A Comparative Study of Radiation Estimation Methodologies.

Daily global radiation data are required in plant growth simulation models for predicting photosynthesis rates, canopy temperatures and evapotranspiration rates (Acock et al., 1984), but these are rarely available from routine weather stations (Supit, 1994). When radiation records are not available, these can be estimated by predictive methodologies. These make use of satellite remote sensing data (Gruter et al., 1986), of spatial interpolation techniques (WMO, 1981) and of correlations found between incoming irradiance and other commonly measured weather variables (Schulze, 1976 and Goldberg et al., 1979). The latter methodologies range from simple empirical formula to extremely complex numerical methods, depending on the available input data. Parameters used as inputs in the various methods include sunshine duration, air temperature, cloudiness, rainfall, relative humidity, precipitable water content, and composition, concentration and size distribution of atmospheric particles. These parameters are used individually or in combination. In this study some of the available procedures to calculate daily global solar radiation (Rg) from commonly measured daily meteorological variables are compared using measured solar radiation data for 18 EU locations. Estimated values are compared with actual measurements and resulting differences are evaluated statistically. Subsequently, values of solar radiation observed and calculated with the different methods are used as input in two plant growth simulation models (grapevine and wheat) and calculated yields are compared.