Comparison of spatial interpolation techniques to generate high-resolution climate surfaces for Nigeria

Climate datagaps are a major challenge for understanding the impacts of climate change on ecosystems, particularly in highly climate vulnerable countries such as Nigeria. The generation of gridded climate data sets in the form of interpolated surfaces may help to fill climate datagaps and in turns enable climate change impact assessments. This article generates climate surfaces of monthly total precipitation and minimum and maximum temperatures for Nigeria at 0.001° spatial resolution by comparing two spatial interpolation techniques, i.e. kriging with external drifts and thin plate splines. Climate data from 43 meteorological stations covering the period of 1960–2012 were used to generate climate surfaces fitting the longitude, latitude, elevation and distance to coastline of the stations as independent variables. Three model error statistics, i.e. root mean square error (RMSE), Nash–Sutcliffe efficiency (NSE) and index of agreement (d), were used to evaluate and compare the performances of interpolation techniques. The second-order and third-order partial thin plate splines were identified as the optimal models for generating precipitation, and minimum and maximum temperatures surfaces, respectively. The best-fit surfaces yielded an average RMSE, NSE and d of 14.98, 0.87 and 0.97 for precipitation, 0.42, 0.91 and 0.98 for minimum temperature and 0.52, 0.89 and 0.97 for maximum temperature. Our high-resolution climate surfaces are freely available from an online repository and widely applicable for climate change analysis as well as for biological, forestry and agricultural studies in Nigeria.