Prototyping NEON’s eddy covariance measurements: Comparison of two closed-path infra-red gas analyzers

Background/Question/Methods

The Fundamental Instrument Unit is designed to integrate the ecological drivers, responses and interactions among the ecosystem-level soil-plant-atmosphere continuum, and enable consistent sampling at the continental scale over many decades, with consistent sampling—to act uniformly as one single integrated observatory enabling multiple scales of inference.  In such a way, NEON plans to measure the Net Ecosystem Exchange of Carbon—a direct estimate of productivity at all 60 NEON sites to understand and predict the carbon balance at continental scale and how productivity interacts with biotic measures, e.g., changes in biomass, plant communities, biodiversity.  The two challenges are i) to apply this technique across the full range of environmental conditions found across the continent, and ii) to minimize the errors and uncertainties for CO₂ measurement between sites over 30-year of NEON lifespan.

For the past decades, researchers had the choice of 2 types of CO₂ analyzers, open- and closed-path.  Open-path analyzers measure the in-situ turbulent environment, low power, and stable, but subject to a density correction that can be several orders of magnitude of the measured flux.  Closed-path analyzers can be automated with calibration gases, but they attenuate the high frequency turbulence, have larger power consumption and climate-controlled demands—yet have lower uncertainties because the errors can be estimated and accounted for.  Here, we test a new analyzer, a hybrid between open- and closed-path environments (Li-7200) against a climate-controlled design of a closed-path analyzer (Li-7000).  This experiment was conducted over a grass prairie, 30 km from Boulder.  A single 3-D sonic anemometer is also deployed to enable flux calculations.  The parameters that will be compared include CO₂/H₂O concentrations, sensible heat, latent heat, as well as CO₂ flux, and stability of calibrations.

Results/Conclusions

The integration of the Li-7200 had better ability to capture the high frequency, turbulent timescales compared to the Li-7000, and both analyzers preformed well sensing the timing and magnitude of larger seasonal patterns.  All parameters are compared well against 1:1 regression line.  In a few cases, the Li-7200 overestimated fluxes compared to the Li-7000 by a few percent, we interpret this result as having better frequency response.  The Li-7200 also had operational advantages: little data loss during precipitation; lower power consumption and higher field stability than Li-7000; can output mixing ratio thus eliminating need for density corrections, which can potentially increase flux data quality and temporal resolution; and calibrations can be automated, like the previous closed-path configuration.