the revised diagnostics of 2m values - motivation, method and impact - m. raschendorfer, fe14...

The revised Diagnostics of 2m Values

- Motivation, Method and Impact -

M. Raschendorfer, FE14

Matthias RaschendorferDWD COSMO Cracow 2008

Matthias RaschendorferDWD COSMO Cracow 2008

There was some significant overestimation of 2m temperature round noon

at grid points with a rough surface (city points)

accompanied by nearly saturated soil and too high 2m dew points

Obviously SMA couldn’t correct this errors

What was the reason?

Could we expect related errors in other variables?

deviation of T2m

compared to SYNOP-measure

parameters controlling evapotranspiration:

-soil moisture

modified relation between sensible and latent heat flux

daily cycle of

boundary layer profiles

for

temperature and humidity

Diagnosis onto 2m level

daily cycle of T2m, TP2m

turbulence model

SMA

-min. stomata resistance, PLCOV, LAI, root depth, sealing, soil type, …

The feed back with soil moisture analysis SMA:

Matthias RaschendorferDWD COSMO Cracow 2008

We assumed a systematic shortcoming in the calculation of 2m temperatures

at grid points with a rough surface (city points)

causing to high values at daytime

SMA reacted by moistening the soil until it was saturated

Artificial soil moistening may have caused artificial evaporation accompanied by

a too moist and too cold boundary layer

At that point SMA couldn’t correct the discrepancy any longer

As a reason of that shortcoming we found a discrepancy between

2m values referring to the whole grid box andthose taken at a standard SYNOP lawn

laminar layer

0z

u0

0A zh

logarithmic Prandtl-layer profile

unstable stable

linear interpolated

Prandtl layer

roughness layer

z

expon. roughness-layer profile

0

Bhm20

m2 S

lowest model main level

upper boundary of the lowest model layer

lower boundary of the lowest model layer

h

0B zh

Ahm10

0

m2

dHz

SYNOP station lawn profile

Mean GRID box profile

Turbulent velocity scale profile

u

Hu

SAH0

SAHz ruz

turbulence-scheme

00 dz

0SH zzu

dzr

:

Az

Bz

A s m2

• Exponential roughness layer profile is valid for the whole grid box,

• but it is not present at a SYNOP station

Sz0

H00 zzd

m2

dz

no storage capacity

Diagnosis of 2m-values with respect to a SYNOP lawn:

m2dzzddd 00H

Accumulated soil moisture increments of SMA for routine and Exp 6343:

The accumulated effect of SMA on soil moisture averaged along a domain that more or less covers Germany is plotted for the fist 6 weeks of Exp. 6343.For about the first 14 days SMA of Exp. 6343 (green and blue lines) takes water out of the soil, whereas SMA of our routine always brings additional water into the soil. During this period soil moisture has adapted to the new formulation of T2m diagnostics. Integrated along the whole 6 week period SMA of the routine has significantly moistened the soil, indicating a systematic error. In contrast SMA of Exp. 6343 shows almost no systematic trend.

Period of soil moisture adoption

m30z0 . ( & ‘plant cover’ 70% & ‘surface height’ 800m) ‘city’ :=

Soi

l de

pth

in [

cm]

Soil moisture profiles:

T2m

of

Exp

. 63

43 –

T2m

of

r th

e ro

utin

e

Roughness length

Joined frequency distribution of roughness length and difference of T2m:

For “summer”

For winter

Mean surface flux densities:

LME ROUT

LME ROUT

Matthias RaschendorferDWD COSMO Cracow 2008

Conclusion:

Discrepancy between the transfer layer of a mean model grid box and a SYNOP lawn caused

The introduction of a SYNOP lawn for deduction of 2m values cured the problem

There are still problems with the daily cycle of near surface values (too warm at midnight)

- daytime overestimation of T_2m compared to SYNOP measurement for rough grid points

- a too wet soil and a too moist and too cold boundary layer due to the action of SMA

- That may be cured by a reformulation of profile functions through the transfer layer

- This step had to be delayed due to other priorities

- We should introduce these steps now with first priority!

Thank You for attention!

Time-Height Cross-section for “errors” in the boundary layer:

Temperature “error” Dew point “error”

Too cold Too moist

800 m a.g.

0 m a.g.

For “summer”

For winter

LME ROUT

LME ROUT

Advection of a warm bias

Too cold from noon to night Still to cold during morning

Downward mixing of positive bias above