Geospatial assessment of forest disturbance and sustainable development goals integration in dry deciduous landscapes using remote sensing and Google Earth Engine

Abstract

Dry deciduous forests are highly susceptible to anthropogenic pressures yet remain underrepresented in global research. This study proposes an integrated geospatial framework that combines remote sensing with Sustainable Development Goal (SDG) assessment to monitor forest dynamics in Purulia District, India. Multi-temporal Landsat data (2003–2023) were processed in Google Earth Engine, applying principal component analysis to 12 vegetation indices. Landscape fragmentation was assessed using edge density, patch density and patch cohesion, while a disturbance index was used to quantify human impacts. The results show an increased bare soil index (–0.17 to 0.17) and reduced vegetation soil moisture (0.26 to –0.03), indicating greater soil exposure and moisture stress. Land-use analysis revealed 22 km² of regeneration versus 15 km² of deforestation. The NDVI peaked in 2005, 2013 and 2021, with notable stress from 2006 to 2010. SDG-based goal level scores indicated the greatest impacts on SDG 15 (1.0), SDG 11 (0.62) and SDG 17 (0.68), offering a transferable model for evidence-driven forest management.