Comments on “Local-Scale Variability of Solar Radiation in a Mountainous Region”

Abstract

Solar irradiation is spatially and temporally variable depending on the atmospheric composition, earth surfacial features, and the location of the point considered. In order to determine these variabilities, solar irradiation is measured at various sites scattered all over the world, all the time. Hence, time series and spatial sampling of the solar irradiation records are obtained, and the main purpose is to depict the complete spatial and temporal variation features from these records. Sometimes the records are not available at desired places and therefore it is a practical problem to develop effective procedures for the prediction of solar irradiation at times and sites where the records are missing. Solar energy system aerial features are necessary for meteorological and agrometeorological purposes. Today, satellite-based methods provide the needed information, but their interpretations are not very accurate for a locality and the results must be supported by the actual measurements within or in the vicinity of this locality. Hence, the spatial variability of the solar irradiance still remains a very significant question awaiting effective techniques for modeling. The regional variability is represented either by contour maps of measured quantities, provided that there are a sufficient number of stations rather uniformly distributed over the area, or the variability is depicted with distance. A very practical question that arises is what is the maximum between-station distance, that is, the radius of influence appropriate for a correct interpolation. Many authors cited by Tovar et al. (1995) have proposed various evaluations of interpolation. Also, different authors have addressed the question of acceptable radius of influence for solar irradiance measurements. Almost all of these studies are concentrated on the analysis of the standard deviation or correlation coefficient of differences in solar irradiance between station pairs as a function of distance. However, the coefficient of variation Cn is defined as