In summary, three types of
afternoon CBL behavior were identified (Fig. 5.14). In the first type (typified by
panel 1), the mixed layer over the Rhine Valley reached heights of about 2,000 m
and the top of the mixed layer was relatively level over the investigation area. In
the second type (typified by panel 3), boundary layer development in the Rhine
Valley was delayed by the presence of fog, and the afternoon mixed layer stayed
well below the mountain height. Consequently, the top of the mixed layer more or
less followed the steep western slope of the Black Forest region in the afternoon.
In the third type (typified by panel 2), the behavior was intermediate between the
other types, with mixed layer heights over the Black Forest region a few hundred
meters higher than over the Rhine Valley. A common feature of all three types
of behavior is the presence of a relatively horizontal capping inversion over the
Black Forest region. The mixed layer top obviously does not follow the individual
small-scale ridges and valleys in the afternoon. Overall, smaller afternoon CBLs
were observed over the mountain top site Hornisgrinde (the highest elevation in
the investigation area) than in the Rhine Valley and the Black Forest region east of
the Hornisgrinde. Applying one-dimensional mixed layer growth rate models for a
valley and the mountain top site showed that for the valley site, relatively good cor-
respondence was found with observations but that the mixed layer growth rates for
280 D.G. Steyn et al.

Fig. 5.13 Vertical cross sections of potential temperature, specific humidity and horizontal wind
vectors from the foothills of the Vosges Mountains (left) through the Upper Rhine Valley (center)
to the Black Forest Mountains (right) at around 1300 Central European Summer Time (CEST)
of 11 September 1992 (De Wekker 1995). White lines in the upper two panels represent balloon
ascent paths or aircraft flight tracks
5 Boundary Layers and Air Quality in Mountainous Terrain 281

Fig. 5.14 Schematic


representation of three
different types of afternoon
mixed layer height (zi )
behaviour along a
cross-section from the Upper
Rhine Valley to the northern
Black Forest derived from
observations of the TRACT
field campaign (De Wekker
et al. 1997)

the mountain top site were clearly overestimated by these models (De Wekker 1995;
Kossmann et al. 1998). De Wekker (1995) and Binder (1997) proposed the use of an
effective sensible heat flux for use in growth rate models that implicitly takes into
account the effects of advection, orography shape and other effects that influence
CBL growth over mountain ridges.
The large density of surface-based meteorological stations during TRACT
allowed the identification of major wind systems in the Upper Rhine Valley and
in some of the adjacent side-valleys (Fig. 5.15). The wind systems in the side
valleys show a pronounced diurnal cycle with upvalley (downvalley) and upslope
(downslope) winds during the day (night). The influence of these flows on the ozone
concentrations during TRACT was studied by Löffler-Mang et al. (1997). Stations
in side-valleys of the upper Rhine Valley often showed less nocturnal ozone removal
(compared to stations on flat terrain) or even a clear secondary ozone maximum in
the night due to down-valley advection of ozone-rich and weakly polluted air from
the mountains.
282 D.G. Steyn et al.

60
50
40
y in km
3020
10
0

0 10 20 30 40 50 60 70
x in km

Daytime wind direction 10%


frequency distribution Height in m MSL

Fig. 5.15 Frequency distribution of daytime wind direction in the period of the TRACT field
campaign (7–21 September 1992), showing channeling and slope flows in the Upper Rhine Valley
(Kossmann 1998)

The behavior of ozone and it precursors was also investigated in the Schauinsland
Ozone Precursor Experiment (SLOPE) in 1996 in the southern part of the Upper
Rhine Valley near Freiburg (Kalthoff et al. 2000; Fiedler et al. 2000). Meteorological
and chemical measurements were carried out at several stations along the expected
transport path between Freiburg in the Rhine Valley and Schauinsland, a mountain
top location in the Black Forest. Results showed that not only mixing layer growth,
but also flow splitting and mountain venting processes, induced on the slopes and
in the side valleys play an important role in explaining the dispersion of pollutants
up the Black Forest Mountains. Kossmann et al. (1999) showed similar results for
5 Boundary Layers and Air Quality in Mountainous Terrain 283

the northern part of the upper Rhine Valley. Another important result from SLOPE
is the strengthening and lowering of an elevated inversion caused by the advection
associated with a thermally driven upvalley wind system. This process confined the
air pollutants to a shallow layer near the surface and degraded the air quality. Similar
interactions between thermally driven wind systems and CBL depth have been noted
by Kossmann et al. (1998) and De Wekker (2008).
The along-valley flow in the Upper Rhine Valley, which is evident in Fig. 5.15,
is mainly due to pressure-driven channeling (Wippermann 1984; Gross and
Wippermann 1987; Kalthoff and Vogel 1992). This type of channeling has been
given much attention in the case of the upper Rhine Valley, and has important
consequences for the dispersion of air pollutants (see Sect. 5.2.3).

5.5 Conclusion

Throughout this chapter we have stressed that air quality in regions of complex,
mountainous topography is frequently (and quite often severely) degraded because
of meteorological processes and phenomena related to the topography. While we
have chosen only two regions as examples to illustrate this, there exist many other
locations that experience degraded air quality because of topographic effects. The
difficulties of both atmospheric measurement and modeling in complex terrain
mean that there will always be substantial uncertainty over air pollution problems
in complex terrain. This uncertainty translates into difficulties in both forecasting
pollution, and providing scientific guidance to policy makers.
The current practice for routine air quality forecasts is to employ numerical
models operating over regional and continental scales, as exemplified by Kang et al.
(2009). The horizontal resolution of such forecasts is generally 10–20 km, a scale at
which most of the phenomena discussed in this chapter are simply not resolved. This
means that anyone wanting to forecast air quality in a mountainous region will have
to take the forecast provided by model output, and interpret or modify it taking into
account the phenomena and effects covered in this chapter. As these models develop
and increase their resolution, and their ability to explicitly capture the phenomena
discussed in this chapter, such forecaster intervention will become less important. It
is unlikely however that it will ever become unnecessary.
It must be emphasized that the quality of an air pollution forecast based on
numerical model output will depend on a number of factors. The first among these
will be the requirement of accurate model initialization and boundary conditions
for both meteorology and atmospheric chemistry. These may be provided through
data assimilation schemes, but chemical data assimilation requires measurements of
many chemical species, which generally are not measured with sufficient detail. The
second requirement is for accurate spatially, temporally and chemically resolved
emissions data. Such data are generally not available in sufficient resolution. This
difficulty is in most cases the major factor limiting the accuracy of air quality
forecasts. The third requirement is for detailed and chemically realistic models of
284 D.G. Steyn et al.

