ucla workshop june 2005 1 issues in boundary layer parametrization for large scale models anton...

UCLA workshop June 2005 1

ISSUES IN BOUNDARY LAYER PARAMETRIZATION FOR LARGE SCALE MODELS

Anton Beljaars*(ECMWF)

• Boundary layer clouds

• Wind turning

• Momentum budget (heterogeneous terrain)

•With contributions from: Andy Brown, Sylvain Cheinet, Hans Hersbach, Martin Koehler, Andrew Orr


Recent implementation of Mass-flux/K-diffusion approach

key ingredients:

• moist conserved variables

• combined Mass-flux/K-diffusion solver

• cloud variability

• transition between stratocumulus and shallow convection


Results: Low cloud cover (new-old)

T511 time=10d

n=1402001 & 2004

old: CY28R4 new PBL


Results: EPIC column extracted from 3D forecasts

Liquid Water Path [g/m2]

16 17 18 19 20 21 22Local Time [October 2001 days]

0

50

100

150

200

250

300

old PBLnew MK PBLEPIC obs

WV Mixing Ratio [g/kg]

0 2 4 6 8 10

1000

950

900

850

800

750

700

Pre

ssur

e [h

Pa]

Mod

el L

evel

s

45

50

55

60


ARM SGP, number of cloudy hours in July 2003

Cloud radar Model

Model lacks afternoon shallow cumulus

Sylvain Cheinet (2005)


ARM SGP, daily 36 hr back trajectories, July 2003

L42, 600 hPa

L54, 950 hPa

Cross section of model moisture and cloud cover for fair weather days in July 2003 (q: shaded; cc: contours at 0.01, 0.02, 0.05)


ARM SGP, surface fluxes (model/obs)

HLE

Monthly domain averaged diurnal cycle

West-East change of LE/H

Time evolution of LE and H for station C1


ARM SGP, 2m T/q, BL wind (model/obs)

28R1 28R1 +MO stable BL


ARM SGP: Sensitivity to shallow convection

Cross section of model humidity difference for fair weather days in July 2003 (Effect of No Shallow Convection)

Cross section of model humidity difference for fair weather days in July 2003 (Effect of halving shallow convection mass flux over land)


Boundary layer cloud issues

•The switching algorithm is unsatisfactory. Unification between stratocumulus and shallow cumulus is desirable. What is the physical mechanism that controls the regime?

•Numerics of inversion handling

•Cloud top entrainment?

•Closure for shallow convection?

•Does a scheme have the correct feedbacks and how can we test this?

•What is the role of shallow convection in more complicated situations (cold air outflow, momentum transport)


