Soil carbon cycle measurement and modelling in two crop managements and two climates.

The potential of soil carbon sequestration (SCS) in arable soil is well-established; however, it is difficult to estimate it with confidence, due to both: soil organic carbon (SOC) large variation in time and space, and large uncertainty in measurement methods. RothC model is a widely used model of C dynamics in soil often used to estimate SCS, but different versions exist for different pedoclimatic and agronomic conditions in which it is applied. Here, two management practices, Conventional Cropping System (CCS) and Efficient Cropping System (ECS, featuring catch crops and optimized field operations according to predicting models), are compared in two Italian sites (Ravenna and Foggia) with different pedoclimatic conditions. All the C input to the soil were measured, as well as the weather conditions, soil temperature, soil moisture, and SOC. Moreover, total soil respiration (Rs) was measured using automated stationary gas chambers. Then, the ability of two different versions of RothC (standard and modified for Mediterranean climate by Farina et al. 2013) to simulate adequately the field observation is investigated; more precisely, the RothC versions were calibrated using field data, and their output for soil water dynamics, SOC stock and soil heterotrophic respiration (Rh) were compared with field observations. In order to compare total soil respiration with RothC estimates of soil heterotrophic respiration, a method to partition Rs into autotrophic respiration (Ra) and Rh was devised. The results of this thesis show that: (a) Rh estimated using the partitioning method and the Rh estimated using RothC compared well, especially in the Ravenna site; (b) none of the RothC versions were able to properly reproduce the soil water regime observed in the two sites; (c) no SOC trend could be determined using field measurements. However, RothC model calibrated using Rh can be used to hypothesize future SOC trends.