Disjunct eddy covariance technique for trace gas flux measurements

A new approach for eddy covariance flux measurements is developed and applied for trace gas fluxes in the atmospheric surface layer. In disjunct eddy covariance technique, quick samples with a relatively long time interval between them are taken instead of continuously sampling air. This subset of the time series together with vertical wind velocity data at corresponding sampling times can be correlated to give a flux. The disjunct eddy sampling gives more time to analyze the trace gas concentrations and thus makes eddy covariance measurements possible using slower sensors. In this study a proton‐transfer‐reaction mass spectrometer with response time of about 1 second was used with a disjunct eddy sampler to measure fluxes of volatile organic compounds from an alfalfa field. The measured day‐time maximum methanol fluxes ranged from 1 mg m−2 h−1 from uncut alfalfa to 8 mg m−2 h−1 from freshly cut alfalfa. Night‐time fluxes were around zero.


Introduction
The eddy covariance method is the most direct method for surface layer flux measurements. Its use has been restricted to relatively few atmospheric compounds due to the requirement of fast response sensors. Other micrometeorological methods for obtaining data on the trace gas exchange between atmosphere and surfaces include gradient and eddy accumulation techniques [Fuentes et al., 1996;Desjardins, 1977;Businger and Oncley, 1990]. These are however more indirect, i.e., they are based on empirical parameterizations, except the true eddy accumulation method. They are also often more labor intensive than the eddy covariance method. The true eddy accumulation method, which is a direct flux measurement technique, has proved to be technically very difficult to realize [e.g., Speer et al., 1985]. The relaxed eddy accumulation method, which is technically easier, relies on empirical parameterizations [e.g., Kramm et al., 1999].
The traditional approach for the eddy covariance technique is to sample air continuously. As it is possible to calculate fluxes using only a subset of the whole time-series, one can also take short grab-samples with a relatively long time  [Haugen, 1978;Kaimal and Gaynor, 1983;Lenschow et al., 1994;Rinne et al., 2000]. This disjunct eddy sampling approach relaxes the requirement for fast concentration measurement and makes eddy covariance flux measurements possible using relatively slow sensors.
In this paper a new measurement system for trace gas fluxes applying the disjunct eddy covariance method (DEC) is described. The analysis of grab-samples was done using proton-transfer-reaction mass spectrometry (PTR-MS) [Lindinger et al., 1998]. The method is demonstrated by field measurement of methanol fluxes above an alfalfa field before and immediately after the harvest.

Methods
The

Conclusions
The results show that by using the disjunct eddy sampling (DES) with proton-transfer-reaction mass spectrometer (PTR-MS) it is possible to measure fluxes of biogenic and by inference other VOCs using direct eddy covariance method. As this method does not rely on empirical parameterizations, they do not introduce systematic errors into the measured fluxes. As the DES approach will often require a rather bulky sampler and high sampling rates, the sampler needs to be displaced from the acoustic anemometer. This sets limitations to lower measurement heights. Future applications of disjunct eddy sampling can be classified into three categories depending on the response time of the analyzer used:

1) Instruments with response time too slow for continuous eddy covariance measurements (3 seconds to 3 minutes).
Within the slower end of this range, averaging times longer than one hour are needed to obtain reasonably reliable flux estimates [Haugen, 1978]. This, however, can lead to nonstationarity of the time series. Instruments in this category includes analyzers for mercury (cold vapor atomic fluorescence spectrophotometry), nitrous oxide (fourier transform infrared spectroscopy), ozone (UV absorption analyzers) and carbon monoxide (reducing gas detector).
2) Instruments with relatively fast response time (0.3-3 seconds). These can be used for continuously sampling eddy covariance measurements under suitable conditions but disjunct sampling can extend their applicability. Instruments in this category include e.g. condensational particle counters, tunable diode laser absorption spectrometers, fast isoprene analyzers and most PTR-MS and similar chemical ionization mass spectrometer instruments.
3) Instruments, which are fast enough for continuous eddy covariance and are able to measure multiple compounds but not simultaneously. The disjunct eddy sampling approach could be used by scanning through the compounds of interest, resulting in a disjunct time series of each compound concentration. Faster PTR-MS instruments could be applied to this method. Even though there are instruments capable of measuring many of the compounds mentioned above fast enough for continuously sampling eddy covariance measurements, the disjunct eddy sampling enables eddy covariance measurements with instruments that are often more stable, easy to operate and inexpensive than the fast instruments. The disjunct eddy sampling also extends the field of eddy covariance measurements to many new compounds.