【作者】 李虎；

【导师】 邱建军；

【作者基本信息】 中国农业科学院 ， 作物生态学， 2009， 博士

【摘要】 随着农田氮肥施用量的不断增加,由氮素造成的农田非点源污染已成为水体污染的主要因素之一。当前世界各国都致力于研究提高农田氮肥利用率和减少非点源氮污染的方法与途径。本文采用田间试验与生物地球化学循环模型相结合的方法,研究了环渤海典型流域——山东小清河流域典型农田非点源氮素的运移机理、排放途径和转化规律。以田间定点观测资料为基础,验证并校正了基于氮循环机理的过程模型——DNDC模型,通过模型模拟提出了农田典型种植方式下的优化管理措施,并与流域GIS数据库相结合,从流域尺度上识别全流域氮流失的潜在负荷及关键源区,定量分析氮污染的主要分布和贡献,进而初步提出基于全流域的调控策略和优化管理措施。主要研究结果如下:（1）以小清河流域典型农作系统冬小麦/夏玉米轮作、冬小麦/大葱轮作、设施蔬菜为研究对象,采用田间原装渗漏计定位观测,研究农田土壤中氮流失特征。试验表明,历城冬小麦/夏玉米轮作地常规施肥处理下小麦生长季0¹ m土壤硝态氮积累量在90.13⁴26.97 kg N hm^-2,玉米季为67.96²04.32 kg N hm^-2。整个轮作生长季常规施肥处理水分渗漏量为176.5mm,由此而带走的氮素淋失量为38.76 kg N hm^-2,占氮素总输入量的6.5%,对照小区为3.54 kg N hm^-2,施肥处理是不施肥处理的10倍以上。章丘冬小麦/大葱轮作地6个不同处理试验表明,大葱生长季硝态氮淋失量随氮肥投入量的增加而增加,常规施肥处理氮淋失量为20.35 kg N hm^-2,占氮肥总输入量的5.2%;对照处理最小,为4.32 kg N hm^-2,前者是后者的约5倍;优化施肥和减量施肥均能有效减少硝态氮淋溶量,分别为17.49和13.46 kg N hm^-2。此外,径流试验表明,该地区氮素径流流失较小,常规施肥处理为0.51 kg N hm^-2,但仅占氮肥总输入量的0.01%,各施肥处理之间N素径流流失与氮肥施用量之间不存在明显的相关关系。寿光设施蔬菜一个生长季的研究表明,常规施肥处理氮肥投入总量为1465 kg N hm^-2,氮淋失量为214.04 kg N hm^-2,占氮肥总输入量的14.6%;对照处理为76.26 kg N hm^-2,前者是后者的约3倍。（2）利用田间实测数据验证和校正模型,结果表明,校正后的DNDC模型能更好地模拟冬小麦/夏玉米轮作系统、冬小麦/大葱轮作系统和设施蔬菜种植系统土壤水分以及氮素淋失的动态变化规律。同时,模型也很好的拟合了冬小麦/夏玉米轮作系统中土壤中NO3--N和NH4+-N的残留量,模拟值和实测值相关性显著。敏感性分析表明土壤pH值、施肥量、土壤孔隙率对氮素的淋失影响较大,是模型需要仔细考虑的输入参数。（3）利用模型提出了基于地块的最优化管理措施,即历城唐王小麦/玉米轮作系统施氮量减小到常规施肥量的60%,并提高玉米秸秆还田率到100%;章丘小麦/大葱轮作系统施氮量减小到常规施肥量的50%;寿光设施蔬菜系统为施氮量减小到常规施肥量的20%,并减小灌溉量为常规的80%。在保证作物产量在可接受范围的条件下,各点通过优化管理措施能显著减少氮素淋失量、N2O排放量以及NH3挥发损失。该评价结果可直接用于指导生产实践。（4）建立了小清河流域分乡镇单元的作物类型、种植面积、农田管理等GIS数据库,利用DNDC模型对流域尺度上氮素淋失量进行了模拟。2006年全流域8个区县年均氮淋失负荷范围为10.44×10³ t ³6.86×10³ t,平均为23.65×10³ t。以氮肥投入总量222.2×10³ t计算,该流域平均氮素流失量占氮肥投入的10.6%。淋失量空间分布表明,大于80 kg N hm^-2的单元主要分布在章丘市和寿光市的部分乡镇;淋失量小于10 kg N hm^-2的单元主要集中在流域的上游济南市历城区,在桓台和广饶县也有零星分布;其他地区氮素淋失量大多为20⁴0 kg N hm^-2和40⁸0 kg N hm^-2。（5）利用DNDC模型初步建立了区域冬小麦/夏玉米轮作体系下优化管理措施。运用该措施可将传统区域氮肥用量减小30%,而且对冬小麦/夏玉米产量没有影响,同时能产生较好的环境效应,氮素淋失量可降低35%,N2O排放降低28%,NH3挥发损失降低了34%。然而,氮肥施用量对作物产量和氮淋失量的影响区域差异极大,在制定区域优化管理措施时,不能全流域一个标准,而应根据当地更小范围农业发展状况和土壤营养元素平衡情况因地制宜地进行制定。更多还原

