Investigation Of Different Parameterizations Of The Mixed Liquid-ice Sub-grid Statistical

Cloud Scheme Over Greece

E. Avgoustoglou (euri@hnms.gr)Hellenic National Meteorological Service

13th COSMO General Meeting, Servizio Meteorolologico, Rome 4th to 9th September 2011

PRESENTATION OUTLINE

Highlights of the sub-grid liquid-ice mixed scheme

Test cases tabulation and model set up

Implementation of the scheme

Results for Radiation temperatures, Cloud Cover,

T_2M,

Conclusions and Comments

Stratiform Cloud Fraction (R) (Sommeria, Deardorff, 1976)

ilc

ice

il

vapour

v

total

w

cp

cli

qqqqqqq

qTc

L

d

,

re temperatupotential water total,

liquid water

heat on condensati phase" mixed" , 1: iilic LrLrL

liwsw dGdqqqHR

)(

0 ,1

0 ,0)(

x

xxH

22

2

22/12 221

1exp

12

1

wwliliwliq

w

q

wlili

q

qqr

rrG

lilili www qqq wli qwliqr

Highlights of the sub-grid liquid-ice mixed scheme

factor icing , :c

ii q

qTr

21

2

1 QerfR

sw qq

Q

2122 2 swsw qqqq

A first order approximation for is assumed in reference to and by using Clausius-Clapeyron equation cloud fraction R is given by

sq sq

Sommeria και Deardorff (1977) further approximated R empirically :

1R0 ,6.1

12

1

QR

In analogy, a SubGgrid Statistical (SGS) cloud scheme is implemented to COSMO model (Raschendorfer) where the stratiform cloud cover is approximated by a two-parameter relation:

1R0 ,1

B

QAR

Parameter A refers to cloud cover at saturation and Β refers to critical value of saturation deficit. The default values of these parameters are set 0.5 and 4.0 respectively.

Test Cases Choice and Model set-upSeven cases are under investigation:

# Starting Date

1 Jan. 1 2005 12UTC (d050101_12)

2 Dec. 24 2007 12UTC (d071224_12)

3 Apr. 25 2006 12 UTC (d060425_12)

4 May 9 2008 12 UTC (d080509_12)

5 May 1 2009 12 UTC (d090501_12)

6 April 27 2011 12 UTC (d110427_12)

7 January 9 2011 12 UTC (d110109_12)

COSMO_4.6 (Modified by M.R) Horizontal grid 0.06250 (~7 Km)

273273 points 40 vertical levels time step: 30 sec GME (Analysis 0.50) 3 hs 48 hs IBM HPC Cluster 1600 (P4+)

m

Model Domain

Implementation of the Sub-grid Cloud Schemes and Test Tabulation

1. cosmo_4.6_rh: (icldm_rad = 4) is a run of COSMO_v4.6 with the default relative humidity scheme

2. R_rh: (icldm_rad = 4) is a run of COSMO_v4.6 with the default relative-humidity scheme and the modifications of the mixed-phase scheme. (It gave a slightly different answer than the default version of cosmo_4.6_rh and was used as reference against the tests of the mixed-phase statistical scheme denoted below. )

3. R_mix_0.5_2.0: (icldm_rad = 2, clc_diag=0.5 and q_crit=2.0) is a run of COSMO_v4.6 with the mixed-phase statistical scheme

4. R_mix_0.5_3.0: (icldm_rad = 2, clc_diag=0.5 and q_crit=3.0) is a run of COSMO_v4.6 with the mixed-phase statistical scheme

5. R_mix_0.5_4.0: (icldm_rad = 2, clc_diag=0.5 and q_crit=4.0) is a run of COSMO_v4.6 with the mixed-phase statistical scheme (This is the standard old run from previous COSMO year with the default values for clc_diag and q_crit and was included for reference purposes)

6. R_mix_0.4_5.0: (icldm_rad = 2, clc_diag=0.4 and q_crit=5.0) is run of COSMO_v4.6 with the mixed-phase statistical scheme

7. R_mix_0.4_4.0: (icldm_rad = 2, clc_diag=0.4 and q_crit=4.0) is run of COSMO_v4.6 with the mixed statistical scheme. (This test was performed only for one test case.)

8. ER_stat_rh_0.5_4.0: (clc_diag=0.5 and q_crit=4.0) is based on a modified COSMO 4.11 version that invokes statistical (not mixed!) scheme when there is no cloud ice at the grid point and the relative humidity scheme when there is cloud ice. (This is an old run related to a personal modification that was demonstrated in the presentation in Moscow.)

