The role of soil structure for the moisture response function of carbon mineralization

Soil organic carbon (SOC) is critical for soil quality and agroecosystem sustainability and soil moisture is a key component in the response of SOC turnover to climate variations. Several processes are involved in the regulation of SOC decomposition by soil moisture, but O₂ transport in large structural pores that drain near saturation is particularly important in wet soils, which may become more common with increasing flood risk under climate change. These effects are poorly understood because standard incubation experiments disrupt soil structure by sieving. We therefore investigated the effects of soil structure on C mineralisation under wet conditions in order to improve models of SOC turnover that take into account soil structure and soil management.

We measured CO₂ emissions from soil cores of contrasting structure in laboratory incubations and derived different parameter sets of moisture response functions. This was done at pressures ranging from saturation to -600 cm, using intact or sieved soils from conventional tillage and no-till treatments. An analytical stochastic model of C mineralisation under different climatic conditions was developed and run using the contrasting parameter sets derived from the incubation data, and the consequences of neglecting soil structure were quantified by differences in model predictions.

The functions describing the response of C emission rates to soil moisture had different shapes for soils of contrasting structure (i.e. between sieved and intact cores or between the two different tillage treatments). The optimum degree of saturation for C emission rates (i.e. where rates were maximal) was closer to one for the more structured soils: 0.90 for intact no-till cores, 0.85 for intact till cores and 0.70 for sieved cores. In addition, sieving increased C mineralisation rates at saturation. Differences between tillage treatments were also evident in the drier range, with C emission rates decreasing more rapidly as the soil was drained from the optimum degree of saturation to the driest pressure head of - 600 cm for the soil from the conventional tillage treatment.

Predictions of C emission rates with the analytical model parameterized using the response curves from sieved or intact soil cores diverged rapidly with increasing rainfall. These differences increased to a plateau as soil conditions became wetter for both tillage treatments, but were always higher for the no-till treatment.

We conclude that neglecting soil structure or changes in soil structure in dynamic predictions of soil organic carbon stocks in response to climatic variations can lead to significant errors. We suggest that a revision of the static view of moisture response functions of C mineralisation is needed. More efforts should be made to establish theoretical or empirical links between soil structural characteristics (in particular the occurrence and distribution of structural pore space) and the parameters of the response function.