Validation of Geospatial model for Biodiversity Characterization at Landscape Level—a study in Andaman & Nicobar Islands, India

Abstract

The present paper deals with Biodiversity Characterization at Landscape Level, preparing thematic maps for spatial representation of landscape parameters and then validation of the geospatial model using a three-tier approach, viz., landscape analysis, field data extrapolation and geospatial analysis, for two groups of islands in Indian subcontinent. On-screen visual interpretation approach based on IRS LISS-III and Landsat TM data sets was used to delineate the vegetation communities. Field survey attributes on species diversity, economical/medicinal uses and endemism were suitably linked with satellite image-derived landscape attributes in a geospatial domain. The landscape parameters, viz., fragmentation and disturbance index, clearly indicate that Nicobar Islands are less disturbed compared to Andaman. The phytosociological data of different vegetation types collected from 544 stratified sample points was recorded with ground coordinates. The plant communities of Andaman show high plant diversity in terms of number of species (523 species) and more heterogeneity compared to those of Nicobar (347 species). Biological richness was estimated as a function of six biodiversity attributes (i.e., spatial, phytosociological, social, physical, economical and ecological) to stratify forest vegetation in two phytogeographically distinct groups of tropical islands, viz., Andaman and Nicobar, using a customized software, SPLAM. Validation of ecological indices had always been a major challenge for ecologist. Here, we discuss a validation technique using field-driven species regression equations for fragmentation (FR), disturbance (DI) and biological richness (BR) indices and propose improvements in the biological richness methodology. The research question of establishment of linkages between expressions of diversity at different scales is sought and thereby validating the existing model outputs has been proposed. The sensitivity of in-field measurements with landscape parameters has been explicitly analyzed. During validation, the species richness, diversity and mean basal area have shown slow decrease with sudden rise in intermediate levels followed by a steep decline at increasing levels of fragmentation and disturbance. The empirically derived regression coefficient has significant correspondence ( p < 0.05) in terms of community response in both island's groups. The vegetation communities in the forested landscape possess significant linkages with landscape scale processes, viz., FR and DI, as evident by the degree of correlation. The heterogeneity measured as diversity in habit group shows community dimensions in different spatial zones. The generation of baseline data on community ecology allows prioritization of planning process and evolves specific management practices. It may be concluded that the improved biodiversity assessment model and the designed validation techniques have enhanced the reliability of model thus widened the scope for its application in the rapid biodiversity assessment.