Spatiotemporal changing pattern of land use and land cover of Sangu- Matamuhuri watershed in the South-Eastern part of Bangladesh

Abstract

This study employed integrated remote sensing (RS) and geographic information system (GIS) approaches to characterize the spatiotemporal changing patterns of LULC in the upper part of Sangu-Matamuhuri Watershed (SMW) from 1988 to 2019 utilizing Landsat imagery of the years 1988, 1999, 2009 and 2019. Geo-spatial object-based supervised classification focusing on five major classes namely forest, agriculture, shrubland, settlement, and water body were applied. The overall classification accuracy was found 92 ~ 94% with a corresponding Kappa coefficient range of 0.90 ~ 0.93. The forest cover showed a major variation with the interval, decreasing from 234634 ha (76%) in 1988 to 168476 ha (54%) in 2019 with an annual loss of 2134.13 ha/year. Conversion results obtained from temporal overlay operation showed that forest patches converted to shrubland due to cutting down trees and forest burning and thereafter this shrubland cleared or re-burned for shifting cultivation which significantly increased agricultural land from 21787 ha (7%) in 1988 to 78704 ha (25%) in 2019 at the rate of 1836.03 ha/year. The area of other land use types including shrubland and settlement also increased. But the area of water body, the only source of drinking water to the indigenous community and irrigation water in the dry season, was found decreased with an average annual rate of 29.84 ha/year which indicate the drying of water body. The observed changing trend of LULC of the studied watershed could have a significant negative impact on the watershed environment as there would be more water shortage in the near future during the dry season and excess water in the rainy season. The findings of this study emphasizes important policy implications for sustainable watershed management in hilly regions of the tropics.