Evaluation of Geostatistical Interpolation Methods for Rainfall data Estimation in Libya

Abstract

Rainfall is considered a highly valuable water resource, particularly in arid and semiarid regions like Libya. Moreover, rainfall data has been shown to be fundamental input for accurate distributed hydrologic modeling. Hydrological Modeling using geostatistical interpolation technique is performed to study hydrological process which is essential for water resource management. This study aim to evaluate six geostatistical interpolation techniques such as Empirical Bayesian Kriging (EBK), Radial Basic Function (RBF), Inverse Distance Weighting (IDW), Global Polynomial Interpolation (GPI), Kernel Interpolation With Barriers (KIB), and Diffusion Interpolation With Barriers (DIW), at that point compare their performance in generating spatial distributions of monthly total rainfall data from 63 meteorological stations located across Libya over 40 years period. Different statistical accuracy measurements such as mean absolute percentage error MAPE%, efficiency factor E, and 95% confidence interval CI95% were used to determine the best geostatistical interpolation methods. Results demonstrate that, RBF, and IDW geostatistical interpolation methods, are both better and more reliable than other geostatistical interpolation methods used in this study. Moreover, they provide a maximum, minimum and average predicted values within the range of 95 CI%