3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Frank Müller, Andreas Chlond

MPI for Meteorology, Hamburg

Treatment of microphysics Treatment of microphysics in the MPI-Hamburg LES modelin the MPI-Hamburg LES model

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

OutlineOutline

The microphysical modules in the MPI-LES model Cloud droplet formation – activation of aerosols Comments on submitted results Outlook

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

The microphysical modules in the MPI-LES modelThe microphysical modules in the MPI-LES model

3 (+ 1) microphysical schemes are implemented in the current LES version:

1.no-rain microphysics: sat. adjustment to form cloud water (including sub-scale condensation), no droplets (Sommeria and Deardorff, 1977)

2.classical 'Kessler' scheme: qc, qr, sat. adjustment, auto-conversion, accre-tion, sedimentation and evaporation of qr (Kessler, 1969)

3.'Lüpkes' scheme: qc, qr, Nr, 2.+ rain drop self-collection, sedimentation of Nr (Lüpkes, 1989; Lüpkes, 1991)

4.two-moment scheme for cloud and rain water, resp.: qc, qr, Nc, Nr, 3.+ self-collection of Nc, sedimentation of qc and Nc, activation (Seifert and Beheng, 2000; Seifert, 2002)

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Cloud droplet formation – activation of aerosolsCloud droplet formation – activation of aerosols

A detailed treatment of aerosol population development is very computer resource demanding.

Since non-activated aerosols are in thermodyn. equilibrium aerosol activa-tion is in general a good approximation to describe the corresponding cloud droplet sources

Aerosol population development Supersaturation

Time [s] Time [s]

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Cloud droplet formation – activation of aerosols Cloud droplet formation – activation of aerosols (cont.)(cont.)

Activation is based on NCCN spectra (Twomey, 1959; Twomey and Woj-ciechowski, 1969; Hudson, 1980)

The power law parameters C and k, given in the legend, were fitted to the activation spectrum based on the given aerosol size distribution.

The activation spectrum is gained from:

N CCN s = C sk

N CCN s = − ∫rap sc=0

rapsc=s

f aprapdrap = ∫rc

∞

f ap rapdrap

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Cloud droplet formation – activation of aerosols Cloud droplet formation – activation of aerosols (cont.)(cont.)

To calculate the no. of newly activated cloud drops the maximum supersaturation has to be determined:

The determination of the cloud water mixing ratio tendency is crucial!

Using the droplet growth equation yields:

dsvwdt

= A1 w A2dqc

dt= f svw = 0

A1 = l21 gRaT

2c pa− gRaT

A2 = pesat ,w

l21

2

RaT2c pa

r drdt

≈ A3 s with A3 = [ wRvTesat ,wDvl21wk aT l21

RvT−1]

−1

r t = [r to2 2 A3∫to

t

s t ' dt ' ]12

dqcdt

= 2wa

2 A32/3 st ∫

0

s

ns ' [∫t ot

st ' dt ' ]12ds '

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Cloud droplet formation – activation of aerosols Cloud droplet formation – activation of aerosols (cont.)(cont.)

According to the original work of Twomey (1959) the activation spectrum is written as:

Taking into account only the loss of water vapor due to condensation of newly activated droplets results in a change of cloud water mixing ratio as follows:

Maximum supersaturation smax and NCCN are given by:

n s = k C sk−1

dqcdt

= 2waA3 k C [ A3

A1w ]12 sk2 B k2,

32

smax = C− 1k2 [ A1w

32

2waA2 A3

32 k B ]

1k2

N CCN ,max = C2k2 [ A1w

32

2waA2 A3

32 k B ]

kk2

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Cloud droplet formation – activation of aerosols Cloud droplet formation – activation of aerosols (cont.)(cont.)

Extending the previous derivation by considering also condensation on already existing cloud droplets yields:

The maximum supersaturation smax is given by:

dqldt

=dq ldt act

dq l

dt cond

= 2waA3 k c B k2 ,32 [ A3

A1w ]1 /2

4w

aA3 N c [ rc

2 A3

A1 ws2 ]

1 /2

smaxn1 = smax

n − f sf ' s

f ' svw = 2waA2 A3 [ k k2 [ A3

A1 w ]1 /2

svwk2 2 N c [ rc

2 A3

A1 w ]1/2

2 N c svwA3

A1 w [ rc2

A3

A1 wsvw

2 ]−1/ 2 ]

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Cloud droplet formation – activation of aerosols Cloud droplet formation – activation of aerosols (cont.)(cont.)

Simulation results from an entraining air parcel model using the two-mo-ment scheme.

..... std. Activation (Twomey, 1959)

___ modified activa tion scheme

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Comments on submitted resultsComments on submitted results

Our submitted results had the following features: Too fast increasing cloud top and cloud base Cloud breakup (Lüpkes scheme) with 1<cc<0.4 Gradually diminishing LWP: t=6h: 70d/m2<LWP<20g/m2

Precipitation only with Kessler scheme

What is the reason for this behaviour? Wrong physics? Bad numerics?

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Comments on submitted results Comments on submitted results (cont.)(cont.)

Std. Kessler scheme

Kessler scheme, ZD

Kessler scheme, GM

Kessler scheme, GMZD

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

Comments on submitted results Comments on submitted results (cont.)(cont.)

Std. Kessler scheme

Kessler scheme, ZD

Kessler scheme, GM

Kessler scheme, GMZD

3rd PAN GCSS Meeting on cloud, climate and models, Athens, May 2005

OutlookOutlook

Perform simulations with the new two-moment schemeImplementation of a detailed (size resolved) cloud microphysical schemeLook at aerosol-cloud interactions

Qr [g/kg]

Time [h]

Hei

ght

[m]