Land Cover Changes and Its Effects on Streamflows in the Malewa River Basin, Kenya

Abstract

Vegetated landscapes are transformed by both natural and human causes. This is thought to influence river flow regimes. It is argued that restored and reforested landscapes increase stream flow. However, studies done to date have been inconclusive on whether or not trees on restored or reforested landscapes increase stream flow. This study aimed to examine the effects of land cover changes on streamflow of the Malewa River Basin in Kenya. Satellite imagery based spatial change detection using ArcGIS 10.1 and ERDAS IMAGINE software was deployed to estimate the land cover changes. Based on projected land cover change data, a multiple regression technique was used to establish the relationship between land cover and streamflow. The results show that at Gauge 2GB01, area under wetland significantly predicted stream flows (b=0.134, t(488) =1.978, p=0.049), with an overall model (R2 =0.018, F(3, 488)=2.976, p=0.031). Area under grassland (b=0.108, t(488)=2.325, p=0.02), shrubland (b=0.112, t(488)=1.976, p=0.049) and amount of rainfall (b=0.533, t(488)=14.048, p=0.000) combined significantly predicted stream flows. Rainfall alone significantly predicted stream flows (b=0.531, t(488)=13.885, p=0.000). Overall, the gains in forest restoration did not specifically influence streamflow except in combination with other vegetation and rainfall. There is need to increase soil cover rather than woody biomass alone in the regulation of stream flows. A systematic response to address the drivers of change in land cover is also needed.