Solar radiation zoning and daily global radiation models for regions with only surface meteorological measurements in China

Abstract Daily global solar radiation is an indispensable variable in developing typical meteorological year and its data are important for constructing energy-efficient buildings. Solar radiation zoning and empirical models are essential for estimating the daily global solar radiation in regions that only have surface meteorological measurements. However, empirical models have become increasingly complicated without significant improvements in their accuracy. Thus, developing elaborate solar radiation zoning with minimal inputs and high accuracy daily global solar radiation models are key issues. In the present study, we develop an optimal criterion for solar radiation zoning by k-means clustering based on data obtained between 1970 and 2017 from 112 solar radiation measuring stations in China. According to the optimal criterion, solar radiation zoning is conducted based on data acquired by 1042 surface meteorological stations between 1970 and 2017. Zone models are established with linear relationships between the clearness index and relative sunshine duration based on data in each zone. Cross-validation is conducted to evaluate the performance of models by comparing individual models based on single station. The results show that the relative sunshine duration is the optimal criterion for solar radiation zoning and nine zones are identified. The zone models obtain good accuracy. Solar radiation zoning with zone models established in present study can be reliable and generalized to other regions with available sunshine duration data.