In mountainous areas, topographic shading effects exert significant influence on the spatial distribution of solar irradiance on the land surface, which should be considered in estimating and modeling physical, biological, and hydrological processes. Various algorithms have been proposed for quantitative assessment of the impact of topography on the arriving solar radiation on slopes in mountainous area based on digital elevation model (DEM). However, different algorithms tend to produce different results even based on the same DEM. In this study, two algorithms for determining the maximum horizontal viewing angle of terrain obstruction were compared in a catchment in the Qilian Mountains, Northwest China, which were identical except for the way in interpreting DEM. Two algorithms for calculating sky view factor were also compared. Then, different algorithm combinations were compared and their sensitivities to the horizontal resolution of DEM were investigated. Results showed that the differences in outputs of different algorithms and combinations were significant and cannot be ignored in estimating the spatial distribution of solar irradiance in mountainous areas. A simple ground validation concluded that the algorithm assuming a linear variation in elevation was a little better than that assuming a stepped variation in elevation in and among neighboring DEM cells. For calculating diffuse solar radiation, the algorithm based on a sloped coordinate system was theoretically better than that based on a horizontal coordinate system. It is worth noting that even though the estimated annual mean of solar irradiance over the study area was almost the same among some algorithm combinations, the difference in spatial distribution was significant, which may lead to consequent difference in assessing or modeling some other environmental variables relevant to surface energy exchange, e.g., evaporation, snowmelt, and ground temperature. However, the analysis in this study was mainly theoretical, and further ground-based validations should be made in future.
Read full abstract