Using raw dissolved oxygen data to infer seasonal patterns of stream metabolism and storm impacts on stream productivity

Abstract:

Mountain forested streams are key contributors to the planet’s carbon cycle. However, estimating stream productivity continues to be challenging due to the low productivity rates of these ecosystems and the uncertainties associated to reaeration and ecosystem respiration rates. We propose the use of high-resolution raw dissolved oxygen data (% saturation, DOsat) and on-site photosynthetically active radiation data (PAR) to develop DOsat~PAR curves as an analogy to the well-known photosynthesis-Irradiance (P-E) curves. The premise of our research is that although these curves are simple, they can provide detailed information of stream ecosystems productivity dynamics. We used data from two streams in the Oregon Coast Range to investigate daily and seasonal trends of stream response to available solar radiation including the identification of patterns of light-saturation and photoinhibition. We used properties of the light-limited portion of the DOsat~PAR regression curves to produce a multivariate linear model to estimate stream productivity. The data from one of the two study sites (Oak Creek) was used for model development while the data from the other site (Mill Creek) was used for model validation. The model was also tested in terms of its ability to quantify the effects of discrete storm events on stream productivity by comparing gross primary productivity estimates calculated through a Bayesian framework (streamMetabolizer) and our raw data driven estimates of productivity which were based on the variability of the DOsat~PAR regression curves. The proposed methodology was developed at mountain-forested streams but may be used at other stream ecosystems. We foresee that our method may be used at any site or group of sites to construct a baseline understanding of productivity dynamics that is independent of the methodological challenges of the current stream metabolism models.