MONITORING FOREST DIEBACK SEVERITY AND PLANT COMMUNITIES DIVERSITY USING MULTI-TEMPORAL SATELLITE DATA IN THE HORTON PLAIN NATIONAL PARK, SRI LANKA

摘要

Forest ecosystems depend largely on surrounding environmental and physical conditions and need long term adaptation. They have a major role in maintaining biological diversity and balancing climatic factors. With emerging trends in climate change and increasing anthropogenic impacts on our environment, forest sustainability has become a major concern worldwide. Conservation of biological diversity operates at different biological and ecological levels and the conservation objectives can be compromised by environmental as well as anthropogenic factors. Identification of the extents of forest plant communities and their habitats, and monitoring changes will help conservation communities. Therefore, an approach is needed to account for changes in forest health and identification of vegetation communities. One area that needs quality information on forest condition is the Horton Plains located in the central hill country of Sri Lanka. Previous studies report that forest diebacks have lead to mortality of some vegetation strands and biodiversity reduction in the study area. The monitoring studies based on satellite data and geographic information systems provide an early warning system of vegetation stresses through the use of a set of indicators. The aim of this study was to determine the vegetation structure of the study area and analyze the spatial patterns, dieback severity and factors associated with it. The data was analyzed using image processing programs supported by GIS techniques. In this study, a relationship was established between diebacks and vegetation indices (e.g. Normalized Difference Vegetation Index (NDVI)). The field investigations were carried out to identify vegetation communities and dieback sites. This information and maps will be very valuable for forest managers and other decision makers at different levels providing early warning on potential risk areas.