Spatio-temporal modeling of global solar radiation dynamics as a function of sunshine duration for Turkey

Abstract

A total of 18 empirical models in linear, quadratic, cubic, logarithmic, exponential and hybrid forms using only sunshine hours, latitude, and altitude were compared to estimate monthly average daily global solar radiation on a horizontal surface for 159 weather stations in Turkey. Gopinathan [Gopinathan, K.K., 1988a. A simple method for predicting global solar radiation on a horizontal surface. Sol. Wind Technol. 5, 581–583] and Jin et al. [Jin, Z., Yezheng, W., Gang, Y., 2005. General formula for estimation of monthly average daily global solar radiation in China. Energ. Convers. Manage. 46, 257–268] models provided the most robust performance based on the error statistics of the coefficient of determination ( R adj. 2 ), the mean percentage error (MPE), the mean bias error (MBE), and the root mean square error (RMSE). A comparison of observed and predicted monthly average daily global solar radiation for a total of 35 sites equally distributed across the seven major climate zones of Turkey resulted in the highest R 2 values of 95.1–94.9%, based on Gopinathan [Gopinathan, K.K., 1988a. A simple method for predicting global solar radiation on a horizontal surface. Sol. Wind Technol. 5, 581–583] and Jin et al. [Jin, Z., Yezheng, W., Gang, Y., 2005. General formula for estimation of monthly average daily global solar radiation in China. Energ. Convers. Manage. 46, 257–268] models, respectively. Spatial variability in mean annual global solar radiation observed was mapped on a 500 m × 500 m grid using universal kriging with the first order trend removal. Cross-validation statistics indicated that universal kriging was sufficiently reliable in prediction of spatial variability in global radiation across the country, particularly, in areas with a lack of global solar radiation data.