EXPANDED DOWNSCALING FOR GENERATING LOCAL WEATHER SCENARIOS

In an attempt to reconcile the capabilities of statistical downscaling and the demands of ecosystem modeling, the technique of expanded down scaling is introduced. Aimed at use in ecosystem models, emphasis is p laced on the preservation of daily variability to the extent that poss ible climate change permits. Generally, the expansion is possible for any statistical model which is formulated by utilizing some form of re gression, but I will concentrate on linear models as they are easier t o handle. Linear statistical downscaling assumes that the local climat e anomalies are linearly linked to the global circulation anomalies. I n expanded downscaling, in contrast, I propose that the local climate covariance is linked bilinearly to the global circulation covariance. This is done by transforming the technique of unconstrained minimizati on of the error cost function into a constrained minimization problem, with the preservation of local covariance forming the side condition. A general normalization routine is included on the local side in orde r to perform the downscaling exclusively with normally distributed var iables. Application of the expanded operator to the daily, global circ ulation works essentially like a weather generator. Using observed geo potential height fields over the North Atlantic and Europe gave consis tent results for the weather station at Potsdam with 14 measured quant ities, even for moisture-related variables. For GCM (general circulati on model) scenarios, satisfactory results are obtained when the origin al variables are normally distributed. If they are not, strong sensiti vity even to small input changes cause the normalization to produce la rge errors. Non-normally distributed variables such as most moisture v ariables are therefore strongly affected by even slight deficiencies o f current GCMs with respect to daily variability and climatology. This marks the limit of applicability of expanded downscaling.