Identification of mangroves and coral reefs in the southern Gulf of Kachchh using principal component analysis and geo-spatial techniques: A case study

Abstract

The major objectives of this study were to extract the coral reef zones and mangroves using IRS P6 LISS IV data by employing the most appropriate technique, suited for the particular feature. Firstly, a few sample locations of corals were superimposed on the imagery to note the behaviour of the coral pixels was observed, the range of the reflectance values was noted in all bands. Since these values were not normally distributed, it was considered best to use the modal value of the dataset to truly represent the reflectance of coral pixels instead of using the entire range. The final range was taken as 1 from the mode, which was used for the extraction of coral pixels by band mathematics in ENVI 5.0. The findings suggested that corals covered 14.01 km sup(2) of the study area. For extracting the mangroves, the technique of Principal Component Analysis (PCA) was applied. This technique was found to give better results by not only identifying the mangrove swamps but also to differentiate between the dense and sparse mangroves. The greatest variance in the dataset could be identified by this method, thereby making it easier to outline the variations existing within the mangroves. Further classification (by supervised method) of the principal components was required to identify Dense and Sparse mangroves. The final findings highlighted that out of the 89.69 km sup(2) of the study area, mangroves covered 38.35 km sup(2) (42.76 percent). Further classification of the mangroves quantified to 7.47 km sup(2) of dense mangroves and 30.88 km sup(2) of sparse mangroves showing a reasonably high prominence of mangroves in the area. The area quantifications were executed using ArcGIS software.