Geostrophic drag law

B

u

V

A

zf

u

u

U

G

om

G

*

*

*

)ln(1

drag

a-geostrophic angle

Wangara data


Operational verification of surface wind

A-geostrophic angle is too small


Operational verification of surface wind

Wind speed is too low during the day and too high at night


Examples of ERA40 forecast veers

260

264

264

264

268

268

268

272

272

272

276

276

276

280

280

284

284

288

288

25.0m/s

40°N40°N

50°N 50°N

60°N60°N

70°N 70°N

60°W

60°W 50°W

50°W 40°W

40°W 30°W

30°W 20°W

20°W 10°W

10°W 0°

0°

2002021312+24 level 55 T and wind

235.9

238

242

246

250

254

258

262

266

270

274

278

282

286

290

293.5

-12

-12

-12

-12

27

27

27

27

27

27

40°N40°N

50°N 50°N

60°N60°N

70°N 70°N

60°W

60°W 50°W

50°W 40°W

40°W 30°W

30°W 20°W

20°W 10°W

10°W 0°

0°

2002021312+24 surface to 850 hPa veer

-180

-27

-12

12

27

180

256260

260

264

264

268

268

268

268272

272

272

276

276

276

280

280

284

284

288

288

25.0m/s

40°N40°N

50°N 50°N

60°N60°N

70°N 70°N

60°W

60°W 50°W

50°W 40°W

40°W 30°W

30°W 20°W

20°W 10°W

10°W 0°

0°

2002021412+24 level 55 T and wind

237.7

240

244

248

252

256

260

264

268

272

276

280

284

288

292

294.0

-12

-12

-12

-12

-12

-12

-12

27

27 27

27

27

27

40°N40°N

50°N 50°N

60°N60°N

70°N 70°N

60°W

60°W 50°W

50°W 40°W

40°W 30°W

30°W 20°W

20°W 10°W

10°W 0°

0°

2002021412+24 surface to 850 hPa veer

-180

-27

-12

12

27

180

250m T andwind

Surface to 850 hPa veer >27o

12o27o

<-27o

-27o-12o

-12o12o

Veering in warm sectorof depression

Backing on and behindcold front

20020213 12Z+24 20020214 12Z+24

Brown et al. 2005


Atlantic sonde stations

• SST, December 1987

278

294

32°N32°N

34°N 34°N

36°N36°N

38°N 38°N

40°N40°N

42°N 42°N

44°N44°N

46°N 46°N

48°N48°N

50°N 50°N

52°N52°N

54°N 54°N

56°N56°N

58°N 58°N

60°N60°N

62°N 62°N

64°N64°N

66°N 66°N

68°N68°N

70°N 70°N

72°N72°N

74°N 74°N

76°N76°N

78°N 78°N

68°W

68°W

64°W

64°W

60°W

60°W

56°W

56°W

52°W

52°W

48°W

48°W

44°W

44°W

40°W

40°W

36°W

36°W

32°W

32°W

28°W

28°W

24°W

24°W

20°W

20°W

16°W

16°W

12°W

12°W

8°W

8°W

4°W

4°W

0°

0°

4°E

4°E

8°E

8°E

ECMWF Mean of 31 Uninitialised Analyses Valid: VT:00UTC 1 December 1987 to 00UTC 31 December 1987 Surface: sea surface temperature

271.0 274 277 280 283 286 289 292 295

SABLE

CHARLIE

LIMA

MIKE


SABLE (All points)

-50 0 50Sonde veer (degrees)

-50

0

50

For

ecas

t vee

r (d

egre

es) N=243

SCA CA NA WA SWA

SABLE (Stable points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=57

SCA CA NA WA SWA

CHARLIE (All points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=266

SCA CA NA WA SWA

CHARLIE (Stable points)


-50

0

50F

orec

ast v

eer

(deg

rees

) N=82

SCA CA NA WA SWA

LIMA (All points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=246

SCA CA NA WA SWA

LIMA (Stable points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=38

SCA CA NA WA SWA

MIKE (All points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=236

SCA CA NA WA SWA

MIKE (Stable points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=14

SCA CA NA WA SWA

SABLE (All points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=243

SCA CA NA WA SWA

SABLE (Stable points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=57

SCA CA NA WA SWA

CHARLIE (All points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=266

SCA CA NA WA SWA

CHARLIE (Stable points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=82

SCA CA NA WA SWA

LIMA (All points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=246

SCA CA NA WA SWA

LIMA (Stable points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=38

SCA CA NA WA SWA

MIKE (All points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=236

SCA CA NA WA SWA

MIKE (Stable points)


-50

0

50

For

ecas

t vee

r (d

egre

es) N=14

SCA CA NA WA SWA

Veering between surface (10m) and 850 hPa wind: comparison of 24 hour ERA40 forecast with 12Z sonde


Composite ERA40 results(4 winters; Sable, Charlie, Lima and Mike combined)

• When sonde shows big veering, model veers less

• When model shows big veering, sonde veers more– Model definitely under-veers

• When sonde shows big backing, model backs less

• When model shows big backing, sonde veers are similar– Model probably under-backs

Surface to 850 hPa

-50 0 50Sonde veer (deg.)

-50

0

50

For

ecas

t vee

r (

deg.

)

Categories by sonde veerCategories by forecast veer

8310121843210849

32

49

240

597

50

23

850 hPa to 700 hPa


-50

0

50

For

ecas

t vee

r (

deg.

