Julio Bacmeister, Peter Lauritzen, and Others

Parameterizing
the effects of

orography (mountains) in global
models
Julio Bacmeister, Peter Lauritzen, and others
DCMIP 10 June 2016

Outline
•  The problem
•  Momentum flux parameterizaDons, i.e., drag
schemes
–  PBL
–  “Wave-‐y”
•  Future direcDons
Raw topography at 3km
resoluDon
Boundary topo for CAM FV
at 0.9x1.25 resoluDon
How does subgrid orography affect
atmospheric momentum ?

turbulent, small-‐scale, …
•  PBL Form drag

“wavy”, small-‐meso scale
•  Mesoscale blocking and low-‐level
nonlineariDes
•  VerDcally propagaDng gravity waves
PBL Form drag
•  Features w/ scales <5km
•  Stable straDficaDon not necessary
•  Flow separaDon increases form drag, but not
necessary for form drag in verDcal shear
(Taylor et al. 1989)
F = hs ∇ps
h`ps://youtu.be/ffrK8LBzt-‐Y
PBL Form drag
Simplest approach – enhance roughness length z0 over rough/hilly terrain, e.g.,
“turbulent mountain stress” (TMS) scheme currently in CESM (Richter et al. 2010)
Fx = CD | U |U(z)
−2
! ! z $$
CD = κ ## ln # &&&
" " z0 % %
z0 is roughness length
z0 is assumed proporDonal to h δ ! 2 where h’δ is topographic variability for
scales λ<3km-‐5km
PBL Form drag
More complex approach integrates over spectrum of topography (for scales
below ~3km-‐5km). Drag from individual components decays in the verDcal
based on scale (Beljaars et al. 2004).
Fx
a2 ∝ hδ"2
Very incomplete bibliography of PBL
form drag schemes
Taylor, P. A. (1977). Numerical studies of neutrally straDfied planetary boundary-‐layer flow above
gentle topography. Boundary-‐Layer Meteorology, 12(1), 37-‐60.
Wood, N., Brown, A. R., & Hewer, F. E. (2001). Parametrizing the effects of orography on the boundary
layer: An alternaDve to effecDve roughness lengths. Quarterly Journal of the Royal Meteorological
Society, 127(573), 759-‐777.
Taylor, P. A., Sykes, R. I., & Mason, P. J. (1989). On the parameterizaDon of drag over
small-‐scale topography in neutrally-‐straDfied boundary-‐layer flow. Boundary-‐Layer
Meteorology, 48(4), 409-‐422.
Beljaars, A., Brown, A. R., & Wood, N. (2004). A new parametrizaDon of turbulent orographic form drag.
Quarterly Journal of the Royal Meteorological Society, 130(599), 1327-‐1347.
Richter, J. H., Sassi, F., & Garcia, R. R. (2010). Toward a physically based gravity wave source
parameterizaDon in a general circulaDon model. Journal of the Atmospheric Sciences, 67(1), 136-‐156.
Mesoscale(λ>5km)/straDfied flow
effects
•  Atmospheric gravity waves
•  High-‐drag configuraDon/downslope wind
(Chinook, Foehn …)
•  Flow-‐spliqng
effects
Stra'fica'on: Gravity waves need stable straDficaDon

g ∂Θ
N = 2
Θ ∂z
where, Θ is basic state potenDal temperature

•  Note, if ∂ z Θ
< 0 , warm air lies below cooler air and
N is imaginary èconvecDve instability
•  Typical free-‐tropo strato values of N are 0.01 s-‐1 to
0.02 s-‐1 so buoyancy period (2π/N) is 300-‐600s
effects
Amplitude/nonlinearity: A basic length scale in gravity
wave/straDfied flow analysis
U
L=
N
where, U
is mean horizontal wind. If U ~10m/s then
L~500-‐1000m
For mountain heights h ~ UN nonlineariDes, blocking,
high-‐drag states become important
For horizontal wavelength ~ 2 π UN (3km-‐6km)
nonhydrostaDc effects and wave trapping become
important (usual argument for separa=on into PBL and
wave-‐based schemes)
High drag
state, do
wnslope w
i nds
Bougeault et al. (1990)

Amospheric Infrared Sounder (AIRS) observaDons of mountain waves in
the stratosphere alDtude 30-‐40km (Joan Alexander)
Northern Hemisphere
Southern Hemisphere
Increasing wind speed with alDtude
ath
Ray p
Trapped lee waves. VerDcal

wavenumber m given by
2
N
m2 = 2 − k 2
U
Waves turn where m2
becomes <0
How do we parameterize this
menagerie of small-‐scale flows in a
global model???
Subgrid momentum fluxes
Momentum EquaDon
∂t ρ u +... + ∂z ρ wu = −∇p − ρ∇φ + F +..., ρ is atmospheric density
Grid box average momentum equaDon

