AN OBJECTIVE METHOD FOR INFERRING SOURCES OF MODEL ERROR

A restricted statistical correction (RSC) approach is introduced to as sess the sources of error in general circulation models (GCMs). RSC mo dels short-term forecast error by considering linear transformations o f the GCM's forcing terms, which produce a ''best'' model in a restric ted least squares sense. The results of RSC provide 1) a partitioning of the systematic error among the various GCM's forcing terms, and 2) a consistent partitioning of the nonsystematic error among the GCM for cing terms, which maximize the explained variance. In practice, RSC re quires a substantial reduction in the dimensionality of the resulting regression problem: the approach described here projects the fields on the eigenvectors of the error covariance matrix. An example of RSC is presented for the Goddard Earth Observing System (GEOS) GCM's vertica lly integrated moisture equation over the continental United States du ring spring. The results are based on the history of analysis incremen ts (''errors'') from a multiyear data assimilation experiment employin g the GEOS model. The RSC analysis suggests that during early spring t he short-term systematic forecast errors in the vertically integrated moisture are dominated by errors in the evaporation held, while during late spring the errors are large in both the precipitation and evapor ation fields. The RSC results further suggest that one-quarter to one- half of the nonsystematic forecast en ors in the vertically integrated moisture may be attributable to GCM deficiencies. Limitations of the method resulting from ambiguities in the nature of the analysis increm ents are discussed. While the RSC approach was specifically developed to take advantage of data assimilation experiments, it should also be useful for analyzing sequences of somewhat longer GCM forecasts (simil ar to 1 day) as long as they are short enough to consider the errors a pproximately local.