Adapting estimation methods of daily solar radiation for crop modelling applications in Canada

Abstract

Estimating daily solar radiation from common meteorological variables plays an important role in agricultural applications, such as driving crop growth models. Relatively simple and accurate estimation methods, which use only daily air temperature together with precipitation, are often required. Based on all available solar radiation data across Canada, the most common and representative solar radiation models were evaluated. All estimation models provided more accurate estimates, in terms of all performance statistics used in this study, than those extracted directly from a high-resolution global dataset of meteorological forcings for land surface modelling. The DS model adapted from one originally developed for the Canadian Prairies performed better than other representative models for all stations. The DS model was then adapted for regional use in southern Canada, mostly the major agricultural regions. We compared simulated crop yields using the CSM–CERES–wheat and CSM–CROPGRO–canola models driven by observed and estimated daily solar radiation data, and we found a difference of approximately 5% for spring wheat (Triticum aestivum L.) and 12% for canola (Brassica napus L.). Based on the results for two locations under different climate regimes with relatively long records (45 and 40 yr, respectively) of solar radiation data, the models using daily temperature range and precipitation were found to be robust for daily solar radiation estimation for the entire time periods of the data records.