Simulation of changes in the twenty-first century maximum temperatures using the statistical downscaling model at some stations in Botswana

Abstract

Simulation of future temperature changes is crucial in the assessment of regional impacts of climate change. This study uses the statistical downscaling Model to simulate changes in monthly maximum temperatures for the selected present period (1981–2010) and future periods (2011–2040, 2041–2070 and 2071–2099) at four weather stations in Botswana based on the Hadley Centre Couple Model Version 3 (HadCM3) output under A2 and B2 scenarios. The four weather stations are Francistown, Gantsi, Shakawe and Tshane. Their observed daily maximum temperatures (predictands) were obtained from the Botswana Department of Meteorological Services. Large scale atmospheric variables (predictors) were obtained from the data portal http://is.gd/hadcm3 . More attention in the analysis is given to simulated monthly maximum temperatures and the 95th percentile of simulated monthly maximum temperatures. The simulated temperatures are assessed against observed mean monthly maximum temperatures for the selected present period. The highest temperature increase is predicted to occur in November at all the selected four stations while the lowest temperature increase is predicted to occur in January at three of the selected four stations. At two of the selected four stations, temperature increase related to A2 predictors is greater than that related to B2 predictors. This is not surprising because the A2 scenario is worse than the B2 scenario.