【作者】 张勇；

【导师】 陈效民；

【作者基本信息】 南京农业大学 ， 土壤学， 2013， 博士

【摘要】 全球正经历着显著的气候变化,最近100年(1906-2005)以来全球平均地表温度上升了0.74℃。气候变暖造成了海平面上升、降水带变化以及极端气候事件频发等问题。农业生产中的各环节对自然条件尤其是气候条件(包括温度、降水、风速和大气C02浓度等)和土壤条件(土壤水分及养分状况)具有很强的依赖性。因此,气候变化势必对农业产生重大影响。为了明确未来气候变化的不确定性给农业生产带来的影响,确保气候变化背景下我国的粮食生产安全,本文以太湖地区典型水稻土(乌栅土)为研究对象,分析了研究区土壤的水力学特征和水环境状况,并且利用农田开放式CO2浓度和温度升高系统(T-FACE:Temperature-Free Air Carbon Dioxide Enrichment)系统平台人工模拟了未来50年C02浓度和温度升高的气侯条件,研究不同C02浓度和温度水平下,小麦不同生育期土壤水分含量、全氮、有机碳、铵态氮、硝态氮、速效磷的变化,并采用多种模型对土壤水分进行了模拟与预测,阐明了气候变化情境下土壤水分和养分含量的动态变化趋势,为气候变化下农田水肥管理提供科学依据。本论文研究结果如下：(1)影响太湖地区典型水稻土水分特征曲线的主要因素是容重和粘粒含量。在低水吸力状况下,耕作层土壤持水量相对较高。乌栅土中有效水含量与容重和粘粒含量呈显著负相关,与孔隙度呈显著正相关。在土壤的各种理化性质中,容重和总孔隙度是影响饱和导水率的主导因素,其它如毛管孔度,pH为次要因素。乌栅土表层(0-14cm)土壤水分总库容达到91.48nm,有效水库容为22.02mm,滞洪库容为20.50mm。因此,乌栅土具有良好的水分状况,能够保证太湖地区粮食作物的稳产高产,缓和气候变化引起的降雨和干旱使土壤水分产生的异常变化,在防治旱、涝灾害方面有着重要意义。(2)在大气C02浓度和温度升高的影响下,不同C02浓度和温度处理使各个试验区内0-14cm土壤平均含水量存在显著性差异(p<0.05),其大小顺序为：CT圈(C02浓度和温度升高)<T圈(温度升高)<C圈(C02浓度升高)<CK(对照圈)。研究区土壤三个层次0-14cm,14-33cm和33-59cm土层中土壤含水量存在显著性差异(p<0.05),且0-14cm土壤含水量比14-33cm和33-59cm土壤含水量高。C圈和CK圈内0-14cm土壤水分含量分别为：273.62g·k分1和279.87g·kg-1,各占研究区耕层土壤田间持水量的72.10%和73.75%；而CT和T圈内0-14cm土壤含水量分别为：253.98g·kg-1和259.72g·kg-,占耕层土壤田间持水量的65.93%和67.44%,说明温度升高能显著降低耕作层土壤含水量,而CO2浓度升高对耕作层土壤含水量的影响较小。通过对研究区三个土壤层次(0-14cm,14-33cm和33-59cm)土壤水分含量的时间稳定性进行分析比较,得出随着土壤层次的加深土壤水分的时间稳定性越来越好。在CO2浓度和温度升高交互作用处理下,土壤水分的时间稳定性表现更好。运用平均相对偏差的方法找到了各土层中具有时间稳定性的代表性测点,并且能很好地反映该层土壤水分的平均动态变化,均方根误差小于3.52%。(3)模拟气候变化条件下,小麦从分蘖期到成熟期,C02浓度升高降低了土壤中NH4+-N含量。因此,在气候变化条件下,增加氮肥用量能保证作物对氮的吸收利用。然而,不同C02浓度和温度处理对土壤中N03-N没有显著影响；C02浓度升高使耕作层土壤速效磷含量增加。在不同CO2浓度和温度处理下,土壤C/N大小依次为：C圈>CT圈>CK圈>T圈。C02浓度升高处理下,土壤水分含量与N03--N含量的相关性达到极显著水平,C圈内土壤水分含量和NH4+-N含量相关性达到显著水平；C02浓度升高和温度升高的交互下,CT圈内土壤水分含量与NH4+-N含量的相关性达到显著水平；温度升高处理下,土壤水分含量与土壤养分含量没有显著影响。从分蘖期到成熟期,CO2浓度升高处理降低了土壤中NH4+-N含量,但农田渗漏水中NH4+-N含量的变化与耕作层和犁底层土壤中NH4+-N含量没有表现出相似的变化趋势,说明气候变化对农田渗漏水中NH4+-N的变化影响不大。(4)利用BP神经网络对土壤含水量与气象因素进行了敏感性分析,从局部和全局敏感性分析的结果可以看出,土壤水分含量对降水量的敏感性最大,蒸发量、平均气温和平均地表温度次之。运用ARIMA, BP-ANN和LS-SVM三种模型对土壤水分进行模拟,结果表明将气象因素作为输入参数能得到较好的模拟效果,而且LS-SVM模拟精度最高。研究区土壤具有良好的水分和养分条件,在模拟气候变化条件下,对小麦生长季的水分和养分变化分析,表明在气候变化条件下,温度升高能够抵消CO2浓度升高增加生物量的正效应；应适当增加氮肥的使用量以保证作物的生长需求；通过三种模型对土壤水分的模拟发现LS-SVM模拟精度最高。由于气候变化对中国农业生产的影响甚为复杂,需要进一步研究温度和CO2浓度升高对N、P转化的影响机制。更多还原

