You are on page 1of 38

10 Oct 2006 12:35 AR ANRV289-EG31-01.

tex XMLPublishSM (2004/02/24) P1: OKZ


10.1146/annurev.energy.30.050504.144308

Annu. Rev. Environ. Resour. 2006. 31:1–31


doi: 10.1146/annurev.energy.30.050504.144308
Copyright 
c 2006 by Annual Reviews. All rights reserved

ABRUPT CHANGE IN EARTH’S CLIMATE SYSTEM


Jonathan T. Overpeck and Julia E. Cole
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

Department of Geosciences, Institute for the Study of Planet Earth, Department


of Atmospheric Sciences, University of Arizona, Tucson, Arizona 85721;
email: jto@u.arizona.edu, jcole@geo.arizona.edu
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

Key Words abrupt climate change, climate dynamics, global warming,


paleoclimatology, paleoceanography
■ Abstract Many aspects of Earth’s climate system have changed abruptly in the
past and are likely to change abruptly in the future. Although abrupt shifts in tem-
perature are most dramatic in glacial climates, abrupt changes, resulting in an altered
probability of drought, large floods, tropical storm landfall, and monsoon rainfall, are all
important concerns even in the absence of significant anthropogenic climate forcing.
Continued climate change will likely increase the probability of these types of abrupt
change and also make abrupt changes in ocean circulation and sea level more likely.
Although global warming may have already triggered abrupt change, current under-
standing and modeling capability is not sufficient to specify details of future abrupt
climate change. Improved adaptation strategies are warranted, as well as efforts to avoid
crossing climate change thresholds beyond which large abrupt changes in sea level,
ocean circulation, and methane-clathrate release could greatly amplify the impacts of
climate change.

CONTENTS
1. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2. COLD-CLIMATE ABRUPT CHANGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2. The Younger Dryas Abrupt Climate Event . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.3. Millennial-Scale Abrupt Climate Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.4. Mechanisms and Theory of Cold-Climate Abrupt Change . . . . . . . . . . . . . . . . . 5
2.5. The Pace of Cold-Climate Abrupt Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3. WARM-CLIMATE ABRUPT CHANGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2. North Atlantic Abrupt Climate Change Before
and During the Holocene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.3. Abrupt Changes in Drought and Drought Regimes . . . . . . . . . . . . . . . . . . . . . . 11
3.4. Abrupt Changes in Flood Regimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.5. Abrupt Change in Tropical Storm Frequency . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.6. Abrupt Change in Asian and African Monsoons . . . . . . . . . . . . . . . . . . . . . . . . 15
3.7. Abrupt Change in Tropical Pacific/ENSO Systems . . . . . . . . . . . . . . . . . . . . . . 15

1543-5938/06/1121-0001$20.00 1
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

2 OVERPECK  COLE

4. ICE SHEET INSTABILITY AND ABRUPT


CHANGE IN SEA LEVEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
5. ABRUPT CLIMATE CHANGE “WILDCARDS” . . . . . . . . . . . . . . . . . . . . . . . . . . 17
6. IMPLICATIONS FOR THE FUTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
6.1. Abrupt Climate Change Is Inevitable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
6.2. Predictability of Abrupt Climate Change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
6.3. What Evidence Exists for Recent Abrupt, Anthropogenic Climate Change? . . . 20
7. CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

20
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

1. INTRODUCTION
The reality of anthropogenic global warming is by now well established, but the
pace and amplitude of future change remains uncertain, particularly on a regional
basis (1). A growing body of evidence from past climate studies argues that even
if radiative forcing changes gradually, the climate system is not likely to respond
smoothly. These studies show that the climate system is capable of extremely rapid
transitions and that this behavior is geographically and temporally widespread (2).
Abrupt climate changes have been observed from North Atlantic to monsoonal cli-
mate systems, from the equator to poles, and from desert to coral reef environments,
all in the absence of abrupt climate forcing. As the resolution, replication, and quan-
tification of paleoclimate studies improve, patterns of abrupt climate change are
emerging along with insights into the underlying mechanisms of abrupt climate
system behavior. Because abrupt changes are not linearly coupled to climate change
forcing, their prediction is currently a significant challenge. The phenomenon of
abrupt climate change makes anticipating and preparing for future climate change
difficult. The potential for abrupt future change argues that we are already courting
“dangerous anthropogenic interference” (3, 4) with the climate system.
What do we mean by abrupt change? Alley et al. (2), in a seminal paper arising
from a U.S. National Academy of Sciences report (5), followed on the original
definition of abrupt change (6): an abrupt climate change occurs when the climate
system is forced to cross some threshold, triggering a transition to a new state at a
rate determined by the climate system itself and faster than the cause. Others have
defined it simply as a large change within less than 30 years (7) or as a transition in
the climate system whose duration is fast relative to the duration of the preceding
or subsequent state (8). We prefer this last definition, with the recognition that
abrupt change can be forced or unforced and is usually a nonlinear or threshold
transition between two more stable regional or global climate states. Initial work
describing abrupt change focused on North Atlantic records of glacial and deglacial
climate change and implied that the influence of large ice sheets was key. More
recent studies illustrate the global imprint of glacial and deglacial abrupt change
and extend this concept to interglacial periods, notably the Holocene [the past
circa (ca.) 11.6 thousand years (kyrs)]. Research on both cold-climate (glacial and
deglacial) and warm-climate (Holocene and previous interglacials) abrupt change
has significant implications for future climate responses. From these studies, we
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 3

know that abrupt change can propagate globally, does not require large ice sheets,
occurs on timescales important to society (years, decades), and is a common mode
of climate system behavior that may or may not be attributable to a specific forcing.
We write this review with several goals: (a) to provide an updated synthesis of
the abrupt change literature useful both within and beyond the paleo- and physical
climate communities; (b) to suggest and refine hypotheses on warm- and cold-
climate abrupt change that will guide future research; and (c) to highlight the
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

relevance of abrupt change to anticipating future climate change and impacts. We


begin with a summary of research on cold-climate abrupt change, highlighting
its global footprint and speculating on potential triggers and amplifiers. We then
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

discuss warm-climate abrupt change and its patterns and potential causes. We
also discuss the potential for abrupt sea level change via the effects of climate
change on large ice sheets (under both deglacial and modern conditions). We briefly
review the potential for abrupt changes forced by “exotic” events (e.g., methane
“clathrate” melting, immense volcanic eruptions, and asteroid impacts). Finally, we
discuss the implications of abrupt change research for understanding future climate
change and its impacts. Our primary conclusions are that abrupt climate change in
the future is inevitable, that continued human forcing of climate change increases
the probability of deleterious abrupt climate change, and that broad policy measures
are needed to reduce the vulnerability of human and ecological systems to an
environment that will likely change in unpredictable ways.

2. COLD-CLIMATE ABRUPT CHANGE


A substantial research effort, both empirical and theoretical, focuses on abrupt
change during glacial and deglacial climates. Global documentation of abrupt
millennial change has emerged, along with energetic debate about the mechanisms
that trigger, amplify, and propagate these changes worldwide. Experiments with
climate models offer promise to elucidate which mechanisms can produce the
observed spatial patterns.

2.1. Introduction
The earliest focus on the issue of abrupt climate change during glacial and deglacial
climates arose out of terrestrial and marine deglaciation records and, in partic-
ular, from high-resolution data from Greenland ice cores. During the most re-
cent deglaciation, an abrupt reversal of warming known as the Younger Dryas
(YD) cold interval (9) provided one of the first well-studied examples of abrupt
change. Corresponding changes in the position of the oceanographic polar front
were documented by North Atlantic paleoceanographic reconstructions (10). In ice
core isotope records, rapid transitions between cold and warm intervals occur on
timescales measurable in human rather than geologic terms, i.e., years to decades
(11). These millennial oscillations characterize the most recent deglaciation and
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

4 OVERPECK  COLE

comprise the dominant pattern of climate variability in these cores—indeed in


many high-resolution climate records—between about 25–75 thousand years (ka)
before present (BP) (Figure 1).
Although large ice sheets clearly play a role in cold-climate abrupt change, such
change is nonetheless highly relevant for understanding potential future abrupt
change. Indeed, some mechanisms involved in cold-climate abrupt change are
likely important even in the absence of large glacial ice sheets. Glacial abrupt
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

changes are not a linear response to external forcing (no sudden forcing has been
identified), so they represent either a threshold response to smaller, more gradual
forcing or an abrupt change in the internal workings of the climate system. Iden-
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

tifying the mechanisms of abrupt climate change is essential to understand the


sensitivity of the climate system and its potential for sudden future shifts. More-
over, the global footprint of cold-climate abrupt change tells us that such changes
can propagate rapidly and globally, a capability not likely limited to cold climates.
This section provides further detail on the distribution, impacts, mechanisms, and
relevance of cold-climate abrupt change.

2.2. The Younger Dryas Abrupt Climate Event


Early work on abrupt climate change focused on the transition into the YD cool
interval, a millennial-scale reversal of deglacial warming to near-glacial conditions
between ca. 11.5–13 ka (9, 12). Although particularly notable in North Atlantic
records, a global footprint of YD impacts has emerged that is largely consistent
with a single abrupt event being rapidly translated worldwide (13). Initial work
attributed the YD anomaly to the catastrophic drainage of glacial Lake Agassiz
through the St. Lawrence seaway and consequent shutdown of the meridional
overturning circulation (MOC) that propagated climate anomalies globally (13,
14). However, a close look at the continental and geophysical evidence for melt-
water routing and ice sheet melt histories suggests that this picture may need
revision. New geologic evidence indicates that Laurentide ice margins may not
have retreated sufficiently to allow Lake Agassiz drainage eastward at YD time
(15). Alternatively, freshwater from the Arctic Ocean, associated with the collapse
of a Keewatin ice dome, might have initiated the observed MOC response (16).
Although the precise trigger remains poorly constrained, paleoceanographic data
leave little doubt that MOC weakening occurred during the YD (17, 18) and that
extreme seasonality probably played a role (19, 20). This weakening would likely
cool the North Atlantic, and global climate models support the transmission of
climate anomalies from a cooler North Atlantic rapidly throughout the Northern
Hemisphere (21). The YD has become something of a paradigm, explaining how
abrupt change in cold climates works. Nonetheless, we still lack important details,
including the geologic component of this event; the sensitivity of the MOC re-
sponse to meltwater timing, location, and amount; and the mechanism for rapid,
near-global propagation of anomalies. As Lowell et al. (15) have noted, these gaps
bode poorly for predicting future abrupt changes.
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 5

2.3. Millennial-Scale Abrupt Climate Events


Abrupt shifts between warm and cold states punctuate the interval between 20 to
75 ka) in the Greenland isotope record, with shifts of 5◦ –15◦ C occurring in decades
or less (Figure 1). These alternations were identified in some of the earliest ice
core isotopic studies [e.g., (22)] and were replicated and more precisely dated by
subsequent work (23). Further analysis of diverse records has distinguished two
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

types of millennial events (13). Dansgaard/Oeschger (D/O) events are alternations


between warm (interstadial) and cold (stadial) states that recur approximately
every 1500 years, although this rhythm is variable. Heinrich events are intervals of
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

extreme cold contemporaneous with intervals of ice-rafted detritus in the northern


North Atlantic (24–26); these recur irregularly on the order of ca. 10,000 years apart
and are typically followed by the warmest D/O interstadials. The glacial sediments
are thickest near the edges of the Laurentide ice sheet and are inferred to originate
from dramatic discharges of sediment-laden icebergs that melt and deposit both
glacial sediment and fresh water in the northern North Atlantic. Surface water
oxygen isotope reconstructions allow the mapping of freshwater anomaly patterns
for certain Heinrich events (26, 27). Heinrich events are associated with a near
shutdown of the MOC, although the varied evidence for such changes associated
with D/O oscillations suggests a more heterogeneous response (28, 29). One study
notes that, although deepwater was generated more or less continuously in the
Nordic seas between 10–60 ka, the mechanism of change varied between stadial
and interstadial (30). Heinrich events correlate with small decreases in atmospheric
CO2 (∼20 ppm), whereas D/O events experience smaller CO2 changes (≤10 ppm)
(31, 32). Methane changes associated with D/O events reach ∼150 parts per billion,
almost half the glacial-interglacial amplitude (33).
Both Heinrich and D/O events exhibit clear global impacts. These patterns
have been summarized in several studies [e.g., (26, 34)]. Although the pattern of
influence appears to differ between these types of anomaly, a clear interpretation of
these differences, particularly in terms of distinguishing physical mechanisms, has
not been developed. As Hemming (26) notes, different global patterns of impact
may simply reflect proxy-specific or site-specific limitations such as sensitivity and
response time. In general, however, a cold North Atlantic corresponds with a colder,
drier Europe, weaker Asian summer monsoon, saltier northwestern tropical Pacific,
drier northern South America, colder/wetter western North America, cooler eastern
subtropical Pacific, and warmer South Atlantic and Antarctic. Table 1 summarizes
the main impacts of a cold North Atlantic (stadial) on key regions and systems.