aerosol chemistry and physics. This matter is a topic for much current research,
and therefore a source of much uncertainty in the context of operational air quality
forecasts. These three requirements are seldom uniformly met, and so efforts to
determine the source of deficiencies in air quality forecasts are made especially
difficult because of interactions between multiple sources of model weaknesses. It is
often unclear whether inadequacies in air quality models are due to meteorological,
chemical or emissions (or all three) weaknesses.
An additional difficulty arises because numerical models needed for air quality
forecasts must perform both meteorological and chemical transformation calcula-
tions, resulting in a great increase in computational demand over pure meteorolog-
ical modeling. The result often is that the spatial resolution is sacrificed in order to
allocate computational resources to the chemical modeling. The resulting reduction
in spatial resolution can result in mountain induced meteorological processes being
inadequately resolved.
Another consequence of the uncertainty surrounding air quality in mountainous
terrain is that most studies of air pollution in such terrain combine both observations
and modeling as approaches to building understanding. Both approaches will always
be needed, with observations to tie model output to measured reality, and models
to fill in the unmonitored spaces. Many phenomena become evident only through
observation, and then subsequently, are better understood through the use of models,
which then become the most important tools for making predictions, and exploring
scenarios. Models and observations are tied in an upward spiraling process of
development, and nowhere is this as evident as in studies of air pollution in
complex terrain. In this way, the increasing demands of air quality forecasting will
inspire further development of new parameterizations of meteorological processes,
which will in turn make possible more accurate air quality forecasts in complex,
mountainous terrain
Acknowledgements The authors would like to thank the three anonymous reviewers and the book
editors for their helpful comments that improved this chapter. We also would like to thank Prof.
Franz Fiedler from the ‘Institute for Meteorology and Climate Research’ of the ‘Karlsruhe Institute
of Technology’ (Germany) for his permission to use Fig. 5.4, Dr. Jerome Fast for his permission to
use Fig. 5.3, and Eric Leinberger who drafted Figs. 5.1 and 5.11.

References

Ainslie, B and D.G. Steyn, 2007: Spatiotemporal Trends in Episodic Ozone Pollution in the Lower
Fraser Valley, B.C. in Relation to Meso-scale Atmospheric Circulation Patterns and Emissions.
J. Appl. Meteorol. 46 (10), 1631–1644.
Arritt, R.W., J.M. Wilczak, and G.S. Young, 1992: Observations And Numerical Modeling Of An
Elevated Mixed Layer. Mon. Wea. Rev.,120, 2869–2880.
Bastin, S., P. Drobinski, A. Dabas, P. Delville, O. Reitebuch, and C. Werner, 2004: Impact of
the Rhône and Durance valleys on sea-breeze circulation in the Marseille area, Atmospheric
Research 74, 303–328.
5 Boundary Layers and Air Quality in Mountainous Terrain 285

Banta, R.M., P.B. Shepson, J.W. Bottenheim, K.G. Anlauf, H.A. Wiebe, A. Gallant, T. Biesenthal,
L.D. Olivier, C-J. Zhu, I.G. McKendry, and D.G. Steyn, 1997: Nocturnal Cleansing Flow In
Tributary Valley. Atmos. Env. 31, 2147–2162.
Bao, J.-W., S.A. Michelson, P.O.G. Persson, I.V. Djalalova, and J.M. Wilczak, 2008: Observed And
WRF-Simulated Low-Level Winds In A High Ozone Episode During The Central California
Ozone Study. J. Appl. Meteorol. Climatol., 47, 2372–2394.
Baumbach, G. and U. Vogt, 1999: Experimental Determination Of The Effect Of Mountain-Valley
Breeze Circulation On Air Pollution In The Vicinity Of Freiburg. Atmos. Env. 33, 4019–4027.
Binder, H.-J., 1997: Tageszeitlich Und Räumliche Entwicklung Der Konvektiven Grenszicht Über
Stark Gegleidertem Gelände. Dissertation, Institut für Meteorologie und Klimaforschung,
Karlsruhe, 310p.
Bischoff-Gauß, I., N. Kalthoff, and F. Fiedler, 1998: The Impact Of Secondary Flow Systems On
Air Pollution In The Area Of Sao Paulo. J. Appl. Meteorol., 37, 269–287.
Chen, R.-R., N. S. Berman, D. L. Boyer, and H. J. S. Fernando, 1996: Physical model of diurnal
heating in the vicinity of a two-dimensional ridge, J. Atmos. Sci., 53(1), 62–85.
Cogliati, M.G. and N.A. Mazzeo, 2006: Air Flow Analysis In The Upper Rio Negro Valley
(Argentina). Atmos. Res., 80, 263–279.
Corsmeier, U., M. Kossmann, A.P. Sturman, and N. Kalthoff, 2006: Temporal Evolution Of Winter
Smog Within A Nocturnal Boundary Layer At Christchurch, New Zealand. Meteorol. Atmos.
Phys., 91, 129–148.
de Foy, B., E. Caetano, V. Magana, A. Zitacuaro, B. Cardenas, A. Ratema, R. Ramos, L.T. Molina,
and M.J. Molina, 2005: Mexico City Basin Wind Circulation During The MCMA-2003 Field
Campaign. Atmos. Chem. Phys., 5, 2267–2288.
De Wekker, S.F.J., 1995: The Behaviour Of The Convective Boundary Layer Height Over
Orographically Complex Terrain. MS thesis University of Karlsruhe, Germany/Wageningen
Agricultural University, The Netherlands, 74 pp.
De Wekker, S.F.J., Kossmann, M., Fiedler, F., 1997: Observations Of Daytime Mixed Layer Heights
Over Mountainous Terrain During The TRACT Field Campaign. 12th AMS Symposium
on Boundary Layers and Turbulence, 28 July – 1 August 1997, Vancouver, B.C., Canada,
498–499.
De Wekker, S.F.J., S. Zhong, J.D. Fast, and C.D. Whiteman, 1998: A Numerical Study Of
The Thermally Driven Plain-To-Basin Wind Over Idealized Basin Topographies. J. Appl.
Meteorol., 37, 606–622.
De Wekker, S.F.J., 2002: Structure and Morphology The Convective Boundary Layer in Mountain-
ous Terrain. PhD thesis The University of British Columbia, Canada, 191 pp.
De Wekker, S.F.J., D.G. Steyn, and S. Nyeki, 2004: A Comparison Of Aerosol Layer- And
Convective Boundary Layer Structure Over A Mountain Range During STAAARTE’97.
Bound.-Layer Meteor., 113, 249–271.
De Wekker, S.F.J., 2008: Observational And Numerical Evidence Of Depressed Convective
Boundary Layer Heights Near A Mountain Base. J. Appl. Meteorol. Climatol., 47, 1017–1026.
De Wekker, S.F.J. and S. D. Mayor, 2009: Observations of atmospheric structure and dynamics in
the Owens Valley of California with a ground-based, eye-safe, scanning aerosol lidar. J. Appl.
Meteorol. climatol., 48, 1483–1499.
Dobrinski, P., S. Bastin, J. Dusek, G. Zängl, and P.H. Flamant, 2006: Flow Splitting At The
Bifurcation Between Two Valleys: Idealized Simulations In Comparison With Mesoscale
Alpine Programme Observations. Meteorol. Atmos. Phys., 92, 285–306.
Eckman, R.M., 1998: Observations And Numerical Simulations Of Winds Within A Broad
Forested Valley J. Appl. Meteorol., 37, 206–219.
Eder, B., D. Kang, R. Mathur, S. Yu, and K. Schere, 2006: An operational evaluation of the Eta-
CMAQ air quality forecast model. Atmos. Env. 40, 4894–4905.
Eder, B., D. Kang, S. Trivikrama Rao, R. Mathur, S. Yu, T. Otte, K. Schere, R. Wayland, S. Jackson,
P. Davidson, J. McQueen, and G. Bridgers, 2010: Using National Air Quality Forecast
Guidance to Develop Local Air Quality Index Forecasts. Bull. Amer. Met. Soc. 2010; 91:
313–326.
286 D.G. Steyn et al.

