【作者】 杜君；

【导师】 白由路；

【作者基本信息】 中国农业科学院 ， 植物营养学， 2011， 博士

【摘要】 小麦是我国主要的粮食作物之一,其总产量和种植面积约占我国粮食作物总产量和面积的1/5和1/4,是我国最重要的商品粮和战略性粮食储备品种。综合考虑不同气候特征、不同土壤类型和土壤养分状况等因素建立我国小麦施肥指标体系,对指导我国小麦的测土配方施肥及改善农田生态环境具有重要意义。本文基于多年多点小麦田间肥效试验数据,综合考虑气温、降水等气象因子及土壤类型、质地等土壤因子对小麦产量、土壤供肥能力和肥料当季利用率的影响,建立以养分平衡原理为依据、土壤养分测定为基础的我国小麦施肥指标体系,并利用ArcGIS提供的二次开发功能,开发出基于GIS的我国小麦施肥决策支持系统。主要结果如下:（1）养分平衡施肥模型基本参数的确定。目标产量在施肥模型中设为自变量,用气候生产潜力产量作为其上限。分析表明,小麦生物产量的养分吸收量与其籽粒产量之间呈极显著直线相关关系;小麦单位产量养分吸收量趋向一个稳定的范围,因此本研究把小麦单位产量养分吸收量定为常数。冬小麦每百千克小麦籽粒产量所需养分量分别为:氮（N）为3 kg、磷（P2O5）为1.2 kg、钾（K₂O）为2.8 kg;春小麦每百千克小麦籽粒产量所需养分量分别为:氮（N）为3 kg、磷（P₂O₅）为1.1 kg、钾（K2O）为2.6 kg。基于试验点数据,分别计算出土壤有效养分校正系数和肥料当季利用率,在全国27个土类上分别研究两参数的变异性,就各土类两参数的平均值来看,不同土类之间差异较大。就同一土类两参数的标准差来看,其空间变异性也均较大,且在大多数土类中都表现出极高的离散程度。小麦生育期内平均温度和平均降水量、土壤养分含量、土壤粘粒含量、土壤pH、及灌溉等因素影响着土壤有效养分校正系数和肥料当季利用率。（2）土壤有效养分校正系数与肥料当季利用率子模型的构建。用土壤有效养分（碱解氮、有效磷和速效钾）含量、土壤粘粒和小麦生育期内平均温度因子分别构建了土壤有效养分校正系数模型。模型的拟合程度均较高,决定系数（R2）在0.55～0.85之间。其中,土壤碱解氮、有效磷和速效钾校正系数模型的拟合精度为:有效磷>速效钾>碱解氮。在不同产量水平下,土壤养分含量与肥料当季利用率呈显著负相关的对数函数关系,分别建立了以土壤养分含量为自变量的肥料当季利用率模型。其模型的拟合程度也较高,相关系数（r）在0.45～0.75之间,均达到了极显著水平。并采用土壤pH值和灌溉因子对两子模型进行了修正。（3）养分平衡施肥模型的建立与验证。目标产量设为输入项,单位产量（每百千克）小麦的养分吸收量定为常数,结合土壤有效养分校正系数和肥料当季利用率两个模型,建立了以目标产量和土壤有效养分测试值为自变量,氮磷钾施肥量为因变量的养分平衡施肥综合模型。并将27个土类归并为15个土类组合,建立了全国范围的小麦推荐施肥指标体系。利用布置在各核心试验区的小麦肥料田间试验结果验证了养分平衡施肥模型推荐的施肥量。与用肥料效应函数法推荐的施肥量相比,养分平衡模型推荐的施肥量位于最高产量施肥量与最佳经济效益施肥量之间,该模型对小麦进行推荐施肥是可行的,模型具有简单、快速、准确等优点。（4）小麦生态环境因子基础空间数据库的构建。基于GIS平台,利用气象资料及土壤类型、质地等土壤资料,建立了影响小麦施肥模型参数的生态环境因子空间数据库。并通过插值、矢栅转换及图层叠置等处理,建立了各种图层,包括小麦生育期内平均温度和平均降水量、土壤粘粒含量和土壤pH、小麦需水量、小麦潜力产量等栅格图层,以及中国县界图与土壤类型图叠置生成的最小施肥单元矢量图层。（5）基于GIS的小麦施肥决策支持系统的开发。在GIS技术框架下,将GIS数据库与小麦施肥模型结合,利用ArcGIS提供的二次开发平台ArcGIS Engine和C#语言,开发出基于GIS的小麦施肥决策支持系统。施肥决策系统实现对空间数据和属性数据管理,以数据库为信息源,施肥模型为决策支持进行施肥推荐。从微观和宏观尺度上,系统分别为农户和县级农业部门提供小麦施肥决策,并为构建其它作物的栽培管理决策支持系统提供了开发框架和思路,为精确农作和数字化农作提供技术支持。更多还原

