Correlation between ground-recorded and GIS modelled solar radiation: Development of regression equations for Australian conditions

Abstract

Solar radiation is a critical driver of many processes on the earth's surface. In fields such as mountain climatology, plant ecology and environmental studies, the estimation of the thermic action of the relief and the calculation of the radiative budget of the surfaces constitute an important prerequisite for any modelling work. GIS-based solar radiation models allow us to take into account, on a fine scale, the complexity of the interactions between the incident solar radiation (direct as well as diffuse) and the local topography. Such models provide rapid, cost-efficient and accurate estimations of radiation over large areas, while considering slope and aspect gradients, and shadowing effects. While GIS-based models predict the potential direct (and in some cases diffuse) radiation and are able to predict quite accurately the relative spatial variation, for a number of applications the actual and not the potential solar radiation is needed. This paper discusses the correlation between the modelled radiation and the actual recorded data and presents regression equations that can be used to convert modelled radiation for use in other environmental models that utilize field recorded solar radiation data. The impacts of different climatic conditions as well as seasons on the regression equations are discussed. Analysis of daily meteorological records spanning eight years for a number of stations around Australia show that the correlations are quite high (r² = 0.95) when the data is pooled and higher still (r² = 0.98) when seasonal effects are taken into consideration.