Optimizing parameters in an atmospheric general circulation model

An efficient method to optimize the parameter values of the subgrid parameterizations of an atmospheric general circulation model is described. The method is based on the downhill simplex minimization of a cost function computed from the difference between simulated and observed fields. It is used to find optimal values of the radiation and cloud-related parameters. The model error is reduced significantly within a limited number of iterations (about 250) of short integrations (5 yr). The method appears to be robust and finds the global minimum of the cost function. The radiation budget of the model improves considerably without violating the already well simulated general circulation. Different aspects of the general circulation, such as the Hadley and Walker cells improve, although they are not incorporated into the cost function. It is concluded that the method can be used to efficiently determine optimal parameters for general circulation models even when the model behavior has a strong nonlinear dependence on these parameters.