Investigation Of Different Parameterizations Of The Mixed Liquid-ice Sub-grid Statistical Cloud Scheme Over Greece E. Avgoustoglou Hellenic National Meteorological Service 13 th COSMO General Meeting, Servizio Meteorolologico, Rome 4 th to 9 th 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) Highlights of the sub-grid liquid-ice mixed scheme

A first order approximation for is assumed in reference to and by using Clausius-Clapeyron equation cloud fraction R is given by Sommeria και Deardorff (1977) further approximated R empirically : 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: 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-up Seven cases are under investigation: #Starting Date 1Jan UTC (d050101_12) 2Dec UTC (d071224_12) 3Apr UTC (d060425_12) 4May UTC (d080509_12) 5May UTC (d090501_12) 6April UTC (d110427_12) 7January UTC (d110109_12)  COSMO_4.6 (Modified by M.R)  Horizontal grid (~7 Km) 273  273 points  40 vertical levels  time step: 30 sec  GME (Analysis ) 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 27 th UTC (mainly) and the “Winter Case” of January 9 th UTC in reference to the default relative-humidity scheme, satellite pictures (MSG Infrared, Water vapor and Cloud Cover) as well as station observations. Results

Deg C Station # 28-April-2011 T2m_min Minimum daily 2-meter Temperatures for 28 th April 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)

28-April-2011 T2m_max Station # Maximum daily 2-meter Temperatures for 28 th April 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 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 :45 UTC % Meteosat/SynesatR_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 :45 UTC Meteosat/SynesatR_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 :45 UTC Meteosat/SynesatR_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 :45 UTC Meteosat/SynesatR_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_mix_0.5_4.0 TCC HCC MCC LCC Cloud Cover for 11 January 2011 ( :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_rhR_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 9 th - 10 th 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-COMMENTS  Due 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.