Wind-induced error in the measurement of soil respiration using closed dynamic chambers

https://doi.org/10.1016/j.agrformet.2005.06.004Get rights and content

Abstract

We assess errors in soil respiration fluxes of CO2 obtained using the closed dynamic chamber method. Particular attention is given to small pressure gradients between the chamber headspace and the external environment that may induce mass flow of soil air, leading to overestimation of soil respiration. These pressure gradients develop as air movement creates a Venturi effect at the vent that is designed to insure pressure equilibration, leading to aspiration of air from within the chamber. During field experiments at the Harvard Forest, the Venturi effect produced pressure gradients of approximately 1 Pa per 1 m s−1 for a chamber sealed to an impermeable plate, but no pressure gradient was observed in an identical system deployed on the forest soil. Mass flow of soil air compensated for the wind-driven pressure gradients, and increases in CO2 fluxes exceeding a factor of 2 were observed in response to wind events even under a dense forest canopy. The high porosity of forest soils allows pressure artifacts induced by winds or by sampling flows to perturb the diffusive flux from soils, potentially affecting virtually all chamber methods. Associated errors in soil respiration measurements must be addressed through chamber design and evaluation.

Introduction

Terrestrial uptake of carbon accounts for a significant fraction of the global emission of fossil fuel CO2. Understanding the processes controlling the terrestrial uptake of carbon is needed to predict the future course of atmospheric CO2. Photosynthesis is reasonably well understood from the leaf to whole canopy scale (Farquhar and Sharkey, 1982), but basic understanding of respiration lags behind. In particular, we do not adequately understand the partitioning between heterotrophic and autotrophic respiration in forest soils, the individual processes that contribute to total soil respiration, or the factors that control them across the landscape (Schimel et al., 2001).

Forest soil respiration, or the sum of live root and soil-microbial respiration and decomposition of forest litter and soil organic matter, can account for 80% or more of total forest respiration (Davidson et al., 1998, Wofsy et al., 1993). Because CO2 production rates in soil are large, and diffusion across the soil–air interface is slow, large concentrations of CO2 build up in the soil (Bajracharya et al., 2000) making it difficult to observe the efflux of CO2 without disturbing the flux. Hence, uncertainties in the measurement of rates for forest soil respiration continue to plague accounting of the forest carbon budget (Andrews et al., 1999).

The key to an unbiased soil respiration measurement is to accurately measure the CO2 evolved from the soil surface without perturbing the diffusion gradients or inducing mass exchange. Eddy covariance measurements near the forest floor (Baldocchi et al., 1988) or flux gradient approaches (Liu and Foken, 2001) have the advantage of minimally disturbing the soil, but it can be very difficult to identify the specific area that contributes to observed fluxes. Furthermore, there are systematic errors in eddy flux measurements, they are expensive to implement, and both sub-canopy eddy flux and gradient measurements include contributions from above ground respiration and photosynthesis by vegetation below the sensor.

This paper examines artifacts associated with measurements of soil CO2 fluxes using “closed dynamic enclosures” placed over the soil. Chamber-based measurements are direct observations and are relatively easy to implement, and therefore are widely used for calculating forest soil respiration. There are three principal types of enclosures used: static, open dynamic and closed dynamic. Static chambers seal a plot of soil and observe the rise of CO2 concentration by chemical absorption and analysis (known as non-flow-through steady-state or NFT-SS) or periodic collection of an air sample by syringe (known as non-flow-through non-steady-state or NFT-NSS) (Hutchinson and Rochette, 2003). This method has been common historically, but the build up of CO2 and H2O in the chamber (or the depletion chemical absorption) and the pressure pulses associated with sampling provide obvious perturbations to the observed rates of CO2 increase in the chamber.

The availability of continuous gas analyzers has made dynamic (“flow-through”) chambers the more common approach recently. An open dynamic chamber (also known as flow-through steady-state or FT-SS; Livingston and Hutchinson, 1995) continuously flushes ambient air through the chamber, allowing calculation of the soil flux from the difference in CO2 concentration between air entering and leaving the chamber (Lund et al., 1999). The build up of CO2 in the chamber is reduced, but air flow through the chamber could introduce pressure differentials that induce mass flow through the soil, and the circulation may perturb diffusion through porous upper soil layers.

