Evaluating Opportunities For Remote Sensing of Tss On Small Rivers

Understanding the spatial and temporal patterns of suspended sediment transport is important for mitigating degraded water quality but is often logistically and financially infeasible to accomplish, especially at scale. Remote sensing offers a promising alternative to in-situ data collection programs, but the applications have mostly been tested on larger rivers (> 150 m in width), limiting our understanding of its applicability to most of the river network. We evaluated the application of widely used single and multi-band Total Suspended Solid (TSS) algorithms within the relatively small rivers of the Lake Champlain Basin in the Northeastern US. We identified Sentinel-2 images captured within a few hours of depth-integrated TSS samples measured in 15 tributaries (17 to 117 m wide) and corrected them for atmospheric conditions. Errors in the relationship between measured and predicted TSS values were used to explore potential limits on the application of existing TSS algorithms to small rivers. We found that remote sensing algorithms had greater success predicting TSS when flow depths were sufficient (approximately greater than 0.5 m), deciduous trees were dormant, and biological activity was limited and on rivers with a relatively fine sediment load (i.e., potentially wash-load dominated). We hypothesized that algorithm performance on smaller rivers would be prone to larger errors due to the greater influence of contrasting reflectance signals from soils and vegetation along river banks. The degree of limitation from this adjacency effect depends on image resolution and season, but across seasons we did not find that wider rivers had less error than narrower ones in our study region using 10-m imagery. Our results highlight the feasibility of extending remote sensing applications to smaller river networks, especially during the colder months, when in-situ monitoring is particularly challenging.