Spatio-temporal variation and prediction of land use based on CA-Markov of southwestern coastal district of Bangladesh

Abstract

This study aims to explore the spatio-temporal variation of land use of Satkhira district, one of the most vulnerable coastal districts of south-western Bangladesh, to monitor environmental degradation. Satkhira district experienced a significant change in its natural setting due to natural disasters, cyclones in precise, and unplanned and unwise human interventions. This study analyzed the Landsat Level 2 images of the years 1989, 2001, and 2018 to detect the change in land use in seven Upazilas of Satkhira district. Integration of remote sensing indices and Supervised Classification was employed to classify the Landsat images and identify the change in four dominant land use classes named ‘Bare Lands’, ‘Built-Up Area’, ‘Vegetation’, and ‘Water Area’. The analysis revealed that ‘Bare Lands’ has been declined at an alarming rate between 1989 and 2001 (67.61%), and between 2001 and 2018 (77.38%). On the other hand, ‘Water Area’ has increased by 291.8% over the last 30 years, grabbing more than 510 sq. km from other land use classes. A substantial change in ‘Water Area’ occurred from 1989 to 2001 (138.13%) compared to the period of 2001–2018 (64.58%). The most significant conversion (more than 240 sq. km) has taken place between ‘Bare Lands’ and ‘Water Area’. This analysis supports the positive growth of Shrimp and Crab farming, and occupational shifting from agricultural production to Shrimp farming in Satkhira. Flourishing Shrimp farming has negatively affected the coastal region by triggering the salinity intrusion and loss of agricultural lands. Population distribution maps of 1981, 1991, 2001, and 2011 indicated that population density has been increasing in the Upper portion of Satkhira, which supports the studies of migration towards urban areas from coastal villages. There is a close interconnection between the distribution of land use types and population among the Upazilas. Land use distribution of 2030 was predicted through the CA-Markov model, which disclosed that ‘Water Area’ will continue to increase by 29.7% and ‘Vegetation’ areas will face a reduction by 16.7% between 2018 and 2030. ‘Bare Lands’ will experience a further 90.4% reduction. The coastal natural setting has already been disrupted due to natural disasters and human interventions. An integrated land use management plan should be formulated immediately to control the unplanned development soon and this study will guide to formulation of the plan.