A Dimensionality Reduction-Based Approach to Select a Suitable Interpolator for the Mapping of Solar Irradiation Across Pakistan

The selection of a suitable gridding method for the interpolation of the input dataset has always been a big challenge. In this study, principal component analysis (PCA) was employed to select a suitable interpolator for the mapping of solar irradiation. For this purpose, sunshine duration data obtained from 50 meteorological stations was used to predict solar irradiation based on the Angstrom–Prescott model. A total of nine gridding methods were utilized to interpolate monthly mean daily global solar irradiation (GHI). Results showed that the GHI values across the Pakistan ranged between 4.50 to 6.00 kWh/m2/day. In the SPSS software package (version 23), PCA was performed and  Kriging was considered as the most suitable interpolation method among all interpolators used in this analysis. To validate PCA results, interpolated GHI maps were compared with the solar atlas developed by the World Bank’s Energy Sector Management Assistance Program (ESMAP). The estimated value of performance indicators showed that the Kriging had the comparatively lowest value of mean absolute error (MAE) and root-mean-square error (RMSE). The outcomes of this study will be useful in the analysis and planning of solar energy system at the local and regional level.