Overestimation of CO2 respiration fluxes by the closed chamber method in low-turbulence nighttime conditions

Knowledge about ecosystem respiration is required for accurate estimates of daily, seasonal and annual carbon exchange between ecosystems and the atmosphere. Measurements by the closed chamber technique were often used to calculate nighttime respiration fluxes during low-turbulence nighttime conditions when the eddy covariance method is problematic. The goal of this study is to investigate if the closed chamber measurements could also be biased during low-turbulence atmospheric conditions. We applied the closed chamber technique for measurements of CO2 flux at flark, lawn and hummock microsites at a boreal peatland (n = 602). The friction velocity (u(*)) was used for screening data into either well developed or low-turbulence conditions (u(*) <= 0.1 m s(-1)). During chamber experiments in low-turbulence nighttime conditions, the CO2 concentrations within the chamber headspace were observed to increase extremely fast and nonlinearly at the start of the chamber deployment period. The possible reason is the abrupt disturbance of the natural concentration gradients in the soils, plants and the near-surface atmosphere. This artifact leads to an overestimation of CO2 fluxes and occurs at all microsite types. However, the strongest overestimation occurs at hummocks. Consequently, we recommend that nighttime chamber measurements have to be carefully checked with respect to prevailing turbulence conditions. Under low-turbulence conditions their results can be overestimated. Since the bias is difficult to correct, chambers are probably not suitable to replace eddy covariance measurements under low-turbulence nighttime conditions, since both measurements tend to be biased.