Zheng Zhang

Document Type

Thesis - University Access Only

Award Date


Degree Name

Master of Science (MS)

Department / School



Solar radiation received at the Earth's surface is the ultimate force that drives the earth-atmosphere system dynamics. Accurate computation of surface solar radiation (SSR) is needed for various environmental studies including surface energy budget, weather processes, vegetation distribution, and precision farming. One of the most important controls on the SSR is the characteristics of terrain, which is often represented by the Digital Elevation Models (DEMs). The resolution and accuracy of DEM data, and the method of the DEM data resampling directly affect the computation of the SSR. This research analyzed the sensitivity of computed SSR to the change of DEM resolution and derived slope and aspect due to resampling by different methods. The factors that govern computation of SSR associated with DEMs resampling across a variety of spatial resolution were examined. The following conclusions were drawn based on results obtained from this research: l) the magnitude of error in computed SSR due to DEM resampling and related change in slope and aspect is rather significant; 2) regardless the resampling method, the computed SSR is most sensitive to change of slope and aspect due to DEM resampling; 3) the rate of change in computed SSR as DEM spatial resolution varies is not linear; 4) the spatial variability of computed SSR is gradually reduced as the DEM grid size becomes coarser regardless of DEM resampling methods; and 5) the computed SSR obtained from DEMs based on different resampling methods show both differences and similarities. Generally speaking, break-line resampling method is superior to the bilinear resampling method in computed SSR for pixels with steep and north facing slope when DEM resampled from the full resolution (30 m) to 100 m and 250 m. As the full resolution DEM is resampled to 500 and 1000 m, there is little difference in computed SSR between the two methods. The conclusions derived from this research have implications for application studies in environmental research. For computation of SSR over small spatial (local) scale useful for micro-climate, agricultural or forest management and urban studies, finer resolution DEM and derived slope/aspect data are highly desirable, and the break-line resampling method is recommended. For computation of SSR over large spatial scale (regional and global scale) useful for regional weather forecast, and global climate or hydrological modeling, the selection of DEM resampling methods (bilinear vs. breakline) is not critical. However, one must be aware of the significant error introduced by DEM data resampling. For future research, the sensitivity of SSR to the DEM representation at other times or seasons, along with the impacts of atmosphere and relief shadow, should be investigated.

Library of Congress Subject Headings

Digital mapping

Solar radiation



Number of Pages



South Dakota State University