Sieving soil before incubation experiments overestimates carbon mineralization but underestimates temperature sensitivity.

The sensitivity of soil organic carbon (SOC) mineralization to temperature could affect the future atmospheric CO levels under global warming. Sieved soils are widely used to assess SOC mineralization and its temperature sensitivity (Q) via laboratory incubation. However, sieved soils cause a temporary increase in mineralization due to the destruction of soil structure, which can affect estimates of SOC mineralization, especially in soils managed with no-till (NT). To identify the effects of soil sieving on SOC mineralization and Q, soil was collected from an 11-year field experiment under a wheat-maize cropping system managed with a combination of tillage [NT and plow tillage (PT)] and residue [residue returning (RR) and residue removal (R0)]. Soil was either sieved or left in an undisturbed state and incubated at 15 °C and 25 °C. SOC mineralization in sieved soils at 25 °C was 47.28 g C kg SOC, 160.1% higher than SOC mineralization in undisturbed soils (P < 0.05). Interestingly, Q values in sieved soils were 1.29, 77.6% lower than Q in undisturbed soils (P < 0.05). Highly significant correlations (P < 0.01) were observed between sieved and undisturbed soils for SOC mineralization (r = 0.85-0.98) and Q (r = 0.78-0.87). Soil macro-aggregates had lower SOC mineralization by 6.1-21.9%, but higher Q values by 4.7-6.5% compared with micro-aggregates, contributing to lower mineralization and higher Q under NT and RR. Furthermore, structure equation and random forest modelling showed that increased SOC contents in NT and RR could not only reduce SOC mineralization, but also constrained the improvement of Q in NT and RR. Overall, these results indicated that although sieved soils overestimated SOC mineralization and underestimated Q due to the destruction of macro-aggregates, the patterns between treatments were similar and sieving soil for incubation is considered as a suitable approach to evaluate the relative impacts of NT and RR on SOC mineralization and Q.