Methods for statistical downscaling of GCM simulations

General Circulation Models (GCMs) are used to study the change of climate due to increases in greenhouse gases in the atmosphere. As GCMs operate on !arge spatial scales, and, furthermore, as the GCM-simulated temporal resolution corresponds to monthly averages at best, the usefulness of GCM data in impact studies and other applications is limited. The present-day free troposphere is modeled relatively well by the coarse GCMs, whereas local or even regional characteristics in surface or near-surface climate variables, their variability and the likelihood of extreme events cannot be obtained directly from GCMs. The same is likely true in the case of climate change experiments with GCMs. The results from GCMs can be superimposed on climatological local­ scale time series or interpreted in some other way in order to address the needs of impact studies. This is known as "downscaling" of GCM simulations. In this survey, five different downscaling methods are introduced. These are the conventional, the statistical, the stochastic, the dynamical and the composite methods. Only the statistical and, to a lesser extent, the stochastic approaches are discussed in detail. This survey is a planning document in the SWECLIM program