Modeling the coupled dynamics of stream metabolism and microbial biomass

Estimating and interpreting ecosystem metabolism remains an important challenge in stream ecology. Here, we propose a novel approach to model, estimate, and predict multiseasonal patterns of stream metabolism (gross primary production [GPP] and ecosystem respiration [ER]) at the reach scale leveraging on increasingly available long-term, high-frequency measurements of dissolved oxygen (DO). The model uses DO measurements to estimate the parameters of a simple ecosystem model describing the underlying dynamics of stream autotrophic and heterotrophic microbial biomass. The model has been applied to four reaches within the Ybbs river network, Austria. Even if microbial biomasses are not observed, that is, they are treated as latent variables, results show that by accounting for the temporal dynamics of biomass, the model reproduces variability in metabolic fluxes that is not explained by fluctuations of light, temperature, and resources. The model is particularly data-demanding: to estimate the 11 parameters used in this formulation, it requires sufficiently long, for example, annual, time series, and significant scouring events. On the other hand, the approach has the potential to separate ER into its autotrophic and heterotrophic components, estimate a richer set of ecosystem carbon fluxes (i.e., carbon uptake, loss, and scouring), extrapolate metabolism estimates for periods when DO measurements are unavailable, and predict how ecosystem metabolism would respond to variations of the driving forces. The model is seen as a building block to develop tools to fully appreciate multiseasonal patterns of metabolic activity in river networks and to provide reliable estimations of carbon fluxes from land to ocean.