)


254511564715676

34

37

120

675

144

54


Errors in ERA40 surface wind direction as f (surface wind direction)

• Southerly flow at Sable and Charlie likely to be associated with warm advection

– Insufficient turning across boundary layer

– Forecast surface wind veered relative to observed wind

• Northerly flow at Sable and Charlie likely to be associated with cold advection

– Forecast wind direction errors become small, and possibly reverse (lack of backing across boundary layer)

– Possible that plotting versus wind direction is not selective enough to pick out the strongest cold advection cases where lack of backing is most likely

SABLE

0 90 180 270 360sonde surf dir

-50

0

50

fc s

urf

dir

- so

nde

surf

dir

CHARLIE


-50

0

50

fc s

urf

dir

- so

nde

surf

dir

LIMA


-50

0

50

fc s

urf

dir

- so

nde

surf

dir

MIKE


-50

0

50

fc s

urf

dir

- so

nde

surf

dir


Impact of resolution and model

• Qualitatively similar results in ECMWF T511 operational model and also in Met Office operational model

ERA40 DJF0002


-50

0

50

For

ecas

t ve

er

(deg

.)


85961974522

31

124

257

23

10

ECMWF DJF0002


-50

0

50

For

ecas

t ve

er

(deg

.)


85952014421

28

104

277

28

9

ECMWF DJF0304


-50

0

50

For

ecas

t ve

er

(deg

.)


48621112315

22

65

154

15

3

MET OFFICE DJF0304


-50

0

50

For

ecas

t ve

er

(deg

.)


48641092016

26

68

144

13

6


Stable boundary layer diffusion

Diffusion coefficients based on Monin Obukhov similarity:

2

2 )(

''

z

U

z

gRi

Riflz

UK

zKw


Impact of MOSBL on veers (1 winter; Sable, Charlie, Lima and Mike combined)

• Typically increased 5o in stable cases

• Further weakening of mixing by reducing length scale from 150m to 50m gives further small increase

UNSTABLE POINTS

-50 0 50sonde veer (deg.)

-20

-10

0

10

20

(mos

bl)-

(con

trol

) fo

reca

st v

eer

(de

g.)

SCA CA NA WA SWA

STABLE POINTS

-50 0 50sonde veer (deg.)

-20

-10

0

10

20

(mos

bl)-

(con

trol

) fo

reca

st v

eer

(de

g.)

SCA CA NA WA SWA


Wind direction from QuikSCAT


Wind direction

• Operational models underestimate a-geostrophic angle (what controls boundary layer veering?)

• Stable boundary layers do occur frequently over the ocean

• Real boundary layers are baroclinic (systematic LES simulations covering a wide range of baroclinicity may be helpful?)


Surface drag, momentum budget

• Surface drag is controlled by boundary layer mixing and surface roughness length

• Less boundary layer mixing (e.g. in stable boundary layer) leads to less surface drag

• Surface drag (or effective roughness) over land is very uncertain due to terrain heterogeneity effects


Z0-table


z0


CD-control


TOD+new z0 table


MO


ZM stress + Ps

Zonal mean West-East turb. stress Zonal mean surf. press. error (step=24)

ControlNew z0-table

ControlMO-stab.f.

ControlNew z0-table

ControlMO-stab.f.


Error tracking

Control

MO-stab.f.


500 hPa RMSerror NH,

200403

500 hPa activityDivided by

Analysis activity NH, 200403

ControlMO-stab.f.

ControlMO-stab.f.


500 hPa RMSerror NH,

200403 ControlNew z0-table

500 hPa activityDivided by

Analysis activity NH, 200403

ControlNew z0-table


BL-Veering

Control-New z0-table

Control

Control-MO


Conclusions

• NWP is sensitive to surface drag. Drag is affected by vegetation roughness, orographic effects and stability.

• It is difficult to verify surface drag.

• Roughness lengths are derived from land use data sets, without considering heterogeneity. Results are uncertain.

Research topics

• Derive geometric land use parameters from satellite data

• Run high resolution canopy resolving models to “measure” effective roughness length and its interaction with stability

• Infer surface drag from synoptic development (variational techniques that minimize short range forecast errors?)

ucla workshop june 2005 1 issues in boundary layer parametrization for large scale models anton...

Documents

ucla workshop

shallow convection slide

land slide

night slide

small slide

cy28r4new pbl slide

momentum transport slide

andrew orr slide