Use of satellite data to assess tropical clouds in climate models
Steve Klein and Jerry Potter Lawrence Livermore National Laboratory
The varying representation of clouds in climate models leads to varying predictions of the nature and magnitude of climate change. Clouds and the associated convection over the tropical ocean can play a large role in determining the sensitivity of climate. In this project, we plan to use both recent and future satellite datasets to assess the nature of tropical clouds in two climate models when run in a forecast mode. This bypasses the need to acquire very long records to assess climatological characteristics. At the Program for Climate Model Diagnosis and Intercomparison (PCMDI), the DOE sponsored Climate Change Prediction Program (CCPP) and the Atmospheric Radiation Measurement (ARM) Program have combined resources to create the CCPP-ARM Parameterization Testbed (CAPT) to facilitate climate model parameterization testing in a weather forecast mode. The CAPT project is currently using the NCAR CAM 3.1 the GFDL AM2 model.
Our work plan is composed of two related studies. The first will focus on the characteristics of the clouds, precipitation, and radiative forcing associated with the 1997/98 El Niño using satellite data from several platforms including CERES and ISCCP. Climate models have trouble simulating the response of circulation, convection, and clouds to the anomalous sea-surface temperature in the tropical Pacific for this unusual period. Of interest is whether these same models will have difficulty simulating the response of convection and clouds when they are give the observed atmospheric state as initial conditions for forecasts. This test will also help to improve understanding of the response of the atmospheric models to El Niño-like conditions when they are coupled to an interactive ocean. Forecasts for selected periods during the peak of the El Niño will be made available though PCMDI and DIME sites. We will also solicit suggestions for creative diagnostics from the CMAI research groups.
In the second study, we will examine the vertical characteristics of tropical clouds from the recently launched CloudSat and humidity structures from AIRS in comparison to those simulated by the NCAR and GFDL climate models when run in forecast mode. Previous work has demonstrated rather large differences between the NCAR CAM3 and GFDL AM2 vertical distributions of relative humidity, cloud fraction, and cloud condensate; contrasting these models’ cloud representation with the satellite observations will presumably highlight deficiencies in the representation of deep convection among other processes. An additional component of this work is the coupling of a previously developed CloudSat simulator to the pre-existing ISCCP simulator. The resultant satellite data simulator will enable direct comparison of model clouds to satellite data and be made available to the broad scientific community.
In summary, we will need satellite data in a coordinated and useful form from CERES, TRMM, CloudSat, AIRS, MISR, and ISCCP to use in comparison with high frequency model output. We also will require access to the CloudSat simulator to make corresponding comparison of model output to observed cloud structure and composition. We are already making effective use of the ISCCP simulator in both the CAM3 and AM2 models and we will provide the CMAI effort with output for collaborative model evaluation.