Elsom, Derek, 1987: Atmospheric Pollution: Causes, Effects, and Control Policies. Blackwell,
Oxford, UK, 1987, 319 p.
Enger, L., D. Koracin, and X. Yang, 1993: A Numerical Study Of Boundary-Layer Dynamics In A
Mountain Valley. Bound.-Layer Meteor., 66, 357–394.
Eugster, W. and F. Siegrist, 2000: The Influence Of Nocturnal CO2 Advection On CO2 Flux
Measurements. Basic Appl. Ecol., 1, 177–188.
Fast, J. D. and S. Zhong, 1998: Meteorological Factors Associated With Inhomogeneous Ozone
Concentrations Within The Mexico City Basin. J. Geophys. Res., 103, 18927–18946.
Fiedler, F.,1983: Einige Charakteristika der Strömung in Oberrheingraben Wissenschafliches
Berichte des Meteorologischen Instituts der Universität Karlsruhe. Nr. 4, 1130123.
Fiedler, F., Adrian, G., Hugelmann, C.P., 1987: Beobachtete Phänomene Während des TULLA-
Experimentes [Phenomena Observed During the TULLA Experiment], 3 rd PEF Statuskollo-
quium, 10–12 March 1987 Karlsruhe, KfK-PEF Vol. 12, Issue 2, 347–365.
Fiedler, F., 1992a: TRACT Operational Plan, Institut für Meteorologie und Klimaforschung,
Karlsruhe, 66p.
Fiedler, F., 1992b: Das Regio-Klima Projekt: Wie Regeln die Natürlichen Energieumsetzungen das
Klima in Einer Region. KfK-Nachr., Jahrg. 24, 3/92, 125–131.
Fiedler, F. and Prenosil, 1980: Das Mesoklip-Experiment, mesoskaliges Klimaprogramm im Ober-
rheintal. Wissenschaftliche Berichte des Meteorologischen Instituts der Universität Karlsruhe,
Nr.1.
Fiedler, F. and Zimmermann, H., 1993: General Report of TRACT, an Overview. in: Annual report
1992 EUROTRAC subproject TRACT, International Scientific Secretariat, Fraunhofer-Institut
für Atmosphärische Umweltforschung, Garmisch-Partenkirchen, 3–34.
Fiedler, F. and P. Borell, 2000: TRACT: Transport of Air Pollutants over Complex Terrain.
In; S, Larsen et al. (Eds.), Exchange and Transport of Air Pollutants over Complex Terrain
and the Sea, Springer-Verlag, 223–268.
Fiedler, F., I. Bischoff-Gauß, N. Kalthoff, and G. Adrian (2000), Modeling Of The Transport
And Diffusion Of A Tracer In The Freiburg-Schauinsland Area, J. Geophys. Res., 105(D1),
1599–1610.
Gross, G., and F. Wippermann, 1987: Channeling and Countercurrent in the Upper Rhine Valley:
Numerical Simulations. J. Appl. Meteorol., 26, 1293–1304.
Frioud, M., V. Mitev, R. Matthey, C. Häberli, H. Richner, R. Werner and S. Vogt, 2003: Elevated
Aerosol Stratification Above The Rhine Valley Under Strong Anticyclonic Conditions, Atmos.
Env. 37, 1785–1797.
Grubišić, V., Doyle, J.D., Kuettner, J., Mobbs, S., Smith, R.B., Whiteman, C.D., Dirks, R., Czyzyk,
S., Cohn, S.A., Vosper, S., Weissmann, M., Haimov, S., De Wekker, S.F.J., Pan, L.L., and Chow,
F.K., 2008: The Terrain-Induced Rotor Experiment – A Field Campaign Overview Including
Observational Highlights. Bull. Amer. Met. Soc., 89, 1513–1533.
Hageli, P., D.G. Steyn and K. Strawbridge, 2000: Spatial and Temporal Variability of MLD and
EZT. Boundary Layer Meteorol., 97 (1), 47–71.
Harada, A., 1981: An Analysis Of The Nocturnal Cyclonic Vortex In The Planetary Boundary
Layer Of The Kanto Plains. J. Met. Soc. Japan 59, 602–610.
Henne, S., M. Furger, S. Ny eki, M. Steinbacher, B. Neininger, S.F.J. De Wekker, J. Dommen,
N. Spichtinger, A. Stohl, and A. S. H. Prévôt, 2004: Quantification Of Topographic Venting Of
Boundary Layer Air To The Free Troposphere. Atmos. Chem. Phys., 4, 497–509.
Huddlestone, B., E. Ataman, P. de Salvo, M. Zanetti, M. Bloise, J. Bel, G. Franceschini and L. Fè
d’Ostiani, 2003: Towards a GIS-Based Analysis of Mountain Environments and Populations.
FAO, UN, Rome. 26p.
Kalthoff, N. and B. Vogel, 1992: Counter-Current And Channelling Effect Under Stable Stratifica-
tion In The Area Of Karlsruhe. Theor. Appl. Climatol., 45, 113–126.
Kalthoff, N., C. Kottmeier, J. Thürauf, U. Corsmeier, F. Sad, E. Fréjafon and P.E. Perros, 2004:
Mesoscale Circulation Systems And Ozone Concentrations During ESCOMPTE: A Case Study
From IOP 2b. Atmos. Res., 74, 355–380.
5 Boundary Layers and Air Quality in Mountainous Terrain 287

