The implications of general circulation model estimates of rainfall for future erosion: a case study from Brazil

One consequence of global change will be shifts in the probability of occurrence of soil erosion by water. This could have serious consequences for those areas of the world which are present-day 'hotspots' for erosion. By means of a case study, this paper suggests an approach to quantifying the change in risk of serious erosion for sites in such areas. The case study focuses on future erosion under intensive soya bean cultivation in the Mate Grosso area of Brazil. On the area's highly erodible latosols, current erosion problems are severe. Scenarios of change future climate change are taken from general circulation models (GCMs) and used to perturb current-climate weather data. These are input to an erosion model (water erosion prediction project (WEPP)-CO2), together with local knowledge regarding current and probable future land use, in order to estimate future changes in erosion rates. WEPP-simulated average annual sediment yield increases in one of the scenarios and decreases in the other two, reflecting the range of uncertainty in predictions of future rainfall. Using the 'best-guess' climate scenario from the UK Meteorological Office's HADCM2 GCM, the increase in mean annual sediment yield is 27%. Increases are disproportionately greater in wetter years. Average rates for individual months increase by over 100%. Erosion increases most on those parts of the hillslope profile which are currently hardest-hit by erosion. At present, an annual sediment yield of 5 t ha(-1) is currently exceeded in about 1 year in 2. The HADCM2 simulations suggest that an equal or greater rate will occur in about 70% of years by around 2050. A rate of at least 10 t ha(-1) yr(-1) is currently exceeded in about 1 year in 5. The HADCM2 simulations suggest that this will rise, to about 1 year in 4. (C) 1999 Elsevier Science B.V. All rights reserved.