Considered Variables

CLCT: Total Cloud Cover

CLCH: High Cloud Cover

CLCM: Medium Cloud Cover

CLCL: Low Cloud Cover

CLC: Cloud Cover over station

T_2M: 2 meter Temperature

Cloud T: Artificial Satellite Images (MSG WV 6.2 microns)

! A larger set is available for further consideration and discussion with COSMO coleagues..

The results for the mixed-phase statistical scheme are presented for the “Spring Case” of April 27th 2011 12 UTC (mainly) and the “Winter Case” of January 9th 2011 12 UTC in reference to the default relative-humidity scheme, satellite pictures (MSG Infrared, Water vapor and Cloud Cover) as well as station observations.

Results

Tmin

6

7

8

9

10

11

12

13

14

15

16

17

18

19

606

614

619

622

624

627

628

632

641

642

643

648

650

654

655

665

667

671

672

675

682

684

685

687

699

706

710

715

716

718

719

723

726

732

734

741

742

744

746

749

750

754

757

759

760

765

766

768

774

R_rh

R_mix_0.5_4.0

R_mix_0.5_2.0

Observations

Deg C

Station #

28-April-2011 T2m_min

Minimum daily 2-meter Temperatures for 28th April 2011. The vast majority of values from the mixed- phase scheme runs (R_mix_0.5_4.0 and R_mix_0.5_2.0) are relatively closer to Observations than the corresponding values of the default relative humidity scheme of COSMO model (R_rh)

12

13

14

15

16

17

18

19

20

21

22

23

24

606

614

619

622

624

627

628

632

641

642

643

648

650

654

655

665

667

671

672

675

682

684

685

687

699

706

710

715

716

718

719

723

726

732

734

741

742

744

746

749

750

754

757

759

760

765

766

774

R_rh

R_mix_0.5_4.0

R_mix_0.5_2.0

Observ ations

28-April-2011 T2m_max

Station #

Maximum daily 2-meter Temperatures for 28th April 2011. The values from the default relative humidity scheme of COSMO model (R_rh) are relatively closer to Observations than the corresponding values of the mixed- phase scheme runs (R_mix_0.5_4.0 and R_mix_0.5_2.0) but the situation is more balanced than the corresponding minimum daily 2-meter Temperatures .

Deg C

28 April 2011 03 UTC

Meteosat/Synesat R_rh R_mix_0.5_4.0

R_mix_0.5_2.0 R_mix_0.4_5.0 ER_stat_rh_0.5_4.0

Deg C

Comparison of cloud radiation temperatures (CRTs) of artificial satellite images at the water-vapor Channel at 6.2 microns against Meteosat satellite image (upper right) manipulated with Synesat software available at HNMS. The general trend (especially in the encircled areas) is that model CRTs for the mixed-phase scheme (R_mix_0.5_4.0, R_mix_0.5_2.0 and R_mix_0.4_5.0) are overall lower to those of the default relative humidity scheme (R_rh) in reference to Meteosat satellite picture (Meteosat/synesat). However, it looks that this difference is mainly due to the treatment of cloud ice by the mixed-phase scheme as it can be infered by a “control” test were the relative-humidity and the statistical scheme are mixed (ER_stat_rh_0.5_4.0).

28 April 2011 02:45 UTC %

Meteosat/Synesat R_rh R_mix_0.5_4.0

Low Cloud Cover (LCC)

Comparison of Low Cloud Cover (LCC) against the corresponding MSG satellite image (Meteosat/Synesat) manipulated with Synesat software available at HNMS. The general trend (especially in the encircled areas) is that the model LCC for the mixed-phase scheme (R_mix_0.5_4.0) show an overall preponderance to the LCC of the default relative-humidity scheme (R_rh) in reference to Meteosat satellite image. Another interesting feature is that in reference to the satellite picture, the mixed-phase scheme provides a better tendency to resolve the cloudiness in contrast to the cloud structure of the relative humidity scheme which has the tendency to remain more compact. The feature is similar to Medium Cloud Cover.

28 April 2011 02:45 UTC

Meteosat/Synesat R_rh R_mix_0.5_4.0

%Medium Cloud Cover (MCC)

Comparison of Medium Cloud Cover (MCC) against the corresponding MSG satellite image (Meteosat/Synesat) manipulated with Synesat software available at HNMS. As in LCC, the general trend (especially in the enclosed areas) is that the model MCC for the mixed-phase scheme (R_mix_0.5_4.0) show an overall preponderance over the MCC of the default relative-humidity scheme (R_rh) in reference to the satellite image. Again, similarly to LCC, we note that in reference to the satellite picture, the mixed-phase scheme provides a better tendency to resolve the cloudiness, while the cloud structure of the relative humidity scheme has the tendency to remain more compact.

