Rising Seas Accelerate Coastal Forest Loss Along North Carolina Coast

Coastal forest loss across North Carolina’s Albemarle-Pamlico Peninsula has accelerated sharply since 2010, with researchers documenting a 21 percent reduction in tree cover between 1985 and 2021. The findings, detailed in a 2026 study published in PLOS One, highlight a critical intersection of climate-driven ecological shift and advanced satellite analytics. Lead author Titilayo Tajudeen of North Carolina State University notes that approximately 64,220 hectares of coastal woodland have been lost over the 36-year period, with more than 40,000 hectares converted into salt marshes, shrubland, and ghost forests. The rate of forest decline is not uniform. Between 2010 and 2021, nearly 24,000 hectares transitioned to non-forested states, a figure 1.5 times higher than the loss recorded during the preceding 25 years. Ghost forests, defined as areas where rising seawater has killed mature trees but left their bleached trunks standing, have expanded even more rapidly, growing by 7,561 hectares post-2010 compared to 3,087 hectares in the prior timeframe. Researchers attribute this acceleration primarily to sea level rise, which drives saltwater intrusion into low-lying coastal zones. The phenomenon is compounded by historical extreme weather events, including the severe 2007 to 2011 drought and Hurricane Irene, which pushed vulnerable ecosystems past ecological recovery thresholds. To capture this multi-decade transformation, the research team deployed artificial intelligence integrated with high-resolution satellite telemetry. The study leveraged data from Landsat 8 and Sentinel-2, training a convolutional neural network to classify land cover changes across grid-based imagery. While Sentinel-2 provides a superior 10-meter spatial resolution, Landsat 8 offers a continuous archival record dating back to 1985, making it indispensable for long-term trend analysis. Validation tests confirmed that Sentinel-2 delivers higher precision for recent years, but the hybrid approach of combining both datasets enabled robust historical tracking. The AI model successfully distinguished between intact forest, emerging marsh, and established ghost forest zones with high accuracy. The study underscores how machine learning and remote sensing are becoming essential tools for monitoring rapid environmental degradation. Concentrated within one kilometer of the shoreline, the affected regions demonstrate how localized ecological stressors can cascade into permanent landscape transformations. Co-authors Marcelo Ardon, Mirela Tulbure, and Katherine Martin of NC State emphasize that these findings provide a quantifiable baseline for coastal resilience planning. As sea levels continue to climb, the integration of AI-driven satellite monitoring will likely play a pivotal role in predicting ecosystem collapse and informing adaptive conservation strategies along the Atlantic coast.