2.4. Mechanisms and Theory of Cold-Climate Abrupt Change


Observations and simulations of cold-climate abrupt change have been used to
argue that the climate system, particularly in the North Atlantic, has two stable
states, characterized by warm/strong MOC and cold/weak MOC conditions (14,
35). Transitions between these states can happen fast and depend on the fresh-
water balance of the North Atlantic. A slightly more nuanced view, based on
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

6 OVERPECK  COLE

Table 1 Global impacts of a cold North Atlantic (stadial) event


System or region Responsea Reference(s)

Europe Colder and drier; vegetation (73, 187, 188)


more xeric
East Asian summer monsoon Summer monsoon weakens (49)
(China) relative to winter monsoon
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

South Asian summer monsoon Weaker upwelling and weaker (50)


(Indian Ocean) summer monsoon intensity
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

Tropical South America (Southern Wetter; southward monsoon (189, 190)


Hemisphere) displacement
Tropical North Atlantic Drier in northern South America; (53, 191)
ITCZ southward, warmer SST
Eastern temperate Pacific Santa Barbara basin well (184, 192)
ventilated, cooler SST
NW tropical Pacific Higher salinity (59)
NE Australia Wetter (58)
North American Great Basin Colder/drier (193)
Southwestern United States Colder/wetter (61)
Antarctic Gradual warming (32, 33)
a
Corresponding to when the North Atlantic is colder.

paleoceanographic observation, suggests three modes of MOC activity associated


with abrupt change (7, 34, 36): modern (deep water forming in the Nordic and
Labrador seas), glacial (open-ocean convection in the subpolar North Atlantic
to relatively shallow depths, ≤2500 m), and Heinrich (convection off; Antarctic
bottom water fills deep basins of North Atlantic). Changes in the MOC, forced
by freshwater anomalies and transmitting surface climate changes through the
atmosphere, offer a sort of “minimum model” of abrupt change.
The involvement of the MOC in cold-climate abrupt change, together with
the apparent antiphase relationship between North Atlantic/Greenland and South
Atlantic/Antarctic temperatures, has led to the concept of a “bipolar seesaw” (37)
in which the MOC moves heat from the southern to northern Atlantic. When the
MOC circulation weakens, the north cools as the south warms. Such a mechanism,
in pure form, would produce strictly antiphase behavior between, say, Greenland
and Antarctic ice core temperatures. Observations show a more complex picture,
with varying leads and lags (see the discussion and references in References 32
and 38). Stocker & Johnsen (38) argue that these complexities can be accounted
for by the addition of a thermal reservoir that convolves the Northern Hemisphere
signal with a time constant of ∼1000 years.
Additional mechanisms and feedbacks are required, however, to explain the
involvement of the global tropics in cold-climate abrupt change. Tropical modes,
such as those associated with the big monsoon systems and El Niño/Southern
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 7

Oscillation (ENSO), are highly energetic and capable of propagating climate


anomalies worldwide. Cane & Clement (39, 40) note that ENSO physics are not
restricted to interannual time periods but can play a role in climate variability
on millennial to Milankovitch timescales. In a model, gradual orbital forcing can
lead to periods of instability in ENSO, whereby phase locking of ENSO to the
annual cycle of Pacific sea surface temperature (SST) leads to more La Niña-
like conditions during certain orbital configurations, notably that of the YD (41).
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

These arguments link a cool tropical Pacific (La Niña) with glacial states and vice
versa; potential teleconnection mechanisms include the effect of North American
temperature anomalies (in the same direction as the tropical Pacific) impacting
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

snow and ice margins directly (39). The response of ENSO to orbital forcing is
clearly shown in model studies and paleoclimate records (42), and in transient
model simulations, it leads to an 11,000-year response period (41, 43). But a clear
mechanistic link to millennial D/O events has proven more elusive.
Additional tropical mechanisms include the influence of North Atlantic vari-
ability on Asian monsoon circulation. Physical climate linkages between the North
Atlantic and Asia have been described for modern variability [e.g., (44)] as well
as for Holocene and deglacial intervals [e.g., (45–48)]. On millennial timescales,
stadials are clearly associated with weaker summer monsoons (49, 50), perhaps re-
lated to enhanced snow cover that diminishes monsoon intensity (51). Regardless
of their origin, abrupt changes that engage the monsoon have the opportunity to
propagate across a large area, even to engage ENSO. Once the monsoon and ENSO
are engaged in the millennial pace of D/O events, the signal can be widely imparted
to teleconnection regions associated with these systems (52). Modern climate rela-
tionships imply that a cold North Atlantic would correspond to a weak Asian mon-
soon and therefore warm El Niño conditions (the latter through mechanisms such
as snow cover, trade wind, and Walker circulation changes, and/or ENSO-related
warming of the tropical Indian Ocean, which weakens Asian monsoon intensity).
Another important tropical connection to cold-climate abrupt change lies in
the Atlantic Intertropical Convergence Zone (ITCZ). Peterson et al. (53), using
data from the Cariaco Basin north of Venezuela, describe enhanced rain in north-
ern South America during interstadials that results from a northward shift of the
ITCZ. If this situation also strengthens the westward moisture transport out of
the Atlantic, it could feed back positively on the MOC by increasing subtropi-
cal Atlantic salinity. This mechanism is supported by modeling studies that link
enhanced vapor transport out of the Atlantic to a vigorous MOC under mod-
ern and future conditions (54, 55). Modeling studies also reproduce the response
of the Atlantic ITCZ to North Atlantic conditions, such that it shifts southward
when the hemispheric gradient increases during cool North Atlantic intervals (56).
These feedbacks reinforce each other to ensure a coupling between tropical and
high-latitude Atlantic variability. A southward-displaced Atlantic ITCZ is also
characteristic of La Niña conditions [e.g., (57)].
Whether cold-climate abrupt change is triggered in the tropics or rapidly im-
ported there from the North Atlantic, it is arguably the engagement of the tropics
that enables the signal to propagate widely and efficiently over the globe. ENSO
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

8 OVERPECK  COLE

is an obvious mechanism for propagating these events, but observational studies


have invoked different scenarios of ENSO’s involvement. Several lines of evi-
dence suggest that during interstadials conditions resembled an El Niño-like state,
for example, dry conditions in northern Australia (58) and a northward-displaced
Atlantic ITCZ (53). But other sources of evidence suggest that interstadials ex-
perienced conditions more comparable to La Niña, with a fresher northwestern
tropical Pacific (59), stronger summer monsoon over South and East Asia (49, 50),
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

warmer/drier conditions in the southwestern United States (60), and warmer/wetter


conditions in the U.S. Great Basin (61).
These discrepancies suggest that the modern ENSO system is not the optimal
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

framework for understanding the global pattern of cold-climate abrupt change.


Global climate relationships can change under varying climate boundary condi-
tions. For example, a large North American ice sheet could change how ENSO
impacts North American climate. Today, western U.S. drought can respond to
both Atlantic and Pacific forcings (62); past millennial variability in the western
United States could reflect either. The Asian monsoon also varies independently of
ENSO, such as in the most recent 30 years (63); downstream impacts from the North
Atlantic are a plausible candidate for imparting past variability [e.g., (46)]. Salinity
variations in the northwest tropical Pacific (59) may reflect monsoonal as much
as ENSO variations. With respect to cold-climate abrupt change, a tropical Pacific
that was somewhat more La Niña-like during stadials, but less globally influen-
tial (i.e., not dictating Asian and North American responses), would be consistent
with much of the existing evidence. Clearly, more data from critical regions of the
tropics are needed to address this issue as well as a comprehensive suite of coupled
model experiments that can credibly assess changes in global teleconnections.

2.5. The Pace of Cold-Climate Abrupt Change


Cold-climate abrupt change occurs with a characteristic timescale of ∼1500 years,
a feature that must be explained by any proposed mechanism. North Atlantic and
the Greenland Ice Sheet Project 2 (GISP2) records exhibit a period of ∼1470 years
(64, 65). However, the adjacent ice core isotope record from the Greenland Ice Core
Project (GRIP) site exhibits periods closer to 1670 and 1130–1330 years, which is
in agreement with the independently dated record from Hulu Cave (49, 66). Time-
series studies generally converge on a picture of a noisy climate system paced by a
regular, perhaps external, forcing, with the sensitivity of the system to the forcing
varying depending on background conditions or stochastic variability [e.g., (67–
69)]. Solar forcing, although subtle, is the leading candidate for external forcing
and has been found to be consistent with either a 1450–1470–year period (70, 71)
or the 1667- and 1130-year periods (66). Alternatively, internal drivers (tropical or
extratropical) may operate on a preferred millennial timescale. Ocean circulation,
ice dynamics, and ENSO have all been suggested to operate with characteristic
timescales near 1500 years (14, 58, 72). Moreover, ice core age models are under
revision as other records suggest systematic chronologic errors in both GRIP and
GISP (49, 73–75).
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 9

3. WARM-CLIMATE ABRUPT CHANGE


Abrupt climate changes are not restricted to cold-climate background states. A
growing body of evidence documents abrupt changes in the climate of the Holocene
(the past 11.6 kyrs) and even in the instrumental record of the past century. Such
changes occur both in temperature and especially in hydrologic balance, with
the abrupt onset of drought a particular concern. These observations confirm that
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

abrupt change must be considered as we work to anticipate how climate will change
in the near future, in response to ongoing anthropogenic forcings.
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

3.1. Introduction
The roots of modern paleoclimatology have origins in studies of late Holocene cli-
mate variability in, and around, the eastern North Atlantic. The so-called Medieval
Warm Period and Little Ice Ages are etched into both the climate change literature
and popular imagination. We now know that parts of Earth were nearly as warm as
the mid-twentieth century between AD 1000 to 1100 and that this generally warmer
period was followed by colder temperatures at least in the Northern Hemisphere
before giving way to the unprecedented global warming of the twentieth century
(76–79). In the instrumental temperature record, perhaps the clearest example of
abrupt change occurs in the late 1970s when global temperature began an unusu-
ally steep rise (http://data.giss.nasa.gov/gistemp). This transition coincides with a
step-like change in the state of the tropical Pacific toward more El Niño-like con-
ditions and in the extratropical Pacific toward generally warmer conditions and
altered circulation (80, 81). Although the Pacific signature of this 1976 shift may
have reversed recently, global temperatures continue to rise unabated. Interglacial
abrupt temperature shifts are large enough to be of significant human interest. For
example, warming temperatures have already had an impact on seasonal snow melt
and consequent changes in water resource availability (which occur independently
of changes in precipitation amounts) (82).
However, abrupt temperature change is not the biggest story nor the most im-
portant likely abrupt change concern for society. More important are potential
“warm-climate surprises” (83, 84) that relate to abrupt shifts in regional climate
variability and moisture availability. Anyone making a livelihood in an arid or semi-
arid region knows that, although temperature change is important, the real focus
is on water. And the paleoclimatic record is rich in evidence for abrupt hydrologic
shifts.
The goal of this section is to describe the most rapidly expanding area of
abrupt climate change research—that focused on past, and potential future, “warm-
climate” abrupt climate change. We review the considerable work focused on
the North Atlantic, highlighting that there is still important uncertainty regarding
the ability of the MOC in the Atlantic to respond abruptly in the absence of
large ice age ice sheets. We then review what is known about abrupt changes in
hydrological regimes. Of particular interest are the patterns and processes related
to megadroughts (droughts that persist from one to several decades) and abrupt
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

10 OVERPECK  COLE

changes in drought frequency, duration, and extent. Although less work has focused
on abrupt warm-climate change related to flood regimes, ENSO, tropical storms
(i.e., hurricanes and typhoons), and monsoon dynamics (e.g., in Africa and Asia),
there is enough literature to review these types of highly relevant abrupt change. As
with the case of abrupt drought and megadrought dynamics, the science associated
with past abrupt flood, hurricane, and monsoon change clearly points to the ability
of the climate system to behave in ways that can have large impacts on human and
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

natural systems.
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

3.2. North Atlantic Abrupt Climate Change Before


and During the Holocene
Early on, concern focused on the ability of the North Atlantic MOC to change
abruptly in warm climates and impact regional (and downstream) climate in dra-
matic ways. Manabe & Stouffer (85) established a challenge to the climate change
community when they used a coupled atmosphere-ocean climate model to suggest
that future anthropogenic climate change could result in an abrupt and long-lasting
cessation of the North Atlantic MOC, an event that would have significant impacts
in at least the regional climate of the circum-Atlantic region. The Intergovernmental
Panel on Climate Change (IPCC) Third Assessment (1) included a comprehensive
multimodel evaluation of possible future greenhouse gas–induced MOC changes
and found that a MOC slowdown, or even complete stop, was still a possible
outcome of continued atmospheric greenhouse gas increases. Some recent works
[e.g., (86)] highlight that the risks of future MOC shutdown are still real. Much
of the current debate centers on what models say. Our goal here is to review how
paleoclimatic data inform this debate along with models.
Although there was plenty of variability in North Atlantic climate over the
Holocene, and some workers have speculated that changes in the North Atlantic
MOC may have helped shape climate of the last millennium (87), there is only one
compelling case of a major abrupt change involving a significant decrease in the
North Atlantic MOC. At approximately 8.2 ka, large freshwater discharges from
North American glacial lakes flooded the North Atlantic and created what has be-
come one of the most well-studied and well-known abrupt climate change events
of the past 10,000 years (48, 88–90). Although there was still a significant residual
ice sheet over northern North America at the time, the “8.2-ka event” provides
valuable lessons for the future. First and foremost, there is a well-documented
and simulated cause and effect: a large rapid freshening of the North Atlantic
likely caused the North Atlantic MOC to weaken by approximately 50% (89, 90).
Second, this change in ocean circulation resulted in the expected large change in
temperature fields over the circum-North Atlantic. These changes were smaller in
magnitude to similar glacial and deglacial events but still resulted in cold temper-
ature anomalies that were greatest in the immediate North Atlantic region, up to
1◦ to 3◦ C over much of the mid- to high-latitude Northern Hemisphere, and that
were likely compensated by some warming in the Southern Hemisphere (89, 90).
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 11

