Impact of Land Cover, Land Use and Climate Change on the Hydrological Regimes of the Mara River Basin

Abstract

The Mara river basin is a serene sub-catchment of Lake Victoria basin and a part of the upper catchments of the Nile Basin. Importantly, it is a lifeline to the internationally renowned Mara - Serengeti ecosystem and a rich mix of physiography, indigenous cultures and land use practices. The need for more land for the rapidly growing population in the basin has resulted in the excision of large parts of the forests that were formerly preserved, and encroachment to other fragile ecosystems, which has affected the flow volumes of Mara River. Little information exists on how variation in the human activities owing to climate change is affecting the hydrology of the basin. Lack of such information hampers the ability of the various resources managers to formulate a comprehensive ecosystems management strategy that would provide for sustainable livelihoods. This study focuses on the nature, extent and rate of change in land cover, land use and climatic scenarios and their impacts on hydrological regimes in Mara River Basin. Land cover dynamics were analyzed from dry weather Landsat TM and ETM+ images for 1984, 1995, 2003, 2011 and 2016 under ENVI 5.0 and ArcGIS 10.4 software environment. Thematic maps comprising eight land cover/use types were created from the imageries using unsupervised land cover classification technique and FAO land use classes. Time series analysis of the thematic maps was done using post classification visual and area comparisons and overlay operations with significance of change tested at α=0.05 using Chi-Square test statistic. Trends in the simulated long-term daily records of maximum and minimum air temperature, rainfall and river discharge from global weather geoportal were analyzed by use of Mann-Kendall test statistic and Student t statistic, testing level of significance at α =0.05. The study analyzed the past land cover, land use and climate scenarios and modelled future variations with respect to the Mara hydrological regimes (flow volumes). SWAT hydrological model was to simulate the land cover/use and climate scenarios and the river discharge. The simulated data from global weather databases used in the simulation exercise because of lack of complete and consistent hydro-climatic data in the basin. The simulated and observed datasets were however, calibrated and validated for reliability and suitability of use and gave a strong correlation, which meant that, the simulated datasets were good enough for use in the study. The SWAT output data analyzed through the Multiple Regression vii and Correlation model. The analysis of land cover scenarios revealed significant changes in which, forestland, shrub land and grassland are all decreasing in spatial coverage while cropland and built-up areas are spatially increasing at accelerated rates with wetlands, water bodies and bare land not having definite trends. Thus, between 1984 and 2016, forestland, shrub land and grassland reduced by 1.37%, 6.61% and 2.99% respectively while cropland and built-up area increased by 10.22% and 0.06% respectively. The analyses of historical data revealed that the climate of the area has progressively become warmer and drier from the 1980s with air temperature indicating gradual upward trends while rainfall and river flow volumes indicate a decline in their trends. When tested at α = 0.05, however, neither the mean monthly nor annual rainfall values recorded significant change. The rains in the Mara are bimodal with the main rainy season experienced between March and May while the second rains come from September to November. The study detected a shifting trend in rainfall in the basin with the usually wet months (MAM and SON) becoming relatively dry while the dry months (DJF and JJA) are becoming relatively wet. On the other hand, the mean monthly maximum and minimum air temperature recorded significant increments in some months while the mean annual values had none at α = 0.05. The trends in land cover, land use and climate expected to continue beyond 2030 with serious impact on the Mara water resources and therefore livelihoods and environmental services. Conservation of this former pristine basin is a critical matter that addressed by both governments of Kenya and Tanzania respectively. This study provides useful methods and information that can apply broadly to inform medium and long-term planning of water resources management not only in the Mara River Basin but also in other basins in and outside Kenya. The results should be incorporated in the mainstream economic strategies and implementation of the National Water Master Plan, 2030.