【Abstract】 The world has been undergoing a significant climate change with temperature rising as main characteristics. Global average surface temperature has increased by0.74℃in the last100years (i.e.,1906to2005). The global warming caused the rising of sea level, precipitation change, the increased occurrence frequencies of extreme weather events etc. Agricultural production closely depends on the natural conditions, especially climatic conditions (including changes of temperature, precipitation, wind speed and atmospheric [CO2](carbon dioxide concentration), etc.), hence climate change has a great influence on it. Soil water and soil nutrients which are indispensable for crop growth would be affected by climate change. For the sake of determining the influence of the uncertainties of climate change on agriculture production, the hydraulics characteristic and water environmental condition in study area were analyzed, and the T-FACE system platform was built to simulate the meteorological conditions (including elevated [CO2] and temperature) in the coming50years and to study the effects of these conditions on the changes of soil total nitrogen (TN), soil organic carbon, ammonium, nitrate, available phosphorus (P) and soil water content at different wheat growth stages in the Taihu Lake region based on the typical paddy soil-wushan soil. The simulation and prediction of soil water were made with three models and the variation trend of soil water and soil nutrients under the future climate changing scenario were clarified, which provided theoretical reference for the management of irrigation and fertilization.The results were as follows:The bulk density and clay content were the main factors affecting the water characteristics curves. While soil water content in plough layer was relatively higher under a lower water suction, the available water content had a significant negative correlation with bulk density and clay content, and had a significant positive correlation with porosity. The bulk density and total porosity were the main impact factors to the saturate hydraulic conductivity, whereas the other factors such as the capillary porosity and pH took the second place. The total reservoir capacity of wushan soil in plough layer (0-14cm) was91.48mm. The available water reservoir capacity was22.02mm, and the flood detention reservoir capacity was20.50mm. Therefore, this soil was in good water condition, which not only guaranteed the realization of stable and high soil productivity but also had a regulatory function of water storage. The data of the water characteristic curve and soil reservoir capacity in this soil will provide the scientific basis for soil water management under the condition of climate change.Under elevated atmospheric [CO2] and temperature conditions, there were obvious differences in the average water content at the depth of0-14cm among different experimental plots. Soil water content increased in the following order:CT(elevated [CO2] and temperature), T(elevated temperature), C(elevated [CO2]) and CK. There were significant differences in soil water content among0-14cm,14-33cm and33-59cm soil layers, with the soil water content in0-14cm soil layer higher than that in14-33cm and33-59cm soil layers.The soil water content of plough layer in C ring and CK ring was273.62g·kg-1and279.87g·kg-1, respectively, accounting for72.10%and73.75%of water holding capacity in study area, and that in CT ring and T ring was253.98g·kg-1and259.72g·kg-1, being65.93%and67.44%of water holding capacity correspondingly, which showed that elevated temperature could significantly reduce soil water content in plough layer and Elevated [CO2] had a little influence on it.The analysis and comparison of temporal stability of soil water content in three soil layers (0-14cm,14-33cm and33-59cm) showed that the temporal stability of soil water increased with the deepening of soil depth. Three measuring points of temporal stability in three soil layers (0-14cm,14-33cm and33-59cm) were determined using the relative mean difference method, which were appropriate to simulate the changes of average soil water in the corresponding soil layers with the RMSE between observed value and simulated value lower than3.52%.When the simulation of climate change was performed, the contents of ammonium decreased significantly with rising [CO2] during the wheat growth season (from tillering to ripening stage) in2011to2012; Therefore, under the condition of climate change, increasing nitrogen fertilizer to ensure meet the nitrogen absorption. Soil NO3--N content showed no clear variation under different [CO2] and temperature conditions; elevated [CO2] enhanced available P in0-14cm soil layer. With elevated [CO2] and temperature, C/N ratio decreased as follows:O CT>CK>T. The correlation between soil water content and nitrate in C ring was highly significant, and that between soil water and ammonium was significant in C ring. There was a significant correlation between soil water and ammonium under the interaction of elevated [CO2] and temperature (CT). No significant correlation was observed between soil moisture and soil nutrients under elevated [CO2].The three treatments of elevated [CO2], elevated temperature, and their interaction (i.e., C, T and CT ring) all decreased the content of ammonium in plough layer during the wheat growth season (from tillering to ripening stage), but the variation of ammonium in farmland leaching water was not consistent with that of soil in plough layer and in plowpan, indicating that there was no significant influence of elevated [CO2] and temperature on ammonium content in farmland leaching water. Elevated [CO2] and temperature treatments all had no obvious effect on nitrate content either in soil or in farmland leaching water.Sensitivity analyses of soil moisture and meteorological factors based on BP neural network were conducted. Models of Auto Regressive Integrated Moving Average (ARIMA), Back Propagation Artificial Neural Network (BP-ANN), and Least Squares Support Vector Machine (LS-SVM) were used to simulate the soil water in soil layers of0-14cm and14-33cm. The results showed that:there were satisfied results when meteorological factors were taken as input variables, and the simulation accuracy of LS-SVM was the highest.The analysis of soil water and nutrient during the whole wheat growth period shows that increased temperature could offset the positive effect on biomass increment which was caused by increased [CO2] under the climate change. The application amount of nitrogen fertilizer should be increased properly in order to meet the need of crops. The output of LS-SVM model has the best simulation precision. In addition, the N, P transformation mechanism under climate change need further study.更多还原