Kalthoff, N., H.-J. Binder, M. Kossmann, R. Vögtlin, U. Corsmeier, F. Fiedler, and H. Schlager,
1998: Temporal Evolution And Spatial Variation Of The Boundary Layer Over Complex
Terrain. Atmos. Env., 32, 1179–1194.
Kalthoff, N., V. Horlacher, U. Corsmeier, A. Volz-Thomas, B. Kolahgar, H. Geiß, M. Möllmann-
Coers, and A. Knaps (2000), Influence Of Valley Winds On Transport And Dispersion
Of Airborne Pollutants In The Freiburg-Schauinsland Area, J. Geophys. Res., 105(D1),
1585–1597.
Kang, D., R. Mathur, and S. Trivikrama Rao, 2009: Assessment of bias-adjusted NAQFC PM2.5 air
quality forecasts over the continental United States during 2007 Geosci. Model Dev. Discuss.,
2, 1–32.
Kimura F., 1986: Formation Mechanism of the Nocturnal Mesoscale Vortex in Kanto Plain. J. Met.
Soc. Japan, 64, 857–869.
Kimura, F. and T. Kuwagata, 1993: Thermally Induced Wind Passing From Plain To Basin Over A
Mountain Range. J. Appl. Meteorol., 32, 1538–1547.
Kossmann, M. and A.P. Sturman, 2003: Pressure-Driven Channeling Effects In Bent Valleys.
J. Appl. Meteorol., 42, 151–158.
Kossmann, M., 1998: Einfluss Orographisch Induzierter Transportprozesse Auf Die Struktur Der
Atmosphärischen Grenzschicht Und Die Verteilung Von Spurengasen. Dissertation, Physics
Faculty, University of Karlsruhe, Germany, 193 pp.
Kossmann, M., R. Vögtlin, U. Corsmeier, B. Vogel, F. Fiedler, H. -J. Binder, N. Kalthoff, and
F. Beyrich, 1998: Aspects Of The Convective Boundary Layer Structure Over Complex Terrain.
Atmos. Env., 32, 1323–1348.
Kossmann, M., Whiteman C.D. and Bian X., 2002a: Dynamic Airflow Channeling Over The Snake
River Plain, Idaho. Proceedings Tenth Conference on Mountain Meteorology, 17–21 June
2002, Park City, Utah, Amer. Meteor. Soc., 360–363.
Kossmann, M., A.P. Sturman, P. Zawar-Reza, H.A. McGowan, A.J. Oliphant, I.F. Owens, and
R.A. Spronken-Smith, 2002b: Analysis of the Wind Field and Heat Budget in an Alpine Lake
Basin during Summertime Fair Weather Conditions. Meteorol. Atmos. Phys., 81, 27–52.
Kossmann, M. and A.P. Sturman, 2004: The Surface Wind Field During Winter Smog Nights In
Christchurch And Coastal Canterbury, New Zealand. Int. J. Climatol., 24, 93–108.
Kossmann, M., U. Corsmeier, S.F.J. de Wekker, F. Fiedler, R. Vögtlin, N. Kalthoff, H. Güsten, and
B. Neininger, 1999. Observations Of Handover Processes Between The Atmospheric Boundary
Layer And The Free Troposphere Over Mountainous Terrain. Contrib. Atmos. Phys. 72,
329–350.
Kurita, H., Ueda, H., and Mitsumoto, S., 1990: Combination Of Local Wind Systems Under Light
Gradient Wind Conditions And Its Contribution To Long-Range Transport Of Air Pollutants.
J. Appl. Meteorol., 29, 331–348.
Kuwagata, T. and M. Sumioka, 1991: The Daytime Heating Process Over Complex Terrain In
Central Japan Under Fair And Calm Weather Conditions. Part III: Daytime Thermal Low And
Nocturnal Thermal High. J. Met. Soc. Japan, 69, 91–103.
Lenschow, D. H., B. B. Stankov, and L. Mahrt, 1979: The Rapid Morning Boundary-Layer
Transition. J. Atmos. Sci., 36, 2106–2124.
Levinson, D.H. and R.M. Banta, 1995: Observations Of A Terrain-Forced Mesoscale Vortex And
Canyon Drainage Flows Along The Front Range of Colorado. Mon. Wea. Rev., 123, 2029–2050.
Lin, Y.-L. and I.-C. Jao, 1995: A Numerical Study Of Flow Circulations In The Central Valley Of
California And Formation Mechanisms Of The Fresno Eddy. Mon. Wea. Rev., 123, 3227–3239.
Löffler-Mang, M., M. Kossmann, R. Vögtlin, and F. Fiedler, 1997: Valley Wind Systems And Their
Influence On Nocturnal Ozone Concentrations. Contr. Atmos. Phys., 70, 1–14.
Lu, R. and R.P. Turco, 1994: Air Pollution Transport In A Coastal Environment. Part I: Two-
Dimensional Simulation Of Sea-Breeze And Mountain Effects. J. Atmos. Sci., 51, 2285–2308.
McGowan, H.A. and A.P. Sturman, 1996: Interacting Multi-Scale Wind Systems Within An Alpine
Basin, Lake Tekapo, New Zealand. Meteor. Atmos. Phys. 58, 165–177.
McGregor, J.L. and F. Kimura, 1989: Numerical Simulation Of Mesoscale Eddies Over Melbourne.
Mon. Wea. Rev., 117, 50–66.
288 D.G. Steyn et al.

McKee, T.B. and R.D. O’Neal, 1989: The Role Of Valley Geometry And Energy Budget In The
Formation Of Nocturnal Valley Winds. J. Appl. Meteorol., 28, 445–456.
McKendry, I.G., D. G. Steyn, J. Lundgren, R. M. Hoff, W. Strapp, K. Anlauf, F. Froude, J. B.
Martin, R.M. Banta and L.D. Olivier, 1997: Elevated Ozone Layers And Vertical Down-Mixing
Over The Lower Fraser Valley, BC. Atmos. Env., 31, 2135–2146.
McKendry, I.G. and J. Lundgren, 2000: Tropospheric Layering Of Ozone In Regions Of Urbanized
Complex And/Or Coastal Terrain: A Review. Prog. Phys. Geog. 24, 329–354.
Millan, M.M., R. Salvador, E. Mantilla, and G. Kallos, 1997: Photooxidant Dynamics In The
Mediterranean Basin In Summer – Results From European Research Project. J. Geophys.
Res., D102, 8811–8823.
Neu, U., T. Künzle, and H. Wanner, 1994: On The Relation Between Ozone Storage In The
Residual Layer And Daily Variation In Near-Surface Ozone Concentration – A Case Study.
Bound.-Layer Meteor., 69, 221–247.
Oettl, D., R.A. Almbauer, P.J. Sturm, M. Piringer, and K. Baumann, 2001: Analysing The
Nocturnal Wind Field In The City Of Graz. Atmos. Env., 35, 379–387.
Oke, T.R. and J.E. Hay, 1994: The Climate of Vancouver. B.C. Geographical Series, Number
50, 84p.
Ohashi, Y. and H. Kida, 2002: Effects Of Mountains And Urban Areas On Daytime Local-
Circulations In The Osaka And Kyoto Regions. J. Met. Soc. Japan, 80, 539–560.
Petkovsek, Z., 1992: Turbulent Dissipation Of Cold Air Lake In A Basin. Meteorol. Atmos.
Phys., 47, 237–245.
Plate, E., 1982: Engineering Meteorology. Elsevier, Amsterdam, 740 p.
Poulos, G.S, J.E. Bossert, T.B. McKee, and R.A. Pielke, 2000: The Interaction Of Katabatic Flow
And Mountain Waves. Part I: Observations And Idealized Simulations. J. Atmos. Sci., 57,
1919–1936.
Poulos, G. and S. Zhong, 2008: An Observational History Of Small-Scale Katabatic Winds In
Mid-Latitudes. Geog. Compass 2, 1–24.
Raga, G.B., D. Baumgardner, G. Kok, and I. Rosas, 1999: Some Aspect Of Boundary Layer
Evolution In Mexico City. Atmos. Env. 33, 5013–5021.
Regmi, R.P., T. Kitada, and G. Kurata, 2003: Numerical Simulation Of Late Wintertime Local
Flows In Kathmandu Valley, Nepal: Implication For Air Pollution Transport. J. Appl. Meteo-
rol., 42, 389–404.
Reuten, C., D.G. Steyn, K.B. Strawbridge, and P. Bovis, 2005: Observed trapping of pollutants in
upslope flow systems. Bound.-Layer Meteorol., 116 (1), 37–61.
Reuten, C., D.G. Steyn, and S. E. Allen. 2007: Water Tank Studies of Atmospheric Boundary
Layer Structure and Air-Pollution Transport in Upslope Flow Systems. J. Geophys. Res., 112,
D11114.
Sakiyama, S.K., 1990: Drainage Flow Characteristics And Inversion Breakup In Two Alberta
Mountain Valleys. J. Appl. Meteorol., 29, 1015–1030.
Salmond, J. A. and I.G. McKendry, 2002: Secondary Ozone Maxima In A Very Stable Nocturnal
Boundary Layer: Observations From The Lower Fraser Valley, BC. Atmos. Env. 36. 5771–5782.
Schmidli, J., G.S. Poulos, M.H. Daniels, and F.K. Chow, 2009: External Influences On Nocturnal
Thermally Driven Flows In A Deep Valley. J. Appl. Meteorol. Climatol., 48, 3–23.
Seibert, P., H. Kromp-Kolb, A. Kasper, M. Kalina, H. Puxbaum, D.T. Jost, M. Schwikowski, and
U. Baltensperger, 1998: Transport of polluted boundary layer air from the Po Valley to high-
alpine sites. Atmos. Env., 32, 3953–3965.
Smedman, A.S., H. Bergström, and U. Högström, 1996: Measured And Modelled Local Wind
Fields Over A Frozen Lake In A Mountainous Area. Contr. Atmos. Phys. 69, 501–516.
Spillane, K., 1978: Atmospheric Characteristics On High Oxidant Days In Melbourne. Clean Air
(Aust) 12, 50–56.
Stensrud, D.J. 1993: Elevated Residual Layers And Their Influence On Surface Boundary-Layer
Evolution. J. Atmos. Sci., 50, 2284–2293.
Steyn D. G., and T. R. Oke, 1982: The Depth Of The Daytime Mixed Layer At Two Coastal Sites:
A Model And Its Validation. Bound.-Layer Meteor., 24, 161–180.
5 Boundary Layers and Air Quality in Mountainous Terrain 289

