Integrating Phenological Features to Enhance Coastal Salt Marsh Monitoring and Infer Vegetation Succession Mechanisms: A Case Study of Yancheng National Nature Reserve

Rapid changes in coastal salt marsh vegetation highlight the need for accurate monitoring to support sustainable management. This study aims to track annual salt marsh landscape dynamics and discuss vegetation succession mechanisms. A novel approach was proposed to improve the discrimination of different plant species by integrating phenological features derived from NDVI time series, thereby enhancing salt marsh classification. The results showed that the proposed method achieves an overall accuracy of 93.6%, significantly outperforming schemes based solely on spectral indices (81.6%) or combined spectral and textural features (85.5%) (p < 0.05). The overall accuracy of annual classification maps for the Yancheng National Nature Reserve wetland from 1994 to 2025 averaged 90.5%, with a minor fluctuation of 2.8%, demonstrating the accuracy and reliability of this method for long-term monitoring of salt marsh landscapes. Annual classification maps revealed that the encroachment of Spartina alterniflora and Phragmites australis has intensified the fragmentation of Suaeda salsa. Notably, eradication efforts targeting Spartina alterniflora appear to have contributed to controlling this invasive species (decreasing by 33.22 km2) and restoring native species (increasing by 2.56 km2). This method serves as a reference for advancing salt marsh landscape classification frameworks.