Abstract:
Water resources systems are important for various uses such as agriculture, water supply, energy, transportation, recreation, ecosystems, among others. Design of water resources systems is of increasing importance due to the ever increasing demand for efficient water management. This study explores water resources assessment in two different regions with contrasting climate, socio-economic development level and data availability. The Thames catchment in England is in the temperate zone, which is data rich and is at a high socioeconomic development level; while, The Tana catchment in Kenya is in the tropics, which is data poor and at a low socioeconomic development level. Accurate estimation of precipitation is a key process in assessing water resources. To furnish optimal input data, point rainfall at un-gauged locations from measurements at surrounding sites is used in obtaining a continuous surface of relevant information. Rainfall interpolation based on the four commonly used methods (namely Thiessen polygon, Thinplate, Inverse Distance Weighting (IDW) and Kriging) is done and validation is achieved by Leave-one out cross validation (LOOCV) method. Different rain gauge densities have been explored to search the optimum interpolation method so that appropriate schemes could be adopted for the catchments in England and Kenya. Other data sources in addition to rain gauges are collected and processed including temperature, stream flow, and solar radiation. Furthermore, hydrological models suited to different catchment characteristics are explored and used to assess different water resources utilisation options. The commonality and differences for the data and the model between the two regions are then analysed. Finally, data fusion techniques are used to integrate data from different sources to quantify data uncertainty and maximise their accuracy. Findings from the study are useful to water resources specialists when assessing water resources across different regions of contrasting climates, geographical zones, socioeconomic development levels and data availability.