An alternative approach is a closed dynamic chamber, which has also been labeled as flow-through non-steady-state or FT-NSS (Livingston and Hutchinson, 1995). Generally, an external CO2 analyzer (such as infrared gas analyzer) is attached to the chamber and air is re-circulated through the chamber and analyzer. In principle, an in situ open-path analyzer could be used inside the chamber, thus eliminating the need for pumping. These systems do not achieve a steady-state; after sealing the chamber, CO2 concentrations increase in the headspace and the time history is used to calculate flux, typically by extrapolating the observed rate of CO2 increase back to the time the chamber was placed on the soil. A small opening to the atmosphere, usually a short section of tubing, is included in order to eliminate pressure gradients between the inside and outside of the chamber and to reduce the effects of placing the chamber on the soil (Hutchinson and Livingston, 2001). By making the sampling duration short (Healy et al., 1996) and extrapolating the rise in CO2 concentration to the starting point, biases due to altering the concentration gradient are minimized.

Closed dynamic chamber systems are widely used (Norman et al., 1992, Davidson et al., 1998), although there remains the potential for pressure gradients to develop between the internal chamber and external environment. Kanemasu et al. (1974) demonstrated that chamber pressure gradients of a few Pascal have the potential to create errors in gas flux measurement as large as an order of magnitude. This error is primarily attributed to the pressure-induced mass flow of soil gases, which are predominantly diffusion-driven under natural conditions (Kimball and Lemon, 1971).

Pressure gradients in normally operating closed dynamic chambers have not been reported, leading many to conclude that these systems are relatively free of systematic measurement error. However, Goulden et al. (1996) and Davidson et al. (1998) both found that chamber-based measurements of soil respiration fluxes were greater by 50% or more than those measured by the eddy correlation method over the same area. This disparity in measured respiration was particularly noteworthy during windy periods according to Goulden et al. (1996). Davidson et al. (1998) concluded that “no plausible explanation” exists for the systematic overestimation of soil respiration by closed dynamic chambers, although several other researchers noted concerns about the effects of wind on chamber-based flux measurements (e.g. Norman et al., 1992, Matthias et al., 1980).

Of particular interest, Conen and Smith (1998) concluded that wind movement around the vent of a closed chamber created a “Venturi effect” leading to overestimated soil gas efflux. Conen and Smith (1998) proposed that wind de-pressurized the chamber by pulling air out of the chamber headspace, leading to the mass flow of soil gases from the permeable soil column into the chamber interior. The chamber de-pressurization was approximated to 0.6V2, where V is equal to wind speed in m s−1, or a 2.4 Pa pressure deficit for a steady 2 m s−1 wind, which resulted in a 233% increase in measured soil emissions. Even under very calm conditions (V < 1 m s−1), systematic errors of 10–50% might be expected. Recently, Davidson et al. (2002) noted chamber pressurizations of 0.9 Pa under windy conditions, raising the question of the accuracy of chamber-based measurements under such conditions and suggesting further research on the phenomenon.

We performed a series of experiments under both controlled and natural conditions to document the pressure gradients induced by wind across a respiration chamber of conventional design, and to examine their affect on observed respiration rates. Pressure gradients by themselves would not create a bias in respiration measurements—it is the mass flow of soil air in or out of the soil column that adds an advective flux of CO2 in addition to the unperturbed diffusive flux. To quantify the true driving force for mass flow in soil chambers, we compared the pressure gradients in a chamber placed over an impermeable surface to the gradients in a chamber over natural soil.

