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