Applicability of Different Interpolation Methods in the Estimation of Potential Evapotranspiration

Abstract

One of the important parameters for the study of climate change is potential evapotranspiration (PET). Accurate estimation of PET is of great importance in the planning and management of agricultural activities and water resources in a basin, in rainfall-runoff models, in determining the water budget, and in combating climate change impacts. Various empirical equations and artificial intelligence models have been widely used to estimate the amount of PET. In this study, the usability of spatial interpolation methods in estimating the monthly PET amount was investigated. For this purpose, firstly, potential evapotranspiration (PET) values were calculated by the Oudin method using the temperature data of 11 meteorological stations located in the Konya Closed Basin between 1971 and 2019. Using the PET values obtained with the Oudin method, the success of spatial interpolation methods in PET estimation was evaluated. In the study where 8 different interpolation methods were used, PET data from 8 of the 11 stations in the basin were used in the areal distribution while 3 stations were used to test the success of the interpolation methods. The PET estimation success of the interpolation methods including geostatistical and deterministic methods was tested with root mean square error (RMSE), mean absolute error (MAE), coefficient of determination (R²), and Nash-Sutcliffe efficiency coefficient (NSE) performance criteria. As a result of the study, it was determined that the NSE values of the most successful interpolation methods for PET estimation for the period 1971 2019 at the test stations were above 0.996 and showed very good success.