Assessment of NASA GISS Post-CMIP5 Single Column Model Simulated Cloud Fraction, Liquid Water Path, and Precipitation Using Ground-based and Satellite Observations

Thursday, 18 December 2014
Lei Zhang1, Xiquan Dong1, Baike Xi1, Aaron D Kennedy1 and Zhanqing Li2, (1)University of North Dakota, Grand Forks, ND, United States, (2)Univ of Maryland College Park, College Park, MD, United States
As the NASA GISS Model E2 GCM (Post-CMIP5) single column model (SCM) recently changed the planetary boundary layer turbulence and moist convection parameterizations, the investigation of any improvements is necessary. In this study, we compared model simulated cloud fraction (CF), liquid water path (LWP), and precipitation with the ground-based and satellite observations at the ARM SGP site from 2002 to 2008 in response of changing the relaxation times. Relaxation techniques are used in simulation to prevent large model error during the simulation. Study revealed that the simulated results agree with observations much better in temperature and q profiles in the 3-hr scale, while the 24-hr scale can produce precipitation and CF closer to the observations. That is, the short relaxation time scale tends to produce dryer atmosphere, which inhibits precipitations and clouds development. For precipitation rate, results of SCM under 24h (3h) are 2.24 (1.35) mm/day, respectively, compared with the ARM observation of 2.43 mm/day. For CF, we have reached the following conclusions: (1) the results of SCM 24h (3h) are 46% (39%)against an observation of 56%; (2) Simulations under 24h and 3h relaxation time can capture the seasonal trend reasonably well, but underestimate the CF, especially for high clouds. The diurnal cycles of simulated clouds tend to peak too early (6 am) compared to the observations (12 am). The mean cloud profiles around 850hPa are improved due to the improvement of PBL turbulence parameterizations. Strong bias occurs in the cloud frequency, simulations predict too much (little) occurrence of mid (high) clouds, and a notable peak in August that is not shown in the observations. For LWP, the simulation under 24h (3h) relaxation time result 157 (171) g/m2, respectively, the one at 24h closes to the ARM value 149 g/m2. The diurnal cycle of simulated LWP peaks later than observations. Results from this study can be used as guidance for the further development of the GISS Model.