Accuracy of eight probability distribution functions for modeling wind speed data in Djibouti

Abstract

In this paper, the regional assessment on the performance of eight different probability distributions functions is investigated and compared for estimating wind speed distributions in Republic of Djibouti; the stations are located in the rural areas which are Ghoubet and Bada Wein, and urban areas which are University of Djibouti and International Airport of Djibouti. To achieve this aim, the statistical test for ranking the selected probability distributions functions, is evaluated based on the coefficient of determination, the root mean square error and the index of agreement. It has been shown from the statistical results that Weibull, Rayleigh and Gamma distributions can generally considered as the appropriate distributions and are generally provide the best fit for all stations; however Nakagami distribution gave the best results for Ghoubet rural station compared to the others used distributions.