A new method of estimating global radiation was developed and tested with a large North-American database. The new method was compared to the widely known Donatelli-Campbell method. Based on the error indexes, it was shown that the new method provides more accurate estimation than the reference method. This good performance mainly stems in the favourable characteristics of the function describing daily heat fluctuation and daily radiation sum. By taking precipitation values into consideration, the accuracy of estimations could be further increased. The estimations could be extended to locations where no radiation measurements are performed. The new method can be effectively used to estimate the input data of crop production simulation models and the uncertainty of the model data resulting from radiation estimation, while the models can also be used in locations where there are no radiation data available.

Keywords: global radiation, crop production model, data supplementation