Developing an Automatic Metric and Diagnostic Package for CAPT Simulations

Description

An automatic diagnostics package has been developed for CAPT simulations to facilitate the process to systematically explore model errors and judge the improvement of model simulations from new parameterizations. The package can also be used to understand how model errors grow with the forecast lead time and find linkage of model errors to particular physical processes. Starting as simple, the current version emphasize on a few key measures for evaluating climate model forecast skills and their simulated precipitation, clouds, and radiation. The metrics and diagnostics are developed for comparing with satellite data at global and regional scales and with field data at limited locations (e.g., ARM sites). The metrics/diagnostics for errors in mean state are shown through global maps of the biases and Taylor diagram using the comparison with satellite data. The time series, diurnal variability, and PDF statistics are also shown using the ARM data. This poster will present preliminary results from implementing this diagnostics package to the 2-year CAPT hindcasts with CAM4 and CAM5 over the Years of Tropical Convection (YOTC) period. Some model errors explored by the diagnostic tool will be discussed. This work is performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.