Evaluation of Deterministic and Geostatistical Methods for Regionalization of Seasonal Precipitation in Hamadan Province

Precipitation information plays an important role in water resources and natural resources studies in any region. With respect to the large changes in the spatial and temporal distribution of precipitation in the arid and semi- arid areas, it is important to regionalize precipitation. One of the methods for regionalization of precipitation is application of deterministic and geostatistical methods. In this research, different methods, namely, inverse distance, local polynomial, radial basis function, simple kriging and ordinary kriging were assessed using different models for regionalization of seasonal precipitation in Hamadan Province utilizing the precipitation statistics of 20 meteorological stations belonging to the Meteorological Organization and Ministry of Energy for a 10-year period (2003-2012). For assessing the performance of the examined methods the cross validation method, the assessment criteria Mean Absolute Error, Root Mean Squared Error and General Standard Deviation index were utilized. The results showed that the radial basis function method with Thin Plate Spline model for the autumn, the inverse distance weighting method for the winter and spring, and the radial basis function with Completely Regularized Spline model for the summer were the most appropriate methods for regionalization of the seasonal precipitation in Hamadan Province.