∂t ρ u +... + ∂z ρ wu = −∇p − ρ∇φ − ∂z ρu$w$ i − ∂z ρ v$w$ j+ F
VerDcal derivaDves of zonal and meridional subgrid verDcal

momentum fluxes produce drag forces
Subgrid momentum fluxes
Let’s turn into coordinates where “x” is perpendicular to wave crests
Our job is then to calculate
τ = ρu!w!
τ = ρu!w!
Wave moves through
without exerDng drag
Drag on mean flow when

wave breaks -‐“saturates”
Wave moves through

Nonlinear low-‐level flow. Early

schemes didn’t worry about this
Complex wave paèrn conceptualized as 2D monochroma<c wave controlled by “saturaDon”
Lindzen, R. S. (1981). Turbulence and stress owing to gravity wave and
Ddal breakdown. Journal of Geophysical Research, 86(C10), 9707-‐9714.
τ = ρu!w!
Wave moves through

Wave moves through

Eliassen-‐Palm theorem:
Non-‐dissipaDng waves conserve τ as
they propagate verDcally

Complex wave paèrn conceptualized as 2D monochromaDc wave controlled by “saturaDon”
Lindzen, R. S. (1981). Turbulence and stress owing to gravity wave and
Ddal breakdown. Journal of Geophysical Research, 86(C10), 9707-‐9714.
τ = ρu!w!
δ = verDcal displacement Wave moves through
of streamline from without exerDng drag
equilibrium level

Wave moves through


How do we calculate τ based on topographic informaDon?

τ = ρu!w!
Orographic gravity wave momentum flux Wave moves through
based on δ and gravity wave dispersion without exerDng drag
relaDonships
u! = Nδ wave breaks -‐“saturates”
w! = kUδ
so momentum flux becomes
τ ≈ C ρ kUNδ 2 Wave moves through
IntuiDvely obvious that δ at source level is

related to mountain heights

Not so obvious how to get δ from
topographic data: Nonlinear low-‐level flow. Early
•  RMS of subgrid topo? schemes didn’t worry about this
•  Residuals leu auer smoothing ?
τ = ρu!w!
Wave moves through

Wave moves through


What about “Satura'on”, i.e.,wave breaking??

τ = ρu!w!
Wave moves through

Wave moves through


Gravity wave saturaDon/breaking occurs when streamlines are verDcal

or overturning èlocal convecDve instability
“SaturaDon hypothesis” holds that
τ = ρu!w!
turbulence conDnually shaves off just Wave moves through
enough energy to keep breaking wave without exerDng drag
exactly at edge of instability (verDcal
streamlines), i.e., Drag on mean flow when
Spilling breaker
Wave moves through
NOT
Plunging breaker
Is satura<on hypothesis actually true? Probably someDmes. Not bad first guess.
τ = ρu!w!
Wave moves through

So when do gravity wave streamlines wave breaks -‐“saturates”
become verDcal?
Wave moves through


τ = ρu!w!
Wave moves through

So when do gravity wave streamlines wave breaks -‐“saturates”
become verDcal?

When
Wave moves through
U
δ !!
=
N
Pseudocode:

At this point you have most of what you need to calculate wave momentum flux

1) EsDmate δ(LM) from topography dataset

2) Calculate τ(LM)=ρkUNδ2

3) Advance to level above: τ(L-‐1)=τ(L)

4) Infer δ(L-‐1)

5) Test for δ(L-‐1)>U/N

if no go to 3)

if yes set δ(L-‐1)=U/N recalculate τ(L-‐1) and go to 3)
Note: Other sources of atmospheric gravity waves exist: fronts, convec<on ….

This basic approach works for all gravity waves with replacement UèU-‐c
τ = ρu!w!

So what about this low-‐level nonlinearity?

