Spatiotemporal dynamics and machine learning-based prediction of aboveground biomass in the Indus delta mangroves

This dataset provides spatially explicit estimates of mangrove aboveground
biomass (AGB) and associated environmental variables for the Indus Delta
mangrove ecosystem. Field-based AGB spatial data were derived from the
NASA CMS Global Mangrove Distribution, Aboveground Biomass, and Canopy
Height dataset and used as reference data for model development.
Multisource remote sensing data, including Sentinel-1 and Sentinel-2
optical imagery, were processed to extract predictor variables such as
vegetation indices and surface characteristics. Additional environmental
variables, including land surface temperature and land use/land cover,
were incorporated to capture ecological controls on biomass distribution.
All satellite datasets underwent standard preprocessing steps, including
atmospheric correction, radiometric calibration, cloud masking, and
spatial resampling. The processed variables were then integrated into
machine learning models (Random Forest, Gradient Boosted Regression Tree,
Support Vector Regression, and Classification and Regression Trees) to
estimate AGB across the study region. The best-performing model (Gradient
Boosted Regression Tree) was used to generate spatially explicit AGB maps
and future projections for 2030, 2040, and 2050. Model outputs were
exported as point-based datasets containing geographic coordinates and
biomass values, along with corresponding spatial layers for mapping and
analysis.