Our approach is to use NWP analysis (or reanalysis) with field observation data to study how a model drifts from observations in a short simulation (0-72 hours). We are developing an initialization method that will allow a climate model to be used in a weather forecast mode. We will then study the model drift from the analysis and from field observations. This approach can be used to improve parameterizations responsible for models errors on longer time scales in much the same way that NWP centers use short-range forecast verifications to improve their models. The goal is to make the techniques developed available to all modeling groups in an effort to improve climate models.

The Atmospheric Radiation Measurement (ARM) data will be compared to the high frequency GCM output. Results presented suggest major model errors in the magnitude and timing of both temperature and humidity anomalies. We will also present diurnal temperature behavior of the forecasts, measures of temperature sensitivity, coupling and surface-troposphere coherence of temperature from both observational estimates (e.g., reanalysis) as well as field data (e.g., ARM data).

This work was performed under the auspices of the U.S. Department of Energy by University of California Lawrence Livermore National Laboratory under Contract No. W-7405-Eng-48.