Perhaps the most interesting aspects of the 8.2-ka event were its apparent global
reach and that many of the impacts were hydrological rather than thermal in nature.
As with the abrupt cold events of glacial and deglacial times, the global climate
system clearly has the ability to propagate abrupt events in one region of the globe
(often the North Atlantic) to far-off locations. Even though the 8.2-ka event was
smaller than its cold-climate cousins, it was able to cause cooling and especially
drying in the tropical North Atlantic and northern Africa (48, 89, 90). Simulations
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

of the event also suggest that there should have been related weakening of parts
of the Asian monsoon system, but observational evidence is sparse and equivocal,
particularly over monsoon Asia (90–92).
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

Lastly, comparisons of simulations and observations of the 8.2-ka event suggest


that our current coupled climate models are able to simulate the general spatial
atmospheric response of a major MOC slowdown fairly well (90). However, there is
still substantial uncertainty with regard to the magnitude and location of freshwater
forcing required to cause a significant weakening of the North Atlantic MOC, along
with associated climate anomalies around the globe. Simulations of future climate
suggest that there will likely be a reduction in the MOC because of precipitation-
related freshening of the North Atlantic (1). However, the range of simulated
future change, and thus sensitivity of coupled models to freshwater forcing, is
large and uncertain. Also unknown is how much a melting Greenland Ice Sheet
could contribute to future changes in the MOC. Otto-Bleisner et al. (93) suggest
this factor might not be large, but this assertion is model dependent and depends on
the future interaction of precipitation-evaporation balance in the North Atlantic,
MOC feedbacks, and melting rates of the Greenland Ice Sheet.
The last important issue is whether abrupt change in the North Atlantic follows
a regular beat, as it apparently did during glacial times (see section 2). This hypoth-
esis is presented by Bond et al. (70) and includes the assertion that a ca. 1500-year
cycle of change in the North Atlantic (including the 8.2-ka event) is driven by
changes in solar output. Evidence for such a regular cycle has been cast in doubt
by more recent work in the North Atlantic (94–97), and although evidence of a
similar pattern of Holocene climate variability has been found in a few locations
around the globe [e.g., (46, 98)], the lack of broader evidence seems to suggest
that a quasi-periodic model of climate change, perhaps driven by the sun, is not
well supported. More research is needed to see if modest changes in solar or other
forcing are able to trigger a succession in warm-climate abrupt climate change.
Otherwise, it seems more likely that processes internal to the climate system are
responsible for the appearance of varying quasi-periodic patterns of behavior in
some paleoclimatic records.

3.3. Abrupt Changes in Drought and Drought Regimes


The possibility of increased drought is among the more ominous potential conse-
quences of anthropogenic climate change (1, 8). Paleoclimate data present a dis-
turbing picture of past conditions that includes sudden onsets of droughts lasting
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

12 OVERPECK  COLE

decades or longer and the occurrence of long (multicentury) regimes in which con-
ditions significantly drier than present prevail (99). One lesson from paleoclimate
data is that the transition to prolonged drought is almost always abrupt (nonlin-
ear with respect to an identifiable forcing such as orbital forcing) on the human
timescales. However, as for much abrupt climate change, the physical mechanisms
of prolonged droughts remain poorly defined. SST patterns play a significant role
in regional drought [e.g., (100, 101)], and the combination of SST anomalies in dif-
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

ferent ocean basins can set up drought anomalies that span one or more continents
(62, 102). A common inference is that droughts are set up by SST anomalies (often
small) and are at least partially maintained through land-surface and atmospheric
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

feedbacks (100), usually in concert with a persistence in SST anomalies.


Examples of persistent drought abound in the paleoclimate record, especially in
the semiarid regions of western North America and Africa. In the western United
States, tree-ring records reveal multidecadal dry periods during the past millennium
in which regional drought prevailed (103, 104). The last half of the sixteenth
century was exceptionally dry over much of the southwestern United States and
northern Mexico (105). Reconstructions of precipitation and river discharge concur
with explicit drought (PDSI) reconstructions [e.g., (106, 107)] and show a strong
southwest U.S. footprint to this event, although dry conditions extended as far east
as Virginia (108). In the thirteenth century, similar multidecadal drought conditions
were felt from the southwest into California, as evidenced by both dendroclimatic
reconstructions and by drowned tree stumps (109). Prior to about AD 1400, much of
the western United States appears to have experienced a drier regime overall, with
prolonged arid anomalies superimposed on drier than modern mean conditions
(104, 110). Aspects of this dry regime appear to be regional (Figure 2). During
this time, droughts more extreme than anything in the twentieth century persisted
for up to a century, and dune fields that are currently vegetated were active (110,
111).
In the mid-Holocene, conditions over much of central and western North
America were more arid than today. Aeolian features (indicating loss of plant
cover and available sediment) are widespread in the Great Plains between 5000–
10,000 years ago and are thought to represent sustained precipitation deficits of
at least 25% (111). Lake sediment cores from the Great Basin (Owens Lake and
Pyramid Lake) also suggest a warmer and drier mid-Holocene between ∼3800
and 6500 years ago, preceded by moister conditions and followed by near-modern,
much wetter conditions during the past ∼3000 years (112). The absence of spring
deposits (“black mats”) in Great Basin alluvial sequences between 2300 and
6300 years ago also implies increased aridity (113).
In the southwestern United States and Mexico, the Holocene pattern of past
aridity seems to differ between summer and winter seasons. Enhanced summer
monsoon activity has been inferred from higher lake level reconstructions and
vegetation reconstructions in parts of Arizona, Texas, and New Mexico (114),
with similar inferences from northern Mexico (115). Marine sediment records
from the Gulf of Mexico also indicate an intensified summer monsoon in the
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 13

mid-Holocene, including abrupt transitions as the monsoon weakened at ∼2500


and ∼4500 years BP (116). But the apparent drying of lakes and springs at mul-
tiple southwestern sites (117) argues for mid-Holocene decreases in winter pre-
cipitation. A speleothem record from Arizona indicates abrupt changes between
prolonged periods of cold/wet and warm/dry conditions as well as changes in the
frequency of such transitions throughout the mid-Holocene (118).
Whether natural or anthropogenic, prolonged drought is almost certain to com-
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

prise a part of the climatic future in the western United States. Past warm periods ap-
pear to have experienced more, longer drought (104). A “perfect ocean for drought”
(102) includes a substantial warming component in the tropical Indo-Pacific, and
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

warming also implies reduced soil moisture, which strengthens drought-enhancing


feedbacks (119). The behavior of ENSO and Pacific decadal variability are also
important for future western U.S. drought; drought in the southwest and Great
Plains is often strongly correlated with cooler tropical to subtropical Pacific SST
anomalies (100, 101).
In the arid and semiarid regions of Africa, abrupt changes into and out of drought
regimes are apparent in both paleoclimatic and instrumental records. Holocene
changes in African aridity were dramatic and abrupt in both the east and west (120–
123). One proposed mechanism is a nonlinear response of land surface processes
to gradually decreasing insolation seasonality (124). In the twentieth century, the
Sahel has experienced perhaps the most dramatic and abrupt shift to drought
of any region (125); explanations include initiation by various patterns of SST
anomalies (126) and maintenance by land surface changes potentially amplified
by anthropogenic desertification (127).
The paleoclimate literature abounds with examples of abrupt drought onset
from other regions [e.g., (47, 128–130)]. In most paleoclimate records, transitions
between wet and dry intervals happen abruptly. Given the physical mechanisms as-
sociated with persistent drought, anthropogenic climate change will likely heighten
the probability of drought, particularly in arid and semiarid regions such as the
western United States and North Africa. The paleoclimate record tells us that the
transitions into dry regimes are likely to be abrupt.

3.4. Abrupt Changes in Flood Regimes


Just as the dry end of the hydrological spectrum is often characterized by abrupt
change, so too is the wet end. Particularly troubling are concerns, grounded in
both theory and observation, that climate change will increase the strength of
the hydrologic cycle, with an associated increase in extreme rain, snowfall, and
flood events in some places and years, offsetting an increase in drought in other
years and locations (1, 131, 132). With ongoing global warming, the snowmelt
season is already moving earlier into the spring in some regions (133, 134) and
resulting in increased flooding (131). Moreover, continued warming will drive an
increase of extratropical convective storm activity earlier into the spring, creating
the opportunity for more frequent coincidence of high rates of convective rainfall
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

14 OVERPECK  COLE

with snowmelt runoff. Lastly, as tropical storms (e.g., hurricanes) continue to


intensify, perhaps abruptly (see section 3.5), another driver of increased flood
frequency will become more common.
The Holocene is replete with evidence that the frequency of large floods changes
through time, often abruptly (135, 136). Strong evidence of this comes from the
upper Mississippi Valley, where large flood frequency increased abruptly between
3000 and 3500 years ago resulting in even larger floods for approximately 200
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

years after AD 1250 (136, 137). The latter change apparently coincided with the
abrupt change to a wetter hydrologic regime across the upper midwestern United
States (110, 136) (Figure 2). In the southwestern United States, flood frequency
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

increased suddenly at around 5 ka and declined at about 3.6 ka; high flood frequency
also occurred between ∼800–1000 years ago and during the past 600 years (135,
138). Although data are limited and chronologies are imprecise, these studies
imply a broad anticorrelation between flood frequency in the southwestern and
the upper midwestern United States (136). Analysis of these records links flood
frequency to changes in mean seasonal or annual precipitation, often associated
with relatively modest changes in atmospheric circulation; flood frequency also
seems to rise during intervals of climate change (136). In the Southwest, flooding
is often related to El Niño conditions, and may reflect long-term variability in that
system (135, 138) (see section 3.7). Although it is beyond the scope of this review
to discuss the global evidence of abrupt changes in flood regimes, North America
is not alone in experiencing such abrupt changes.

3.5. Abrupt Change in Tropical Storm Frequency


A scientific consensus is emerging that wind speeds and rainfall amounts associated
with tropical storms (hurricanes and typhoons) will increase with continued global
warming (1, 139). Satellite and instrumental records indicate that intensification of
tropical storms has likely already begun and has been associated with an increase
in storm frequency in the North Atlantic (140–142). These changes will factor into
future flood frequency and magnitude, particularly in regions already experiencing
the effects of tropical storms or their remnant disturbances (see section 3.4), and
will also compound the impacts of abruptly rising sea level (see section 4).
The paleoclimatic study of past tropical storm variability, or “paleotempestol-
ogy,” is a relatively new field, but already evidence is emerging that abrupt changes
in tropical storm landfall have occurred in the past (143). The largest concentration
of paleostorm records come from the East and Gulf Coasts of the United States.
Evidence supports the existence of a hyperactive hurricane period on the Gulf
Coast between 3800 and 1000 years ago, followed by a less active period lasting
until the twentieth century (143). The reverse pattern characterizes the East Coast,
suggesting that the storm record may be one of changing storm tracks, rather
than changing number of overall storms, and that changes in atmospheric circula-
tion may have been the main factor in producing shifts in the regional number of
large storm landfalls (144, 145). Given available data, the changes in atmospheric
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 15

circulation and regional large storm landfall frequency appear to have been abrupt
(143).
The longest documentary record of past tropical storm frequency comes from
China, where a 900+ year record of past typhoon activity indicates multiple
decades-long periods (e.g., AD 1660–1680 and AD 1850–1880) with anoma-
lously high typhoon-landfall frequencies (146). As in the case with the North
Atlantic, atmospheric circulation–driven shifts in storm tracks are the most likely
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

mechanism behind the storm variability reflected in this long record. Changes in
typhoon tracks over the South China Sea and eastern Asia are also evident in the late
twentieth century, associated with the changing strength of the regional subtropical
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

high (147).

