Every study on this page began with a place — a hillslope in Khagrachari, an urban lake in Dhaka, a reef flat on Saint Martin's Island. Explore the map to see where the fieldwork, remote sensing, and modelling took place, or filter by category below.

A machine learning–based approach to map landslide susceptibility in Khagrachari district, integrating 15 conditioning factors and comparing RF, BRT, and KNN. The boosted regression trees model achieved an AUC of 0.95 and 82% overall accuracy, with maximum rainfall and elevation as the dominant drivers.
DOI: 10.1007/s11356-024-34949-5
Dynamic vegetation-cover trends across the Chattogram division using NDVI time series and physio-climatic drivers, revealing greening patterns interspersed with localized browning tied to climate and land-use change.
DOI: 10.1007/s10668-024-05505-5
Machine learning combined with the Analytical Hierarchy Process to evaluate rainwater harvesting potential across hilly Bangladesh, identifying optimal zones for sustainable water resource management.
DOI: 10.1016/j.jaesx.2024.100189
Multi-year remote sensing and statistical analysis quantifying how urban lakes reduce local temperatures in Dhaka, mitigating the urban heat island effect.
DOI: 10.1016/j.hydres.2025.01.001
Trends in ecological environmental quality across Dhaka assessed through the Remote Sensing based Ecological Index (RSEI).
DOI: 10.3390/land14061258
Explainable AI methods uncover the key drivers of landslide occurrence in northeastern Bangladesh, adding interpretability to susceptibility modelling.
DOI: 10.1016/j.icee.2026.100007
An integrated bivariate–multivariate GIS approach combining AHP with frequency ratio models to delineate high-risk landslide zones across the Himalayan region.
DOI: 10.1007/978-981-97-4680-4_11
A data-driven, remote-sensing and machine-learning approach to assessing coastal vulnerability along Bangladesh's coastline, presented internationally in New Delhi.
A new ecological index built from hyperspectral imagery, thermal data, and radar backscatter to assess mangrove health, resilience, and carbon sequestration in the Sundarbans.
LiDAR-driven machine learning models of forest canopy height for biomass estimation and carbon stock assessment in the hill forests.

A hybrid machine learning framework integrating climatic and geospatial parameters to predict ecotourism-suitable zones on Saint Martin's Island. Currently under journal review.

Thesis project examining how land surface albedo variation affects land surface temperature, contributing to understanding of the urban heat island effect.

Remote sensing–based mapping of environmental and infrastructural indicators to assess livability degradation and inform urban improvement policy in Dhaka.

Reviewed a manuscript on groundwater quality assessment, multivariate statistical analysis, and human health risk in Northwest Bangladesh — evaluating methodology, data quality, and environmental significance.

Reviewed a manuscript on landslide hazard prediction and received a Certificate of Reviewing for the contribution to the journal's evaluation process.