Forecasting shoreline dynamics and land use/land cover changes in Balukhand-Konark Wildlife Sanctuary (India) using geospatial techniques and machine learning
Manoranjan Mishra, Debdeep Bhattacharyya, Brihaspati Mondal, Suman Paul, Richarde Marques da Silva, Celso Augusto Guimarães Santos, Rajkumar Guria
Abstract
Coastal ecosystems play a critical role in biodiversity conservation, climate regulation, and supporting human livelihoods. This study investigates future shoreline and land use/land cover change (LULCC) projections in the Balukhand-Konark Wildlife Sanctuary (BKWS). Shoreline and LULCC dynamics were quantified using satellite imagery, geospatial tools, and predictive modeling. An ensemble machine learning approach, combining Cellular Automata-Artificial Neural Networks (CA-ANN) and Random Forest techniques, was employed to enhance prediction accuracy. The shoreline dynamics analysis indicated persistent erosion in the northern zones and accretion in the southern zones, trends projected to continue through 2033 and 2043. LULC assessments showed a high reduction in dense vegetation, from 41.8 % in 1993 to 37.1 % in 2023, alongside an increase in sparse vegetation to 50.3 %. Predictions suggest continued degradation in the northern sections of BKWS, stabilization of sparse vegetation, and expansion of sandy areas by 2043. The results highlight high shoreline erosion and LULC changes. This study shows the need to integrate shoreline dynamics and LULC analyses into coastal management practices and provides actionable recommendations for habitat restoration, sustainable land-use planning, and climate adaptation strategies.