Development of Original Remote Sensing-Based Methodologies for Monitoring Coastal Fisheries Habitat Change in the Niger Delta

The rapid and large-scale degradation of coastal fisheries habitats in the Niger Delta through oil pollution, mangrove deforestation, and coastal erosion requires monitoring methodologies capable of detecting change at ecologically relevant spatial and temporal scales, yet no remote sensing analytical framework validated for Niger Delta coastal fisheries habitats exists in the literature. This study develops and validates original remote sensing-based methodologies for multi-temporal monitoring of three critical coastal fisheries habitats, specifically mangrove forest extent, seagrass bed distribution, and intertidal mudflat area, across the 36,000-square-kilometre Niger Delta coastal zone. The methodological contribution centres on the development of a machine learning habitat classification algorithm optimised for the spectral confusion inherent in the oil-contaminated, sediment-rich optical environment of Niger Delta coastal waters, trained on 2,400 ground-truth points collected by boat-based field surveys in Rivers and Delta States. The classifier was applied to multitemporal Sentinel-2 and Landsat 8-9 image stacks spanning 2015 to 2024, generating annual habitat extent maps. Validation against independent field reference data yielded overall classification accuracy of 89.3% across all habitat classes. Change detection analysis revealed a net mangrove loss of 4.2% over the study decade, with losses concentrated near petroleum infrastructure. Seagrass extent declined by 17.3% in the eastern Niger Delta, associated with increased turbidity. The developed methodology is packaged as an open-source Python toolbox for use by Nigerian environmental monitoring agencies and fisheries research institutions.

Keywords: remote sensing, habitat monitoring, Niger Delta, mangrove, coastal fisheries