CR 28:199-212 (2005)  -  doi:10.3354/cr028199

ABSTRACT: statistical models were defined for explaining the spatial distribution of rainfall in eastern Africa (southern Kenya and NE Tanzania) on the basis of various topographical descriptors. The procedure is first presented for the month of April, and subsequently month-to-month changes in the role of topography on rainfall distribution are discussed. Predictors included quantitative estimates of slope, mean and standard deviation of elevation, and principal components that describe the topographical environments around the meteorological stations. Different window sizes were considered in order to account for processes of different scales. Linear models were defined for each mean monthly rainfall field, and cross-validated. Inclusion of interpolated residuals yield explained variances in the range of 53 to 89%. Results were compared to a simple interpolation of gauge rainfall. The models are particularly useful in areas with few rain gauges. Selected predictors, as well as their scales, change seasonally. North–south exposure contrasts are the main factor of rainfall variation, except in the northern summer. East-facing stations are wetter in the short rains season (October to December), and drier in the monsoon season. These variations coincide with seasonal atmospheric circulation changes above the region. Slopes also contribute to rainfall variability. At the scale of the study mean elevation had little effect; other predictors, especially exposure, actually describe the interaction between rainfall and topography more adequately.

KEY WORDS: Rainfall · Geostatistical modelling · Orography · Kenya · Tanzania