Demars et al. present a review of stream ecosystem metabolism theory that provides the basis for a set of best practices. Here we clarify two technical and conceptual errors in reviewing the current state of research regarding inverse modeling of ecosystem metabolism. Specifically, the assertions by Demars et al. that inverse modeling of metabolism is limited to low reaereation, high productivity ecosystems and, when feasible, measurement of gas transfer by volatile tracer injection is the preferred method, are based on an incomplete conceptualization of current ecosystem metabolism models and does not recognize the unconstrained biases in tracer injection methods. In addition to these corrections, we highlight how inverse model fitting is a robust methodology to constrain key metabolic parameters by capturing the information contained in the dynamics of diel O 2 data. When done in a Bayesian context, inverse model fitting also provides a means to incorporate multiple sources of available data, including those from tracer studies, and formally propagate uncertainties in parameter estimates. Last, we describe how model fitting of diel O 2 data can also provide information on temperature sensitivity of ecosystem respiration.Sixty years of research in stream ecosystem metabolism using diel O 2 measurements has demonstrated that subdaily biological responses of streams and lakes to physical drivers such as light and temperature provide great insight into how these ecosystems function. The review by Demars et al. (2015) aims to highlight recent advancements in stream metabolism research and identify current challenges. The authors make a number of valid points about the utility of experimenting with diel O 2 models, the need to devise methods for separating autotrophic vs. heterotrophic respiration, and issues of scale and heterogeneity in real ecosystems. However, in providing recommendations about gas transfer and the role of inverse modeling in analyzing diel O 2 data, Demars et al. (2015) misrepresent key conceptual principles behind these models and their application. We seek to clarify some of the more consequential misconceptions and provide a comprehensive picture of how to integrate models with diel O 2 data to determine aquatic ecosystem metabolic parameters.
Parameterization of gas transferDemars et al. (2015) make two errors with regards to gas exchange parameterization that are important for how diel O 2 data are interpreted and used to quantify ecosystem metabolism. The first is a misrepresentation of the technique presented in Holtgrieve et al. (2010), where the gas transfer velocity (K L , using the conventions of Demars et al. 2015) is determined by inverse fitting of a dynamic O 2 mass balance model to diel changes in O 2 concentrations. Demars et al. (2015) make the incorrect assertion that estimating gas transfer from diel O 2 only works in low K L systems by stating these methods are "restricted to sinusoidal diel curves (i.e., K L < 0.5 m h 21 and high productivity)." Whether model fitting can r...