Blocking, low-‐level turning
(e.g. Scinocca&McFarlane 2000)
i – verDcally propagaDng waves ∂zτ via saturation

ii -‐ downslope wind layer ∂zτ ~ ∂zτ sat (1+ γ )

iii – low-‐level flow turning ∂zτ ~ U 2
! U $
min #α , h &
! U $ " N %
min # β , h &
" N %
i.
h ii.
iii.
Anisotropy
isotropic Anisotropic “ridge-‐y”
wave
Drag force
ParDal history of orographic drag
schemes
Isotropic topography, no low-‐level blocking or other nonlineari<es
McFarlane, N. A. (1987). The effect of orographically excited gravity wave drag on the general circulaDon of the
lower stratosphere and troposphere. Journal of the Atmospheric Sciences, 44(14), 1775-‐1800.
Anisotropy, low-‐level blocking, high-‐drag states

Pierrehumbert, R. T., & Wyman, B. (1985). Upstream effects of mesoscale mountains. Journal of the atmospheric
sciences, 42(10), 977-‐1003.
Lo`, F., and M. J. Miller (1997). A new subgrid-‐scale orographic drag parametrizaDon: Its formulaDon and
tesDng. Quarterly Journal of the Royal Meteorological Society 123.537: 101-‐127.
Gregory, D., Shu`s, G. J., & Mitchell, J. R. (1998). A new gravity-‐wave-‐drag scheme incorporaDng anisotropic
orography and low-‐level wave breaking: Impact upon the climate of the UK Meteorological Office Unified Model.
Quarterly Journal of the Royal Meteorological Society, 124(546), 463-‐493.
Scinocca, J. F., & McFarlane, N. A. (2000). The parametrizaDon of drag induced by straDfied flow over anisotropic
orography. Quarterly Journal of the Royal Meteorological Society, 126(568), 2353-‐2393.
Alpert, J. C. (2004) Sub-‐grid scale mountain blocking at NCEP. Proceedings of 20th Conference on WAF, 16th
conference on NWP.
TMS added to CAM (par<ally compensa<ng for missing mesoscale drag?)
Richter, J. H., Sassi, F., & Garcia, R. R. (2010). Toward a physically based gravity wave source parameterizaDon
in a general circulaDon model. Journal of the Atmospheric Sciences, 67(1), 136-‐156.
Future direcDons
Trapping effects not actually included
in current parameterizaDons.
Horizontal propagaDon of waves

across grid boxes (Dme-‐dependence
also? Ray-‐based? Super-‐param.?)
Future direcDons
Wave cloud radiaDve effects and chemical effects
Nacreous ice-‐clouds in stratosphere

Future direcDons
Strong orographic precipitaDon biases in the tropics.
•  Terrain-‐following coordinates to blame?
Final QuesDon
At which resoluDon can we live without parameterizaDons of
orographic drag
•  Wave-‐y type (not 25km, … 5km?)
•  PBL type (probably less than 1km)

Julio Bacmeister, Peter Lauritzen, and Others

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Julio Bacmeister, Peter Lauritzen, and Others

Uploaded by

Copyright:

Available Formats

Parameterizing

the eﬀects of

DCMIP 10 June 2016

Bougeault et al. (1990)

Trapped lee waves. VerDcal

Grid box average momentum equaDon

VerDcal derivaDves of zonal and meridional subgrid verDcal

Our job is then to calculate

Drag on mean ﬂow when

Wave moves through

Nonlinear low-­‐level ﬂow. Early

Drag on mean ﬂow when

Wave moves through

Nonlinear low-­‐level ﬂow. Early

Drag on mean ﬂow when

Wave moves through

Nonlinear low-­‐level ﬂow. Early

How do we calculate τ based on topographic informaDon?

IntuiDvely obvious that δ at source level is

Drag on mean ﬂow when

Wave moves through

Nonlinear low-­‐level ﬂow. Early

What about “Satura'on”, i.e.,wave breaking??

Drag on mean ﬂow when

Wave moves through

Nonlinear low-­‐level ﬂow. Early

Gravity wave saturaDon/breaking occurs when streamlines are verDcal

Drag on mean ﬂow when

Wave moves through

Nonlinear low-­‐level ﬂow. Early

Drag on mean ﬂow when

Nonlinear low-­‐level ﬂow. Early

So what about this low-­‐level nonlinearity?

isotropic Anisotropic “ridge-­‐y”

Anisotropy, low-­‐level blocking, high-­‐drag states

Horizontal propagaDon of waves

Nacreous ice-­‐clouds in stratosphere

You might also like

Nonlinear low-‐level ﬂow. Early

Nonlinear low-‐level ﬂow. Early

Nonlinear low-‐level ﬂow. Early

Nonlinear low-‐level ﬂow. Early

Nonlinear low-‐level ﬂow. Early

Nonlinear low-‐level ﬂow. Early

Nonlinear low-‐level ﬂow. Early

So what about this low-‐level nonlinearity?

isotropic Anisotropic “ridge-‐y”

Anisotropy, low-‐level blocking, high-‐drag states

Nacreous ice-‐clouds in stratosphere