Evaluating the suitability of remote sensing lake turbidity in validating model sediment loadings: An application in Lake Tana basin

Validating water quality model applications is challenging due to data gaps in in-situ observations, especially in developing regions. To such a challenge, remote sensing (RS) has provided an alternative to monitor the water quality of inland waters due to its low cost, spatial continuity and temporal consistency. However, limited studies have exploited the option of validating water quality model outputs with RS water quality data. With sediment loadings regarded as a threat to the turbidity and trophic status of Lake Tana in Ethiopia, this study aims at using existing RS lake turbidity data to validate the seasonal and long-term trends of sediment loadings in and out of Lake Tana. A hydrologically calibrated SWAT+ model is used to simulate river discharge and sediment loadings flowing in and out of Lake Tana basin. Together, with a remote sensing dataset of lake turbidity from Copernicus Global Land Service (CGLS), seasonal and long term correlations between lake turbidity and sediment loadings at the river mouths of Lake Tana are estimated. Results indicate a strong positive correlation between sediment load from inflow and out flow rivers with RS lake turbidity (r2 > 0.5). Other strong positive relations were observed between the stream flow from inflow rivers and the lake turbidity (r2 > 0.5). These indicate that river streamflow accounts for significant responses in river sediment loads and lake turbidity which likely occurred from a combination of overland transport of sediment into streams due to erosion of the landscape, scouring of streambanks, and resuspension of sediment from channel beds. We conclude that RS water quality products can potentially be used for validating seasonal and long term trends in simulated SWAT+ water quality outputs especially in data scarce regions.