Exploring ML‐Driven Insights on the Impact of Rising Soil Salinity on Sundarbans Mangrove Ecosystems and Ecological Sustainability Through Nature‐Based Solutions

Soil salinization presents a critical threat to coastal ecosystems and agricultural productivity, particularly within the Sundarbans mangrove region. This research addresses the limitations of traditional in situ and remote sensing methods by evaluating an integrated machine learning framework for regional and global soil salinity mapping. Utilizing data from the Harmonized World Soil Database alongside localized in situ salinity measurements, the methodology integrates satellite-derived time-series imagery and detailed soil characteristics. Multiple ensemble models, including Random Forest, Artificial Neural Networks, and a hybrid Least Absolute Shrinkage and Selection Operator-Genetic Algorithm-Backpropagation Neural Network (LASSO-GA-BPNN) model, are implemented to capture complex seasonal variations. The predictive framework achieves high accuracy across diverse agricultural and industrial

zones, facilitating targeted land-use planning and ecological restoration strategies. By modeling the physiological stress of tropical mangroves under hypersaline conditions, this research provides a robust computational approach to support nature-based solutions and sustainable coastal zone management. The integration of multi-temporal remote sensing indices with advanced optimization algorithms enhances the spatial resolution and reliability of salinity hazard mapping, offering critical decision-support tools for environmental engineers.

Message: Hi FE, I need help with “Exploring ML‐Driven Insights on the Impact of Rising Soil Salinity on Sundarbans Mangrove Ecosystems and Ecological Sustainability Through Nature‐Based Solutions” in “Environmental Engineering”