Solar radiation modelling using ANNs for different climates in China

Artificial neural networks (ANNs) were used to develop prediction models for daily global solar radiation using measured sunshine duration for 40 cities covering nine major thermal climatic zones and sub-zones in China. Coefficients of determination (R²) for all the 40 cities and nine climatic zones/sub-zones are 0.82 or higher, indicating reasonably strong correlation between daily solar radiation and the corresponding sunshine hours. Mean bias error (MBE) varies from -3.3 MJ/m² in Ruoqiang (cold climates) to 2.19 MJ/m² in Anyang (cold climates). Root mean square error (RMSE) ranges from 1.4 MJ/m² in Altay (severe cold climates) to 4.01 MJ/m² in Ruoqiang. The three principal statistics (i.e., R², MBE and RMSE) of the climatic zone/sub-zone ANN models are very close to the corresponding zone/sub-zone averages of the individual city ANN models, suggesting that climatic zone ANN models could be used to estimate global solar radiation for locations within the respective zones/sub-zones where only measured sunshine duration data are available. © 2007 Elsevier Ltd. All rights reserved.