Application of a novel remote sensing ecological index (RSEI) based on geographically weighted principal component...

In evaluating the integrated remote sensing-based ecological index (RSEI_PCA), principal component analysis (PCA) has been extensively utilized. However, the conventional PCA-based RSEI (RSEI_PCA) cannot accurately evaluate component indicators' spatially shifting relative significance. This study presented a novel RSEI evaluation strategy based on geographically weighted principal component analysis (RSEI_GWPCA) to address this deficiency. Second, compared to the classic RSEI_PCA, RSEI_GWPCA was tested at English Bazar and surrounding areas using two-fold validation. In this regard, the Jaccard test from a different setting and correlation analysis were utilized to examine the geographical distribution of RSEI derived by PCA and GWPCA. The validation output revealed better effectiveness of GWPCA over PCA in assessing the RSEI. The findings revealed that (i) in RSEI assessment, the spatial heterogeneity of the dataset helped to formulate individual weights by GWPCA that was not performed by PCA; and (ii) the areas having higher RSEI were primarily located around the Chatra wetland of this study area, and the areas with lower RSEI were located mainly in the industrial part. It has been concluded that RSEI_GWPCA is a helpful approach in the RSEI evaluating for the regional and local scale like English bazaar city and its neighbourhood.