Intercomparison of Solar Maps Derived from an Ensemble-ANN Model and a Semi-empirical Model for a Desert Environment

Abstract

Increasing interest is devoted to the development of solar powered plants in the countries of the Arabian Peninsula. However, the unique climatological conditions of the region, characterized by high dust concentrations, high humidity and modest cloud coverage, cause a challenge in solar radiation estimations. This study presents and compares yearly direct normal, diffuse horizontal and global horizontal irradiation maps for the years 2008 to 2010 over the United Arab Emirates using an ensemble artificial neural network model and a cloudiness index corrected semi-empirical model, both of which were derived from satellite images. A validation against ground measurements from one station is also performed.