3.6. Abrupt Change in Asian and African Monsoons


The monsoon systems of Asia and Africa strongly impact a large proportion of
humankind. Monsoon rains are the basis for agriculture and other regional water
needs; either excessive or deficient rain can spell disaster. Monsoon rains are ex-
pected to generally increase with global warming (1), just as they tracked changes
in summer insolation over recent Earth history (148, 149). However, paleoclimate
data make clear that the monsoons of Asia and Africa, more often than not, respond
to altered forcing by changing abruptly rather than gradually. The mechanisms for
observed past abrupt changes in monsoon strength are not well understood, which
means that prediction of future abrupt monsoon shifts will be difficult.
The abrupt nature of South Asian monsoon change during the Holocene is
suggested in both terrestrial and marine paleoclimate records (46, 91, 150), with
abrupt decreases in monsoon strength and precipitation at ca. 4500 to 5000 years
ago and around AD 1300, and a recent intensification in the driest reaches of the
monsoon (149, 151). The record from northern Africa is even more clear, with
dramatic and widespread shifts throughout the Holocene apparently in response to
regional climate (123, 152).The record of drought and megadrought (see section
3.3) offers abundant evidence that abrupt shifts in regional hydrologic regime occur
regularly in northern Africa. The ultimate causes of such abrupt change are not well
understood, but these changes seem to have occurred throughout the Holocene and
under a range of past climate forcing. It thus appears that abrupt shifts in monsoon
precipitation will continue to be a regular, if unpredictable, feature of monsoon
Asia and Africa.

3.7. Abrupt Change in Tropical Pacific/ENSO Systems


The ENSO system is the dominant pattern of interannual variance in global climate,
whose influence pervades the tropics and extends well into the midlatitudes of
much of the world. Through its physical climate influence, ENSO also impacts
ecosystems, the carbon cycle, human health, agriculture, and other systems. Thus,
unexpected change in this highly influential system has the potential to cause
significant disruptions in important natural and human systems. The role of ENSO
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

16 OVERPECK  COLE

in cold-climate abrupt change is discussed in section 2.4; here, we explore how


ENSO changed during the Holocene.
ENSO refers to a pattern of variability in time and space, so conceptualizing
abrupt change in ENSO requires that we think about changes in its dominant
frequencies, teleconnection patterns, or mean state about which the variations
occur. Paleoclimate data and modeling studies indicate that ENSO has the potential
to change abruptly in many of these aspects, even in response to gradual forcing.
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

Clement et al. (41) demonstrate how orbital forcing of a coupled Pacific model leads
to sudden, nonlinear responses in ENSO frequency and amplitude, as the system
abruptly locks to (or unlocks from) the seasonal cycle and interannual variance
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

shuts down for a century or more. Paleoclimate data support the idea of ENSO’s
response to orbital forcing and other slowly varying boundary conditions [e.g.,
(42)], but continuous records that allow the characterization of abrupt transitions
are rare. One site in Ecuador has yielded records indicating sudden changes in the
frequency and amplitude of strong El Niño events, superimposed on a pattern of
increasing frequency through the Holocene (72).
In the last millennium, reconstructions of ENSO from corals indicate that the
amplitude of interannual variability may have changed significantly even in the ab-
sence of large forcing changes (153). These variations are likely the result of both
significant internal variance and the system’s response to small radiative forc-
ing changes, particularly volcanic (153, 154). The variance spectrum of ENSO
changes significantly even over shorter intervals; for example, in the late nine-
teenth century, decadal variance appears to dominate tropical Pacific records, and
interannual variance was weak (155). The result of such a change is that ENSO-
related anomalies (such as drought) (156) persist longer, with potentially more
damaging consequences as the mechanisms for coping with briefer impacts are
overwhelmed. Spatial patterns of drought linked to ENSO have also changed in the
past century, perhaps as a consequence of changing conditions outside the tropical
Pacific (157). It is uncertain how ENSO will change in the future in response to
anthropogenic forcing, but this system clearly has the capability for abrupt changes
in frequency, amplitude, and teleconnection pattern.

4. ICE SHEET INSTABILITY AND ABRUPT


CHANGE IN SEA LEVEL
The most recent IPCC consensus regarding future ice sheet melting and resulting
global sea level rise suggests that both these changes would be slow and gradual
over the coming centuries (1). However, new evidence has emerged that ice sheets,
and thus global sea level, can respond more quickly to climate change, perhaps in
an abrupt manner.
Two primary types of evidence support the possibility that ice sheet collapse
could proceed fast enough to generate up to, and perhaps more than, 1 m of global
sea level rise per century. First, a significant number of new observations indicate
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 17

that both the Greenland and Antarctic ice sheets are responding more rapidly to
recent climate warming than anticipated and in ways that are more dynamically
complicated than represented in state-of-the-art ice sheet models (158, 159). This
raises concern that both ice sheets, but particularly the Antarctic ice sheet, could
lose mass more quickly than by simple surface melting alone.
Second, new research focusing on the last time global sea level was higher than
today (the last interglaciation, or LIG, approximately 129–116 ka) demonstrates
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

that relatively modest warming can precipitate abrupt ice sheet and sea level re-
sponses (Figure 3). During the LIG, sea level was 4–6 m higher than present.
These new studies suggest that ∼2.2 to 3.4 m of this rise came from Greenland
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

and other Arctic ice fields (93) and that the remainder likely came from Antarc-
tica (160)—most likely the West Antarctica ice sheet (WAIS) (161)—despite only
minor high-latitude Southern Hemisphere warming at the time. These findings
resurrect the original assertion of Mercer (162) that the WAIS could be quite sus-
ceptible to collapse in the future. Moreover, several lines of evidence suggest that
the rates of WAIS-driven sea level rise were fast, possibly exceeding 1 m per cen-
tury (160, 163). Given that warming temperatures and other factors could soon be
more conducive to ice sheet melting and collapse than at 129 ka (160), it is prudent
to be concerned about sea level rise starting abruptly in the twenty-first century
and lasting for centuries.

5. ABRUPT CLIMATE CHANGE “WILDCARDS”


Abrupt climate change and abrupt sea level change are real concerns for the future
and are worthy of effective and timely climate change mitigation and adaptation.
Here we mention three other exotic types of abrupt climate change, which, although
perhaps unlikely in the near-term future, played important roles in Earth’s history
and provide important climate change lessons. These are abrupt changes associated
with asteroid impacts, exceptionally large volcanoes, and clathrate melting. Al-
though we have little sway over the first two, we can work actively to avoid the latter.
The largest abrupt climate change event in the last 100 million years occurred
in association with the Chicxulub meteorite impact at the Cretaceous-Tertiary
boundary 65 Mya. Famous for wiping out the dinosaurs and driving one of the
largest mass extinctions in Earth’s history, the impact likely generated a significant
abrupt climate change characteristic of the five or six large meteorite impacts that
have occurred since multicellular life evolved on Earth over 500 Mya (164). In the
case of the Chicxulub impact event, the local geology of the Yucatan Peninsula
may have favored large emissions of greenhouse gases that dominated the climatic
impact (with warming) after the abrupt cooling impact of mineral, NOx and SO2
aerosols became limited by the rapid removal of these cooling agents from the
atmosphere.
Exceptionally large volcanic eruptions are somewhat more common than as-
teroid impacts and are possible drivers of abrupt climate change. However, to
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

18 OVERPECK  COLE

date, we know of no well-documented case in which a large volcanic eruption


triggered prolonged abrupt climate change. Climate change associated with vol-
canic eruptions is typically limited to the immediate fallout time of the strato-
spheric aerosol associated with a large eruption—a few years for a Pinatubo-sized
eruption (1991) to a decade or more for an exceptionally large eruption. This
could be due to a self-limiting effect of large eruptions (165). One exception-
ally large eruption, Toba, apparently did play a role related to abrupt North-
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

ern Hemisphere cooling at approximately 73 ka, although it appears that the


eruption probably did not trigger the abrupt changes, but rather intensified the
impact of a particular D/O event (see section 2.3) (166, 167). Nonetheless, a
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

similar magnitude eruption in the future would no doubt have serious climatic
implications.
Perhaps the most serious concern in terms of extreme abrupt climate change
threats is that associated with the potential release of methane hydrates from
clathrates, frozen methane trapped in the sediments along ocean margins (168).
Recent estimates suggest that as much as 2000 to 4000 gigatons (Gt) of carbon
could be released from these clathrate reservoirs if we allow atmospheric con-
centrations of CO2 to reach levels three times higher than preindustrial (168). If
such a release occurred, it would amplify tremendously the anthropogenic climate
forcing, perhaps in a very short time span; current fossil fuel emissions are approxi-
mately 6 Gt/yr (1). Paleoclimate evidence indicates that a clathrate methane release
along with subsequent abrupt climate and ocean change apparently did occur ap-
proximately 55 Mya, with the onset of the Paleocene-Eocene thermal maximum
(PETM). The abrupt global warming associated with this event lasted ∼100,000
years (169–171). The PETM increase in atmospheric CO2 led to a decrease in
ocean pH and widespread dissolution of seafloor carbonates (172); this process is
expected for ongoing and future increases in CO2 (173, 174). Unlike the case of
meteorite impacts and unusually large volcanic eruptions, humans have the power
to precipitate or preclude a large release of seafloor sediment carbon, along with
the abrupt climate and ocean changes that would result. Minimizing the causes
of anthropogenic climate change also minimizes the risk of clathrate methane
release.

6. IMPLICATIONS FOR THE FUTURE


Paleoclimate data provide both lessons and warnings regarding ongoing global
warming. The lessons include recognizing the likelihood that the climate system
is “wired” to respond to gradual forcing through abrupt transitions; the warn-
ings are obvious in that so much of our well-being depends on climate, from
agriculture and ecosystems to health, natural disasters, and coastal inundation.
We lack skilled predictive capability for abrupt change, and the prudent course
of action is therefore to do what we still can to minimize the possibility of its
occurrence.
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 19

6.1. Abrupt Climate Change Is Inevitable


The rapid growth of abrupt climate change science stems from the fact that such
change could be the biggest environmental concern for the future. It is the wildcard
of future climate change, regardless of the degree to which humans are transform-
ing Earth’s climate system. As this review makes clear, the climate system tends
to change abruptly, particularly when forced, but also perhaps owing to internal
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

climate system processes alone. The recent U.S. National Academy of Sciences
review of abrupt climate change science (2, 175) suggests that anthropogenic forc-
ing may increase the odds of abrupt climate change, based on the likelihood that
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

accelerated climate change forced by greenhouse gases will cause more abrupt
change thresholds to be crossed. This is evident for many of the types of abrupt
change reviewed above. Global warming and the associated enhanced hydrological
cycle will likely increase the odds of abrupt transitions to more drought- and flood-
prone regimes in some regions, most likely those areas that already experience
regular drought or floods. Rapid climate warming is expected to drive ice sheet
melting and collapse and, in doing so, cause abrupt sea level rise. An anthropogeni-
cally increased flux of freshwater into the North Atlantic from ice sheet melting is
the most likely way to accelerate the current freshening of the high-latitude North
Atlantic and to drive an abrupt decrease in Atlantic MOC. Theory and observations
support the likelihood of continued intensification of tropical storms and chang-
ing storm tracks associated with circulation changes; paleoclimatic observations
suggest storm track changes can occur abruptly.
In a geological context, the accelerating warming of Earth represents an abrupt
change that is unprecedented, perhaps, since the onset of the PETM, ∼55 Mya.
Of course, as discussed in section 5, the PETM was likely triggered by an abrupt
release of marine sediment methane hydrates (much of which was probably oxi-
dized to CO2 ). Humans appear to have the opportunity to push Earth past a similar
threshold as well and, in doing so, possibly ensure that that human-driven abrupt
warming is truly unprecedented in the history of life on Earth.

6.2. Predictability of Abrupt Climate Change


The biggest obstacle to reliable abrupt change prediction is the limited state of our
coupled atmosphere-ocean and ice sheet modeling capability. Coupled climate
modeling has reached the point where it is useful for assessing many aspects of
climate change, but these same models are still limited, or at least unproven, in
their ability to simulate realistic abrupt climate change in hydrologic regimes over
land as well as in the response of North Atlantic MOC to freshwater and wind
forcing. As discussed in section 4, ice sheet models are also unable to simulate
the full suite of dynamic processes that control how fast sea level can rise in the
future. A major challenge to the scientific community is to build models that can
simulate the observed record of past abrupt climate change in a realistic manner.
For the near future, however, it is possible only to say that abrupt changes of the
types discussed in this review are going to happen, but not when or where.
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

20 OVERPECK  COLE

6.3. What Evidence Exists for Recent Abrupt,


Anthropogenic Climate Change?
Based on paleoclimatic observations, modeling, and theory, there is reason to be-
lieve that anthropogenic climate change is already causing abrupt climate change
in some cases and is driving the climate system toward thresholds beyond which
other abrupt change is inevitable. The recent convergence of theory and observa-
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

tions (see section 3.5) suggests that an abrupt increase in strong tropical storms has
already begun. As discussed in section 3.6, Asian and African monsoon systems
may be starting to change abruptly. Although statistically robust attribution is not
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

yet possible, the ongoing near-decadal drought of western North America appears
linked to tropical SSTs that are highly suggestive of anthropogenic forcing (176).
Moreover, the severity of this current drought has been made worse by concurrent
record warmth, highlighting that this western U.S. drought may be the first anthro-
pogenic drought of the twenty-first century, with significant human and ecological
impacts [e.g., (177)].
As discussed in section 5, ice sheet melting is accelerating to rates greater
than anticipated. This change, plus the general unprecedented meltdown of the
Arctic (178), highlights that humans could be on track to push—during the current
century—Earth’s system across a threshold beyond which abrupt ice sheet collapse
and sea level rise is inevitable (178). Although less certain, the same could be true
for changes in the North Atlantic MOC. There are signs that rapid freshening
of the North Atlantic could soon trigger important changes in ocean circulation
(179–181).

