Peeters, Frank; Hofmann, Hilmar; Fernandez, Jorge Encinas

Metabolic transformations have a major impact on the development of primary producers in aquatic systems and thus affect the dynamics of the entire aquatic food web. Furthermore, metabolic transformations contribute to the carbon budget and thereby influence CO2 emissions from aquatic systems. Several techniques have been developed that aim at an easy assessment of metabolic rates over long time periods or in many systems. The 18/16O technique, which utilizes the isotopic fractionation between 18O and 16O isotopes due to metabolic transformations, is receiving increasing popularity in studies comparing the metabolism in many different lakes and served as basis for the conclusions that production increases with increasing atmospheric CO2 and that surprisingly little terrestrial carbon is recycled in lakes of the arid circumpolar landscape. However, we demonstrate here that the steady state assumptions underlying the 18/16O technique cause large uncertainties in the estimated metabolic rates. This conclusion is based on a sensitivity analysis using a numerical model of dissolved oxygen, DO, and of dissolved 18O, 18ODO, but is also confirmed by published metabolic rates estimated from the 18/16O and the diel O2 techniques. Metabolic rates obtained from the 18/16O technique appear unsuited for correlation analyses between lakes but may provide reasonable estimates in systems with low and long-term stable production. In addition we illustrate that the combination of few 18O measurements with the diel O2 technique and an inverse fitting procedure can improve estimates of metabolic rates and in particular of respiration rates.