Downscaling of current and future rainfall climatologies for southern Morocco. Part I: Downscaling method and current climatology

The use of a statistical-dynamical downscaling approach to obtain a high-resolution rainfall climatology for a subtropical region in southern Morocco is analysed. The statistical part of the downscaling uses Circulation Weather Types (CWTs) as a measure for near-surface wind fields calculated from sea-level pressure (SLP) data. The daily CWTs are correlated with daily rainfall data at three different climate stations in the region for the period 1978-1997 and systematic differences are discussed. Results are viewed for one extreme dry year (1984) and one extreme wet year (1989) to show the limits of purely statistical downscaling. Dynamical downscaling is realized for representative days to enable statistical-dynamical downscaling. Comparison of statistically-dynamically downscaled rainfall with measured rainfall for the year 2002 and climatological results show that the method is capable of capturing the relevant mechanisms triggering rainfall in the area. For extreme dry (wet) years, rainfall is overestimated (underestimated) by the method and errors might occur even for normal years. However, application to a climatology calculated for the period 1958-1997 via analysing daily SLP fields from NCEP Reanalyses gives satisfactory results. For sufficiently long periods, the method is well capable of producing a reliable high-resolution rainfall climatology. We will therefore apply the method to climate change simulations in part II of this paper. Copyright (C) 2007 Royal Meteorological Society.