Error assessment of grid-based direct solar radiation models

This research assessed two grid-based direct solar radiation models, ESRI’s Solar Analyst (SA) and Kumar’s model (KM), using artificial surfaces. Mathematically derived radiation on the surfaces was compared with grid-based model results. While both models showed good consistency with theoretical derivations, they both underestimated direct radiation at daily, most seasonal, and annual scales. KM performed better than SA at all the scales except at annual scale and in the summer. Horizon angle calculation and numerical integration are the common error sources in both models. Interpolation in horizontal angles and the use of a sky size parameter are the additional error sources in SA. Larger errors were found in SA when the sky size parameter and modeling time interval were not compatible. Overall, KM is a better choice for direct solar radiation modeling as it is more accurate and computationally efficient, easier to understand, and needs fewer parameters.