7. CONCLUSIONS
The state-of-the-art understanding of abrupt climate change indicates the following
primary concerns for the broader climate and global change community:
 Paleoclimate records demonstrate that abrupt climate change is not excep-
tional but rather is a normal way in which the climate system responds to
changes in forcing.
 Regional forcing can trigger widespread (global and far-field remote) changes.
 Abrupt climate change occurs naturally in aspects of climate that are highly
relevant to society (e.g., droughts, floods, tropical storms, and sea level
change), and many of these aspects are changing now in ways consistent
with projected responses to anthropogenic forcing. Thus, the potential for
abrupt change in important components of the climate system is very real
and potentially immediate.
 The emerging theory of abrupt climate change suggests that we should ex-
pect the climate system to respond quickly without warning to current and
future climate forcing in ways whose specifics are difficult to predict with
geographic and temporal specificity.
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 21

 Anthropogenic climate change likely increases the threat of major abrupt


climate change and also likely increases the severity of some abrupt change
impacts. For example, future droughts will be hotter and thus more stressful
on human and natural systems. Also, increased chances of abrupt sea level
rise coincident with more powerful tropical storms is likely to result in
significantly greater multiple stress impacts in coastal regions (particularly
when growing population and investment in coastal regions are considered).
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

 Some abrupt climate change threats are plausible in coming years to decades
(e.g., drought and tropical storms), and these changes warrant public discus-
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

sion of no-regrets adaptation strategies (2) that would prove cost-effective


regardless of the magnitude of anthropogenic change (i.e., because other
factors, including natural climate variability and human population changes,
warrant action).
 Other abrupt climate change threats (e.g., major sea level rise and large-
scale clathrate methane release) are next generational in that human actions
of the twenty-first century could commit humankind to major abrupt climate
change events in future centuries. These potential changes warrant public
discussion of both adaptation and mitigation strategies.

Society has two primary response strategies for coping with abrupt climate
change.

1. There are many no-regrets steps society can make to reduce vulnerability
to abrupt climate system change particularly because greenhouse gas emis-
sions to date have already committed Earth to additional climate change.
For example, reducing vulnerability to prolonged drought, sea level rise, and
severe tropical storms is prudent regardless of whether future anthropogenic
climate change is small or large.
2. Reducing anthropogenic greenhouse gas forcing will likely reduce the prob-
ability of several high-impact types of abrupt climate change (i.e., ice sheet
disintegration, rapid large sea level rise, hot megadrought, intensified trop-
ical storms, abrupt change in ocean circulation, and ocean acidification).
Climate change is accelerating and may soon take us across thresholds be-
yond which Earth is committed to increased probabilities of substantial and
abrupt climate change. Effective action to reduce greenhouse gas emissions
should become an immediate priority if we wish to avoid crossing these
thresholds.

ACKNOWLEDGMENTS
We thank the many colleagues and students who have discussed abrupt climate
system change with us over the last decade. Special thanks to David Anderson,
Diana Liverman, and David Schimel for their advice and helpful reviews of the
manuscript.
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

22 OVERPECK  COLE

The Annual Review of Environment and Resources is online at


http://environ.annualreviews.org

LITERATURE CITED
1. Houghton JT, Ding Y, Griggs DJ, Noguer crease in Greenland snow accumulation
M, van der Linden, et al. 2001. Cli- at the end of the Younger Dryas event.
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

mate Change 2001: The Scientific Ba- Nature 362:527–29


sis: Intergovernmental Panel on Climate 12. Peteet D. 1995. Global Younger-Dryas.
Change (IPCC) Working Group I to the Quat. Int. 28:93–104
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

Third Assessment Report. Cambridge UK: 13. Broecker WS. 2003. Does the trigger for
Cambridge Univ. Press 881 pp. abrupt climate change reside in the ocean
2. Alley RB, Marotzke J, Nordhaus or in the atmosphere? Science 300:1519–
WD, Overpeck JT, Peteet DM, et al. 22
2003. Abrupt climate change. Science 14. Broecker WS, Denton GH. 1989. The
299:2005–10 role of ocean-atmosphere reorganizations
3. United Nations. 1992. United Nations in glacial cycles. Geochim. Cosmochim.
Framework Convention on Climate Acta 53:2465–501
Change 15. Lowell TV, Fisher TG, Comer GC, Hajdas
4. Schellnhuber HJ, Cramer W, Nakicenovic I, Waterson N, et al. 2005. Testing the
N, Wigley T, Yohe G, eds. 2006. Avoiding Lake Agassiz meltwater trigger for the
Dangerous Climate Change. Cambridge, Younger Dryas. EOS 86:365–72
UK: Cambridge Univ. Press. 406 pp. 16. Tarasov L, Peltier WR. 2005. Arctic fresh-
5. Natl. Res. Counc. 2002. Abrupt Climate water forcing of the Younger Dryas cold
Change: Inevitable Surprises. Washing- reversal. Nature 435:662–65
ton, DC: Natl. Acad. Press. 244 pp. 17. Hughen KA, Overpeck JT, Lehman SJ,
6. Rahmstorf S. 2001. Abrupt climate Kashgarian M, Southon J, et al. 1998.
change. In Encyclopedia of Ocean Sci- Deglacial changes in ocean circulation
ences, ed. J Steele, S Thorpe, K Turekian, from an extended radiocarbon calibration.
pp. 1–6. London: Academic Nature 391:65–68
7. Clark PU, Pisias NG, Stocker TF, Weaver 18. Muscheler R, Beer J, Wagner G, Finkel
AJ. 2002. The role of the thermohaline cir- RC. 2000. Changes in deep-water forma-
culation in abrupt climate change. Nature tion during the Younger Dryas event in-
415:863–69 ferred from Be-10 and C-14 records. Na-
8. Trenberth K, Overpeck J, Solomon S. ture 408:567–70
2004. Exploring drought and its implica- 19. Denton GH, Alley RB, Comer GC,
tions for the future. EOS 85:27 Broecker WS. 2005. The role of season-
9. Dansgaard W, White JWC, Johnsen S. ality in abrupt climate change. Quat. Sci.
1989. The abrupt termination of the Rev. 24:1159–82
Younger Dryas event. Nature 339:532– 20. Lie Ø, Paasche Ø. 2006. How extreme was
34 Northern Hemisphere seasonality dur-
10. Ruddiman WF, McIntyre A. 1981. The ing the Younger Dryas? Quat. Sci. Rev.
North-Atlantic Ocean during the last 25:404–7
deglaciation. Palaeogeogr. Palaeoclima- 21. Rutter NW, Weaver AJ, Rokosh D,
tol. Palaeoecol. 35:145–214 Fanning AF, Wright DG. 2000. Data-
11. Alley RB, Meese DA, Shuman CA, Gow model comparison of the Younger Dryas
AJ, Taylor KC, et al. 1993. Abrupt in- event. Can. J. Earth Sci. 37:811–30
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 23

22. Dansgaard W, Clausen HB, Gundestrup Dome West Antarctica, during the last
N, Hammer CU, Johnsen SF, et al. 1982. glacial period. Quat. Sci. Rev. 24:1333–
A new Greenland deep ice core. Science 43
218:1273–77 33. Blunier T, Brook EJ. 2001. Timing
23. Johnsen SJ, Clausen HB, Dansgaard W, of millennial-scale climate change in
Fuhrer K, Gundestrup N, et al. 1992. Ir- Antarctica and Greenland during the last
regular glacial interstadials recorded in a glacial period. Science 291:109–12
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

new Greenland ice core. Nature 359:311– 34. Rahmstorf S. 2002. Ocean circulation and
13 climate during the past 120,000 years. Na-
24. Heinrich H. 1988. Origin and conse- ture 419:207–14
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

quences of cyclic ice rafting in the 35. Stocker TF, Marchal O. 2000. Abrupt cli-
Northeast Atlantic Ocean during the past mate change in the computer: Is it real?
130,000 years. Quat. Res. 29:142–52 Proc. Natl. Acad. Sci. USA 97:1362–65
25. Bond G, Heinrich H, Broecker W, 36. Alley RB, Clark PU. 1999. The deglacia-
Labeyrie L, McManus J, et al. 1992. Evi- tion of the Northern Hemisphere: a global
dence for massive discharges of icebergs perspective. Annu. Rev. Earth Planet. Sci.
into the glacial North Atlantic. Nature 27:149–82
360:245–49 37. Broecker WS. 1998. Paleocean circula-
26. Hemming SR. 2004. Heinrich events: tion during the last deglaciation: A bipolar
massive late pleistocene detritus layers of seesaw? Paleoceanography 13:119–21
the North Atlantic and their global climate 38. Stocker TF, Johnsen SJ. 2003. A mini-
imprint. Rev. Geophys. 42:RG1005 mum thermodynamic model for the bipo-
27. Cortijo E, Duplessy JC, Labeyrie L, lar seesaw. Paleoceanography 18:1087;
Duprat J, Paillard D. 2005. Heinrich doi:10.1029/2003PA000920
events: hydrological impact. C. R. Geosci. 39. Cane MA, Clement A. 1999. A role
337:897–907 for the tropical Pacific coupled ocean-
28. Elliot M, Labeyrie L, Duplessy JC. 2002. atmosphere system on Milankovich and
Changes in North Atlantic deep-water for- millennial time scales. Part II: Global im-
mation associated with the Dansgaard- pacts. See Ref. 194, pp. 373–83
Oeschger temperature oscillations (60–10 40. Clement A, Cane MA. 1999. A role
ka). Quat. Sci. Rev. 21:1153–65 for the tropical Pacific coupled ocean-
29. Boyle EA. 2000. Is ocean thermoha- atmosphere system on Milankovich and
line circulation linked to abrupt stadial/ millennial time scales. Part 1: A modeling
interstadial transitions? Quat. Sci. Rev. study of tropical Pacific variability. See
19:255–72 Ref. 194, pp. 363–71
30. Dokken TM, Jansen E. 1999. Rapid 41. Clement AC, Cane MA, Seager R. 2001.
changes in the mechanism of ocean con- An orbitally driven tropical source for
vection during the last glacial period. Na- abrupt climate change. J. Clim. 14:2369–
ture 401:458–61 75
31. Indermuhle A, Monnin E, Stauffer B, 42. Tudhope AW, Chilcott CP, McCulloch
Stocker TF, Wahlen M. 2000. Atmo- MT, Cook ER, Chappell J, et al. 2001.
spheric CO2 concentration from 60 to 20 Variability in the El Niño Southern Oscil-
kyr BP from the Taylor Dome ice core, lation through a glacial-interglacial cycle.
Antarctica. Geophys. Res. Lett. 27:735– Science 291:1511–17
38 43. Clement AC, Seager R, Cane MA. 2000.
32. Brook EJ, White JWC, Schilla ASM, Ben- Suppression of El Niño during the mid-
der ML, Barnett B, et al. 2005. Timing of Holocene by changes in the Earth’s orbit.
millennial-scale climate change at Siple Paleoceanography 15:731–37
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

24 OVERPECK  COLE

44. Goswami BN, Madhusoodanan MS, 55. Latif M, Roeckner E, Mikolajewicz U,


Neema CP, Sengupta D. 2006. A physi- Voss R. 2000. Tropical stabilization of the
cal mechanism for North Atlantic SST in- thermohaline circulation in a greenhouse
fluence on the Indian summer monsoon. warming simulation. J. Clim. 13:1809–
Geophys. Res. Lett. 33:LO2706 13
45. Overpeck JT, Anderson DM, Trumbore 56. Broccoli AJ, Dahl KA, Stouffer RJ.
SE, Prell WL. 1996. The southwest Indian 2006. Response of the ITCZ to Northern
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

monsoon over the last 18,000 years. Clim. Hemisphere cooling. Geophys. Res. Lett.
Dyn. 12:213–25 33:LO1702
46. Gupta AK, Anderson D, Overpeck J. 57. Giannini A, Kushnir Y, Cane MA. 2000.
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

