Understanding the spatio-temporal dynamics of mangrove species of Bhitarkanika and Pichavaram ecosystems, India

Despite extensive research on mangrove extent and health, a unified understanding of long-term health dynamics, interspecies transitions, and species-specific phenological responses across varying hydrological regimes remains crucial for effective conservation and management. This study leverages Landsat 5–9 and Sentinel-2 to characterize the mangrove ecosystem dynamics in Bhitarkanika and Pichavaram, India, focusing on interspecies changes, broad ecological trends, and detailed species-specific phenological patterns. Various machine learning algorithms were employed to perform species-level classification on Sentinel-2 multispectral data and to assess the spatio-temporal changes in the species distribution between 2019 and 2023, with Random Forest and Gradient Tree Boosting outperforming other classifiers (overall accuracy >90 %, kappa >0.9). Long-term (1988–2023) Normalized Difference Vegetation Index (NDVI) trends were analyzed using the non-parametric Theil-Sen's Slope. Furthermore, harmonic regression modelled species phenology from Sentinel-2 derived vegetation indices and demonstrated enhanced species differentiation using temporal parametric model metrics. The change detection analysis revealed significant interspecies interactions, indicating species competition and degradation in the Bhitarkanika and Pichavaram ecosystems. Long-term NDVI trends indicated stress in Bhitarkanika (−0.008 < TSS < −0.002, post-1997), contrasting with stability in Pichavaram. Species-specific analysis highlighted distinct bimodal phenological patterns with variations influenced by local hydrological conditions. Crucially, this phenology-based approach significantly enhanced the differentiation of spectrally similar species. These findings highlight mangrove ecosystem dynamics and reveal species-specific responses, crucial for targeted conservation and adaptive management. Future research should address the relationship between phenological patterns and local environmental conditions for robust vulnerability assessments and improved management practices, targeting different mangrove species. • Interspecies transitions reveal dynamic competition and degradation. • Bhitarkanika mangroves show long-term stress; Pichavaram remains stable. • Monsoon patterns drive distinct site-specific mangrove phenological cycles. • Phenology metrics significantly enhance discrimination of mangrove species. • Novel framework aids targeted mangrove conservation and adaptive management.