Enhanced spatial–spectral fusion for mapping invasive Spartina alterniflora in mangrove forests

Spartina alterniflora (S. alterniflora) has rapidly invaded mangrove forests, posing serious threats to biodiversity and ecosystem functioning. Accurate mapping is essential for invasion management but remains limited because of the lack of available imagery that offers high spatial and spectral resolution. To address this gap, we propose an Enhanced Spatial–Spectral Fusion Model (ESSFM), which improves upon the conventional Robust and Adaptive Spatial–Spectral Fusion Model (RASSFM) by incorporating correlation-guided band simulation and edge fusion to effectively combine the spatial resolution of PlanetScope (PS) and the spectral information of Sentinel-2 (S2) imagery. ESSFM produced fused images with higher spectral fidelity and sharper edge detail than RASSFM, enhancing the separability between S. alterniflora and mangroves. Random Forest classification based on ESSFM achieved the highest accuracy (overall accuracy = 92.44%), with the Blue, Red and NIR bands identified as the most critical for invasion detection. These results demonstrate that ESSFM has improved the single-date mapping of invasive S. alterniflora in mangrove ecosystems.