Precipitation downscaling in climate modelling using a spatial dependence function

Climate change impacts are among the future significant effects on the socioeconomic, cultural and political affairs of any society, whether as towns, cities, regions, countries and the world at large. It is therefore necessary to predict future impacts by modelling logical and rational scenarios. Such scenarios are available internationally in the Special Report on Emission Scenarios (IPCC) and their inclusion within numerical oceanic-atmospheric models produce future synthetic data concerning various meteorological variables (temperature, precipitation, relative humidity, wind speed, irradiation, etc.). Unfortunately, the end products of General Circulation Models (GCMs) are available at a set of nodes which are about 300 km apart from each other. Hence, it is necessary to downscale these data to smaller scales so that future planning can be achieved at local scales. It is the purpose of this paper to present Spatial Dependence Functions (SDFs), which consider the regional dependence between the ground measurements, say precipitation, and the GCM output data. For this purpose 299 precipitation records are considered from Turkey with their SDFs. As an example, the SDF for the city of Istanbul is presented and the necessary monthly rainfall downscalings are generated from 2000 to 2100 on the basis of the National Center for Atmospheric Research (NCAR) centre's GCM data using the A2 SRES scenario.