High-frequency attenuation in eddy covariance measurements from the LI-7200 IRGA with various heating and filter configurations – a spectral correction approach

• A new method to determine the cut-off frequency for flux corrections is presented. • High frequency filtering of EC data causes flux underestimation. • Use of the LI-7200 IRGA heated tube reduced filtering effects. • Filtering effects of CO 2 were small or negligible for our site and set-up. • Filtering effects of H 2 O worsen with increasing relative humidity. The use of (en)closed-path Infrared Gas Analysers (IRGA) in the measurement of Eddy Covariance (EC) fluxes results in inadvertent high-frequency tube attenuation due to diffusion and mixing of sampled gas inside the tube. The application of tube heating and installation of particulate filters along the tube length also contributes to high-frequency attenuation. The goal of this research is first, to quantify the attenuation effects of different tube heating and filter configurations on CO 2 and H 2 O fluxes. And second, to present a modified power spectral approach (PSA) based on theoretical power spectra to calculate the effective cut-off frequency fc . Measurements for each experimental configuration were performed at an Integrated Carbon Observation System (ICOS) station equipped with the standard LI-7200 enclosed-path IRGA and Gill HS-50 3D sonic anemometer. Correction factors for each dataset were determined and implemented in post-processing. We found only very small attenuation effects of CO 2 fluxes between the examined configurations. In agreement with previous studies, we found attenuation worsens with increasing relative humidity rH , in the fluxes of H 2 O. As expected, the highest (best) f c for H 2 O was found in the lowest examined rH class of 45-50 % with the configuration of heating on, no filter. The lowest (worst) f c for H 2 O was in the highest rH class of 90-95 % with the configuration of heating off with the 7 μm filter. Our results confirm that tube attenuation effects for the standard ICOS setup are negligible for CO 2 and small for H 2 O, depending on tube heating settings and use of particulate filters. We also show that the post-processing of attenuation effects, especially for H 2 O, could improve the accuracy of long-term EC measurements. We recommend that this novel approach be considered by users of datasets collected with the LI-7200 enclosed-path IRGA.