Development of statistical methods for estimating hourly direct and diffuse solar radiation using public data for precise cultivation management

Abstract

　High resolution data for solar radiation are particularly useful for precise cultivation management of hilly or mountainous terrains and hourly minima for both direct and diffuse solar radiation are needed to determine global solar radiation in such terrain. Estimation methods based on statistical models for determining hourly direct and diffuse solar radiation on a horizontal surface have been developed using low-cost public data. These methods use a statistical model for atmospheric transmittance. Direct solar radiation was estimated from hourly atmospheric transmittance data using a multiple regression equation with two meteorological variables, i.e., sunshine duration and extraterrestrial solar radiation on a horizontal surface. Diffuse solar radiation was estimated using a dimensionless parameter K_ds based on direct solar radiation, whenever sunshine duration was greater than zero. When sunshine duration was zero, diffuse solar radiation was estimated using a multiple regression equation with two other meteorological variables, i.e., precipitation and normal extraterrestrial solar radiation. The root mean square errors of estimated data for both hourly direct and diffuse solar radiation were ~0.3 MJ m^-2 h^-1. Hourly direct and diffuse solar radiation estimated using the proposed method will be useful throughout Japan to facilitate better agricultural crop management. In addition, development of estimation methods for hourly direct and diffuse solar radiation using a numerical weather prediction model was also explored. It proved to be difficult to use estimated hourly data in the numerical model, but this numerical approach appears to be feasible for daily values.