Section snippets

Materials and methods

Two parallel, closed dynamic chamber systems were constructed for comparison during field experiments, arranged as shown in Fig. 1. Each closed chamber system utilized a 5 L PVC chamber with a vertically oriented vent tube (0.19 cm vent tube inner diameter, 3.56 cm vent tube length). A pump circulated air though the chamber at approximately 0.5 L/min in a closed loop through a LI-COR (Lincoln, NE) infrared gas flux analyzer (IRGA)—one system used a LI-6252 CO2 Analyzer and the other used a LI-6262

Results

Under fan-controlled wind conditions in the field, the chamber system deployed on an impermeable plate consistently developed a slight vacuum relative to the outside in response to horizontal wind—the “Venturi effect” proposed by Conen and Smith (1998) (Fig. 2) that de-pressurizes the chamber relative to the soil underneath. The Venturi effect was demonstrated across 13 field trials (Fig. 3), yielding an approximately 1 Pa drop in pressure per 1 m s−1 increase in horizontal wind at low wind

Discussion

It is evident that natural wind conditions can induce selective systematic error in vented, closed dynamic chamber systems through the Venturi effect, providing a plausible explanation for the overestimation of forest soil respiration by chambers as discussed by Goulden et al. (1996). Our results highlight two notable difficulties in quantifying error caused by the Venturi effect: quantification of the “true” internal pressure of soil-deployed chambers and turbulent pressure anomalies.

Several

Conclusion

These results provide evidence for systematic overestimation of soil respiration by the dynamic chamber method when exposed to even weak surface winds, as suggested by Goulden et al. (1996), and attributed to the “Venturi effect” by Conen and Smith (1998). But these results also highlight greater potential biases in the chamber methodology, noting potential error in measuring the “true” chamber pressure gradient over a soil column and demonstrating anomalous pressure effects unexpected in the

Acknowledgements

We especially thank Dave Saum of the Infiltec Corporation for assistance engineering the DM4 Micromanometer for the purposes of this project. We thank the Harvard Forest and its staff for hosting and supporting our research site. We thank the Harvard College Research Program and Federal Work-Study program for supporting William Bain and the Harvard Forest Summer Undergraduate Internship Program for supporting David Patterson. This research was supported by the Office of Science, Biological and

References (22)

  • E.A. Davidson et al.

    Minimizing artifacts and biases in chamber-based measurements of soil respiration

    Agric. For. Meteorol.

    (2002)
  • J.A. Andrews et al.

    Separation of root respiration from total soil respiration using C-13 labeling during free-air carbon dioxide enrichment (FACE)

    Soil Sci. Soc. Am. J.

    (1999)
  • R.M. Bajracharya et al.

    Erosion effects on carbon dioxide concentration and carbon flux from an Ohio Alfisol

    Soil Sci. Soc. Am. J.

    (2000)
  • D.D. Baldocchi et al.

    Measuring biosphere–atmosphere exchanges of biologically related gases with micrometeorological methods

    Ecology

    (1988)
  • C.C. Barford et al.

    Factors controlling long- and short-term sequestration of atmospheric CO2 in a mid-latitude forest

    Science

    (2001)
  • F. Conen et al.

    A re-examination of closed flux chamber methods for the measurement of trace gas emissions from soils to the atmosphere

    Eur. J. Soil Sci.

    (1998)
  • E.A. Davidson et al.

    Soil water content and temperature as independent or confounded factors controlling soil respiration in a temperate mixed hardwood forest

    Global Change Biol.

    (1998)
  • G.D. Farquhar et al.

    Stomatal conductance and photosynthesis

    Annu. Rev. Plant Physiol.

    (1982)
  • M.L. Goulden et al.

    Measurements of carbon sequestration by long-term eddy covariance: methods and a critical evaluation of accuracy

    Global Change Biol.

    (1996)
  • R.W. Healy et al.

    Numerical evaluation of static-chamber measurements of soil–atmosphere gas exchange: identification of physical processes

    Soil Sci. Soc. Am. J.

    (1996)
  • G.L. Hutchinson et al.

    Vents and seals in non-steady-state chambers used for measuring gas exchange between soil and the atmosphere

    Eur. J. Soil Sci.

    (2001)
  • Cited by (0)

    View full text