Scholarlink

<p>This unique study develops a GIS and Machine Learning integrated hybrid Multi-Criteria Decision Making (MCDM) framework employing Shannon Entropy and Combined Compromise Solution (CoCoSo) models to map mangrove habitat suitability for regeneration on Dhanchi Island, a Bay-facing continental island in the Indian Sundarbans Biosphere Reserve, which has shrunk from 38.73 km<sup>2</sup> (1970) to 33.14 km<sup>2</sup> (2020) due to erosion and tidal dynamics. From 27 initial parameters across physical-scape (slope, land cover, elevation, distance to saltpans), hydro-climatic-scape (rainfall, LST, inundation, drainage density, distance to creeks, TWI), chemical-scape (soil salinity), and environmental-scape (canopy height, DBH, NDVI), 14 were selected post-multi-collinearity testing (all VIF &lt; 10, tolerance &gt; 0.1), with hydro-climatic factors dominating weights (30.64%) followed by environmental (29.65%) and physical (28.68%) scapes. Shannon Entropy objectively derived data-driven weights, while CoCoSo combined weighted sum/product measures for compromise solutions; both classified suitability into five zones (critically unsuitable to optimally suitable) using ArcMap 10.8 and R Studio processing of Copernicus DEM, Landsat 8, NASA rainfall, and field data, validated against &gt; 100 ground-truthed points via ROC-AUC (Shannon Entropy AUC = 0.734; CoCoSo = 0.697). Shannon Entropy allocated larger highly/optimal zones (9.66 km<sup>2</sup> combined) in southern/western fringes and erosion hotspots, outperforming CoCoSo's conservative estimates (8.10 km<sup>2</sup>), with central saltpans (~ 5 km<sup>2</sup>) unsuitable due to hyper-salinity in the saltpans. Dumbbell plots highlighted Shannon Entropy's smoother patterns for regeneration planning of native species (e.g., <i>Avicennia</i> spp.<i>, </i><i>Aegiceras corniculatum</i>, <i>Acanthus ilicifolius</i>, <i>Bruguiera gymnorrhiza</i>, <i>Ceriops tagal</i>). These findings advance nature-based solutions for coastal resilience, prioritizing mudflat restoration via sediment trapping and salinity mitigation to counter shoreline retreat, offering a robust, objective MCDM tool for Sundarbans mangrove conservation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Mangrove habitat suitability mapping integrating hybrid MCDM modelling approach of bay facing Dhanchi Island, Indian Sundarbans

  • 期刊论文

  • 发表日期：

    2026年7月7日

查看全文
  • Sweta Chatterjee,
  • Gupinath Bhandari

摘要

This unique study develops a GIS and Machine Learning integrated hybrid Multi-Criteria Decision Making (MCDM) framework employing Shannon Entropy and Combined Compromise Solution (CoCoSo) models to map mangrove habitat suitability for regeneration on Dhanchi Island, a Bay-facing continental island in the Indian Sundarbans Biosphere Reserve, which has shrunk from 38.73 km2 (1970) to 33.14 km2 (2020) due to erosion and tidal dynamics. From 27 initial parameters across physical-scape (slope, land cover, elevation, distance to saltpans), hydro-climatic-scape (rainfall, LST, inundation, drainage density, distance to creeks, TWI), chemical-scape (soil salinity), and environmental-scape (canopy height, DBH, NDVI), 14 were selected post-multi-collinearity testing (all VIF < 10, tolerance > 0.1), with hydro-climatic factors dominating weights (30.64%) followed by environmental (29.65%) and physical (28.68%) scapes. Shannon Entropy objectively derived data-driven weights, while CoCoSo combined weighted sum/product measures for compromise solutions; both classified suitability into five zones (critically unsuitable to optimally suitable) using ArcMap 10.8 and R Studio processing of Copernicus DEM, Landsat 8, NASA rainfall, and field data, validated against > 100 ground-truthed points via ROC-AUC (Shannon Entropy AUC = 0.734; CoCoSo = 0.697). Shannon Entropy allocated larger highly/optimal zones (9.66 km2 combined) in southern/western fringes and erosion hotspots, outperforming CoCoSo's conservative estimates (8.10 km2), with central saltpans (~ 5 km2) unsuitable due to hyper-salinity in the saltpans. Dumbbell plots highlighted Shannon Entropy's smoother patterns for regeneration planning of native species (e.g., Avicennia spp., Aegiceras corniculatum, Acanthus ilicifolius, Bruguiera gymnorrhiza, Ceriops tagal). These findings advance nature-based solutions for coastal resilience, prioritizing mudflat restoration via sediment trapping and salinity mitigation to counter shoreline retreat, offering a robust, objective MCDM tool for Sundarbans mangrove conservation.