【作者】 王迎春；

【导师】 王道龙；

【作者基本信息】 中国农业科学院 ， 作物生态学， 2009， 博士

【摘要】 本研究立足于华北平原小麦玉米两熟区典型农田生态系统,以“国家土壤肥力与肥料效益长期监测基地网”河南郑州实验站肥料长期试验、中国农大曲周实验站氮磷配施长期施肥试验和山东禹城实验站水氮耦合试验数据为支撑,利用澳大利亚农田生态系统模型（APSIM）作为系统工具将农田生产管理的诸要素有机结合、综合分析,揭示农田生态系统氮磷的循环与平衡规律,并将养分平衡、作物生产力、环境效应置于气候波动背景下,分析长期不同施肥水平与田间管理方式对农田生产力与氮磷平衡的影响。主要研究结果如下:（1）通过构建农田生态系统模型（APSIM）氮磷养分平衡模拟系统,利用华北平原3个试验点（河南郑州、河北曲周、山东禹城）的实测数据对模型进行了参数修正和全面验证。验证结果表明:3个实验点小麦的生物量和产量的模拟值和实测值之间的决定系数（r²）均大于0.73,玉米生物量和产量的模拟值和实测值之间的r²分别大于0.61和0.81,小麦和玉米植株含氮量多年平均模拟值和实测值之间r²为0.91和0.97、植株含磷量r²为0.96和0.88,土壤全氮、有机碳和土壤速效磷的决定系数r²分别为0.70、0.78和0.64。证明APSIM模型能够在华北平原氮磷养分循环分析中应用并进行有关的情景分析。（2）APSIM模型模拟得出郑州点15年间长期施肥条件下,采用长期施肥试验中玉米郑单8号品种和豫麦13号小麦品种模拟的潜在产量均值为6.9 t ha^-1和8.3 t ha^-1,而小麦和玉米所需的最大吸氮量多年分别是170 kg N ha^-1和189 kg N ha^-1,多年平均最大吸磷量是32 kg P ha^-1和34 kg P ha^-1;改善灌溉等田间管理措施最高肥处理（1.5NPKM）处理仍然有0.7 t ha^-1和1.6 t ha^-1的产量空间可以增长。小麦的氮磷钾配合施肥处理（NPK）比磷钾处理（PK）的AE和REN提高了4.26倍和11.2倍,在小麦各处理中居于最高;小麦秸秆还田可以显著提高玉米的AE和REN,氮磷钾配施并秸秆还田处理（NPKS）的玉米氮肥利用率最高,AE和REN分别比磷钾处理（PK）提高3.1倍和2.3倍。小麦和玉米的磷肥农学效率（AE）以氮磷配施处理（NP）最高,分别为45.2 kg^-1P2O5和25.3 kg^-1P2O5,小麦和玉米的REP以氮磷配施（NP）和氮磷钾配施（NPK）两个处理最高,分别为25.3%和14.7%。在河南郑州点轻壤质潮土种植区,明确了氮和磷是制约小麦玉米生产的第一要素、第二要素,玉米季施肥配合小麦秸秆还田可以显著提高氮磷肥料利用率。（3）利用APSIM模型对华北平原郑州点小麦-玉米轮作系统多年（1961年-2007年）不同施肥水平下的模拟结果表明,小麦氮肥施用量在180kgNha^-1以下,玉米施氮量150 kgNha^-1以下,为氮肥施用的敏感节点,超过该值后产量对施肥量的增加响应不敏感。46年间施氮水平在150kgNha^-1时小麦和玉米产量达6t和7t的保证率分别为75%和46%,施氮水平在120 kgNha^-1时玉米6 t ha^-1产量的保证率是81%。此外,模拟结果还显示两季作物秸秆全部还田条件下施氮水平越高有机碳的矿化越大;在施氮量为150kgNha^-1水平下,不同秸秆还田水平下保证率在75%～25%之间的矿化量为80～132 kgNha^-1,平均年矿化量为104 kgNha^-1,矿化水平对农田施肥管理与土壤培肥具有重要意义。通过长期模拟也显示,作物产量、作物吸氮、吸磷量以及土壤有机碳的矿化也在随着气候波动而变化,不同的气候年型,各项指标的变异比较大。（4）通过对不同氮磷肥及秸秆还田不同比例综合措施的长期（46年）模拟发现,当氮肥施用量为50kgNha^-1以下或者氮肥施用100 kgN ha^-1配合磷肥20 kgPha^-1水平时发生硝氮淋溶的概率为0;当单施氮肥用量在100 kgNha^-1以上而不施加磷肥时发生硝态氮淋溶的概率与秸秆还田比例有关,秸秆还田比例越大淋溶风险越高;当氮肥用量提高到150 kgNha^-1磷肥施用40 kgPha^-1时淋溶风险最小。从环境效应和作物生产力两者均衡看,配合施肥组合在氮肥150 kgNha^-1、磷肥40 kgPha^-1、秸秆还田一半或不还田时硝态氮淋溶可以被控制。在充分灌溉条件下,土壤NO3--N淋溶与降水仍有较强的正相关性（相关系数r为0.65）,该结论对现实生产中减少氮肥的淋溶风险具有指导意义。（5）利用APSIM模型对华北平原2006年298个县域实际施肥水平下的小麦玉米产量进行了模拟测试,模拟结果显示小麦和玉米产量模拟值与统计值的相关系数r分别为0.74,0.62。通过APSIM模型模拟华北平原小麦-玉米轮作系统0-40cm土壤层中全氮含量在1755-8624 kgNha^-1,平均值为2417 kgNha^-1,0^-160cm土层NO3--N累积量11^-1839 kgNha^-1之间,平均值为214.16kgha^-1,有60%的区域NO₃--N累积量超过100 kgNha^-1;模拟结果也显示华北平原稳定态磷在0～40cm土层含量平均值为1193 kgha^-1,有效磷在0～40cm土层的含量在57-200kgha^-1之间,在安徽北部、河南东北部地区含磷量较高、环境风险比较明显。更多还原

