Sensible heat bias in open-path eddy covariance carbon dioxide flux measurements

Abstract:

The widely observed differences between net carbon dioxide (CO₂) flux estimates derived from eddy covariance systems deploying open- and closed-path infrared gas analyzers (IRGAs) pose a major challenge for site intercomparison studies. Our limited knowledge about potential systematic biases in the derivation of CO₂ flux estimates by these two types of systems hampers our ability to detect significant differences in CO₂ flux measurements made at contrasting ecosystems. Here we explore potential systematic biases in CO₂ fluxes measured with two open-path IRGAs. Comparison of fluxes from open- (EC150 & IRGASON, Campbell Scientific Inc.) and (en)closed-path IRGAs (LI7000 & LI7200, LI-COR Biosciences) at a northern peatland and a northern boreal forest site revealed consistent differences in CO₂ flux estimates across a wide range of environmental conditions. These differences directly scaled with the magnitude of the sensible heat flux indicating a selectively systematic bias in open-path CO₂ flux measurements due to the temperature sensitivity of the CO₂ density measurements. We present two empirical correction procedures: the “direct” approach requires data from a limited period of concurrent CO₂ flux measurements by open- and closed-path IRGA-based eddy covariance systems, whereas the second approach only requires wintertime CO₂ flux data from the open-path IRGA. The “direct” approach effectively removes the bias in the open-path CO₂ flux measurements and results in remaining differences with the closed-path CO₂ fluxes smaller than 0.5 µmol m^-2 s^-1. In contrast, the “wintertime” approach seems to overcompensate for the sensible heat effects with differences remaining between 0.9 µmol m^-2 s^-1 and 1.8 µmol m^-2 s^-1. When a high-frequency air temperature is used to compensate for the temperature sensitivity of the CO₂ density measurements, open- and closed-path CO₂ flux agree within ±0.5 µmol m^-2 s^-1, similar to the “direct” post-processing correction. These findings highlight the importance of corrections for systematic errors in open-path CO₂ flux measurements to ensure intercomparability among years at a given ecosystem and among contrasting ecosystems