Quantifying nighttime ecosystem respiration of a meadow using eddy covariance, chambers and modelling

Aim of the present paper is to quantify the ecosystem respiration of a mountain meadow in the Austrian Alps during the vegetation period 2002 by constraining nighttime eddy covariance measurements with ecosystem respiration derived from (i) daytime eddy covariance, (ii) ecosystem chamber and (iii) scaled up leaf and soil chamber measurements. The study showed that the discrimination of valid nighttime eddy covariance measurements based on friction velocity (u*), the so-called u*-correction, is very sensitive to the imposed quality control criteria. Excluding half-hourly nighttime data, which deviate more than 30% from the stationarity and integral turbulence tests caused the magnitude of the u*-correction to be significantly reduced. Based solely on nighttime eddy covariance data, we are currently unable to decide whether the observed high CO2 fluxes during intermittent turbulence represent artefacts and should be screened out, or whether these reflect a genuine transport of CO2 not accounted for by the storage term and must be retained. Evidence against the inclusion of these data is derived from soil respiration rates measured in situ and calculated inversely from the other approaches, which were significantly lower as compared to soil respiration calculated from inversion of the half-hourly nighttime data inclusive of the observations which failed to meet the specified quality control criteria. Seasonal (8 March-8 November 2002) nighttime carbon balances simulated based on the parameters derived from the remaining approaches agreed with each other to within 35%, which is of the order of the uncertainty of each individual approach.