【Abstract】 This study combined experimental data analysis and systems modeling to investigate the crop productivity and nitrogen and phosphorous balances in the wheat-maize double cropping system in the North China Plain （NCP）, and their responses to different management practices under the background of variable climate. Experimental data used include: 15 years of crop and soil data from Zhengzhou station in Henan province under the National Soil Fertility and Fertilizer Efficiency Long-term Monitoring Network, 5 years of data from a long-term fertilization experiments at Quzhou, Hebei province, and 2 years of data from Yucheng, Shandong province. The main findings are:（1） The Agricultural Production Systems Model （APSIM） was calibrated and validated using long-term and short-term experimental data and the performance of APSIM model was evaluated by its ability to predict: （1） biomass, grain yield, and N/P uptakes of both wheat and maize crop, （2） soil nitrogen and phosphorus dynamics, and （3） soil carbon dynamics, in response to different N/P application rates. The results showed that APSIM is able to capture the response of crop growth, yield and N/P uptakes to different fertilizer application levels. The determination values （r²） were all above 0.73 between simulation and measured data for wheat biomass and yield, r² also were above 0.61 and 0.81 for that of maize biomass and yield respectively. It also simulated the soil C, N and P changes with acceptable accuracy, their r were 0.70, 0.78 and 0.64 for total N, soil organic carbon and labile phosphrous in 0-20cm soil, respectively.Therefore, it can be used to conduct scenario analysis for different fertilization treatments.（2） Analysis of the long-term data at Zhengzhou showed that both N and P were limiting factors for crop growth. Without additional N and P fertilisation,only a very low yield level (ca 2t ha^-1 for wheat and 3 t ha^-1 for maize) could be maintained. To achieve the potential productivity （i.e. yield level free of water and nutrient stresses） of wheat (6.9 t ha^-1) and maize (8.3 t ha^-1), wheat would need, on average,170 kg N ha^-1and 32 kg P ha^-1, while maize would need 189 kg N ha^-1,34 kg P ha^-1.There was still a gaps of 0.7 t ha^-1 and 1.6 t ha^-1 for wheat and maize which can be increased in under the high fertilizer treatment depending on the improvement of management. On average the AE and REN of NPK treatment could be 4.26 and 11.2 times than PK treatment and the highest among all treatments. Treatments with nutrient inputs higher than the NPK treatment and treatments without combination of N and P have led to accumulation of N and P in the soil profile. Returning wheat straw to field the AE and RE would be significantly increased for maize.（3） At Zhengzhou, long-term （1961-2007） simulation results revealed that the optimal nitrogen application rates for wheat and maize are 180kgNhm-2 and 150kgNhm-2 respectively, above those change point yield did not increase anymore with increased N applications. These rates are significantly lower than the actual N application rates in many part of NCP. The probability of exceedance with 6t and 7t of yields of wheat and maize were 75% and 46% respectively applied 150 kgNha^-1fertilizer and the probability of exceedance with 6t yield of maize were 81% applied 120 kgNha^-1fertilizer.Furthermore, the simulation results also been showed that under the situation of returning all of straw into field the mineralization would be increased with more nitrogen fertilizer input. With the leavel of 150 kgNha^-1 nitrogen, the mineralization amount were 80^-132 kgNha^-1 with 75%～25% probability of exceedance. According to the simulation results, the yield, nitrogen and phosphrous uptake and mineralization could be variable and be changed with the varies climate.（4） Simulation results also showed that if N application rates were below the optimal rates, little N leaching would occur below the crop root zone. However,if N application rate was higher than the optimal rates,nitrate N starts to accumulate in the soil profile and significant leaching will occur subsequently. N leaching mainly occurred in the maize season due to the concentrated summer monsoon rainfall. N leaching events were significantly related to rainfall events. Combined 150kgNha^-1 and 40 kgPha^-1 for each crop were optimal fertilizer levels due to the balances of leaching risk and crop yiled. N leaching events were significantly related to rainfall events （correlation coefficient r = 0.65） although the farm was irrigated enough.（5） The APSIM model was also appled to simulate the wheat-maize cropping system in 298 counties of NCP. The correlation coefficient between the statistical and simulated yield were 0.74 and 0.62 for wheat and maize respecively. Under the current nitrogen application rate, the total nitrogen content was in 1755 8624 kgNha^-1, average 2417 kgNha^-1 in 0-40cm soil layers, NO3-N accumulation was 11 1839 kgNha^-1, average 214 kgNha^-1 in 0^-160cm soil. In NCP 60% of the areas with NO3 - N accumulation were more than100 kgNha^-1. Simulation results also showed that Unavailable-P content averaged 1193 kgP₂O₅ha^-1 in 0 - 40cm soil layer. The high labile P content was distributed in northern Anhui province, north-east of Henan province with high risk of environment.更多还原