【作者】 曹静；

【导师】 曹卫星；

【作者基本信息】 南京农业大学 ， 生态学， 2008， 博士

【摘要】 农业生产过程的精确管理模型及决策支持系统是变量处方生成的数字化基础,是整个精确农作信息流管理中的核心环节。本研究以小麦、水稻为对象,运用系统学分析方法和定量建模技术,基于作物-土壤-气候-技术的动态关系,建立了精确农业管理模型的框架结构及共性算法,重点构建了水氮运筹设计模型、生长指标动态模型,进而建立了基于作物和土壤实时养分及水分状况的精确诊断与动态调控模型；最后,基于农田信息获取技术,结合GIS和数据库,运用软构件设计方法,研制和实现了基于模型和GIS的精确农作决策支持系统。研究成果为构建精确生产管理模型提供了基本框架,也为精确衣作技术的应用奠定了技术基础。通过综合分析作物生产管理系统的相关成分及影响因子,提出了精确农作管理模型的结构及框架,包括播前栽培方案设计、产中适宜生育指标预测以及基于实时苗情信息的作物生长精确诊断与动态调控三大模块。在此基础上,利用作物栽培理论与技术方面的已有研究资料,借助动态建模技术,对作物生育及栽培管理指标与品种类型、生态环境及生产力水平之间的关系进行解析、提炼和综合,建立了各功能模块的算法框架,为有效利用精确农作管理模型中的通用函数构件,提高精确农作模型系统的开发效率提供了技术指导和基本框架。通过定量分析和提炼作物水氮管理方案与栽培技术、品种类型、生态环境之间的动态关系,建立了具有时空适应性的作物水氮运筹设计模型。模型包括氮肥运筹和水分管理两个子模块。氮肥运筹模型借鉴养分平衡原理,综合考虑了气候、土壤、品种等多种因子的影响,通过将不同类型稻麦品种按产量和品质进行分类,确定了不同品种类型的百公斤籽粒吸氮量,并根据目标产量和空白区产量计算目标产量需氮量和土壤供氮量,确立了氮的合理用量及基追比。水分管理模型基于水分平衡原理,采用土壤水势作为灌溉指标,动态计算了各主要生育期的适宜水分灌溉量。利用历史文献资料及大田试验资料对氮肥运筹模型和水分管理模型的可靠性和适应性进行了测试与检验,结果表明本模型具有较好的指导性和广适性。根据作物生理发育时间(PDT)恒定的原理,以PDT为生育期预测器,通过动态计算到达不同环境条件下各主要生育时期所需的累积生长度日(GDD),然后将生长指标和GDD进行归一化处理,以相对GDD和相对生长指标为参数建立了适宜生长指标动态的相对变化曲线。根据有理方程,建立了作物相对叶面积指数动态和相对茎蘖数动态;根据Logistic增长模式,建立了相对干物质积累动态和相对植株氮积累动态。在建立相对生长指标动态模型的基础上,通过进一步计算适宜条件下的最大生长指标值,从而得到实际的适宜生长指标动态。利用不同品种、不同栽培处理的试验资料对模型设计值进行了对比分析,表明模型对于高产条件下的作物生长指标适宜动态具有较好的预测性和指导性。通过分析作物生长动态与栽培措施、生态环境等之间的关系,确定了作物生长的诊断指标,并结合作物适宜生长指标动态模型及作物水氮运筹设计模型,建立了基于作物和土壤实时养分及水分状况的精确诊断与动态调控模型。模型以适宜生长指标动态模型设计的适宜生长指标为标准“专家曲线”,当田间实时苗情明显偏离标准的“专家曲线”时,模型通过计算作物长势调控因子,再结合作物水氮运筹设计模型,给出适宜的施肥灌溉量及调控时期。经田间资料测试表明,模型对于不同条件下的作物生长动态优化管理具有较好的科学性和适用性,克服了传统作物栽培模式及农业专家系统中经验性知识规则的时空适应性差等弱点,为作物生长动态的优化管理提供了定量化工具和标准化模式。在构建精确农作管理模型的基础上,应用面向对象的程序设计与构件化技术,并有效耦合GIS技术,在.NET平台上运用C#语言设计和实现了基于模型和GIS的精确农作决策支持系统(MGPFDSS),实现了不同尺度下精确农作生产管理决策的精确化和动态化,为精确农业体系中主要农作物的全过程变量处方生成提供了模型化和系统化的决策工具,为精确农作和数字农作的发展奠定了技术基础。更多还原

