Metabolism is an integrative measurement of stream and river ecosystem functioning, and thus, could be used to assess impairment. Stream metabolism is measured by different methods which often yield contrasting results. Furthermore, open-channel measurements of metabolism, which offer the best potential for continuous monitoring of stream functioning, rely on calculations of gas exchange with the atmosphere, for which a plethora of methods exists. Therefore, to incorporate metabolism in stream monitoring programs, it is necessary to determine which methods yield comparable results under a given set of environmental conditions. We studied 21 streams in the Basque Country (northern Spain), ranging widely in physical characteristics and water quality. We calculated reaeration during summer baseflows using three different approaches: the night-time drop in oxygen, the lag between noon and peak oxygen concentration, and ten empirical equations relating depth and velocity with reaeration coefficients obtained from the literature. Differences among methods were very large, especially at the shallower sites. The results obtained with most empirical equations were highly correlated, but showed little agreement with the night-time and peak lag methods. We then analyzed the response of reaeration rate to river stage: reaeration calculated by the night-time method during 1 year of continuous monitoring was regressed against discharge at each site, and the resulting model was compared to the results of empirical equations, using software HecRas 2.2 to model hydraulic conditions at different river stages. The shape of reaeration-discharge plots differed greatly and in a site-dependent manner, and there was little agreement between methods. Finally, we investigated the effects of reaeration rate on estimates of metabolism. The choice of method greatly affected the estimates of both primary production and respiration. The empirical equations, except E-7 and E-10, yielded the most unrealistic estimations of stream metabolism. Overall, the night-time method, especially when regressed against discharge, seems to be the most robust and reliable among those tested, with the energy dissipation method (E-10) appearing to be a viable alternative when the night-time method does not work.
