Estimating Riverine Air-Water Gas Exchange and Metabolism from Long Oxygen Time Series

Abstract:

To accurately depict the role of streams and rivers in carbon cycling requires estimating air- water gas exchange, productivity, and respiration. It is possible to estimate gas exchange and metabolism (gross primary production and ecosystem respiration) simultaneously from oxygen data themselves, but estimates from any single day often contain a substantial (and unknown) amount of parameter error. Here we developed a statistical method to leverage the extra information in a long time series to better estimate daily rates of gas exchange and metabolism. Such time series are ubiquitous in water quality monitoring programs, and these data are readily available over broad spatial scales. We developed a hierarchical model that estimates gas exchange as a function of discharge for a year-long time series of dissolved oxygen data. Gas exchange, and therefore metabolism, had much lower temporal variability than if we estimated parameters on separate days. Rates of gas exchange were positively related with discharge, but the relationship was river specific and often nonlinear. Our approach provides a robust means to estimate gas exchange and metabolism from the many rivers that have oxygen time series collected as part of water quality monitoring.