STOCHASTIC DOWNSCALING OF GCM-OUTPUT RESULTS USING ATMOSPHERIC CIRCULATION PATTERNS

Summary

ABSTRACT A methodology is presented for downscaling GCM-output results to regional scale precipitation using atmospheric circulation patterns. A sequence of observed daily air pressure distributions is used to define circulation patterns. The classification of the circulation patterns is done with the help of a neural network. A multivariate stochastic model describes their link to observed daily precipitation amounts at a number of selected locations. To assess precipitation under changed climate circulation patterns derived from GCM output pressure values are used to condition the stochastic precipitation model. The methodology is demonstrated by results obtained for a selected location (Essen, Germany).

INTRODUCTION

Climate change will have a major influence on the hydrological cycle. It is of vital importance to assess the possible impacts as soon as possible, in order to find strategies to adapt to these changes.

The only physically based tools in predicting climate change effects are General Circulation Models (GCM). GCMs deliver meteorological variables in a fine time resolution (30 minutes to a few hours) but in a very coarse spatial grid (200–500 km × 200–500 km). Many climatic parameters like temperature, precipitation, wind, clouds, radiation, snow cover and soil moisture can be simulated by these models.

Unfortunately precipitation, which is the main input in hydrological models, cannot be well modelled by the GCMs. GCM control runs for the present climate indicate that single grid values cannot be taken as representative rainfall amounts of the corresponding area.