Mapping, Dynamics, and Future Change Analysis of Sundarbans delta using Cellular Automata and Artificial Neural Network Modeling

The harmful effects of anthropogenic activities are felt by low-lying deltas, which have grown significantly during the past century. Due to its ecological and economic significance, the Sundarbans Delta is a crucial tidal-influenced region for Bangladesh and India. The study used a Random Forest machine learning technique to assess the land use and land cover (LULC) changes of the low-lying Sundarbans delta, and it used a Cellular Automata approach combined with an Artificial Neural Network (CA-ANN) to model LULC. The accuracy of both user and producer ranges from 100-85% for the Mangrove and 100-79% for the surrounding waterbodies. Results show that over the decades, the thick, dense mangrove coverage of the Sundarbans has receded while the width of the intricate network of channels has increased, and the spatial coverage of water bodies has risen. The spatial vegetation index shows that the average minimum Normalized Difference Vegetation Index (NDVI) for all months from 1990 to 2020 is not consistent, varying from -0.72 to 0.85. The month of November has the lowest negative NDVI (-0.716), while June has the least low negative NDVI (-0.48). This negative NDVI primarily indicates the extent and severity of flooding in the Sundarbans Delta during the monsoon season. In 1970, the mangrove coverage was at 65%, but by 2020 it had dropped to 59%. In the period between 1990 and 2015, the coverage of channels in the Sundarbans increased from 33% to 43%. The study also showed that built-up and barren land areas were slowly increasing. Water bodies were encroaching on the mangrove forest, which could be due to rising sea levels or deforestation. The analysis revealed that mangroves in the Sundarbans Delta have decreased from 65% to 59% of the total area of the delta during the study period. Temporary human settlements were also visible in the area. The most significant changes were seen in the southern part of the delta, which is more vulnerable to sea level rise. The water bodies fluctuated across the mangrove forests because many rivers and channels crisscross the Sundarbans delta. Future simulation for the years 2030-2050 revealed a decline in mangrove coverage from about 63.5% to 56.7%, while the total area of water bodies will increase from 33.9% to 39.8%. By 2080, mangrove coverage would decrease by about 269 km(2), the waterbodies would increase by 159 km(2), built-up areas would increase by similar to 3.89 km(2), and barren areas or beaches would become 371 km2. For the simulation of 2100, the mangroves would decrease by similar to 48%, the waterbodies would increase to 45%, and the built-up and barren areas would reach up to 117 and 517 km(2). The study also reported an increase in sea level in the Sundarbans by an average of +5.82, mm/year at different tide gauge stations. The findings of this study will assist the Bangladesh and Indian governments and policymakers in developing optimal land use plans and implementing sustainable mangrove forest management and conservation plans for the region.