2003. Abrupt changes in the Holocene Interannual variability of Caribbean rain-


Asian southwest monsoon and their links fall, ENSO, and the Atlantic Ocean. J.
to the North Atlantic Ocean. Nature Clim. 13:297–311
421:354–57 58. Turney CSM, Kershaw AP, Clemens
47. Morrill C. 2002. Century-scale changes SC, Branch N, Moss PT, Fifield LK.
in Asian summer monsoon precipitation 2004. Millennial and orbital variations of
during the Holocene. PhD thesis. Univ. El Niño/Southern Oscillation and high-
Ariz., Tucson. 173 pp. latitude climate in the last glacial period.
48. Morrill C, Jacobsen RM. 2005. How Nature 428:306–10
widespread were climate anomalies 8200 59. Stott L, Poulsen C, Lund S, Thunell
years ago? Geophys. Res. Lett. 32:L1970 R. 2002. Super ENSO and global cli-
49. Wang YJ, Cheng H, Edwards RL, An ZS, mate oscillations at millennial time scales.
Wu JY, et al. 2001. A high-resolution Science 297:222–26
absolute-dated Late Pleistocene monsoon 60. Wagner JD, Cole JE, Beck JW, Patchett
record from Hulu Cave China. Science PJ, Henderson GM. 2005. Abrupt mil-
294:2345–48 lennial climate change in the Arizona
50. Schulz H, von Rad U, Erlenkeuser H. desert inferred from a speleothem isotopic
1998. Correlation between Arabian Sea record. EOS 86 (Abstr.):PP31A; 1499
and Greenland climate oscillations of the 61. Benson L, Lund S, Negrini R, Lins-
past 110,000 years. Nature 393:54–57 ley B, Zic M. 2003. Response of
51. Barnett TP, Dümenil L, Schlese U, Roeck- North American Great Basin Lakes to
ner E. 1988. The effect of Eurasian snow Dansgaard-Oeschger oscillations. Quat.
cover on global climate. Science 239:504– Sci. Rev. 22:2239–51
7 62. McCabe GJ, Palecki MA, Betancourt JL.
52. Trenberth KE, Branstator GW, Karoly D, 2004. Pacific and Atlantic Ocean influ-
Kumar A, Lau NC, Ropelewski CF. 1998. ences on multidecadal drought frequency
Progress during TOGA in understanding in the United States. Proc. Natl. Acad. Sci.
and modeling global teleconnections as- USA 101:4136–41
sociated with tropical sea surface temper- 63. Kumar KK, Kleeman R, Cane MA, Ra-
atures. J. Geophys. Res. 103:14291–324 jagopalan B. 1999. Epochal changes in
53. Peterson LC, Haug GH, Hughen KA, Rohl Indian monsoon-ENSO precursors. Geo-
U. 2000. Rapid changes in the hydrologic phys. Res. Lett. 26:75–78
cycle of the tropical Atlantic during the 64. Bond G, Showers W, Elliot M, Evans
last glacial. Science 290:1947–51 M, Lotti R, et al. 1999. The North
54. Schmittner A, Appenzeller C, Stocker TF. Atlantic’s 102 kyr climate rhythm: re-
2000. Enhanced Atlantic freshwater ex- lation to Heinrich events, Dansgaard/
port during El Niño. Geophys. Res. Lett. Oeschger cycles and the Little Ice Age.
27:1163–66 See Ref. 194, pp. 35–58
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 25

65. Mayewski PA, Meeker LD, Twickler MS, the Greenland time scale: implications for
Whitlow S, Yang QZ, et al. 1997. Ma- Antarctic time scales and for Delta C-14.
jor features and forcing of high-latitude Quat. Sci. Rev. 23:1513–22
Northern Hemisphere atmospheric circu- 76. Mann ME, Bradley RS, Hughes MK.
lation using a 110,000-year-long glacio- 1999. Northern Hemisphere temperatures
chemical series. J. Geophys. Res.-Ocean. during the past millennium: inferences,
102:26345–66 uncertainties, and limitations. Geophys.
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

66. Clemens SC. 2005. Millennial-band cli- Res. Lett. 26:759–62


mate spectrum resolved and linked to 77. Pollack HN, Huang SP. 2000. Climate
centennial-scale solar cycles. Quat. Sci. reconstruction from subsurface temper-
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

Rev. 24:521–31 atures. Annu. Rev. Earth Planet. Sci.


67. Alley RB, Anandakrishnan S, Jung P. 28:339–65
2001. Stochastic resonance in the North 78. Cook ER, Esper J, D’Arrigo RD. 2004.
Atlantic. Paleoceanography 16:190–98 Extra-tropical Northern Hemisphere land
68. Schulz M. 2002. The tempo of climate temperature variability over the past 1000
change during Dansgaard-Oeschger inter- years. Quat. Sci. Rev. 23:2063–74
stadials and its potential to affect the man- 79. Osborn T, Briffa KR. 2006. The spatial
ifestation of the 1470-year climate cycle. extent of twentieth-century warmth in the
Geophys. Res. Lett. 29:1002 context of the past 1200 years. Science
69. Rahmstorf S. 2003. Timing of abrupt cli- 311:841–44
mate change: a precise clock. Geophys. 80. Trenberth KE, Hurrell JW. 1994. Decadal
Res. Lett. 30:1510 atmosphere-ocean variations in the Pa-
70. Bond G, Kromer B, Beer J, Muscheler R, cific. Clim. Dyn. 9:303–19
Evans MN, et al. 2001. Persistent solar in- 81. Graham NE. 1995. Simulation of re-
fluence on North Atlantic climate during cent global temperature trends. Science
the Holocene. Science 294:2130–36 267:686–71
71. Braun H, Christl M, Rahmstorf S, 82. Barnett TP, Adam JC, Lettenmaier DP.
Ganopolski A, Mangini A, et al. 2005. 2005. Potential impacts of a warming
Possible solar origin of the 1,470-year climate on water availability in snow-
glacial climate cycle demonstrated in a dominated regions. Nature 438:303–9
coupled model. Nature 438:208–11 83. Overpeck J. 1996. Warm climate sur-
72. Moy CM, Seltzer GO, Rodbell DT, prises. Science 271:1820–21
Anderson DM. 2002. Variability of El 84. Overpeck J, Webb R. 2000. Nonglacial
Niño/Southern Oscillation activity at mil- rapid climate events: past and future. Proc.
lennial timescales during the Holocene Natl. Acad. Sci. USA 97:1335–38
epoch. Nature 420:162–65 85. Manabe S, Stouffer RJ. 1994. Multiple-
73. Genty D, Blamart D, Ouahdi R, Gilmour century response of a coupled ocean-
M, Baker A, et al. 2003. Precise dating of atmosphere model to an increase of
Dansgaard-Oeschger climate oscillations atmospheric carbon-dioxide. J. Clim. 7:5–
in western Europe from stalagmite data. 23
Nature 421:833–37 86. Rahmstorf S. 2000. The thermohaline
74. Burns SJ, Fleitmann D, Matter A, ocean circulation: A system with dan-
Kramers J, Al-Subbary AA. 2003. Indian gerous thresholds? An editorial comment.
Ocean climate and an absolute chronol- Clim. Chang. 46:247–56
ogy over Dansgaard/Oeschger events 9 to 87. Rind D, Overpeck JT. 1993. Hypothe-
13. Science 301:1365–67 sized causes of decade-to-century-scale
75. Shackleton NJ, Fairbanks RG, Chiu TC, climate variability: climate model results.
Parrenin F. 2004. Absolute calibration of Quat. Sci. Rev. 12:357–74
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

26 OVERPECK  COLE

88. Alley RB, Mayewski PA, Sowers T, Stu- ice in the northern North Atlantic: evi-
iver M, Taylor KC, Clark PU. 1997. dence for different spatial and temporal
Holocene climatic instability: a promi- modes. Paleoceanography 21:2017; doi:
nent, widespread event 8200 years ago. 10.1029/2005PA001214
Geology 25:483–86 98. Hu FS, Kaufman D, Yoneji S, Nelson D,
89. Alley RB, Agustsdottir AM. 2005. The 8k Shemesh A, et al. 2003. Cyclic variation
event: cause and consequences of a ma- and solar forcing of Holocene climate in
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

jor Holocene abrupt climate change. Quat. the Alaskan subarctic. Science 301:1890–
Sci. Rev. 24:1123–49 93
90. LeGrande AN, Schmidt GA, Shindell DT, 99. Woodhouse CA, Overpeck JT. 1998. 2000
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

Field CV, Miller RL, et al. 2006. Con- years of drought variability in the central
sistent simulations of multiple proxy re- United States. Bull. Am. Meteorol. Soc.
sponses to an abrupt climate change event. 79:2693–714
Proc. Natl. Acad. Sci. USA 103:837–42 100. Schubert SD, Suarez MJ, Pegion PJ,
91. Morrill C, Overpeck JT, Cole JE. 2003. Koster RD, Bacmeister JT. 2004. Causes
A synthesis of abrupt changes in the of long-term drought in the US Great
Asian monsoon since the last deglacia- Plains. J. Clim. 17:485–503
tion. Holocene 13:465–76 101. Seager R, Kushnir Y, Herweijer C, Naik
92. Morrill C, Overpeck JT, Cole JE, Liu KB, N, Velez J. 2005. Modeling of tropical
Shen CM, Tang LY. 2006. Holocene vari- forcing of persistent droughts and pluvials
ations in the Asian monsoon inferred from over western North America: 1856–2000.
the geochemistry of lake sediments in cen- J. Clim. 18:4065–88
tral Tibet. Quat. Res. 65:232–43 102. Hoerling MP, Kumar A. 2003. The per-
93. Otto-Bliesner BL, Marshall SJ, Overpeck fect ocean for drought. Science 299:691–
JT, Miller GH, Hu AX, CAPE Last Inter- 94
glaciation Project Members. 2006. Simu- 103. Hughes MK, Graumlich LJ. 1996. Mul-
lating Arctic climate warmth and icefield timillennial dendoclimatic records from
retreat in the last interglaciation. Science western North America. In Climatic Vari-
311:1751–53 ations and Forcing Mechanisms of the
94. Andrews JT, Smith LM, Preston R, Last 2000 Years, ed. PD Jones, RS
Cooper T, Jennings AE. 1997. Spatial Bradley, J Jouzel, pp. 109–24. Berlin:
and temporal patterns of iceberg rafting Springer-Verlag
(IRD) along the East Greenland margin, 104. Cook ER, Woodhouse CA, Eakin CM,
ca 68◦ N, over the last 14 cal ka. J. Quat. Meko DM, Stahle DW. 2004. Long-term
Sci. 12:1–13 aridity changes in the western United
95. Risebrobakken B, Jansen E, Andersson States. Science 306:1015–18
C, Mjelde E, Hevroy K. 2003. A high- 105. Stahle DW, Cook ER, Cleaveland MK,
resolution study of Holocene paleocli- Therrell MD, Meko DM, et al. 2000.
matic and paleoceanographic changes Tree-ring data document sixteenth cen-
in the Nordic seas. Paleoceanography tury megadrought over North America.
18(1017):764–77 EOS Trans. Am. Geophys. Union 81:121,
96. Schulz M, Paul A, Timmermann A. 2004. 125
Glacial-interglacial contrast in climate 106. Grissino-Mayer HD. 1996. A 2129-year
variability at centennial-to-millennial annual reconstruction of precipitation for
timescales: observations and conceptual northwestern New Mexico USA. In Tree-
model. Quat. Sci. Rev. 23:2219–30 Rings Environment and Humanity, ed.
97. Moros M, Andrews JT, Eberl DE, Jansen DM Meko, TW Swetnam, JS Dean,
E. 2006. The Holocene history of drift pp. 191–204. Tucson: Univ. Ariz. Press
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 27

107. Salzer MW, Kipfmueller KF. 2005. Re- 116. Poore RZ, Pavich MJ, Grissino-Mayer
constructed temperature and precipitation HD. 2005. Record of the North
on a millennial timescale from tree-rings American southwest monsoon from
in the southern Colorado Plateau USA. Gulf of Mexico sediment cores. Geology
Clim. Chang. 70:465–87 33:209–12
108. Stahle DW, Cleaveland MK, Blanton DB, 117. Anderson RS, Betancourt JL, Mead JI,
Therrell MD, Gay DA. 1998. The Lost Hevly RH, Adam DP. 2000. Middle- and
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

