Utilizing Machine Learning and DSAS to Analyze Historical Trends and Forecast Future Shoreline Changes Along the River Niger

Abstract This study investigates shoreline changes along the River Niger in Nigeria over a 70-year period (1974–2044). We employ remote sensing data, machine learning, and the DSAS tool to analyze historical changes (1974–2024) and predict future trends (2024–2044). Landsat imagery obtained from the United States Geological Survey (USGS) through the Google Earth Engine API is analyzed using ArcGIS, DSAS 5.0 and 6.0 software and rainfall data acquired from the Center for Hydrometeorology and Remote Sensing (CHRS). Findings indicate notable spatial and temporal variations in shoreline dynamics across Bayelsa, Delta, and Anambra States. Around 51.47% of the transects experienced erosion, while 48.53% underwent accretion, with an average annual shoreline change rate of 1.66 meters. Despite the equilibrium, erosion exhibits a more significant impact, with a mean rate of -2.26 meters per year compared to an accretion rate of 3.92 meters per year. The study identifies a total shoreline change envelope (SCE) of 442.86 meters and a net shoreline movement (NSM) of 92.33 meters, indicating substantial overall shoreline advancement. Looking ahead, projections for 2024–2044 show varying erosion and accretion patterns across different sections. Section D faces the most significant threat, with 80% of transects experiencing erosion at a rate of -2.96 meters/year. Rainfall data analysis suggests a strong correlation (R² = 0.7576) between precipitation and shoreline change, highlighting the crucial influence of climate on coastal dynamics. These findings emphasize the need for integrated coastal management strategies that account for rainfall variability and prioritize mitigating erosion, particularly in vulnerable sections.