Biodiversity of benthic habitats is among the highest of all ecological communities. This study was conducted to model benthic habitat biodiversity indices using a remote sensing approach in optically shallow waters in Karimunjawa Islands-Indonesia. These islands have a wide variety of benthic environments. Two multispectral imagers, namely Sentinel-2A and Landsat 8 OLI, were used. A series of statistical tests were applied in the empirical modeling using the pixel values of both images with in situ Shannon index (H), Simpson index (D), and Shannon's Equitability (EH) calculations. The modeling inputs were sunglint-corrected bands, water column-corrected bands, PCA-transformed bands, MNF bands, and occurrence texture bands. The results indicate that multispectral remote sensing images can be used to map benthic habitat biodiversity indices. However, the difference between the concepts of H, D, and EH calculations and the reflectance value recorded by the sensor remove the possibility of obtaining higher accuracy. H, D, and EH maps derived from Sentinel-2A had varying levels of accuracy, namely 46.8%, 59.1%, and 54.5%, respectively, while Landsat 8 OLI produced these three maps with 45.81%, 57.34%, and 53.81% accuracy.
Read full abstract