| Quantifying soil organic carbon（SOC）dynamics is critical to monitoring agricultural sustainability and mitigating global climate change.With the increasing concern of global warming,it has been a hotspot whether soil is carbon（C）source or C sink.Model simulation has been widely used in studying C turnover process and source-sink effects.How to improve the prediction accuracy of these models is an important issue in the current academic community.Denitrification-Decomposition（DNDC）model was adopted in this study because of its friendly interface,simple input parameter and powerful simulation.This study took Yangtze-Huai River Basin as the study area,which is eco-fragile and medium-low yield land.The parent material was homogenous with Xiashu loess,making it an ideal area to study the anthropogenic influencing factors（e.g.,agricultural management practices）of SOC.Based on high-density sampling data,detailed agricultural management parameter,meteorological data,topographic and vegetation data,this study assessed SOC dynamics by DNDC over a large region with high spatial resolution,elucidated in detail the possible influence of natural and anthropogenic factors on the spatial variability of SOC,and quantified the effects of various agricultural management practices on soil C sequestration.The main results of this study were as follows:（1）Soil fertility in the study area was low on the whole,with the mean SOC content lower than the national and provincial level.The SOC contents showed distinct spatial variability with different land uses and agricultural management practices.（2）Since the 1980s,the SOC content in the study area has shown an increasing trend,which accounted for 57.7%of the whole study area.Specifically,the proportion of level 5（lacking）soil decreased from 45.9%in 1981 to 0.06%in 2003.The proportion of soil in level 6（very lacking）soil was 10.6%in 1981,but no longer existed in 2003.In 2003,the proportion of soil in level 4（relatively lacking）accounted for 96.6%of the whole study area.（3）Among the environmental factors,soil texture and p H played more important role in SOC content than elevation,slope,normalized difference vegetation index（NDVI）and topographic wetness index（TWI）.To increase fertilizer application and straw return had significant effects on SOC accumulation.（4）The pattern of the simulated SOC content matches well with the observations.Both the correlation coefficient（R2）and the concordance index（CI）were high between observed values and simulated values(R2=0.74**,p<0.01;CI=0.87).Therefore,it could be concluded that the batch-site simulation,which was based on large-quantity and high-density,facilitated improving the simulation accuracy and reducing the uncertainty caused by limited sites.（5）There were 71%of the study area identified as C sink areas,29%with C balance,and no areas experienced net C loss.Therefore,the current management practice facilitates the sequestration of C in agricultural soil.The SOC stock at 0-20 cm increased from 6020 Gg C in 2003 to 6301 Gg C in 2016,with the mean sequestrating rate of 38.3 kg C hm-2 y-1.The C sink areas were mainly distributed in paddy soil used as paddy land,while the C balance areas were found mainly in yellow-cinnamon soil,which is usually used as upland.（6）Comprehensive practices,such as organic manure with straw return and no tillage,are more efficient than single management to sequestrate more C.Optimizing agricultural management practices can further improve SOC stock,among which straw return is the most significant factor.The annual increase rate of SOC under 60%straw return reached as high as 402%,much higher than those of other single management practices.Therefore,straw return played the most important role in C sequestration among these agricultural management practices. |