Steyn, D.G., J.W. Bottenheim, and R.B. Thomson, 1997: Overview Of Tropospheric Ozone In The
Lower Fraser Valley, And The PACIFIC ‘93 Field Study. Atmos. Env., 31 (4), 2025–2036.
Stull, R.B., 1998: Introduction to Boundary Layer Meteorology. Kluwer, Dordrecht. 666 p.
Sun, J., R. Desjardins, L. Mahrt, and I. MacPherson, 1998: Transport of carbon dioxide, water
vapor, and ozone by turbulence and local circulations, J. Geophys. Res., 103(D20), 25,
873–25,885.
Troude, F., E. Dupont, B. Carissimo, and A. Flossmann, 2002: Relative Influence Of Urban And
Orographic Effects For Low Wind Conditions In The Paris Area. Bound.-Layer Meteor., 103,
493–505.
Whiteman, C. D., 1982: Breakup Of Temperature Inversions In Deep Mountain Valleys: Part I.
Observations. J. Appl. Meteorol., 21, 270–289.
Whiteman, C.D., T.B. McKee, and J.C. Doran, 1996: Boundary Layer Evolution Within A
Canyonland Basin. Part 1. Mass, Heat, And Moisture Budgets From Observations. J. Appl.
Meteorol., 35, 2145–2161.
Whiteman, C.D., S. Zhong, X. Bian, J.D. Fast, and J.C. Doran, 2000: Boundary Layer Evolution
And Regional-Scale Diurnal Circulations Over The Mexico Basin And Mexican Plateau.
J. Geophys. Res., 105, 10081–10102.
Whiteman, C.D., S. Zhong, W.J. Shaw, J.M. Hubbe, and X. Bian, 2001: Cold Pools In The
Columbia Basin. Weather and Forecasting 16, 432–447.
Wilczak, J.M. and T.W. Christian, 1990: Case Study Of An Orographically Induced Mesoscale
Vortex (Denver Cyclone). Mon. Wea. Rev., 118, 1082–1102.
Wippermann F., 1984: Air Flow Over And In Broad Valleys: Channelling And Counter-Current.
Contr. Atmos. Phys., 57, 92–105.
Zaremba, L.L. and J.J. Carroll, 1999: Summer Wind Flow Regimes Over The Sacramento Valley.
J. Appl. Meteorol., 38, 1463–1473.
Zawar-Reza, P. and A.P. Sturman, 2006 Two-Dimensional Numerical Analysis Of A Thermally
Generated Mesoscale Wind System Observed In The Mackenzie Basin, New Zealand.
Australian Meteorol. Mag. 55, 19–34.
Zhong, S., X. Bian, and C.D. Whiteman, 2003: Time Scale For Cold-Air Pool Breakup By
Turbulent Erosion. Meteorolog. Zeit. 12, 229–233.
Zhong, S., J. Li, C.D. Whiteman, X. Bian, and W. Yao, 2008: Climatology Of High Wind Events
In Owens Valley, California. Mon. Wea. Rev., 136, 3536–3552.
Chapter 6
Theory, Observations, and Predictions
of Orographic Precipitation

Brian A. Colle, Ronald B. Smith, and Douglas A. Wesley

Abstract There have been rapid advances in the understanding of orographic


precipitation during the past two decades given the advent of high resolution
mesoscale modeling and numerous field studies around the world. This chapter
begins with introducing the fundamental ingredients, processes, and scaling param-
eters for orographic precipitation. If the low level air is stable, terrain flow blocking
can occur, which shifts the precipitation enhancement upwind of the barrier. Low-
level diabatic cooling from evaporation and melting can enhance the near-surface
stability and flow blocking. In contrast, latent heating within the cloud reduces
the effective stability and the potential for flow blocking, which in turn results in
stronger vertical motions over the ridges and increases the riming and precipitation
fallout.
For less stable flows, there are mountain gravity waves over the mountain that
can change the depth and upstream extent of the orographic cloud, and lead to
precipitation enhancements over individual windward ridges. Barrier dimension is
also important, with wide barriers favoring precipitation enhancement upwind of
the crest and more efficient water vapor removal from the ambient flow. If the ridges
are very narrow, the mountain waves decay with height and have less impact on the
precipitation. For unstable flow, the upslope flow can trigger convective systems or
quasi-stationary banded precipitation.
The various orographic processes are illustrated using several field study and
modeling results. The prediction of orographic precipitation is discussed using a

B.A. Colle ()


School of Marine and Atmospheric Sciences, Stony Brook University/SUNY, Stony Brook,
NY 11794-5000, USA
e-mail: brian.colle@stonybrook.edu
R.B. Smith
Geology and Geophysics Department, Yale University, New Haven, CT 06520-8109, USA
D.A. Wesley
Compass Wind, 1730 Blake St., Ste. 400, Denver, CO 80202, USA

F. Chow et al. (eds.), Mountain Weather Research and Forecasting, 291


Springer Atmospheric Sciences, DOI 10.1007/978-94-007-4098-3 6,
© Springer ScienceCBusiness Media B.V. 2013
292 B.A. Colle et al.

1.33-km grid spacing simulation and sensitivity results from a linear orographic
precipitation model for a flooding event over southwest Washington in early
November 2006. The chapter concludes with a summary and areas for future work.

