Long-term eddy covariance measurements of the isotopic composition of the ecosystem–atmosphere exchange of CO2 in a temperate forest

Abstract

Measurements of the isotopic composition of the net ecosystem–atmosphere exchange of CO2 (NEE) have been desired as a means to probe ecosystem carbon cycling and in particular to partition NEE into gross ecosystem photosynthesis and respiration. Several attempts at such measurements have combined eddy covariance (EC) measurements of the total net CO2 flux with flask measurements of the isotopic composition of CO2 in ambient air – an indirect method that has never been validated. Direct EC measurements of the isotopic composition of NEE (i.e. of the net exchanges of 12C16O2, 13C16O2, and 18O12C16O) have been made only twice, in short-term (2-month and 1-month) campaigns.Here we present a full growing season of direct EC measurements of the isotopic composition of NEE in a temperate deciduous forest, and we use these data: (1) to rigorously assess their limiting sources of error, (2) to test the indirect EC/flask method, and (3) to give an indication of the potential for ecological analyses. We describe the method and instrumentation, including the new cryogen-free, continuous-wave, quantum cascade laser spectrometer, which can determine δ13C and δ18O in ambient CO2 with unprecedented noise levels of ±0.02‰ and ±0.03‰ (1 standard deviation), respectively, for a 100s integration time and a 1-h calibration interval. We find: (1) that precision is jointly limited by the instrumentation and by horizontal ecosystem or landscape heterogeneity, so that there is little to be gained by further improvements to instrument performance; (2) that the isotopic composition of NEE obtained by the EC/flask method can be biased, on a monthly timescale, by 2‰; and (3) that the present measurements are precise enough to elucidate biological mechanisms controlling the ecosystem-scale carbon balance.