Abstract :
[en] The escalating pressures of food demand and climate change have heightened global interest in sequestering carbon (C) in agricultural soils. Soil organic C (SOC) increases with C input, but the C sequestration efficiency (CSE), i.e., the conversion ratio of C input to SOC, varies with the type and amount of C input. Comparing CSE across regions and/or fertiliser treatments is challenging due to complex interactions between climate, soil properties, and fertiliser practises. Therefore, a comprehensive study of the individual effects of these factors on CSE under long-term fertilisation is urgently needed. Furthermore, limited C inputs do not fully reflect the linear-to-asymptotic behaviour of SOC dynamics, and classical long-term experiments with substantial C input accumulation are essential for understanding soil CSE characteristics. Model predictions are necessary for investigating CSE in soils that are far from C saturation, and future climate change, a pivotal factor affecting crop growth and C cycling, must be considered in these predictions. This research first studied the soil CSE and its key drivers in the plough layer under different fertilisations based on data from eight long-term experiments established in 1980s in China and four classical long-term experiments conducted over a century in the UK and USA. Then, the SPACSYS model’s performance in simulating long-term SOC dynamics in the Broadbalk continuous winter wheat system was validated. The impact of future climate change on wheat yield and SOC stock was also examined. Finally, four long-term experiments with the same soil type (i.e., Mollisol) and similar cropping systems were chosen to compare CSE at different C input accumulation levels between Northeast China and the USA. Fertiliser practises favourable for improving soil CSE and mitigating future climate change were recommended.
The eight long-term experiments in China indicated that the CSE has remained relatively constant over nearly four decades of fertilisation, with SOC linearly increasing with C input. Variance partitioning analysis (VPA) illustrated that the CSE of the main dryland region of China was mostly controlled by edaphic characteristics, especially the soil C/N ratio and clay content, followed by C input and climate. Manure amendment had the most significant fertilisation effect on SOC sequestration with an average CSE of 14.9%, followed by chemical fertilisation (9.0%) and straw return treatment (7.9%). VPA and structural equation modelling (SEM) showed that the improvement of soil nutrients and clay content controlled CSE, highlighting the main positive direct effect of soil chemical properties in the manure amendment treatment. Soil C/N ratio and pH were main explanatory factors influencing CSE in the long-term chemical NPK treatment. The negative impact of C input was the main driver of the CSE under straw return treatment due to the low humification for crop straw and low stabilization of straw-derived C.
Based on the four classical long-term experiments with SOC showing nonlinear changes over time, the negative exponential dynamics of soil CSE was fully captured by the model (CSE=b+a*e-k*Cinput, P < 0.01). The CSE decreased quickly when C input started to pile up, until the accumulated C input reached 123–315 t C ha–1 in typical Chromic Luvisols and 93–145 t C ha–1 in typical Mollisols. Subsequently, it continued to decrease more slowly towards an asymptotic value (i.e., b) as a large amount of C input had been accumulated. Extreme gradient boosted tree (XGBoost) regression modelling showed that C input, MAT and TN stock were crucial factors of soil CSE at all sites with more than a century of fertilisation. Practises with manure application had more significant fertilisation effects on SOC sequestration rates and efficiency than did those with chemical fertilisation. A larger CSE asymptotic value was also observed in the manure applied treatments (2.45–10.25%) than in the chemical fertiliser treatments (0.48–3.85%).
Parametrized by data from the longest and well-documented Broadbalk continuous wheat experiment over a century, the SPACSYS model was capable of simulating the grain and straw yields of various tall and modern short-strawed varieties of winter wheat, as well as the dynamics of SOC and TN stocks in the plough layer. Without changing cultivars and managements, wheat yield could significantly enhance by 5.8–13.5% under future climate conditions compared to the baseline scenario. Meanwhile, the SOC stock was found to increase for all studied fertiliser practises under the scenarios by 0.19–3.99%, except for the NPK fertiliser practises under RCP2.6, which showed a decrease of 0.27–1.08%. Increased C input through “CO2-fertilisation effects” can compensate C losses by soil respiration under the RCP scenarios. Manure application practises can be considered as a sustainable strategy for enhancing wheat yield and soil C sequestration under future climate change.
Verified by the Harbin and Gongzhuling Mollisols experiments in Northeast China, the SPACSYS model well simulated the yields of corn, spring wheat and soybean, as well as the SOC stock. Without changing cultivars and managements, future climate change, particularly increased temperatures, reduced corn (14.5% for Harbin and 13.3% for Gongzhuling) and soybean yields (10.6%). SOC stocks were projected to decrease by 8.2% for Harbin and 7.6% for Gongzhuling by 2100 under the RCP scenarios. Future climates could also lead to average CSE reductions of 6.3% for Harbin and 12.1% for Gongzhuling. Manure combined with chemical fertiliser is recommended as a fertiliser practise to mitigate negative future climate effects on crop yield and SOC sequestration. Mollisol CSE also showed an exponential decrease with C input in the future. Additionally, the average Mollisol CSE at two sites in the USA (13.6–27.4%) due to manure application for all fields was higher than that in Northeast China (9.2–27.1%) during the rapid decline phase of CSE. During the asymptotically stable stage, the Mollisol CSE at two sites in the USA (0.2–6.8%), with higher temperatures and large C inputs from straw return, was lower compared to that in Northeast China (3.8–10.6%).
In conclusion, soil CSE remained constant across fertiliser treatments when SOC linearly increased with C input. Edaphic factors were more important CSE controllers than C input and MAT. However, when SOC showed non-linear changes with the accumulation of a large amount of C input, the CSE exponentially decreased and eventually approached an asymptote. The importance of C input and MAT to soil CSE promoted ahead of soil properties. Additionally, manure application was a recommended fertiliser practise not only for improving CSE under long-term fertilisation but also for mitigating adverse future climate impacts on SOC sequestration. These findings are of great significance for guiding sustainable agricultural practises, enhancing C sequestration, and addressing climate change in the study area.