6.1 Introduction

Orographic precipitation can be defined as the modification of rain, snow, and other
hydrometeors resulting from the interaction of moist flow with topography. As
noted by Sawyer (1956), the three important factors for orographic precipitation
are: (1) the interaction of the ambient flow with terrain, (2) cloud microphysical
processes, and (3) larger-scale atmospheric circulation. For enhanced precipitation
growth to occur over the windward slope (Fig. 6.1), there is often a synergistic
relationship between the large-scale ascent associated with an approaching baro-
clinic or tropical cyclone, and the moist dynamical flow (upslope) response, which
is dependent on the magnitude of the upstream flow orientated perpendicular to the
mountain, moisture, and stratification (Fig. 6.1). The microphysical processes and
the time scale for available ice/water growth within the cloud can also modify the
precipitation distribution. Atmospheric models and operational forecasters need to
consider all of these factors in order to forecast precipitation accurately in areas of
steep terrain.
There has been a rapid advance in the understanding of orographic precipitation
with the application of high-resolution models and new observational field datasets
collected over various topographic barriers around the world (Fig. 6.2). During the
last few decades these field studies have quantified various orographic precipitation
processes for a broad spectrum of ambient environments and barrier dimensions
(Table 6.1). For example, the Cascade Project in the early 1970s examined oro-
graphic airflow and microphysics over the Washington Cascades using aircraft
(Hobbs et al. 1973; Hobbs 1975). During the 1980s, the Sierra Cooperative Pilot
Project (SCPP) studied the airflow and precipitation processes over the Sierra
Nevada Mountains of California using both ground-based radar and aircraft in situ
sensors (Reynolds and Dennis 1986).
The mesoscale flow and precipitation structures of storms approaching the U.S.
West coast were also studied during the Coastal Observations and Simulation with
Topography (COAST) experiments in December 1993 and 1995 (Bond et al. 1997)
and in the California Land-Falling Jets (CALJET) and Pacific Land-Falling Jets
(PACJET) experiments in 1998 and 2000–2002, respectively (Ralph et al. 1999;
Neiman et al. 2002). Using airborne dual-Doppler radar data around the Olympic
Mountains in COAST, Colle and Mass (1996) showed that three-dimensional winds
can be synthesized from airborne Doppler radar over steep terrain. This helped put
the precipitation in context with the three-dimensional flow, which motivated other
subsequent orographic field experiments to use this airborne Doppler technology.
For example, the IMPROVE II (Improvement of Microphysical PaRameterization
through Observational Verification Experiment II) field program investigated the
6 Theory, Observations, and Predictions of Orographic Precipitation 293

Fig. 6.1 Schematic diagram of the elements of orographic precipitation: upstream water vapor
flux (Fw ), large-scale ascent, windward ascent, streamlines of theta-E, snow and graupel fallout,
and windward precipitation within a distance P

Fig. 6.2 Location of orographic precipitation field projects around the world that involved radar,
aircraft, and/or isotope measurements in areas of terrain (shaded). See Table 6.1 for a detailed list
of the experiments and their focus
Table 6.1 Major orographic precipitation field studies around the world (shown on Fig. 6.2) that utilized radar, aircraft, and/or isotope measurements
294

Number Project title Location Year Objective


1 N/A South Wales 1970s Precipitation enhancement over small hills
2 Sierra – Cooperative Project Sierra Mountains late 1970s–1980s Precipitation evolution, thermodynamics, and
microphysics over a wide barrier with
landfalling storms
3 Taiwan Area Mesoscale Experiment Taiwan 1987 The effects of orography on the Mei-Yu front and
(TAMEX) on mesoscale convective systems
4 Hawaiian Rainband Experiment (HARP) Hawaii 1990 Diurnal variability of precipitation around
isolated topography
5 Coastal Observation and Simulation with Olympic Mountains 1993 Precipitation and kinematic evolution over
Topography (COAST) isolated barrier
6 Winter Icing Storms Project (WISP) Colorado Front 1994 Ice nucleation, super-cooled water, and
Range precipitation evolution along the Front Range
7 Southern Alpine Precipitation Experiment New Zealand Alps 1996 Understanding the processes through which the
(SALPEX) Southern Alps (narrow steep barrier)
influence precipitation
8 Mesoscale Alpine Programme (MAP) Alps 1999 Linkage of moist dynamics (blocked and
unblocked flow) with precipitation
distribution and processes
9 California Landfalling Jets Experiment California coastal 1998, 2000–2001 Coastal precipitation enhancement, warm rain
CALJET/Pacific Jets (PACJET) mountains processes, and atmospheric rivers
10 Improvement of Microphysical Central Oregon 2001 Precipitation processes and microphysics, and
Parameterization through Cascades the role of gravity waves on precipitation
Observational Verification Experiment
(IMPROVE-2)
11 Intermountain Precipitation Experiment Wasatch Mountains 2001 Precipitation processes, diabatic impacts, and
(IPEX) microphysics for narrow barrier
12 Southern Andes Project Southern Andes 2005 Air mass transformation and isotope analysis
13 Convective Orographic Precipitation Southwest Germany 2007 Study orographically-induced convective
Experiment (COPS) precipitation
B.A. Colle et al.
6 Theory, Observations, and Predictions of Orographic Precipitation 295

microphysics and orographic cloud structures for frontal systems crossing the
Oregon Cascades (Stoelinga et al. 2003). Baroclinic storm systems and moist
southerly flow impinging on the Alps were documented in autumn 1999 during the
Mesoscale Alpine Programme (MAP, Bougeault et al. 2001; Rotunno and Houze
2007). Meanwhile, the impact of more narrow and steep barriers on orographic
precipitation, such as the Wasatch in Utah (IPEX, Schultz et al. 2002) and southern
Andes (Smith and Evans 2007) have also been investigated. More tropical envi-
ronments associated with three-dimensional island geometries were investigated
in Taiwan during TAMEX (Kuo and Chen 1990) and Hawaiian Rainband Project
(HARP) (Chen and Nash 1994). More recently, convective initiation over terrain was
investigated over southwest Germany during the Convective and Orographically-
induced Precipitation Study (COPS) experiment (Wulfmeyer et al. 2008).
Significant loss of life and disruptions to the local economy can occur when
flash flooding, avalanches, and transportation delays occur in association with heavy
orographic precipitation as outlined in Chap. 1. There have been several flash
convective flood events in areas of steep terrain, such as the southern slopes of
the European Alps (Lin et al. 2001), southern Appalachians (Pontrelli et al. 1999),
and the Colorado Front Range (Petersen et al. 1999). Along the mountainous U.S.
West Coast, a plume of subtropical moisture associated with a landfalling baroclinic
wave can interact with the steep coastal terrain to produce flooding over Pacific
Northwest (Colle and Mass 2000; Lackmann and Gyakum 1999) and the California
coast (Ralph et al. 2003). Meyers et al. (2003) and Poulos et al. (2002) investigated
a heavy snow and blizzard event over the Front Range of the Rocky Mountains in
1997 that severely disrupted transportation during the 24-h snowfall event.
Understanding the fine-scale precipitation structures over a topographic barrier
is also important for hydrologic streamflow and river modeling. For example,
Westrick et al. (2002) illustrated that the streamflow forecasts over the Washington
Olympics and Cascades were sensitive to the observed and simulated precipitation
distributions on the ridge and valley scale. Ralph et al. (2003) also showed that
the peak river flows during CALJET were highly sensitive to the exact location of
the windward enhancements and rain shadows. Landfalling hurricanes often create
heavy rainfall and flooding in areas of terrain, such as hurricane Agnes along the
Appalachians (Carr and Bosart 1978), and hurricane Dean over the mountainous
island of Dominica (Smith et al. 2009b).
Mesoscale atmospheric models, such as the Penn State – National Centers for
Atmospheric Mesoscale Model (MM5; Grell et al. 1994) and Weather Research and
Forecasting model (WRF; Skamarock et al. 2005), have been shown to realistically
predict precipitation structures when run at <5 km grid spacing (Mass et al. 2002),
since the terrain (upslope) forcing for relatively wide barriers (Sierras, Cascades,
etc : : : ) can be satisfactorily represented at these grid spacings (Chap. 10 of this
book; Colle and Mass 2000; Colle et al. 2000). However, there are uncertainties
and deficiencies in model microphysics (Garvert et al. 2005b; Colle et al. 2005a,
b), upstream initialization uncertainties (McMurdie and Mass 2004), and intrinsic
predictability limits on the mesoscale precipitation over terrain (Hohenegger et al.
2006).
296 B.A. Colle et al.

