Analysis of landcover change in southwest Bengal delta due to floods by NDVI, NDWI and K-means cluster with landsat multi-spectral surface reflectance satellite data

Abstract

Floods in the southwest coastal area of Bangladesh are very much regular. Tidal floods are coming at every high tidal period twice in a day. The landcovers of the study area are mainly water bodies or wetlands – for saltwater shrimp and crab cultivation, and vegetation – for agricultural activities and the Sundarbans mangrove forest. There are significant changes and alteration in landcovers due to floods events in May 2009. Seasonal tidal floods are taking place at the wet season in the study area because of the damage of Polder-32 at several points, which had been breached during cyclone Aila and causing a prolonged inundation period. After cyclone Aila, it takes 4 years till 2013 November to recover the land from flood water. Due to cyclone Aila, a 43.7 sq. km flooded surface emerged and 43.6 sq. km vegetation covers disappear in 2009 November. These significant landcover changes are studied well with two spectral indexes –NDVI (Normalized Difference Vegetation Index) and NDWI (Normalized Difference Water Index), and with a K-Means Cluster unsupervised and supervised land classification. The study found NDVI and NDWI are prominent to identify vegetation and water covers considering their individual constrain, along with the validation by K-means clustering unsupervised and supervised land classifications. Supervised K-means clustering is used to estimate the land cover areas. The NDVI and NDWI were used to identify the vegetation and water covers, and K-means cluster was used to estimate the vegetation and water covers.