The objective of the Research is to utilize high-precision satellite microwave measurements of atmospheric temperature as a base for validation and analysis of output from global climate models involved in the Atmospheric Model Intercomparison project (AMIP) being carried out at Lawrence Livermore National Laboratory (LLNL).

The Atmospheric Model Intercomparison project (AMIP) was an experiment in which nearly all of the available atmospheric General Circulation Models (GCMs) were systematically driven by a common set of sea surface temperatures (SSTs) over the 1979-88 period. The GCMs of AMIP, given this common forcing, were thus more easily intercompared with one another, and compared with actual observations.

In the following study gross measures of tropospheric temperatures were studied in comparison with the troposheric temperatures observed by the Microwave Sounding Units (MSUs) for the same 1979-88 period. Two measures of temperature from AMIP GCMs were utilized, 1) Zonal mean pressure-level temperatures from which a simulated MSU temperature is derived and 2) the 850-200hPa thickness (gridpoint).

GCM simulated temperatures were found to have mimicked the general interannual differences, related to E1 Nin~o events, quite well. However, a more subtle measure of model variation-decadal trend-showed that in nearly all GCMs, the trend was greater than observed. Most GCMs also demonstrated a thickness response that was less than what was observed-a consequence of GCN atmospheric heating that was deposited too low in the modeled atmospheric column.

The appropriateness of using these GCMs as regional climate predictors is questioned as it was found that regional anomalies in the GCMs are often more persistent and sever than those observed in the real world. Therefore, while providing, relatively good global and tropical response to SSts, further GCm improvements are required to more accurately represent decadal trends and regional variation-two key parameters for studying global climate change.