General models for estimating daily global solar radiation for different solar radiation zones in mainland China

• Five solar radiation (Rs) zones were identified by long-term mean daily Rs.
• Temperature- and sunshine-based models for Rs estimation were given for each site.
• General models were obtained using geographical factors and site-specific equations.
• Results showed that the general model for each zone had an acceptable accuracy.

Empirical models, proposed to estimate solar radiation (Rs) in various areas, were site-specific in essence. However, it is questionable when they are applied to other stations where there is no record of Rs. This study aimed to develop general models to estimate daily Rs for different solar radiation zones in mainland China. Daily weather data including Rs, sunshine duration, relative humidity, maximum and minimum temperatures were collected and analyzed from 83 stations. Two types of simple empirical equations, namely, temperature- and sunshine-based models, were obtained for each site. Five Rs zones were determined by k-means clustering algorithm based on long-term mean daily Rs. For each zone, the general model for Rs estimation was developed based on geographical factors (latitude, longitude and altitude) and site-specific models. Coefficient of residual mass (CRM), mean bias error (MBE), mean percentage error (MPE), root mean square error (RMSE) and percent root mean square error (%RMSE) were used to investigate the model performance. The comparative results between measured and estimated daily Rs showed that the general models had an acceptable accuracy. It is believed that the general models developed in this work can be reliable and applicable for the locations without available Rs data in mainland China.