Abstract
Intensified agriculture practices contribute to nutrient enrichment in freshwater lakes, causing eutrophication, algal blooms, and water hyacinth infestations. Eutrophication in Lake Tana, the source of the Blue Nile in Ethiopia, necessitates effective monitoring due to rapid infestation of water hyacinths. While traditional monitoring is costly and limited in spatial and temporal coverage, remote sensing offers a promising alternative. This study develops a regression model to estimate Chlorophyll-a (Chl-a) concentration using in situ and remote sensing reflectance data. Field measurements from 143 locations across Lake Tana were used to validate the correlation equations. Results show that the Moderate Resolution Imaging Spectroradiometer (MODIS) in near-infrared reflectance exhibits the strongest linear relationship with in situ Chl-a measurements for August 2016 (r2 = 0.53), December 2016 (r2 = 0.56) and March 2017 (r2 = 0.61). The developed models were validated with a root-mean-square error of 2.76 μg/L, 5.89 μg/L, and 8.04 μg/L for August, December, and March, respectively. Applying the developed model from 2008–2018, the Chl-a concentration of the lake indicated an increasing trend, likely driven by non-point sources from surrounding watersheds, causing infestation of the lake by hyacinths since 2011. The agreement between MODIS and in situ Chl-a data, coupled with the satisfactory performance of the linear regression model, underscores that developing a regression model for Chl-a estimation from remote sensing in water hyacinth-infested lakes is a useful method in tracking spatiotemporal variations. This study will serve as a foundation for future Chl-a variation studies in Lake Tana and other similar lakes.