【Abstract】 As one of the main grain crops in China, wheat is the most important commodity grain and strategic variety for grain reserve, which yield and planting area accounted for 1/5 and 1/4 of the total yield and planting area in total food crops. It is of great significance to establish N, P and K fertilizer recommendation system of wheat in China based on different climatic characteristics, soil types and soil nutrients status in guiding crop fertilization and improving farmland ecological environment. This paper studied the effect of climate factors such as temperature, precipitation and soil factors such as soil type, texture on wheat yield, soil fertility capacity and fertilizer use efficiency according to the field trials of wheat in multiple years and sites. Based on nutrient balance and soil tests, a wheat fertilizer recommendation system was established, and wheat fertilization decision-supporting system was also developed using secondary development of ArcGIS. Main conclusions were obtained as follows:（1） Determination of basic parameters in nutrient balance fertilization modelTarget yield was used as an independent variable in the fertilization model and the upper limit of yield was limited by climatic potential productivity. Statistical analysis showed that a significant linear correlation was reached between nutrient uptake of wheat biological yield and its corresponding grain yield. The nutrient uptake per unit yield was trended to a stable range. Therefore, we can set the nutrient uptake per unit yield as a constant in this study. For per 100 kg yield of winter wheat, 3 kg N, 1.2 kg P₂O₅ and 2.8 kg K2O nutrient were respectively needed. And for per 100 kg yield of spring wheat, 3 kg N, 1.1 kg P₂O₅ and 2.6 kg K2O were respectively needed. The calibration coefficient of soil available nutrients and the fertilizer utilization efficiency were respectively calculated based on the experiment data. It was studied the variability of the two parameters on 27 soil types in China. The average value of the two parameters for each soil type was different and the variability which indicated by the standard deviation in the same soil type was also significant. Extremely high discrete degree also showed the high variability in most soil types. The calibration coefficient of soil available nutrients and the fertilizer utilization efficiency were influenced by average temperature and precipitation during wheat growth, soil clay content, pH value, irrigation, and so on.（2） Establishment of two sub-models for calibration coefficient of soil available nutrient and fertilizer utilization efficiencyIt was established the sub-model for calibration coefficient of soil available nutrient by using the factors including soil available nutrient （alkaline hydrolytic nitrogen, available phosphorus and available potassium）, soil clay content and average temperature during wheat growth period. The goodness-of-fit was higher which the coefficients of determination （R2） were from 0.55 to 0.85. The fitting accuracy of sub-model for calibration coefficient of soil available nutrient was available phosphorus higher than available potassium, and than alkaline hydrolytic nitrogen. It showed a significant negative correlation with logarithm function between soil nutrient contents and fertilizer utilization efficiency. The sub-model for fertilizer utilization efficiency was established by using soil nutrient content as the independent variable at different yield level. The goodness-of-fit was also higher which the related coefficients （r） were from 0.45 to 0.75 and showed a significant difference. The two sub-models were revised by using soil pH and irrigation factors, respectively.（3） Establishment and validation of nutrient balance fertilization modelIn the nutrient balance fertilization model, the target yield parameter was set as input item, the nutrient uptake of per unit yield was set as a constant. Combining with the two sub-models of calibration coefficient for soil available nutrient and fertilizer utilization efficiency, a comprehensive nutrient balanced fertilization model was established. Target yield and soil available nutrient testing value were set as independent variables, NPK fertilizer application rates as dependent variable. A nationwide fertilizer recommendation system of wheat was established based on 15 soil groups incorporated from 27 soil types. The recommended fertilization from nutrient balance fertilization model was validated by the wheat field trial results arranging the main trial area. The recommended fertilizer rate based on the nutrient balance fertilization model was between the rate for the maximum yield and the optimized economical benefit yield recommended by fertilizer response function model. It indicated the model was feasible for fertilizer recommendation for wheat with simple, fast and accurate characteristics.（4） Building basic spatial database of the ecological and environmental factors for wheatOn the GIS platform, the basic spatial database including each ecological and environmental factor which impacted the wheat fertilization model parameters was built based on meteorological data and soil data such as soil types and soil texture etc. Different kinds of layers were constructed by interpolation, conversion between vector and grid layer, and layer superposition treatment etc. The layers included the grid layers of average temperature, average precipitation during wheat growth, soil clay content and pH value, wheat water requirement and the climatic potential productivity yield etc, and vector layer of the minimum fertilization unit by superposing county boundary map with soil type map in China.（5）Development of wheat fertilization decision-making system based on GISCombining GIS database and wheat fertilization model, the wheat fertilization decision-making system was developed by using ArcGIS Engine and C# language under GIS environment. The layer management of the spatial data and attribute data was realized through the fertilization decision-making system. The fertilizer recommendations were performed based on the database as information source and the fertilization model as decision-making support. The system could support wheat fertilization decision-making for farmers and agriculture departments at the micro-sale and macro-scale level respectively. The development of this system could provide a framework and way to develop the decision-making system for other crops, and technical support for precision agriculture and digital griculture.更多还原