Modeling Whole-river Metabolism with In-stream Hydrology and Water Quality Dynamics in a Lowland River

We develop a new approach to estimate ecosystem metabolism by linking it with process-based modeling of in-stream flows and water quality. We test this new approach along a 62 km long stretch in the lower River Thames in England using high-frequency water quality observations of two years (2013-2014). Conventional methods for estimating river metabolism rates do not specifically account for oxygen transport under varying flows and oxygen transformations due to biogeochemical processes. Our approach overcomes both these challenges and thus, allows us to quantify specific metabolic pathways including photosynthetic production and respiration, reaeration, benthic oxygen demand, biochemical oxygen decay (BOD), and nitrification. We find that the river is dominantly autotrophic during biomass growing season, and mainly heterotrophic during the rest of the year. Ecosystem respiration at the upstream end of the network is mainly driven by oxygen loss through BOD decay, autotrophs and nitrification, which together comprise 97% of total respiration. At the downstream end, in addition to BOD decay, autotrophs and nitrification losses (80%), benthic communities also show a significant contribution (19%). We also analyse the sensitivity of our estimated metabolism rates to multiple environmental controls using empirical modeling techniques such as random forests and generalised least squares regression. We find that primary productivity in the river is highly influenced by light intensity, water temperature and flow, whereas ecosystem respiration is mainly sensitive to flow, inorganic phosphorus and suspended sediment concentrations. We also identify important interactions between these environmental variables that influence metabolism rates within the river network. Overall, our study demonstrates the potential of combining process-based and empirical modeling approaches to estimate metabolism in large river networks and evaluate its sensitivity to multiple environmental controls.