Factoring spatial variability of rainfall in the design of optimum rain gauge network for Uganda.

Abstract

Weather and climate monitoring is a strategic undertaking by the global community for purposes of understanding and planning with the climate natural resource. The challenge of climate change calls for accurate meteorological data, information and/or advisories especially while undertaking climate trend analysis for planning purposes. For rainfall, a well-designed rain gauge- network addressing spatial variability is critical for accurate and reliable estimates of the areal or point average rainfall estimates at any desired location. This study therefore sought to assess the current rain gauge network and design a feasible optimum rain gauge network for Uganda. The study also sought to validate the global datasets (GPCC and CRU-TS, 1901-2013) against the UNMA data using statistical measures, like; Long-term mean; coefficient of correlation for Seasonal rainfall, extreme events, as well as the demonstration of spatial variation of the reanalyzed data within a climatological zone (CZ). However, after realizing the insensitivity of the global datasets to spatial variation within a grid, UNMA observed data was used for consistence with the topic of the study. Basing on the 16 HCZs and quality controlled data for 136 stations from previous studies, the WMO recommended formula was applied to determine the Feasible Optimum Rain Gauge Network (FORN) for each CZ through suitability analysis using AfcGIS VI 0.3 software. A 7% maximum allowable error (g) for. rainfall estimation, was subjected to the coefficient of variation of every CZ to keep the error as low as possible but also to factor in affordability and sustainability. Results indicated very low functionality of 5.2% for the current rain gauge network compared to the colonial time coverage of 1075 stations. The buffer analysis yielded the land area left for locating rain gauges, which when divided by the number of stations in a HCZ, gives the pixel size, translating to the gauge density per that particular zone. The 7% rainfall estimation error therefore resulted in 1,057 rain gauges (921 new Rain gauge stations and the 136 Reference stations), which is close to 1,075 rain gauges that have ever been operated though with subjective distribution and hence the term 'Feasible’. Once the resultant network design is fully implemented, poor coverage and generation of adequate rainfall data shall be addressed; which will further help in comprehensive hydrological analyses to support water resources management plans, Ito boost the national climate change adaptation and mitigation efforts especially within the agriculture and energy sectors.