Environmental coherence framework for multi-sensor remote sensing: water hyacinth assessment in Lake Tana

Abstract

Abstract Invasive water hyacinth threatens Lake Tana, Ethiopia—the source of the Blue Nile. This study presents the first comprehensive 11-year (2013–2024) remote sensing assessment of the water hyacinth invasion dynamics. It introduces a novel environmental coherence framework—grounded in established indirect validation paradigms—to evaluate 11 multisensor algorithms (using Landsat 8/9, Sentinel-1, and Sentinel-2 data) against hydro-meteorological drivers, providing an ecologically grounded validation alternative when systematic field surveys are impractical. This environmental coherence framework identified Sentinel-2 NDVI and NDVI + FAI as the best indicators, achieving the highest relative environmental coherence scores among the 11 indicators tested. The time series generated by this framework follows a complex ‘boom-bust’ invasion cycle, with a peak phase (2018–2019), a subsequent decline, and a recent resurgence. Our findings establish a replicable, Google Earth Engine based workflow for monitoring aquatic invasions in data-scarce tropical regions and provide critical, data-driven insights for targeted environmental management.