【作者】 串丽敏；

【导师】 何萍；

【作者基本信息】 中国农业科学院 ， 植物营养学， 2013， 博士

【摘要】 合理推荐施肥是促进粮食高产、肥料高效的有效途径，而如何确定适宜的施肥量一直是施肥技术的核心和难点。本研究提出了将热带土壤肥力定量评价（Quantitative Evaluation of the Fertilityof Tropical Soils，QUEFTS）模型与改良的实地养分管理（Site-specific Nutrient Management，SSNM）技术相结合的小麦养分专家推荐施肥系统，其方法是在多年多点试验数据基础上，基于产量反应和农学效率进行施肥推荐。该方法不但满足了对不同大小田块推荐施肥的适用性，而且考虑了土壤基础养分供应量和N、P、K养分间的相互作用。与传统的测土施肥、植株营养诊断施肥等一系列测试技术和方法相比，本理论与技术具有时效性强、简便经济、易于掌握、适用广泛等优点，特别是在测土和植株诊断等条件不充分时采用本技术显得尤为重要。因此，基于产量反应和农学效率的推荐施肥系统是当前集约化农业生产条件下一种重要的施肥方法。本文在2000~2011年大量文献数据收集与已有试验数据库获取基础上，分析了当前中国小麦产量和养分吸收特征，运用QUEFTS模型模拟了一定目标产量下小麦地上部最佳养分吸收以及籽粒养分最佳移走量，评价了当前小麦产量与养分吸收特征，分析了基于产量反应和农学效率的小麦推荐施肥理论基础，建立了小麦养分专家系统，并进行两年的田间试验，验证其应用效果。研究获得的主要结论如下：（1）2000~2011年中国小麦平均产量为5950kg/ha，平均N、P2O5和K2O施用量分别为172、102和91kg/ha。冬小麦产量为<4.0、4.0~6.0、6.0~8.0、8.0~10.0和10.0~12.0t/ha所对应的生产每吨籽粒产量所需要的N养分分别为22.9、24.4、24.8、25.0和27.6kg，P养分分别为6.5、6.4、6.5、6.8和7.6kg，K养分分别为15.2、17.5、19.0、22.6、36.3kg。春小麦<4.0、4.0~6.0、6.0~8.0t/ha产量范围吨粮N养分吸收分别为24.3、25.4、27.0kg，P养分吸收分别为3.3、3.4和4.1kg，K养分吸收分别为49.8、32.3和17.3kg。（2） QUEFTS模型利用两条边界线描述了N、P和K养分的最小和最大养分内在效率。小麦N、P和K养分最小和最大养分内在效率分别为28.8和62.6kg籽粒/kg N，98.9和487.4kg籽粒/kg P，以及23.0和112.9kg籽粒/kg K。随着目标产量的增加，QUEFTS模型预估的作物养分平衡需求呈线性―抛物线―平台曲线。在产量潜力60%~70%范围内呈线性关系，此时生产1000kg籽粒所需要的N、P和K养分分别为22.8、4.4和19.0kg，对应的N:P:K比例为5.18:1:4.32，对应的氮、磷和钾养分内在效率分别为43.9、227.0和52.7kg籽粒/kg N、P和K；QUEFTS模型模拟的形成1000kg籽粒产量作物籽粒N、P和K养分最佳移走量分别为18.3、3.6和3.5kg，对应的N:P:K比例为5.08:1:0.97。约有80%、82%和18%地上部的N、P和K累积在籽粒中并被移出土壤。（3）土壤基础养分供应是表征土壤养分状况的重要参数，我国小麦产区土壤基础养分供应分别为122.6kg N/ha、38.0kg P/ha和120.2kg K/ha。产量反应是评价地力对施肥效应的重要指标，农学效率是反映施肥增产效应的重要参数。我国小麦产区施氮、磷和钾肥的产量反应平均分别为1.67、1.00和0.80t/ha，氮、磷和钾肥农学效率平均分别为9.4、10.2和6.5kg/kg。不施氮、磷和钾肥的相对产量分别为0.76、0.85和0.90。相对产量数值可以用于对土壤基础养分供应强度的分级。土壤基础N养分供应低、中、高级别所对应的相对产量参数分别为0.60、0.77和0.88；土壤基础P养分供应低、中、高级别所对应的相对产量参数分别为0.79、0.87和0.93；土壤基础K养分供应低、中、高级别所对应的相对产量参数分别为0.84、0.90和0.94。（4）产量反应和土壤速效养分之间存在显著负相关关系，和土壤基础养分供应之间存在显著负指数关系，和相对产量之间存在显著负线性关系。产量反应（x）和农学效率（y）之间存在显著一元二次函数关系。N产量反应（xN）与N农学效率（yN）之间函数关系为：y_N=-0.3729x_N~2+6.1333x_N+0.1438(R~2=0.76, n=601)，P产量反应（x_P）与P农学效率（y_P）之间关系为：y_P=-0.5013x_P~2+8.3209x_P~2+2.3907(R=0.65, n=288)，K产量反应（x_K）与K农学效率（y_K）之间关系为：y_K=-1.6581x_K~2+9.099x_K+0.7668(R~2=0.58, n=379)。上述相关性是基于产量反应和农学效率进行推荐施肥的重要理论基础。（5）田间试验结果表明，QUEFTS模型可以用于预估一定目标产量下N、P和K养分最佳需求量，表征籽粒产量与养分吸收之间的关系。基于产量反应和农学效率的小麦养分专家施肥推荐与农民习惯相比，降低了7.6%~55.8%的氮肥用量，调整了磷肥和钾肥用量，使小麦产量增加了0.1%~4.3%，纯收益增加了533~1734元/ha，氮、磷和钾肥农学效率以及回收率多表现出增加趋势或达显著水平。基于产量反应和农学效率的小麦养分专家系统作为一种可供选择的推荐施肥方法在理论支撑和技术实践上是可行的，可以在中国小麦产区进行施肥推荐。更多还原