There have been a few review summaries of orographic precipitation in recent


years given the rapid progress of this field (Roe 2005; Smith 2006; Rotunno and
Houze 2007; Lin 2005). Most of these previous write-ups have focused on the
theory (Smith 2006), one particular field study (Rotunno and Houze 2007), or
have emphasized one particular aspect of orographic precipitation. In this chapter,
we synthesize results from several field studies and compare with recent theory,
and then attempt to subsequently link the results back to the forecast process.
Chapter 7 highlights many of the microphysical issues associated with orographic
precipitation. Section 6.2 will summarize some of the key processes of orographic
precipitation, which will incorporate results from several field programs and high
resolution model simulations. In Sect. 6.3, some of the latest forecast tools for
orographic precipitation will be described, ranging from nested models to linear
downscaling models. These tools will be applied for a major flooding event over the
Pacific Northwest on 5–7 November 2006, in which 25–50 cm of rain fell over the
Cascade mountains of southwest Washington, resulting in a 1 in 100-year flood in
some locations. The predictability of orographic precipitation will be discussed in
general and related to the November 2006 flooding event. Finally, some of the future
areas of research in orographic precipitation will be discussed.

6.2 Key Processes in Orographic Precipitation

6.2.1 Water Vapor Flux, Terrain-Forced Ascent,


and Vapor Condensation

Three important concepts in orographic precipitation (OP) are water vapor flux,
terrain-forced ascent, and vapor condensation. The horizontal water vapor flux (FW )
is the rate at which water vapor is carried horizontally past a point in a layer above
the earth’s surface. This horizontal vector quantity can be computed from a balloon
sounding according to
ZzTOP ZTOP
P
1
FEW D qV UE d z D qV UE dp; (6.1)
g
0 P0

with units kg  m1  s 1 . In (6.1), qV is the specific humidity (kg/kg), U is the


wind speed normal to the barrier (m s1 ), ¡ is the air density (kgm3 ). In the special
case of uniform wind with height, FEW is the product of the wind speed and the
precipitable water integrated in the column, i.e. FEW D UE  PWAT where PWAT is
given in units kg  m2 D mm of liquid. If the air is saturated, qV D qVSAT .T; p/
in (6.1).
While it is not yet the custom to routinely compute FEW for each balloon sounding
(e.g. as is done for CAPE), this practice would call attention to the importance
6 Theory, Observations, and Predictions of Orographic Precipitation 297

of this variable for orographic precipitation, as highlighted in previous studies


of orographic precipitation (Smith et al. 2003; Lin and Colle 2009; Smith et al.
2009).Other things being equal, OP is roughly proportional to FEW . When a model
underestimates FEW , it is likely to underestimate OP as well. For example, if the
forecast cross barrier wind of 20 m s1 has an error of 5 m s1 (25%), the OP will
also be in error of 25%. This is why it is imperative for a model to accurately
simulate the moisture and cross barrier flow correctly in order to have a chance to
properly forecast the OP amounts.
For the case of a fully saturated atmosphere, the moist adiabatic lapse rate forms
a lower bound on the column integrated water and the water vapor flux. In this
case there is a unique relationship between surface temperature, surface saturation
specific humidity, and the column integrated water vapor. If the earth’s surface is at
p D 1,000 hPa, these are

qVSAT0 0:0038 exp.0:068T / (6.2a)

PWAT 9 exp.0:088T / (6.2b)

where T is the surface temperature in Celsius. It is remarkable how rapidly the


precipitable water and the water vapor flux increase with environmental tempera-
ture. With a wind speed of 10 m s1 and a temperature of 0ı C, (6.2b) gives FW D
U PWAT D 90kgm1 s 1 while a temperature of 20ı C gives FW D 523 kgm1 s 1 .
The equivalent moist layer depth (Dw ) also increases with temperature for a moist
adiabatic profile. From (6.2a, 6.2b)

PWAT
DW D 2400 exp.0:02T /; (6.2c)
qVSAT0

and from (6.1), the water vapor flux can be expressed as FEW D qVSAT0 DW UE .
The second important concept is terrain-forced ascent. When the low-level flow
is not blocked, ascent or descent along the local terrain slope can be approximated
by the product of the horizontal wind speed and the terrain slope (Fig. 6.1).

w D UE rh.x; y/ (6.3)

where h.x; y/ is the terrain description. In practice, formula (6.3) is slightly


ambiguous. First, what wind should we use for UE ; the ten meter wind, the 925 hPa
wind, or the wind at some other level? Second, what slope should we use, the local
terrain slope or some average slope over a 10 km patch? Third, we remain ignorant
of how the vertical motion will change with height. It might increase with height if
the horizontal wind speed does, or decrease with height if the wind speed decreases
aloft or if the streamlines flatten out.
Putting aside these questions for now, we can introduce the thermodynamics of
moist ascent. When air is forced upwards, it moves to lower pressure, expands, and
298 B.A. Colle et al.

cools. The maximum amount of water in the vapor state qVSAT .T; p/ decreases at a
rate d qVSAT =d zjPARCEL. The column integrated condensation is then

Z1
d qVSAT
C.x; y/ D w jPARCELd z (6.4)
dz
0

If w is independent of height and the sounding is moist adiabatic, the integral


(6.4) reduces to

C.x; y/ D w  qVSAT0 ; (6.5)

where qVSAT0 is the saturation specific humidity at the ground (6.2a). C(x,y) has the
same units .1 kg  m2  s 1 D 1 mm  s 1 D 3600 mm  hr 1 / as rain rate P(x,y)
assuming all the condensed water dropped instantly to the ground.
Combining (6.3) and (6.5)

C.x; y/ D qVSAT0 UE ı rh.x; y/ (6.6)

which is the simplest possible model of orographic precipitation (Colton 1976; Rhea
1978). With a temperature of T D 0ı C ,  D 1 kg  m3 , U D 10 m  s 1 and a
slope of S D 0.05, (6.2a, 6.6) give C D 0:0019 kg  m2  s 1 D 6:8 mm  hr 1 :
For the northern California coastal mountains during CALJET (Table 6.1),
Neiman et al. (2002) showed that the upslope flow is highly correlated (R  0.9,
where R is the correlation coefficient) with the hourly rain rate when the low-level
flow is not blocked (Fig. 6.3a). The rain rate is correlated best with the average
wind just above the mean mountain top (1 km above MSL for the California
coastal mountains), but this correlation drops to near 0 at 3 km above MSL. Falvey
and Garreaud (2007) also found strong correlation between precipitation and the
upstream wind and moisture flux over a 10-year period in central Chile. They further
emphasized that the positive correlation between rainfall and moisture flux is largely
determined by the variability in the wind velocity.
Although the rate of condensation above sloping terrain is related to the
horizontal water vapor flux and the terrain slope, more knowledge is needed of the
vertical motion above the terrain. Also, there are other factors that determine the
percent of the condensed water that is converted to hydrometers and falls out (i.e.
precipitation efficiency). For example, during conversion and fallout, condensed
water is advected downwind. If the condensed water is carried to the lee of the
barrier, a fraction of the falling water may evaporate before reaching the ground.
One useful measure of orographic precipitation for an entire mountain range
is the total water removed from the airstream (PT ). Integrating precipitation rate
downwind
6 Theory, Observations, and Predictions of Orographic Precipitation 299

