Development and comparison of parametric models to predict global solar radiation: a case study for the southern region of Saudi Arabia

Global and diffuse solar radiation data (measured) on a flat surface and the bright sunshine hours quantity, relative humidity, wind speed, dew point temperature, and total sky coverage for the southern city Najran (18.0 °N 45.4 °E), Saudi Arabia, for 15 Years (2003–2017) are analyzed. The data are collected from the weather station and from “Climate.OneBuilding.Org”. This research aimed to develop, review, and compare performance for the models (empirical) developed to predict global solar radiation. Overall, 121 empirical models are built to define in three categories [Sunshine Duration Fraction (SDF), Modified Sunshine Duration Fraction (MSDF), Non-Sunshine Duration (NSD)] with all possible metrological parameters available for predicting Global Solar Radiation (GSR) that correlated the clearness index to the relative sunshine time and other metrological parameters. Using DataFit® software, the models were created by fitting a model form and extracting the coefficients from all the collected data. A model comparison has been performed, and performance evaluation has been established using some well-known statistical indicators. Therefore, the Global Performance Indicator (GPI) is to be measured using scaled value statistical indicators. The model’s GPI lays within the range of − 6.35 to 0.1585, which demonstrates the top achieving prototype represented by the greatest significance. The developed generalized models (in order Model-63, Model-64, Model-69, Model-57, and Model-22) were included among the top-ranking models as prototypes of high level.