Evaluating the accuracy of VEMAP daily weather data for application in crop simulations on a regional scale

Weather plays a critical role in eco-environmental and agricultural systems. Limited availability of meteorological records often constrains the applications of simulation models and related decision support tools. The Vegetation/Ecosystem Modeling and Analysis Project (VEMAP) provides daily weather variables on a 0.5 latitude–longitude grid across the conterminous USA. Daily weather data from the VEMAP (1961–1990) for the state of Georgia were compared with data from 52 individual ground stations of the National Weather Service Cooperative Observer Program (COOP). Additionally, simulated crop grain yields of soybean (Glycine max) were compared using the two data sources. Averaged daily maximum and minimum temperatures (Tmax and Tmin, respectively), solar radiation (SRAD), and precipitation (PPT) differed by 0.2 °C, −0.2 °C, 1.7 MJ m−2 d−1, and 0 mm, respectively. Mean absolute errors (MAEs) for Tmax, Tmin, SRAD, and PPT were 4.2 °C, 4.4 °C, 4.4 MJ m−2 d−1, and 6.1 mm, respectively, and root mean squared errors (RMSEs) for Tmax, Tmin, SRAD, and PPT were 5.5 °C, 5.9 °C, 5.8 MJ m−2 d−1, and 13.6 mm, respectively. Temperature differences were lowest during summer months. Simulations of grain yield using the two data sources were strongly correlated (r = 0.68, p < 0.01). The MAE of grain yield was 552 kg ha−1. The RMSE of grain yield was 714 kg ha−1. Hybrid analyses indicated that the variation of simulated yield was mainly associated with the differences in rainfall. The results showed that the VEMAP daily weather data were able to be adequately applied to crop growth simulation at spatial and temporal scales, especially for long-term climate change research. Overall, the VEMAP weather data appears to be a promising source for crop growth modeling concerned with scale to 0.5° coordinate grid.