【Abstract】 The precision crop management model and decision support system is the digital base of making variable prescription, and is the core of precision farming information management. On the basis of the time-course relationships of crop-soil-climate-technology, using the system analysis method and mathematical modeling technique, this study developed a basic framework for design and development of precision crop management model and mainly developed three models including water and nitrogen strategy design model, dynamic of suitable growth indices model and growth dynamics regulation model. By further adopting the component method and integrating the GIS and database technology, a model and GIS-based decision support system for precision farming was established. This work has provided a basic platform for development of precision crop management model, and laid a technical foundation for the practical application of precision farming system.By combining with the experts of crop cultivation and analyzing the crop management problems, a framework and structure of precision crop management model was developed under the guide of system engineering. The model included three modules as pre-sowing plan deign, suitable dynamic indices prediction and real-time regulation according to actual crop growth status. Based on the analysis and extraction of the newest researches on crop cultivation theories and technologies, using the system analysis theory and method, the arithmetic framework of each module was developed by quantifying the time-course relationships of crop growth and cultivation management indices to variety traits, eco-environments and production levels. This work provided the technology guide and framework for effectively using the general function components in the precision crop management model, and for improving the development efficiency of the precision crop management model.Based on the analysis and extraction of the newest researches on crop cultivation theories and technologies, a design model for water and nitrogen strategy in crop was developed driven by quantifying the time-course relationships of water and nitrogen management plan to production levels, variety traits and eco-environments. The model included two modules as nitrogen strategy and water management. The submodel for nitrogen strategy was developed with the principle of nutrient balance and by combining the effects of climate, soil and variety. In the submodel, N demand per 100 kg grain was calculated by classifying the rice and wheat varieties according to yield and quality, and the N demand for grain yield target and indigenous soil N supply were quantified by the grain yield target and the grain yield in the plot without N fertilizer. The submodel for nitrogen strategy can make decisions on the suitable total N rates and the ratio of base to topdressing N. Based on the principle of water balance, the submodel for water management was developed by adopting soil water potential as irrigation index. The model was evaluated using the data from literatures, field experiments of diverse eco-sites and years. The results indicated the model’s good guidance and wide applicability.According to the concept of constant physiological development time (PDT), the development stages under different environment were predicted with PDT. Then through dynamically calculating growing degree days required by main development stages under different environments, the dynamic model for growth and nutrient diagnosis and management in crop driven by physiological development time-based relative growing degree-days and relative growth indices was developed.The relative dynamic curve of leaf area index and stem and tiller number was according to the Raditional Function, and the relative dynamic of dry matter accumulation and plant nitrogen accumulation was established according to the growth characteristic of Logitic curve. On the basis of the relative growth indices dynamics, the model for design of growth indices dynamics in crop was developed by calculating the maximal indices under suitable conditions. Case studies with the data from field experiments at Nanjing in three years indicate a good performance of the model system for prediction and wide applicability.Based on the analysis and extraction of the newest researches on crop cultivation theories and technologies, a model for growth dynamics regulation in crop was developed combining with suitable dynamics of indices model, water and nitrogen strategy model. The growth diagnosis indices were confirmed by analyzing the time-course relationships of crop growth dynamics to production levels and eco-environments. The model made the suitable growth and nutrient indices designed by suitable dynamic of indices model as standard expert curve. When the actual crop growth condition departed from the expert curve, the model analyzed the reason and studied out the suitable dressing N application and irrigation amount by calculating the regulation factor and combining with the water and nitrogen strategy model. Case studies with the data sets of field experiments indicated the model’s wide applicability and good guidance. The model avoided the deficiencies of regional limitation and empirical decision with traditional crop cultivation patterns and expert systems, and thus provided a promising quantitative tool for crop growth management.On the basis of the precision crop management model, a model and GIS-based decision system for precision farming (MGPFDSS) was established on the platforms of.NET with the programming language of C# by using the techniques of object-oriented and component-based programming and incorporating the GIS. The MGPFDSS provided a modeling, parametric, dynamic and systematic tool for making variable prescription during the process of crops growth and development, and lay a foundation for the development of precision farming and digital farming.更多还原