A unified code for conventional and disjunct eddy covariance analysis of trace gas measurements: An urban test case

Abstract. We describe and test a new versatile software tool for processing eddy covariance and disjunct eddy covariance data. We present an evaluation based on urban NMVOC measurements using a Proton-transfer-reaction quadrupole interface time of flight mass spectrometer (PTR-QiTOFMS) at the Innsbruck Atmospheric Observatory. The code is based on MATLAB ® and can be easily configured to process high frequency, low frequency and disjunct data. It can be applied to a wide range of analytical setups for NMVOC as well as other trace gas measurements, and is tailored towards the application of noisy data, where lag-time corrections become challenging. Several corrections and quality control routines are implemented to obtain the most reliable results. The software is open-source, so it can be extended and adjusted to specific purposes. We demonstrate the capabilities of the code based on a large urban dataset collected in Innsbruck, Austria, where ambient concentrations of non-methane volatile organic compounds (NMVOC) and auxiliary trace gases were sampled with high temporal resolution above an urban canopy. Concomitant measurements of 12C and 13C isotopic NMVOC fluxes allow testing algorithms used for determinations of flux LODs and lag time analysis. We use the high frequency NMVOC data set to generate a set of disjunct data and compare these results with the true eddy covariance method. The presented analysis allows testing the theory of DEC in an urban environment. Our findings confirm that the disjunct eddy covariance method can be a reliable tool, even in complex urban environments, when fast sensors are not available, but that the increase in random error impedes the ability to detect small fluxes due to higher flux LODs.