Comparison of different chamber techniques for measuring soil CO2 efflux

Twenty chambers for measurement of soil CO2 efflux were compared against known CO2 fluxes ranging from 0.32 to 10.01 mumol CO2 m(-2) s(-1) and generated by a specially developed calibration tank. Chambers were tested on fine and coarse homogeneous quartz sand with particle sizes of 0.05-0.2 and 0.6 mm, respectively. The effect of soil moisture on chamber measurements was tested by wetting the fine quartz sand to about 25% volumetric water content. Non-steady-state through-flow chambers either underestimated or overestimated fluxes from -21 to +33% depending on the type of chamber and the method of mixing air within the chamber's headspace. However, when results of all systems tested were averaged, fluxes were within 4% of references. Non-steady-state non-through-flow chambers underestimated or overestimated fluxes from -35 to +6%. On average, the underestimation was about 13-14% on fine sand and 4% on coarse sand. When the length of the measurement period was increased, the underestimation increased due to the rising concentration within the chamber headspace, which reduced the diffusion gradient within the soil. Steady-state through-flow chambers worked almost equally well in all sand types used in this study. They overestimated the fluxes on average by 2-4%. Overall, the reliability of the chambers was not related to the measurement principle per se. Even the same chambers, with different collar designs, showed highly variable results. The mixing of air within the chamber can be a major source of error. Excessive turbulence inside the chamber can cause mass flow of CO2 from the soil into the chamber. The chamber headspace concentration also affects the flux by altering the concentration gradient between the soil and the chamber. (C) 2003 Elsevier B.V. All rights reserved.