【Abstract】 The inappropriate application of fertilizer has become a common phenomenon in wheat plantingsystem in China and has led to nutrient imbalances, inefficient use and large losses to the environment.Having access to a science-based fertilizer recommendation is critical for improvement of fertilizer useefficiency in high yielding crops. However, how to establish fertilizer recommendation suitable forsmall-holder farmers remains a challenge. This paper described a new fertilizer recommendationmethod named Nutrient Expert for Wheat decision support system based on yield response andagronomic efficiency using datasets from2000to2011. This system applies Site-specific NutrientManagement (SSNM) principles, which includes the use of the Quantitative Evaluation of the Fertilityof Tropical Soils (QUEFTS) model to determine crop nutrient uptake for a specific yield. Fertilizerrecommendation based on yield response and agronomic efficiency is an alternative approach developedfor use, which meets the adaptability of different sizes of field plots and considers the soil indigenousnutrient supply and N, P and K interactions. This methodology has the advantages of good timeliness,simple and economical, and easy to master and widely to apply, especially when soil testing and plantdiagnosis technology are not available. It is believed that Nutrient Expert for wheat decision supportsystem will become the popular method of fertilizer recommendation in China. The results of this studyshowed as follows.(1) The statistic of data from all experiments showed that the grain yield was averaged at5950kg/ha,the rates of N, P2O5, and K2O fertilizer application were172kg N/ha,102kg P2O5/ha, and91kgK2O/ha, respectively. At the yield ranges of <4,4~6,6~8,8~10, and10~12t/ha, the N requirementfor per ton of grain yield for winter wheat were22.9,24.4,24.8,25.0, and27.6kg, and P were6.5,6.4,6.5,6.8, and7.6kg, and K were15.2,17.5,19.0,22.6, and36.3kg, respectively. However, forspring wheat at the yield ranges of <4,4~6, and6~8t/ha, N for per ton of grain yield required were24.3,25.4, and27.0kg, P were3.3,3.4, and4.1kg, and K were49.8,32.3, and17.3kg,respectively.(2) In the QUEFTS model, two boundary lines described the minimum and maximum internalefficiencies (IEs, kg grain per kg nutrient in above-ground plant dry matter) of N, P and K. Theminimum and maximum IEs for wheat were28.8and62.6kg grain per kg N,98.9and487.4kggrain per kg P, and23.0and112.9kg grain per kg K. The QUEFTS model predicted alinear-parabolic-plateau curve for balanced nutrient uptake with target yield increasing. The linearpart continued until the yield was approximately at60%~70%of the potential yield, and22.8kg N,4.4kg P and19.0kg K were required to produce1000kg grain. The corresponding N:P:K ratio was5.18:1:4.32, and the corresponding IEs were43.9,227.0and52.7kg grain per kg N, P and K,respectively. The QUEFTS model simulated balanced N, P and K removal by1000kg grain were18.3,3.6and3.5kg, respectively, with a N:P:K ratio of5.08:1:0.97. Approximately80%,82%and18%of N, P and K in total above-ground plant material were presented in the grain and removedfrom the field. (3) The soil indigenous nutrient supply is an important parameter to characterize the soil nutrient status,the yield responses to N, P and K are most important indices for evaluation of fertilizer effect indifferent soil fertility, and the agronomic efficiency is a most important parameter to evaluate theyield increase by fertilizer application. The result showed that the average soil indigenous nutrientsupplies in China were122.6kg N/ha,38.0kg P/ha, and120.2kg K/ha, respectively. The meanyield responses of wheat to N, P and K were1.67,1.00and0.80t/ha. The mean agronomicefficiency values of N, P, and K were9.4,10.2, and6.5kg/kg, respectively. The mean relative yieldof N, P and K were0.76,0.85and0.90. The relative yield could be used to classify the soilindigenous nutrient supply. The values of relative yield for N were0.60,0.77and0.88for low,medium, and high indigenous N supply, respectively. The values of relative yield for P were0.79,0.87and0.93for low, medium, and high indigenous P supply, respectively, and values of relativeyield for K were0.84,0.90and0.94for low, medium, and high indigenous K supply, respectively.(4) There was a significant negative relationship between yield response and soil available nutrient, andsignificant negative exponential relationship between yield response and indigenous nutrient supply,and a significant negative linear correlation between yield response and relative yield. It was alsodemonstrated a quadratic equation between yield response (x) and agronomic efficiency (y)(P<0.05). The relationship between yield response and agronomic efficiency for N wasyN=-0.3729x_N~2+6.1333xN+0.1438(R~2=0.76, n=601), for P was y_P=-0.5013x_P~2+8.3209x_P+2.3907(R~2=0.65, n=288), and for K was y_K=-1.6581x_K~2+9.099x_K~2+0.7668(R=0.58, n=379). Theseequations were all incorporated as part of the Nutrient Expert for Wheat fertilizer recommendationdecision support system.(5) Field validation based on Nutrient Expert for Wheat confirmed that the QUEFTS model could beused as a practical tool for this decision support system to make fertilizer recommendation. Alsomultiple field experiments helped to validate the feasibility and showed that fertilizerrecommendation based on yield response and agronomic efficiency could reduce the N fertilizerapplication by7.6%~55.8%, and adjust the P and K fertilizer, the gain yield increased by0.1%~4.3%, and net profit was improved by522~1734yuan/ha. The agronomic efficiency andrecovery efficiency of N, P and K mainly showed a trend to increase or were improved significantly.It concluded that Nutrient Expert for Wheat could be used as an alternative method to makefertilizer recommendations in China.更多还原