Colony and Jamestown droughts. Science late-Wisconsin paleobotanic and paleocli-


280:564–67 matic records from the southern Colorado
109. Stine S. 1994. Extreme and persistent Plateau USA. Palaeogeogr. Palaeoclima-
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

drought in California and Patagonia dur- tol. Palaeoecol. 155:31–57


ing medieval time. Nature 369:546–49 118. Wagner JD, Cole JE, Beck JW, Patchett
110. Laird KR, Cumming BF, Wunsam S, PJ, Henderson GM. 2005. Early to mid-
Rusak JA, Oglesby RJ, et al. 2003. Holocene climate in southern Arizona in-
Lake sediments record large-scale shifts ferred from speleothem stable isotopes.
in moisture regimes across the northern Presented at GSA Annu. Meet., Salt Lake
prairies of North America during the past City, Utah
two millennia. Proc. Natl. Acad. Sci. USA 119. Rind D. 1988. The doubled CO2 cli-
100:2483–88 mate and the sensitivity of the mod-
111. Forman SL, Oglesby R, Webb RS. 2001. eled hydrologic-cycle. J. Geophys. Res.-
Temporal and spatial patterns of Holocene Atmos. 93:5385–412
dune activity on the Great Plains of 120. de Menocal P, Ortiz J, Guilderson T, Ad-
North America: megadroughts and cli- kins J, Sarnthein M, et al. 2000. Abrupt
mate links. Glob. Planet. Chang. 29:1–29 onset and termination of the African Hu-
112. Benson L, Kashgarian M, Rye R, Lund mid Period: rapid climate responses to
S, Paillet F, et al. 2002. Holocene multi- gradual insolation forcing. Quat. Sci. Rev.
decadal and multicentennial droughts af- 19:347–61
fecting northern California and Nevada. 121. Verschuren D, Laird KR, Cumming BF.
Quat. Sci. Rev. 21:659–82 2000. Rainfall and drought in equatorial
113. Quade J, Forester RM, Pratt WL, Carter East Africa during the past 1,100 years.
C. 1998. Black mats, spring-fed streams, Nature 403:410–14
and late-glacial-age recharge in the south- 122. Stager JC, Cumming BF, Meeker LD.
ern Great Basin. Quat. Res. 49:129– 2003. A 10,000-year high-resolution di-
48 atom record from Pilkington Bay Lake
114. Thompson RS, Whitlock C, Bartlein PJ, Victoria East Africa. Quat. Res. 59:172–
Harrison SP, Spaulding WG. 1993. Cli- 81
matic changes in the western United 123. Shanahan T, Overpeck JT, Wheeler CW,
States since 18,000 years B.P. In Global Beck JW, Pigati JS, et al. 2006. Pa-
Climates Since the Last Glacial Maxi- leoclimatic variations in West Africa
mum, ed. HE Wright Jr, JE Kutzbach, WF from a record of late Pleistocene and
Ruddiman, FA Street-Perrott, T Webb III, Holocene lake level stands of Lake Bo-
PJ Bartlein, pp. 468–513. Minneapolis: sumtwi Ghana. Palaeogeogr. Palaeocli-
Univ. Minn. Press matol. Palaeoecol. In press
115. Metcalfe SE, O’Hara SL, Caballero 124. Claussen M, Kubatzki C, Brovkin V,
M, Davies SJ. 2000. Records of Late Ganopolski A, Hoelzmann P, Pachur HJ.
Pleistocene-Holocene climatic change 1999. Simulation of an abrupt change in
in Mexico—a review. Quat. Sci. Rev. Saharan vegetation in the mid-Holocene.
19:699–721 Geophys. Res. Lett. 26:2037–40
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

28 OVERPECK  COLE

125. Dai AG, Lamb PJ, Trenberth KE, Hulme 137. Knox JC. 1993. Large increases in flood
M, Jones PD, Xie PP. 2004. The recent magnitude in response to modest changes
Sahel drought is real. Int. J. Climatol. in climate. Nature 361:430–32
24:1323–31 138. Ely LL, Enzel Y, Baker VR, Cayan DR.
126. Giannini A, Saravanan R, Chang P. 2003. 1993. A 5000-year record of extreme
Oceanic forcing of Sahel rainfall on in- floods and climate-change in the south-
terannual to interdecadal time scales. Sci- western United States. Science 262:410–
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

ence 302:1027–30 12
127. Nicholson S. 2000. Land surface pro- 139. Knutson TR, Tuleya RE. 2004. Impact
cesses and Sahel climate. Rev. Geophys. of CO2 -induced warming on simulated
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

38:117–39 hurricane intensity and precipitation: sen-


128. Haug GH, Hughen KA, Sigman DM, sitivity to the choice of climate model
Peterson LC, Rohl U. 2001. Southward and convective parameterization. J. Clim.
migration of the intertropical conver- 17:3477–95
gence zone through the Holocene. Science 140. Emanuel K. 2005. Increasing destructive-
293:1304–8 ness of tropical cyclones over the past
129. Hodell DA, Curtis JH, Brenner M. 1995. 30 years. Nature 436:686–88
Possible role of climate in the collapse 141. Webster PJ, Holland GJ, Curry JA, Chang
of the classic Maya civilization. Nature HR. 2005. Changes in tropical cyclone
375:391–94 number, duration, and intensity in a warm-
130. Overpeck J, Cole J, Bartlein P. 2005. A ing environment. Science 309:1844–
“paleoperspective” on climate variability 46
and change. In Climate Change and Bio- 142. Hoyos CD, Agudelo PA, Webster PJ,
diversity, ed. T Lovejoy, L Hannah, pp. Curry JA. 2006. Deconvolution of the fac-
91–108. New Haven: Yale Univ. Press tors contributing to the increase in global
131. Milly PCD, Wetherald RT, Dunne KA, hurricane intensity. Science 312:94–97
Delworth TL. 2002. Increasing risk of 143. Liu KB. 2004. Paleotempestology: prin-
great floods in a changing climate. Nature ciples, methods, and examples from Gulf
415:514–17 Coast lake-sediments. In Hurricanes and
132. Trenberth KE, Dai AG, Rasmussen RM, Typhoons: Past, Present, and Future, ed.
Parsons DB. 2003. The changing charac- R Murnane, K Liu, pp. 13–57. New York:
ter of precipitation. Bull. Am. Meteorol. Columbia Univ. Press
Soc. 84:1205–17 144. Liu KB, Fearn ML. 2000. Reconstruc-
133. Groisman PY, Knight RW, Easterling DR, tion of prehistoric landfall frequencies of
Karl TR, Hegerl GC, Razuvaev VAN. catastrophic hurricanes in northwestern
2005. Trends in intense precipitation in Florida from lake sediment records. Quat.
the climate record. J. Clim. 18:1326–50 Res. 54:238–45
134. Stewart IT, Cayan DR, Dettinger MD. 145. Scott DB, Collins ES, Gayes PT, Wright
2005. Changes toward earlier streamflow E. 2003. Records of prehistoric hurricanes
timing across western North America. on the South Carolina coast based on
J. Clim. 18:1136–55 micropaleontological and sedimentolog-
135. Ely L. 1997. Response of extreme floods ical evidence, with comparison to other
in the southwestern United States to cli- Atlantic Coast records. Geol. Soc. Am.
matic variations in the late Holocene. Ge- Bull. 115:1027–39
omorphology 19:175–201 146. Liu KB, Shen CM, Louie KS. 2001.
136. Knox JC. 2000. Sensitivity of modern and A 1,000-year history of typhoon land-
Holocene floods to climate change. Quat. falls in Guangdong, southern China,
Sci. Rev. 19:439–57 reconstructed from Chinese historical
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 29

documentary records. Ann. Assoc. Am. 157. Cole JE, Cook ER. 1998. The changing re-
Geogr. 91:453–64 lationship between ENSO variability and
147. Wu LG, Wang B, Geng SQ. 2005. Grow- moisture balance in the continental United
ing typhoon influence on east Asia. Geo- States. Geophys. Res. Lett. 25:4529–32
phys. Res. Lett. 32:L18703 158. Alley RB, Clark PU, Huybrechts P,
148. Overpeck J, Anderson D, Trumbore S, Joughin I. 2005. Ice-sheet and sea-level
Prell W. 1996. The southwest Indian mon- changes. Science 310:456–60
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

soon over the last 18000 years. Clim. Dyn. 159. Domack E, Duran D, Leventer A, Ishman
12:213–25 S, Doane S, et al. 2005. Stability of the
149. Overpeck J, Liu K, Morrill C, Cole Larsen B ice shelf on the Antarctic Penin-
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

J, Shen C, et al. 2005. Holocene en- sula during the Holocene epoch. Nature
vironmental change in the Himalayan- 436:681–85
Tibetan Plateau region: lake sediment 160. Overpeck JT, Otto-Bliesner BL, Miller
and the future. In Global Change and GH, Muhs DR, Alley RB, Kiehl JT.
Mountain Regions: An Overview of Cur- 2006. Paleoclimatic evidence for future
rent Knowledge, ed. UM Huber, HKM ice-sheet instability and rapid sea-level
Bugmann, MA Reasoner, pp. 83–92. rise. Science 311:1747–50
Dordrecht: Kluwer Acad. 161. Scherer RP, Aldahan A, Tulaczyk S, Poss-
150. Anderson DM, Overpeck JT, Gupta AK. nert G, Engelhardt H, Kamb B. 1998.
2002. Increase in the Asian southwest Pleistocene collapse of the West Antarctic
monsoon during the past four centuries. ice sheet. Science 281:82–85
Science 297:596–99 162. Mercer JH. 1978. West Antarctic ice sheet
151. Brauning A, Mantwill B. 2004. Summer and CO2 greenhouse effect—threat of dis-
temperature and summer monsoon his- aster. Nature 271:321–25
tory on the Tibetan plateau during the last 163. McCulloch MT, Esat T. 2000. The coral
400 years recorded by tree rings. Geophys. record of last interglacial sea levels and
Res. Lett. 31:L24205 sea surface temperatures. Chem. Geol.
152. Gasse F. 2000. Hydrological changes in 169:107–29
the African tropics since the last glacial 164. Kring D. 2000. Impact events and their
maximum. Quat. Sci. Rev. 19:189–211 effect on the origin, evolution, and distri-
153. Cobb KM, Charles CD, Cheng H, bution of life. GSA Today 10:1–6
Edwards RL. 2003. El Niño/Southern Os- 165. Pinto JP, Turco RP, Toon OB. 1989. Self-
cillation and tropical Pacific climate dur- limiting physical and chemical effects
ing the last millennium. Nature 424:271– in volcanic-eruption clouds. J. Geophys.
76 Res.-Atmos. 94:11165–74
154. Mann ME, Cane MA, Zebiak SE, Clement 166. Zielinski GA. 2000. Use of paleo-records
A. 2005. Volcanic and solar forcing of the in determining variability within the
tropical Pacific over the past 1000 years. volcanism-climate system. Quat. Sci. Rev.
J. Clim. 18:447–56 19:417–38
155. Urban FE, Cole JE, Overpeck JT. 2000. 167. Schulz H, Emeis KC, Erlenkeuser H, von
Modification of tropical Pacific variabil- Rad U, Rolf C. 2002. The Toba volcanic
ity by its mean state inferred from a 155 event and interstadial/stadial climates at
year coral record. Nature 407:989–93 the marine isotopic stage 5 to 4 transition
156. Cole JE, Overpeck JT, Cook ER. 2002. in the northern Indian Ocean. Quat. Res.
Multiyear La Niña events and persis- 57:22–31
tent drought in the contiguous United 168. Archer D, Buffett B. 2005. Time-
States. Geophys. Res. Lett. 29:1647; doi dependent response of the global ocean
10.1029/2001GL013561 clathrate reservoir to climatic and
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

30 OVERPECK  COLE

anthropogenic forcing. Geochem. Geo- Recent changes in the North Atlantic.


phys. Geosyst. 6:Q03002 Philos. Trans. R. Soc. London Ser. A
169. Kennett JP, Stott LD. 1991. Abrupt deep- 361:1917–33
sea warming palaeoceanographic changes 180. Curry R, Mauritzen C. 2005. Dilution of
and benthic extinctions at the end of the the northern North Atlantic Ocean in re-
Paleocene. Nature 353:225–29 cent decades. Science 308:1772–74
170. Zachos JC, Wara MW, Bohaty S, Delaney 181. Bryden HL, Longworth HR, Cunningham
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

