Study On The Evolution Of Soil Physicochemical Properties, Carbon Stabilization Mechanisms And Bacterial Community Succession Under Long-term Cultivation After Reclamation Of Estuarine Wetlands

Agricultural soil development is one of the critical drivers of long-term agroecosystem evolution and greatly influences agricultural sustainability. Understanding how soil would change under long-term agricultural use would provide scientific bases for making wise land use strategies and help to predict the dynamics of agroecosystem development. However, agricultural soil development has been rarely investigated from ecological perspectives, while the traditional pedologic studies mainly focus on the pedogenesis in soil development, with little attention paid to soil biogeochemistry and microbiology during this process. This study, from aperspective of ecosystem ecology, aimed to explore how the mineral part of soil, soil carbon and microbial communities might function together to drive the long-term agricultural soil development. The basic pedogenetic processes, carbon stabilization mechanisms and soil bacterial community succession were studied to depict the pattern of soil development during500years of cultivation after wetland reclamation on the Chongming Island in the estuary of the Yangtze River. The effects of recent land use (paddy-dry rotation vs. upland cropping) on soil development were also investigated.On the Chongming Island, soils that have been reclaimed for8,16,40,75,120,300and500years were extensively sampled (mainly in topsoil) in the wetlands and in the paddy/upland croplands that had persisted for8to100years to date. Main soil physicochemical properties were analyzed to understand the basic pedogenetic processes. Soil carbon fractionation methods were used to study the carbon stabilisation mechanisms. To study the taxonomic composition, diversity and community structure of soil bacteria, the techniques of Terminal Restriction Fragment Length Polymorphism (T-RFLP) and454pyrosequencing of16S rRNA genes were applied. The main results are as follows:Soil physicochemical properties in croplands have significantly changed with time since wetland reclamation, and clear patterns of soil development were observed. Major pedogenetic processes included decreases in the mean soil particle size and improvements of soil aggregation, increased contents and crystallinity of iron oxyhydrates, accumulation of soil organic matter, leaching of cations, and interestingly, possible N limitation in young soils and P limitation in old soils as consistent with the Walker and Syers (1976) model. Variations in most soil variables during the period of500years were significantly related to that of soil organic matter, suggesting that soil organic matter might be a key factor indicating and regulating the overall development of cultivated soils. The overall soil fertility has improved after500years of soil development, but the overall deterioration in soil fertility was observed at the early stage (16years in this study) after wetland reclamation, with declines in most important soil physiochemical properties (eg., SOC, TN, CEC and Fe/Al oxyhydrates) due to a rapid loss of SOC induced by wetland reclamation. Principal Component Analysis (PCA) indicated an orderly pattern of the evolution of soil physiochemical properties during500years of cultivation, with the land use causing a displacement in the trajectories but not the overall direction of soil physicochemical development.The organic carbon content in bulk soil (SOC) decreased rapidly within a short period after reclamation and increased thereafter, along with evident achange in SOC stabilization mechanisms during soil development. The short term loss of SOC induced by reclamation was mainly related to declines in the Fe/Al-bound C and light fraction C, whose contributions to the loss of SOC were60%and30%, respectively. Decline in the Fe/Al-bound C was probably related to the leaching of amorphous Fe oxyhydrates, while the light fraction C was not protected by soil and thus subject to rapid microbial decomposition and loss. The long-term accumulation of SOC mainly due to the chemical protection of C by soil mineral and physical protection of particulate C by microaggregates, whose increases with time accounted for75%and19%of SOC accumulation, respectively. In addition, the importance of chemically protected C to SOC increased with time, while that of physically protected C slightly decreased. All these suggested that the chemical protection mechanism was more important to SOC accumulation than physical protection in croplands of the Chongming Island. However, the formation of microaggregates promoted not only the physical protection of particulate C but also the chemical protection of C in the silt-clay part of microaggregates, suggesting that the physical and chemical stabilization of soil C might be not independent each other. This can be seen from the higher C concentrations of the silt-clay fraction inside microaggregates (S+C_μ) than the free silt-clay outside microaggregates (S+C_f). Also the contribution of S+C_μ to SOC increased with soil development, compared to that of S+C_f which remained constant or decreased. For the chemically protected C, the mean proportions of Ca-bound and Fe/Al-bound C to SOC were11%and49%, respectively. The total contribution of Ca-and Fe/Al-bound C to SOC stabilization showed a decreasing trend with time. In the wetlands and soils younger than16years, Ca-and Fe/Al-bound C together accounted for74-88%of SOC, while, they only accounted for31-36%of SOC in the oldest cropland soils. Therefore, the importance of Ca and Fe/Al to the chemical protection of C progressively decreased with time, whereas that of other components of soil mineral increased.Paddy fields consistently had more SOC than upland fields, and this defference became more evident at the late stage of soil development. Physical protection explained25%of the differences in SOC between paddy and upland fields, while chemical protection by soil minerals can explain60%. Hence, chemical protection should be the main mechanism responsible for the higher C sequestration in paddy fields. For chemically protected C, there was no significant difference in Ca-bound C between paddy and upland fields; Fe/Al-bound C in paddy fields was10%higher than in upland fields, while C protected by other components of soil minerals was23%higher in the former, suggesting that the protection provided by Fe/Al oxyhydrates was not the major mechanism underlying the stronger chemical protection of C in the paddy fields. In addition, the light fraction C in paddy fields was28%higher than in upland fields, suggesting that the higher C input might be another important cause of stronger SOC sequestration in paddy fields.Evident bacterial succession in cultivated soils was observed along the500-year soil chronosequence. Both the T-RFLP and UniFrac analysis indicated directional changes in soil bacterial community structure. Based on the results of454pyrosequencing, there were evident temporal trends in the relative abundance of at least four phyla(Proteobacteria, Actinobacteria, Acidobacteria, Verrucomicrobia), of which Proteobacteria (especially the class a-Proteobacteria) and Verrucomicrobia were most closely correlated with bacterial succession. The temporal pattern of bacterial succession resembled that of soil physicochemical evolution, i.e., the changing rates were relatively high in the short term after reclamation and became slow thereafter. Therefore, soil bacterial succession can be divided into two phases, a short "rapid succession phase" after wetland reclamation and a long "gradual succession phase" thereafter. At the "rapid succession phase", the overall diversity (indicated by the Shannon index), richness (indicated by the Chaol estimator) and evenness (indicated by the Simpson index) of soil bacteria evidently increased, while in the long "gradual succession phase", these indices slightly decreased although microbial biomass increased. In addition, both the T-RFLP and UniFrac analyses showed that land use did not change the overall direction of bacterial succession but caused displacements in the trajectories ofsuccession. During the succession, paddy fields consistently had more Acidobacteria, Chloroflexi and Bacteroidetes, compared to upland fields where there were more Actinobacteria and Firmicutes.To conclude, this research provided a preliminary investigation on how the soil mineral, carbon processes and microbial communities might promote or feedback to each other during the long-term agricultural soil development. On one hand, changes in the mineral part of soil profoundly influenced carbon stabilization. For example, increases in microaggregates with time improved the C pool of S+C_μ, decalcification led to declines in Ca-bound C, and increases in the crystallinity of Fe oxyhydrates caused drops of Fe/Al-bound C. In turn, most changed soil variables during soil development was related to soil organic matter, highlighting the possible feedback of carbon processes to the physicochemical evolution of soil minerals. On the other hand, the results of Canoial Correspondence Analysis (CCA) showed that60%of variation in bacterial community structure during soil development can be explained by soil physicochemical properties. The shifts of soil bacterial communities induced by wetland reclamation was mainly related to the decreases in soil salinity and amorphous Fe oxyhydrates, while the bacterial succession under long-term cultivation was most closely related to such chemical variables as total phosphorous (TP)、potentially mineralizable nitrogen (PMN)、activity of iron (Feo/Fed)、ammonium and pH. Hence, it can be concluded that soil bacterial succession was driven by temporal shifts in the chemistry of the mineral and organic components of soil. In summary, this research highlighted that soil mineral, organic matter and soil biota may interact with each other to drive the soil development under long-term agricultural use.