Eddy covariance carbonyl sulfide flux measurements with a quantum cascade laser absorption spectrometer

摘要

The trace gas carbonyl sulfide (COS) has latelyreceived growing interest from the eddy covariance (EC) community due to itspotential to serve as an independent approach for constraining gross primaryproduction and canopy stomatal conductance. Thanks to recent developments offast-response high-precision trace gas analysers (e.g. quantum cascade laserabsorption spectrometers, QCLAS), a handful of EC COS flux measurements havebeen published since 2013. To date, however, a thorough methodologicalcharacterisation of QCLAS with regard to the requirements of the EC techniqueand the necessary processing steps has not been conducted. The objective ofthis study is to present a detailed characterisation of the COS measurementwith the Aerodyne QCLAS in the context of the EC technique and to recommendbest EC processing practices for those measurements. Data were collected fromMay to October 2015 at a temperate mountain grassland in Tyrol, Austria.Analysis of the Allan variance of high-frequency concentration measurementsrevealed the occurrence of sensor drift under field conditions after anaveraging time of around 50 s. We thus explored the use of two high-passfiltering approaches (linear detrending and recursive filtering) as opposedto block averaging and linear interpolation of regular backgroundmeasurements for covariance computation. Experimental low-pass filteringcorrection factors were derived from a detailed cospectral analysis. TheCO and HO flux measurements obtained with the QCLAS were comparedwith those obtained with a closed-path infrared gas analyser. Overall, ourresults suggest small, but systematic differences between the varioushigh-pass filtering scenarios with regard to the fraction of data retained inthe quality control and flux magnitudes. When COS and CO fluxes arecombined in the ecosystemrelative uptake rate, systematic differences between the high-pass filteringscenarios largely cancel out, suggesting that this relative metric representsa robust key parameter comparable between studies relying on differentpost-processing schemes.