Paper Title
Modeling Health Vulnerability of Coral Reefs of Gulf of Mannar by Machine Learning using Statistical Program ‘R’ and Arc Gis Module

Abstract
Tissue degradation and mediated mortality turn into a major threat to coral reef systems around the world. Detailed knowledge on interactions of prime biological factors that mediates tissue loss and mortality is of paramount importance in understanding the prevailing reef health scenario and to trial management actions. In the present study, a series of benthic surveys were conducted in selected islands of Gulf of Mannar and gelocated information on ecological stressors was gathered. The information obtained was fed to the statistical program R to estimate the hierarchy of interaction of environmental variables to coral mortality. The observations from the hierarchical analysis were used to derive vulnerability maps using machine learning by weighted overlay analysis module in Arc Gis 10.1. Vulnerability maps derived can be used as a baseline observation to identify areas of very high vulnerability and specific stressor prevalent in those sites. This will be helpful in defining stressor and site-specific management plans. Similar machine learning models can be executed to other islands of GOM to develop a coral health vulnerability atlas. Keywords - Coral reefs, Statistical modeling, Vulnerability model, Hierarchical analysis, Mortality