Relative humidity effects on water vapour fluxes measured with closed-path eddy-covariance systems with short sampling lines

It has been formerly recognised that increasing relative humidity in the sampling line of closed-path eddy-covariance systems leads to increasing attenuation of water vapour turbulent fluctuations, resulting in strong latent heat flux losses. This occurrence has been analyzed for very long (50 m) and long (7 m) sampling lines. To date, only a few analytical or in situ analyses have been proposed to quantify and correct such effects, among which the comprehensive method by Ibrom et al. (2007) was proved effective for the very long sampling line of a forest eddy-covariance setup.Here we analyze data from eddy-covariance systems featuring short (4 m) and very short (1 m) sampling lines running at the same clover field and show that relative humidity effects persist also for these setups, and should not be neglected. Starting from the work of Ibrom and co-workers, we propose a mixed method, a composite of two existing approaches, for correcting eddy-covariance fluxes. By means of a comparison with parallel open-path measurements, we show that the mixed method leads to an improved estimation of latent heat fluxes, with respect to the method described by Ibrom et al. (2007). The quantification and correction method proposed here is deemed applicable to closed-path systems featuring a broad range of sampling lines, and indeed applicable also to passive gases as a special case. The methods described in this paper are incorporated, as processing options, in the free and open-source eddy-covariance software packages ECO2S and EddyPro.

► RH-induced attenuations of H2O in closed-path instruments using short tubes is relevant.
► Procedures to assess and correct such attenuation are proposed and evaluated.
► The procedures are proved accurate for a broad range of sampling lines.
► The procedures are proved especially suited for measurements close to smooth surfaces.
► The procedures are available for free in the eddy covariance software EddyPro.