【Abstract】 Modern agricultural practices are strongly linked to fertilizer application for maintaining optimum yields. However, inefficient fertilizer use has led to a significant portion of the nitrogen （N） applied to farm fields reaching surface or underground water systems. Non-point source pollution caused by Nitrate from cropland is major factor seriously causing water quality degradation with the increasing application of N fertilizers. The scientists from all over the world are committing themselves on improving the utilization of N fertilizer and decreasing the non-point source N pollution from farmland. By combining the field experiment methods with biogeochemical model in this paper, non-point source pollution of nitrogen in farmland was studied on its diffusion and movement mechanisms, releasing routes and transferring characteristics in the typical watershed, Xiaoqing River basin, around Bohai Bay. Observations of N leaching from typical croplands in Xiaoqing River basin were used for the validation and modification of the N cycling process-based model, the DNDC or DeNitrification-DeComposition model. The optimum management practices for different crop systems at site scale were proposed based on analyzing the simulated results in typical croplands. And then coupled with Geographic Information System （GIS）, the DNDC model estimated the N leaching load and identified key pollution areas at regional scale with the support of the GIS database, and quantified the distribution and contribution of N pollution, and evaluated impacts of current management practices on N accumulation and loss. At last, some suggestions were further put forward. The main results were as follows:（1） The lysimeter method was used to study the characters of N leaching under summer maize-winter wheat rotation system, winter wheat-green onion rotation field, and vegetables-greenhouse fields in typical farmlands of Xiaoqing River Basin. For summer maize-winter wheat rotation field in Licheng site, the results showed that NO3--N was the dominant N style remained in the soil. The soil NO3--N accumulation for conventional treatment ranged from 90.13 to 426.97 kg N hm^-2 during the winter wheat growth period, and from 67.96 to 204.32 kg N hm^-2 during summer maize growth period. During the whole rotation growth season, the amount of water leaching was 176.5mm, and N leaching caused by water leaching was 38.76 kg N hm^-2 occupying 6.5% of the total N fertilizer input for the conventional treatment, which was over 10 times than that for the control treatment with 3.54 kg N hm^-2. Six different treatments were set up in winter wheat-green onion rotation field in Zhangqiu site. The amount of N leaching during green onion growth season was significantly increased with N fertilizer application increasing. The N leaching was 20.35 kg N hm^-2 for the conventional-fertilizer treatment accounting for 5.2% of the total N fertilizer applied, while it was 2.28 kg hm^-2 for the control treatment. The N leaching amount was 17.49 and 13.46 kg hm^-2 for the optimum-fertilizer and decrease-fertilizer treatment, respectively, they both can reduce N leaching effectively. Moreover, the runoff experiment showed that the maximum amount of N runoff was 0.51 kg N hm^-2 for the conventional-fertilizer treatment, only accounting for 0.01 % of the total N fertilizer applied, it was far less than the amount of N leaching, proving that the N leaching was dominant during green onion growth season. The relative relationship between the N runoff loss and fertilizer applied showed insignificantly. For the greenhouse vegetable field in Shouguang, The results showed that the N leaching amounted to 214.04 kg N hm^-2 for the conventional treatment, accounting for 14.6% of the total N fertilizer applied with the amount of 1465 kg N hm^-2 during one crop growth season, while it was 76.26 kg N hm^-2 for the control treatment. The former amount of N leaching was 2 times than latter.（2） Observations of field N leaching were used for model validation and modification, the results showed that the modified model can simulate the soil water movement and the change of N leaching well in winter wheat-summer maize rotation field, winter wheat-green onion rotation field, and greenhouse vegetables field. Moreover, the model generally had acceptable performances in the model simulations for soil residual NO3--N, NH4+-N in winter wheat-summer maize rotation field. Sensitivity tests showed that the simulated great impacts of soil pH, fertilization, and soil porosity on nitrate leaching were consistent with observations reported by other researchers.（3） The optimum management practices for different crop systems at site scale were proposed based on analyzing the simulated results in typical croplands, which was the decrease of fertilizer application to 60% of conventional fertilizer rate, coupled with increase the maize residue returned rate to 100% for the wheat-summer maize rotation field at Licheng site; the decrease of fertilizer application to 50% of conventional fertilizer rate for the wheat-green onion rotation field at Zhangqiu site; the decrease of fertilizer application to 20% of conventional fertilizer rate, coupled with decrease the irrigation rate to 80% of conventional irrigation rate for greenhouse vegetable field at Shouguang site. Under maintaining the grain yields, the optimum management practices for each site can reduce the N loss by N leaching, N2O emissions, and NH3 volatilization. These results can guide directly for the agricultural production.（4） The DNDC model was used to estimate the N potential leaching load at regional scale with the support of the GIS database which was established by compiling the local climate/soil condition and agricultural census data. The results showed that the N potential leaching load ranged from 10.44×10³ t to 36.86×10³ t, with the average of 23.65×10³ t in Xiaoqing River Basin in 2006. Taking the amount of total N fertilizer applied 222.2×10³ t, the average N leaching accounted for 10.6% of the N fertilizer applied. The spatial distribution of N leaching showed that the cells （i.e. towns） with N leaching more than 80 kg N hm^-2 mainly distributed in Zhangqiu City and Shouguang City; the cells with N leaching less than 10 kg N hm^-2 mainly distribued in Licheng District, Jinan City, and Huangtai county and Guangrao county. The amount of N leaching in other regions mainly concentrated between 20 and 40 kg N hm^-2, and also between 40 and 80 kg N hm^-2.（5） Based the regional DNDC model, the regional optimum management practices for winter wheat summer maize rotation system had been established preliminarily. The management practices can reduce the N fertilizer rate 30% compared with conventional N fertilization with no negative impact on agricultural production, meanwhile, the N leaching can reduced 35%, N2O emission reduced 28%, and NH3 volatilization decreased 34%. However, the effects of N fertilization on crop yields and N leaching varied greatly in different regions. Therefore, consideration of the N pollution control effectiveness, it was not effective or efficient to implement a uniform policy thoughout the whole watershed. Instead, it should be site-specific to make management practices.更多还原