Coupled oxygen-carbon dioxide modelling to partition potential external contribution to stream carbon dioxide concentrations.

Recent modelling studies in large catchments have estimated that in excess of 74% of the dissolved carbon dioxide found in first and second order streams originate from allochthonous sources. Stable isotopes of carbon-13 in carbon dioxide have been used to identify ground water seeps in stream systems, where decreases in δ13CO2 occur along gaining stream reaches, suggesting that carbon dioxide in ground water is more depleted than what is found in surface water due to fractionation of CO2 during emissions across the air water interface. Although isotopes represent a chemical tracer in stream systems for potential groundwater contribution, the temporal resolution of discrete samples make partitioning allochthonous versus autochthonous sources of CO2 difficult on hydrologically relevant time scales. Here we show results of field deployments of high frequent dissolved CO2, O2, PAR, Temperature and pH from the Thornton Creek Watershed, the largest urban watershed in Seattle, WA. We present an exploration into using high resolution time series of dissolved oxygen and carbon dioxide in a dual gas approach to separate the contribution of in stream respiration from external sources. We extend upon previous efforts to model stream metabolism across diel cycles by incorporating simultaneous direct measurements of dissolved oxygen, PCO2, and pH within an inverse modeling framework and Bayesian parameter estimation. With an initial assumption of a stoichiometric ratio of 1:1 for O2 and CO2 for autochthonous driven metabolism, we investigate positive or negative departures from this ratio as an indicator of external CO2 to the stream (terrestrial or atmospheric) and factors contributing to this flux.