Inverse Methods for Climate Model Parameter Estimation and Climate Sensitivity

点击翻译按钮获取中文摘要

The proposed research will explore the application of systematic techniques for parameter estimation based on inverse methods to climate models having realistic dynamical complexity but with ideal physical parameterizations. These techniques seek to find parameters and empirically derived forcing corrections that minimize the deviation of the climate model solution from analyzed atmospheric phase space trajectories in the NCEP/NCAR reanalysis data. This approach may be viewed as using atmospheric circulation data in order to improve subsequent forecasts and climate model simulations. The goals of the proposed research are (i) to assess the potential for the application of these techniques in this idealized setting with an eye toward application to more realistic climate models, (ii) to study and understand the skill and reasons for loss of skill for climate models with empirical physical parameterizations in AMIP-style simulations forced by atmospheric boundary conditions, and (iii) to use these techniques to attribute cause and effect in the problem of understanding the catalytic role of midlatitude atmospheric dynamical processes in the response of the climate to imposed forcing. The work is important because it will yield insights into improving climate forecasts through more realistic climate model simulation.