28 April 2010 02:45 UTC

Meteosat/Synesat R_rh R_mix_0.5_4.0

%High Cloud Cover (HCC)

Comparison of High Cloud Cover (HCC) against the corresponding MSG satellite image (Meteosat/Synesat) manipulated with Synesat software available at HNMS. In reference to MSG picture, both cloud schemes overestimate HCC, especially over the West/South-West parts, addressing the issue of proper accounting of cloud-ice content. In contrast to LCC and MCC and in reference to the satellite picture, the relative-humidity scheme provides a better tendency to resolve the cloudiness, while the cloud structure of the mixed-phase scheme has the tendency to remain more compact.

28 April 2010 02:45 UTC

Meteosat/Synesat R_rh R_mix_0.5_4.0

%

Comparison of Total Cloud Cover (TCC) against the corresponding MSG satellite image (Meteosat/Synesat) manipulated with Synesat software available at HNMS. Both cloud schemes are in very good agreement with each other as well as with MSG picture. This feature demonstrates that the relative differences in Low, Middle and High cloud-cover layers between the two schemes are converging towards very similar Total Cloud Cover.

Total Cloud Cover (TCC)

%R

_rh

R_m

ix_0

.5_4

.0

TCC

TCC

HCC

HCC

MCC

MCC

LCC

LCC

Cloud Cover for 11 January 2011 (12 + 06:00 UTC)

Model Cloud Cover (TCC, HCC, MCC, LCC) against the corresponding “common” MSG satellite image (MSG 3.9 microns), as it is available in the University of Dundee Database and for a winter case due to unavailability of Synesat data. In reference to MSG image, we may see that TCC is overestimated for both schemes but the mixed-phase scheme performs better over Eastern parts. This preponderance is mainly due to the differences in LCC and MCC between the two schemes. Regarding HCC both cloud schemes are in very good agreement with each other but in contrast to the previous “spring” test case, HCC is overestimated for both schemes as is shown in MSG picture and this has a strong effect against the agreement with the satellite picture.

MSG 3.9 microns

%

R_rh R_mix_0.5_4.0

Low Cloud Cover

MSG 3.9 microns

Low Cloud Cover

For low clouds in particular, and for the “winter” test case (January 9th-10th 2011), it was noticed that the bulk of low cloud cover for the COSMO model default relative humidity (R_rh) scheme interchanges position with the mixed statistical scheme (R_mix_0.5_4.0).

CONCLUSIONS-COMMENTSDue to the nature of the project, satellite data are of prime importance regarding the evaluation of the investigated sub-grid cloud schemes. There was an effort to make use of MSG data as they can be manipulated at HNMS through the SYNESAT system available locally.Due to limitations regarding older data, a new recent “Spring” test case was chosen, for which satellite data were at the time available in the local SYNESAT station.SYNESAT provided a nice flexibility regarding the comparison of satellite data and model output and quite a few relative differences of the sub-grid cloud schemes under consideration were highlighted. It looks that, in reference to satellite data, the mixed-phase scheme pushes the model towards colder radiation temperatures than the default relative-humidity scheme. The overall cloud cover is decently compared with the satellite figures for both schemes.Regarding medium and low clouds, it looks that the mixed-phase scheme preponderates in areas with relatively low cloudiness and the relative humidity scheme preponderates in more cloudy areas.It should be noted that regarding minimum 2m Temperatures the mixed-phase scheme performs better comparing with observations while the situation for the maximum 2m Temperatures is more balanced between the two schemes.For low clouds in particular, by carefully choosing another “winter” recent test case it was detected that the bulk of low cloud cover for the COSMO model default relative humidity scheme interchanges position with the mixed-phase statistical scheme. On the basis of the current as well as past works, the overall effect regarding the operational implementation of the sub-grid mixed-phase statistical cloud scheme to COSMO model over the wider geographical area of Greece may considered overall neutral. However, the overall slight improvement of minimum daily 2m-temperature as well the middle and low cloudiness over marine areas should be considered decent assets. The implementation of the SGS mixed-phase scheme as an option to a later COSMO model version will be of value for both operational and research purposes in the COSMO community and is recommended although a systematic investigation will be of value to fully evaluate the mixed-phase scheme.