Estimation of maximum possible daily global solar radiation

Abstract

The estimation of maximum possible daily global solar radiation is important in many applied sciences. This study develops and evaluates climatic extreme based modifications of two single-atmospheric-layer broadband solar shortwave irradiance models for the purpose of estimating a dynamic upper boundary for global solar radiation at any given location. Climatic component models were developed for five rural locations in the central United States: Ames, IA, Bismarck, ND, Columbia, MO, Dodge City, KS, and Wooster, OH. Each site had long-term (30 or 31 years) records of daily global solar radiation data available. Aerosol optical depth, precipitable water, and surface albedo were the input variables. Data for the first two inputs were obtained from the SAMSON database (Solar and Meteorological Observation Network). Albedo interpolating curves were estimated from the predecessor of SAMSON. For each site, precipitable water and aerosol optical depth daily data were used to develop annual trends in the climatic lows and normals for each variable. The normals were based on median daily values. Nonlinear generalized least squares regression analyses were used to develop the interpolating curves. To evaluate the global radiation models, maximum daily global solar radiation values were selected for each day in the year from the entire period of record for each site. In either broadband model, the use of the climatic normals in each input variable either interpolated or underestimated the selected radiometer data. The use of the climatic lows, however, yielded a reasonable upper boundary for the selected data. The result may be partially due to the fact that the climatic minima curves for each turbidity variable were generally significantly different throughout the year from the climatic normals ( P < 0.05). The global solar radiation models were most sensitive to aerosol optical depth. Although it is more sensitive to input variation and it is somewhat less conservative, the simpler of the two broadband models is adequate for most applications. While results are site-specific, the methodology is general and provides a climatic-based definition for maximum possible daily global solar radiation.