Abstract ATM-9520579 Schneider, Edwin K. Kinter, James L. Institute of Global Environment and Society Inc. Title: Variability of the Climate System The project consists of a series of numerical experiments designed to increase understanding of the natural variability of the climate system. It is necessary to have accurate estimates of the natural variability to identify climate signals produced by external forcing and to evaluate climate predictions. This research is relevant to the goals of TOGA and CLIVAR/GOALS and is expected to contribute to IPCC/USGCRP. The proposed research will define and carry out an experimental procedure to isolate and estimate the contributions to natural climatic variability from atmospheric, land and oceanic processes. Specific processes that will be examined are atmospheric chaos, stochastic forcing of the land surface, stochastic forcing of the ocean mixed layer, coupled ocean/atmosphere chaos, the annual cycle of solar forcing. Results will be obtained by analysis of several multi-decade to century long numerical integrations made with variants of a numerical climate model. The basic version of the model will include global atmosphere and ocean general circulation submodels, land surface processes including effects due to vegetation, and sea ice. First, the errors of the fully interactive coupled model in simulating the observed annual mean and annual cycle of sea surface temperature will be ascertained. Modifications to the model will be made to reduce the magnitude of these errors. A control integration without flux corrections will be carried out to determine the equilibrium climate. In subsequent integrations the interactions between the component submodels will be restricted in order to either eliminate or reduce atmosphere/land and atmosphere/ocean coupling. The changes, which will be introduced in various combinations, include replacement of the dynamical ocean with a specified climatological sea surface temp erature or by mixed layer ocean, specification of climatological soil mixture and snow cover, and elimination of the annual cycle. The climatology for the restricted model simulations will be that of the control integration. Intercomparsion of the results and Nature will provide estimates of the contributions of the atmosphere/land and atmosphere/ocean interactions to the modeled and actual climate variability. Related experiments will also be done to study the influence of flux correction and oceanic initial conditions in climate simulation and prediction. Changes in natural variability introduced by the flux correction procedure will be evaluated using an equilibrium simulation in which the full climate model before improvement is flux corrected to produce the same climatology as the control integration. An experiment will be done to estimate the departure of the current climate from equilibrium due to the changing greenhouse forcing.