Abstract:
Mangroves are one of the most productive and biologically diverse ecosystems, but they have
gradually been declined due to anthropogenic pressure. They are highly threatened in most
developing countries, including Sri Lanka. The coastal development activities make mangrove
ecosystems seriously vulnerable. There is an increased need to assess and monitor this ecosystem in
order to help guide conservation and restoration efforts. In this study, an attempt was made to
investigate mangrove stands in Pottuvil coastal zone bio-shield. The objective of the study is to
define the image texture in discriminating mangrove species with landsat imageries in Pottuvil
coastal area in Ampara District. The graphical visual pattern is very complex in the texture analysis
with important quality characteristics of various image products. The image classification was
executed by the quality of characteristics of the image. The mangrove studies typically involve
intensive field surveys. However, geospatial technology is a cost-effective means of mapping and
monitoring mangrove bio-shield at large scale. K-mean classification method was employed to
define vegetation indices by Normalized Difference Vegetation Index (NDVI), which is an
important quality characteristic of various image products. The study revealed that the overall
accuracy assessment of the texture image classification was 5 percent more than the non-texture
image classification. Also, the analysis revealed that the extent of the mangrove vegetation has
reduced by 34.5 ha over the 10 year’s period. This approach will be important in the process of
decision-making and policy implementation in mangrove ecosystem conservation efforts.