Fig. 6.3 Cross section perspective of the orographic rainfall distribution during (a) unblocked and
(b) blocked low-level flow (From Neiman et al. 2002). The representative coastal profiles of wind
velocity (full barb D 10 kts) and correlation coefficients (based on the magnitude of the upslope
flow at the coast versus the rain rate in the coastal mountains) shown on the left. The variable h in is
the scale height of the mountain barrier. The spacing between the rain streaks in (b) is proportional
to rain intensity. The symbol “˝” within the blocked flow in (b) portrays a terrain-parallel barrier
jet (© American Meteorological Society. Reprinted with permission)
300 B.A. Colle et al.

Z1
PT D Pds (6.7)
1

where ds is the increment of distance parallel to the wind direction. This quantity
can be normalized by the incoming water vapor flux (FW ) to form the Drying
Ratio (DR),
PT
DR D (6.8)
FW

If, as (6.6) suggests, the precipitation is proportional to FW , the DR may be


independent of FW , and thus is a useful quantity to compare different mountain
ranges. Several methods for estimating DR from conventional data and models are
discussed by Smith et al. (2003) and Smith (2006) and involve calculating the water
budget across the mountain range using estimates of the surface precipitation and
incoming and outgoing water vapor fluxes. For example, one can calculate the DR
for the upslope region of a barrier by selecting a boxed region along the windward
slope, and then calculate the average Fw entering the box along the edges over some
time period divided by the total precipitation that reaches the surface within the
box (Lin and Colle 2009). DR can also be estimated using stable isotopes of water
obtained from streams and sapwood (Smith et al. 2005; Smith and Evans 2007). The
relative concentration of deuterium to hydrogen isotopes decreases across a barrier,
since the heavy isotopes are progressively rained out. Thus, the DR can be estimated
by a ratio of these isotopes upwind and downwind of the barrier according to:
 1=.˛1/
DR D 1  Rp =Rp0 ; (6.9)

where Rp0 and Rp are the isotope ratios in precipitation upwind and downwind of the
barrier, and ’ is the fractionation factor (Smith 2006). Smith’s studies have yielded
estimates of DR of 35% for the Alps, 44% for the Oregon Cascades, and 50%
for the southern Andes.
Another way to quantify microphysical efficiency is to determine the characteris-
tic residence time for all water and ice generated aloft to fall as precipitation (Smith
et al. 2003). The residence time (RT) is defined within the box region as:

RT D Total ice and water amount aloft=Precipitation rate (6.10)

Typical RTs vary between 1,500 and 2,500 s (Smith et al. 2003; Colle 2008; Lin
and Colle 2009). Residence time typically decreases as the ice/snow becomes more
rimed, which increases the terminal fall speed of the particles. The RT has been
applied to determine how efficient bulk microphysical schemes are in removing
precipitation aloft (Lin and Colle 2009). For example, a scheme that has more
graupel than snow aloft, and thus larger fallspeeds, will have a smaller RT.
6 Theory, Observations, and Predictions of Orographic Precipitation 301

6.2.2 The Role of Stability

The amount of forced ascent over an orographic barrier is strongly dependent on the
static stability around the mountain. For nearly saturated conditions, a moist static
stability (Nm ) should be diagnosed upstream of the mountain, which is given by
(6.11) (Durran and Klemp 1982),
  
g dT Lq g dqw
Nm2 D C m 1C  (6.11)
T dz RT 1 C qw dz

where qv and qc are the mixing ratios of water vapor and cloud water, respectively;
qw D qv C qc is the total water mixing ratio; R is the ideal gas constant for dry air;
L is the latent heat of evaporation; and  m is the moist adiabatic lapse rate, since
it includes latent heat of condensation and the vertical gradient of liquid water.
A useful parameter to determine the mountain flow response is the non-dimensional
mountain height (inverse Froude number), as given by:

M D Nm hm =U (6.12)

in which hm is the height of the terrain and U is the upstream flow velocity
approaching the terrain averaged below crest level. Jiang (2003) found that the
upstream deceleration ratio (u/U), in which u is the perturbed flow, collapses onto a
single curve for a critical threshold for flow blocking (M  1.2) when M is computed
with the moist static stability frequency. This illustrates that the character of the
windward flow response can still be determined even when one includes moisture.
The parameter M in (6.12) can be useful to determine the flow response over the
barrier if Nm does not vary as the flow ascends the mountain; however, the moist
static stability (Nm ) can change as a result of differential latent heating, radiation,
moisture depletion over the windward ridges, changes in the vertical gradients
of liquid water, and melting processes. Recent idealized studies by Reeves and
Rotunno (2008) showed that a simple substitution of Nm for N in (6.12) for moist
flows does not provide reasonable predictions of the steady-state flow response for
all cases. They show that including other parameters in the non-dimensional phase
space (lifted condensation level and freezing level height) can better determine the
flow perturbations and therefore the precipitation changes over the barrier. There
are also horizontal gradients in the moisture that complicate the orographic flow
and precipitation response. For example, Rotunno and Ferretti (2001) showed that
the temporal and spatial distribution of precipitation for the 1994 Piedmont flood
was dependent on horizontal gradients in the moisture, which helped create a
convergence boundary between blocked and unblocked flow. Galewsky and Sobel
(2005) showed that localized regions of latent heating were important in reducing
the effective stability and thus enhancing the uplift over the mountain. Future
research is needed to determine the impacts of moisture gradients on orographic
precipitation, which clearly can lead to some predictability problems.
302 B.A. Colle et al.

Fig. 6.4 Three-hour average cross section of (a) reflectivity (shaded in dBZ) and (b) radial
velocity (shaded in m s1 ) for a blocked flow case during MAP (IOP8 – 21 October 1999). (c)
and (d) Same as (a) and (b) except for the unblocked IOP2b event (Adapted from Medina and
Houze 2003. © John Wiley and Sons Publishing. Reprinted with permission)

6.2.2.1 Flow Blocking

Chapter 3 highlights several dynamical mountain flows, such as flow blocking and
terrain-forced ascent and decent. Several field studies have showed the importance
of the moist dynamical flow response over a barrier on the orographic precipitation
distribution, since the sloping ascent within this blocked flow can represent an
“effective” barrier that extends upwind of the windward slope. However, upstream
flow blocking also reduces the effective slope of the barrier (Marwitz 1980),
therefore reducing the intensity of the upward motion and precipitation over the
windward slope. For example, for the nearly two-dimensional Sierra Mountains,
Marwitz (1987b) and Rauber (1992) noted an upstream shift of precipitation and
high liquid water contents associated with the upstream convergence between the
ambient westerly flow and the terrain-parallel (southerly) barrier jet within the
blocked flow. For the flow blocking events, Neiman et al. (2002) showed little corre-
lation between the orographic precipita