Temperature influences the rates of virtually all biochemical processes and, as a result, rates of important ecosystem processes such as productivity and respiration. Quantifying how ecosystem process rates vary across broad ranges of temperature is essential to testing and improving ecosystem carbon-flux models. Soil respiration is one such process: it is the second largest land-atmosphere carbon flux, is temperature sensitive, is widely studied, and many algorithms exist to describe its variation with temperature. Nevertheless, a long-standing question persists: “Will climate warming stimulate soil respiration such that a positive feedback develops, with warmer temperatures stimulating soil-CO2 production, thereby increasing atmospheric CO2 concentrations, warmer temperatures, and yet more soil-CO2 production?” There is sufficient soil carbon to allow such a feedback to persist, should it develop. Abundant evidence supports the proposition that soil respiration rates do increase with temperature in all cases except where water limitations exist. The concern that warmer temperatures could stimulate further warming cannot be ruled out. I address this hypothesis again, using large, multiannual and multisite datasets of in situ soil-CO2 emissions collected from highly scattered locations within sites with modern soil gas-exchange instruments, and applied statistical analyses of temperature-flux relationships to define best-fit relationships between soil-CO2 emissions and temperature.
Results/Conclusions
As has been found elsewhere, I found that (i) soil-CO2 emissions (Rsoil) strongly and positively correlated with surface-soil temperatures, (ii) that linear relationships were significantly positive, (iii) and that use of ln(Rsoil) instead of Rsoil as the dependent factor improved model fits. These are common findings. However, large datasets include broad ranges of independent factors and allow for the evaluation of warm-end fluxes. In my case, addition of a quadratic (temperature-squared) term resulted in higher model R2 and lower AIC, which is unusual. I thus tested a second, entirely independent dataset from a very different place, with the same results. These results indicate that Rsoil per unit of soil area does increase more slowly at warmer soil temperatures.