SGER: A Stochastic Single Column Modeling Framework for Diagnosing Community Climate System Model (CCSM) Tropical Variability

University Of Colorado At Boulder, Boulder CO

Investigators

Abstract

This SGER grant supports the construction of a computationally inexpensive numerical climate model that reproduces tropical variability from the full and computationally expensive Community Climate System Model (CCSM) Atmospheric Model (CAM). The idea is that the inexpensive model would then be used as a diagnostic tool for testing new parameterizations in the CAM. This is an innovative idea that examines a strategy for assessing the quality of climate forecasts made years to centuries in advance, where, unlike for weather predictions, validation observations at forecasted times are not available in a few days or weeks. The traditional method of evaluating a climate model's ability to predict long into the future is to estimate its error statistics for retrospective "forecasts" or simulations of, e.g., the climate of the 19th and 20th centuries, for which global observations exist. Climate models are complex and expensive to execute on computers. As a result comprehensive error statistics are not practical. Instead, the PI's computationally inexpensive numerical model can be used to diagnose the full CAM's ability to simulate the important features of the tropical variability. This project has the potential to provide a large number of CCSM/CAM developers and users with an efficient and effective tool for advancing their research.