High-frequency productivity estimates for a lake from free-water CO.sub.2 concentration measurements

Lakes are important actors in biogeochemical cycles and a powerful natural source of CO.sub.2 . However, they are not yet fully integrated in carbon global budgets, and the carbon cycle in the water is still poorly understood. In freshwater ecosystems, productivity studies have usually been carried out with traditional methods (bottle incubations, .sup.14 C technique), which are imprecise and have a poor temporal resolution. Consequently, our ability to quantify and predict the net ecosystem productivity (NEP) is limited: the estimates are prone to errors and the NEP cannot be parameterised from environmental variables. Here we expand the testing of a free-water method based on the direct measurement of the CO.sub.2 concentration in the water. The approach was first proposed in 2008, but was tested on a very short data set (3 days) under specific conditions (autumn turnover); despite showing promising results, this method has been neglected by the scientific community. We tested the method under different conditions (summer stratification, typical summer conditions for boreal dark-water lakes) and on a much longer data set (40 days), and quantitatively validated it comparing our data and productivity models. We were able to evaluate the NEP with a high temporal resolution (minutes) and found a very good agreement (R.sup.2 ≥ 0.71) with the models. We also estimated the parameters of the productivity-irradiance (PI) curves that allow the calculation of the NEP from irradiance and water temperature. Overall, our work shows that the approach is suitable for productivity studies under a wider range of conditions, and is an important step towards developing this method so that it becomes more widely used.