SAMPLING SOIL CO2 FOR ISOTOPIC FLUX PARTITIONING: NON STEADY STATE EFFECTS AND METHODOLOGICAL BIASES

Abstract:

Measurements of δ¹³C of soil CO₂ are used to partition the surface flux into autotrophic and heterotrophic components. Models predict that the δ¹³CO₂ of the soil efflux is perturbed by non-steady state (NSS) diffusive conditions. These could be large enough to render δ¹³CO₂ unsuitable for accurate flux partitioning. Field studies sometimes find correlations between efflux δ¹³CO₂ and flux or temperature, or that efflux δ¹³CO₂ is not correlated as expected with biological drivers. We tested whether NSS effects in semi-natural soil were comparable with those predicted. We compared chamber designs and their sensitivity to changes in efflux δ¹³CO₂.

In a natural soil mesocosm, we controlled temperature to generate NSS conditions of CO₂ production. We measured the δ¹³C of soil CO₂ using in situ probes to sample the subsurface, and dynamic and forced-diffusion chambers to sample the surface efflux. Over eight hours we raised soil temperature by 4.5 ^OC to increase microbial respiration. Subsurface CO₂ concentration doubled, surface efflux became ¹³C-depleted by 1 ‰ and subsurface CO₂ became ¹³C-enriched by around 2 ‰. Opposite changes occurred when temperature was lowered and CO₂ production was decreasing. Different chamber designs had inherent biases but all detected similar changes in efflux δ¹³CO₂, which were comparable to those predicted. Measurements using dynamic chambers were more ¹³C-enriched than expected, probably due to advection of CO₂ into the chamber.

In the mesocosm soil, δ¹³CO₂ of both efflux and subsurface was determined by physical processes of CO₂ production and diffusion. Steady state conditions are unlikely to prevail in the field, so spot measurements of δ¹³CO₂ and assumptions based on the theoretical 4.4 ‰ diffusive fractionation will not be accurate for estimating source δ¹³CO₂. Continuous measurements could be integrated over a period suitable to reduce the influence of transient NSS conditions. It will be difficult to disentangle biologically driven changes in soil δ¹³CO₂ from physical controls, particularly as they occur on similar timescales and are driven by the same environmental variables, such as temperature, moisture and daylight.