Comparison of spatial interpolation methods for estimating the precipitation distribution in Portugal

Precipitation has a strong and constant impact on different economic sectors, environment social activities all over the world. An increasing interest for monitoring estimating precipitation characteristics can be claimed in last decades. However, some areas, ground-based network is still sparse spatial data coverage insufficiently addresses needs. In decades, interpolation methods provide an efficient response describing distribution of precipitation. this study, we compare performance seven used retrieving mean annual mainland Portugal, as follows: local polynomial (LPI), global (GPI), radial basis function (RBF), inverse distance weighted (IDW), ordinary cokriging (OCK), universal (UCK) empirical Bayesian kriging regression (EBKR). We generate using from 128 rain gauge stations covering period 1991 to 2000. The results were evaluated cross-validation techniques each method was error (ME), absolute (MAE), root square (RMSE), Pearson’s correlation coefficient (R) Taylor diagram. indicate that EBKR performs best distribution. order determine accuracy generated by methods, calculate prediction standard (PSE). PSE result Portugal increases south north.