Evaluation of spatial interpolation methods for annual rainfall on the highlands of Eritrea

This article evaluates different spatial interpolation methods to form a continuous surface based on measured annual rainfall totals. Interpolation methods examined in this study include inverse distance weighted (IDW), Spline and Kriging. The annual rainfall totals measurements were taken at 22 discrete locations in the highlands of Eritrea. Root mean square error (RMSE) and index of agreement (D) quantified suitability. Judging from the statistical results, the standard implementation of Kriging provides the smallest prediction errors than IDW and Spline. IDW prediction errors were almost similar to Kriging while Spline produced the largest prediction errors.