ML, Petrizzo MR, et al. 2003. A tran- SA. 2005. Slowing of the Atlantic merid-
sient rise in tropical sea surface tempera- ional overturning circulation at 25◦ N. Na-
ture during the Paleocene-Eocene thermal ture 438:655–57
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

maximum. Science 302:1551–54 182. Grootes PM, Stuiver M, White JWC,


171. Tripati A, Elderfield H. 2005. Deep-sea Johnsen S, Jouzel J. 1993. Comparison of
temperature and circulation changes at oxygen isotope records from the GISP2
the Paleocene-Eocene thermal maximum. and GRIP Greenland ice cores. Nature
Science 308:1894–98 366:552–54
172. Zachos JC, Rohl U, Schellenberg SA, 183. Shackleton NJ. 2000. The 100,000-year
Sluijs A, Hodell DA, et al. 2005. ice-age cycle identified and found to lag
Rapid acidification of the ocean during temperature, carbon dioxide, and orbital
the Paleocene-Eocene thermal maximum. eccentricity. Science 289:1897–902
Science 308:1611–15 184. Hendy IL, Kennett JP. 2000. Dansgaard-
173. Kleypas JA, Buddemeier RW, Archer D, Oeschger cycles and the California Cur-
Gattuso JP, Langdon C, Opdyke BN. rent System: planktonic foraminiferal re-
1999. Geochemical consequences of in- sponse to rapid climate change in Santa
creased atmospheric carbon dioxide on Barbara basin Ocean Drilling Program
coral reefs. Science 284:118–20 hole 893A. Paleoceanography 15:30–42
174. Orr JC, Fabry VJ, Aumont O, Bopp L, 185. Berger AL. 1978. Long-term variations of
Doney SC, et al. 2005. Anthropogenic caloric insolation resulting from Earth’s
ocean acidification over the twenty-first orbital elements. Quat. Res. 9:139–67
century and its impact on calcifying or- 186. Peltier WR, Fairbanks RG. 2006. Global
ganisms. Nature 437:681–86 glacial ice volume and last glacial maxi-
175. Alley RB, Marotzke J, Nordhaus WD, mum duration from an extended Barbados
Overpeck JT, Peteet DM, et al. 2002. sea level record. Quat. Sci. Rev. In press
Abrupt Climate Change: Inevitable 187. Genty D, Nebout NC, Hatte C, Blamart
Surpises. Washington, DC: Natl. Acad. D, Ghaleb B, Isabello L. 2005. Rapid cli-
Press. 230 pp. matic changes of the last 90 kyr recorded
176. Hoerling M, Kumar A. 2003. The perfect on the European continent. C. R. Geosci.
ocean for drought. Science 299:691–94 337:970–82
177. Breshears DD, Cobb NS, Rich PM, Price 188. Allen JRM, Brandt U, Brauer A, Hub-
KP, Allen CD, et al. 2005. Regional berten H-W, Huntley B, et al. 1999. Rapid
vegetation die-off in response to global- environmental fluctuations in southern
change-type drought. Proc. Natl. Acad. Europe during the last glacial period. Na-
Sci. USA 102:15144–48 ture 400:740–43
178. Overpeck JT, Otto-Bliesner BL, Kiehl 189. Arz HW, Patzold J, Wefer G. 1998. Corre-
JT, Miller GH, Alley RB. 2006. Paleo- lated millennial-scale changes in surface
climatic evidence for future ice sheet in- hydrography and terrigenous sediment
stability and rapid sea level rise. Science yield inferred from last-glacial marine de-
311:1747–50 posits off northeastern Brazil. Quat. Res.
179. Dickson RR, Curry R, Yashayaev I. 2003. 50:157–66
10 Oct 2006 12:35 AR ANRV289-EG31-01.tex XMLPublishSM (2004/02/24) P1: OKZ

ABRUPT CLIMATE CHANGE 31

190. Cruz FW, Burns SJ, Karmann I, Sharp NE Pacific, during the past 60 kyr. Nature
WD, Vuille M, et al. 2005. Insolation- 379:243–46
driven changes in atmospheric circulation 193. Benson LV, Burdett JW, Kashgarian M,
over the past 116,000 years in subtropical Lund SP, Phillips FM, Rye RO. 1996. Cli-
Brazil. Nature 434:63–66 matic and hydrologic oscillations in the
191. Ruhlemann C, Mulitza S, Muller PJ, Owens Lake basin and adjacent Sierra
Wefer G, Zahn R. 1999. Warming of the Nevada California. Science 274:746–
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

tropical Atlantic Ocean and slowdown of 49


thermohaline circulation during the last 194. Clark PU, Webb RS, Keigwin LD,
deglaciation. Nature 402:511–14 eds. 1999. Mechanisms of Global Cli-
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

192. Behl RJ, Kennett JP. 1996. Brief inter- mate Change at Millennial Time Scales.
stadial events in the Santa Barbara basin, Washington, DC: Am. Geophys. Union
HI-RES-EG31-01-Overpeck.qxd 10/10/06 12:47 PM Page 1

ABRUPT CLIMATE CHANGE C-1

10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000


-34
YD H1 H2 H3 H4 H5 H6
-36
A. GISP2 δ18O
δ18O (‰)

-38

-40 0
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

-42 0.5
1

δ18O (‰)
B. SST,
-44 Iberian margin
1.5
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

2
7
Reflectance (% at 550 nm )

2.5
9
11 C.Precipitation, 3
13 N. South America 3.5
15
17
-0.5
19
0
21

δ18O (‰)
23 0.5
25 D. SST, Santa 1
Barbara basin
1.5
2
-9.5 2.5
-8.5 3
δ18O (‰)

E. East Asian 3.5


-7.5 monsoon
strength, China
-6.5

-5.5

-4.5
YD H1 H2 H3 H4 H5 H6

10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000


Years before present

Figure 1 (See legend on next page)


HI-RES-EG31-01-Overpeck.qxd 10/10/06 12:47 PM Page 2

C-2 OVERPECK ■ COLE

Figure 1 A subset of the many records worldwide that show cold-climate abrupt change,
particularly Dansgaard/Oeschger (D/O) events. (A) Greenland ice core oxygen isotope
record (182), usually interpreted as local temperature. (B) Sea surface temperature (SST)
reconstructed from a planktonic foraminiferal oxygen isotope record off the Iberian margin
(183). (C) Cariaco Basin sediment reflectance, representing productivity and river-borne
terrigenous material (northern South America) (53). Increased productivity and terrigenous
input are driven by regional rainfall, linked to Inter-Tropical Convergence Zone (ITCZ)
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

position; lower reflectance values thus correspond to more organic-rich and darker sedi-
ments, higher productivity, higher rainfall, and a relatively northward mean position of the
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

ITCZ. (D) SST reconstructed from planktonic foraminiferal oxygen isotopes from the
Santa Barbara basin (184). Results from two species are shown here, Globigerina bulloides
(orange) and Neogloboquadrina pachyderma (pink). (E) Speleothem oxygen isotope
record from Hulu Cave, near Nanjing, China (49). Values represent the strength of summer
relative to the winter monsoon. Bars (gray) indicate selected D/O event visible in all
records, and the thicker gray bar at left, labeled YD, highlights the Younger Dryas event.
Heinrich events are shown by boxes at the top and bottom, labeled H1 through H6.
HI-RES-EG31-01-Overpeck.qxd 10/10/06 12:47 PM Page 3

ABRUPT CLIMATE CHANGE C-3

Drought area index


A. B. C. 15 25 35 45 55 E.
2000

1800

1600
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

1400

1200
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

Ye ar A.D.

1000

800
D.
600

400

200
Wet ter Dr ier
0

-200

-400
-2 -1 0 1 2 -2 -1 0 1 -0.2 0 0.2 22 21 20 19 18
Nor malized deviation Nor malized Reconstr ucted
from mean log salinity devia tion from precipitation
mean total P (cm/year)

Figure 2 An example of abrupt change in hydrologic regime during the Holocene, in the
western United States. Records A-C each reflect precipitation minus evaporation and were
derived from radiocarbon-dated records of diatom species abundance in northern Great
Plains lakes (110). Records A and B are reconstructions of salinity from Coldwater Lake
and Moon Lake, respectively, and record C is a reconstruction of total phosphorus (P)
(related to salinity) from Elk Lake. Records D and E are based on tree rings (thus precise-
ly dated). Record D is the drought area index over the western United States, which reflects
the number of grid spares (out of 103) experiencing drought (104). Record E is a precipi-
tation reconstruction from the White Mountains in southeast California (103). The lake
records suggest a hydrologic transition around AD 1200, with the sign of the transition
reflecting that one area became abruptly drier at the same time the other area turned abrupt-
ly wetter. The tree-ring records indicate a transition around AD 1300, with greater inci-
dence of drought and fewer wet intervals before this time. These shifts may be the same
event, given the error in radiocarbon lake sediment chronologies, or the western U.S.
abrupt change may have occurred following the Great Plains shift.
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.
C-4

Time (kyrs before present) Time (kyrs before present)


0 3 6 9 12 15 18 120 126 132 138
D D
0
O 0 -20 0 -20 -40 O
-20 20 -40 -20 0
A 20 A
J 40 40 60 J

Month
40
Month
A 20 60 40 20
0 A
0
HI-RES-EG31-01-Overpeck.qxd

F 0 0 F
0 0
Modern sea level Modern sea level
OVERPECK ■ COLE

20 20
10/10/06

40 40

60 60
12:47 PM

80 80
Page 4

100 100
? ?

Depth below modern sea level (m)


Depth below modern sea level (m)

120 120

Super-interglacial (includes threshold above which GIS melts; possible sea level rise > 1 m/year)

Deglaciation/interglacial (GIS stable at Holocene configuration; deglacial sea level rise up to ~1 m/year)

Initial deglaciation (GIS begins to melt slowly from glacial configuration; sea level rise << 1m/year)

Figure 3 (See legend on next page)


HI-RES-EG31-01-Overpeck.qxd 10/10/06 12:47 PM Page 5

ABRUPT CLIMATE CHANGE C-5

Figure 3 Abrupt ice sheet retreat and sea level rise. Insolation (at 70°N, anomaly from
present) and reconstructed global sea level rise over the past two major deglaciations in
Earth’s history, illustrating how the rate of sea level rise depends on the magnitude and rate
of radiative forcing change. Over the last 20 kyrs, summer radiative forcing peaked at
50 W m2 owing to changes in Earth’s orbit and did not cause significant retreat of the
Greenland ice sheet (GIS) from the present day configuration of the ice sheet. In contrast,
the insolation anomaly was 40% larger (70 W m2) over the penultimate deglaciation
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

as well as into the subsequent interglaciation (the Eemian) when global sea level was 4 to
6 m above present (178). Sea level rise was significantly more abrupt over this earlier peri-
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

od, and retreat of Arctic ice fields (mostly the GIS) was substantial—between 2.2 and
3.4 m of sea level rise resulted from this ice field melting (93), with the rest apparently
coming from the Antarctic ice sheet (178). Thus, summer warming over Greenland in
excess of the Holocene, and less than the last previous interglaciation, could cause future
abrupt sea level change—a threshold that could be reached in the twenty-first century
(178). Sources: insolation data (185), sea level summaries for the last 20 kyrs (186) and for
the period 140 to 120 ka (163).
P1: JRX
September 8, 2006 9:50 Annual Reviews AR289-FM

Annual Review of Environment and Resources


Volume 31, 2006

CONTENTS
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

I. EARTH’S LIFE SUPPORT SYSTEMS


Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

Abrupt Change in Earth’s Climate System, Jonathan T. Overpeck and Julia


E. Cole 1
Earth’s Cryosphere: Current State and Recent Changes, Claire L. Parkinson 33
Integrated Regional Changes in Arctic Climate Feedbacks: Implications
for the Global Climate System, A. David McGuire, F.S. Chapin III, John
E. Walsh, and Christian Wirth 61
Global Marine Biodiversity Trends, Enric Sala and Nancy Knowlton 93
Biodiversity Conservation Planning Tools: Present Status and Challenges
for the Future, Sahotra Sarkar, Robert L. Pressey, Daniel P. Faith,
Christopher R. Margules, Trevon Fuller, David M. Stoms, Alexander
Moffett, Kerrie A. Wilson, Kristen J. Williams, Paul H. Williams, and
Sandy Andelman 123
II. HUMAN USE OF ENVIRONMENT AND RESOURCES
Energy Efficiency Policies: A Retrospective Examination, Kenneth
Gillingham, Richard Newell, and Karen Palmer 161
Energy-Technology Innovation, Kelly Sims Gallagher, John P. Holdren,
and Ambuj D. Sagar 193
Water Markets and Trading, Howard Chong and David Sunding 239
Biotechnology in Agriculture, Robert W. Herdt 265
III. MANAGEMENT, GUIDANCE, AND GOVERNANCE OF RESOURCES
AND ENVIRONMENT
Environmental Governance, Maria Carmen Lemos and Arun Agrawal 297
Neoliberalism and the Environment in Latin America, Diana M. Liverman
and Silvina Vilas 327
Assessing the Vulnerability of Social-Environmental Systems, Hallie
Eakin and Amy Lynd Luers 365
Environment and Security, Sanjeev Khagram and Saleem Ali 395

vii
P1: JRX
September 8, 2006 9:50 Annual Reviews AR289-FM

viii CONTENTS

Sustainability Values, Attitudes, and Behaviors: A Review of Multinational


and Global Trends, Anthony A. Leiserowitz, Robert W. Kates, and
Thomas M. Parris 413
IV. INTEGRATIVE THEMES
Linking Knowledge and Action for Sustainable Development, Lorrae van
Kerkhoff and Louis Lebel 445
by UNIVERSITY OF WASHINGTON - HEALTH SCIENCES LIBRARIES on 01/18/07. For personal use only.

INDEXES
Cumulative Index of Contributing Authors, Volumes 22–31 479
Annu. Rev. Environ. Resourc. 2006.31:1-31. Downloaded from arjournals.annualreviews.org

Cumulative Index of Chapter Titles, Volumes 22–31 483

ERRATA
An online log of corrections to Annual Review of Environment and
Resources chapters (if any, 1997 to the present) may be found at
http://environ.annualreviews.org

You might also like