作物生长模型和遥感数据同化的双尺度作物氮素预测方法研究

Crop nitrogen status is an important indicator in field management and crop yield and quality prediction. Difficulties exist in crop nitrogen dynamic monitoring and prediction by conventional methods with destructive sampling and expensive laboratory analysis and by remote sensing with instantaneous satellite imageries. The objective of the project is to achieve the dynamic monitoring and prediction of two-scale nitrogen status (i.e. leaf nitrogen content for leaf scale and crop nitrogen accumulation for canopy scale) based on assimilation of DSSAT-CERES-Wheat model and remotely sensed data. With field experiments of winter wheat, the main work of this study are as follows: 1) localizing the parameters of DSSAT-CERES model, especially those of the nitrogen simulation module, and then carrying out the two-scale nitrogen status simulation with the crop model; 2) constructing the inversion models of two-scale nitrogen by combining crop spectral information and the agronomy knowledge and models; 3) establishing the approaches for three assimilation ways of crop model and remotely sensed nitrogen data, which using the leaf nitrogen content for leaf scale, the crop nitrogen accumulation for canopy scale and both the two-scale nitrogen parameters respectively as the assimilation state variables; 4) validating the above proposed models and approaches at different sites and quantitatively assessing their stability and performances. The expected outcomes of this project will further enrich the monitoring method system of crop nitrogen status, which will be useful in carrying out the dynamic monitoring and prediction of crop nitrogen status and improving its practical application value.

作作物氮素状况是进行田间管理和产量品质预报的重要指标。田间调查取样分析法和遥感反演方法均难以实现作物氮素状况的连续动态监测与预测。项目拟以冬小麦为研究对象，基于田间实验观测，借助作物生长模型和遥感同化手段，开展叶片尺度氮含量和群体尺度氮素累积量（即双尺度氮素）的连续动态监测与预测研究。主要内容包括：① 本地化DSSAT-CERES作物模型参数，特别是氮素模块参数的校正，利用作物生长模型开展作物双尺度氮素水平的精确模拟研究；② 综合利用光谱信息和农学知识模型，构建作物双尺度氮素指标的遥感精确反演模型；③ 分别采用以叶片尺度氮含量、群体尺度氮素累积量及双尺度氮素指标共同作为状态变量等3种同化方式，开展作物模型和遥感氮素同化研究；④对所构建的模型方法进行多站点验证，定量评价其稳定性和实际效用。研究成果将进一步丰富作物氮素状况监测方法体系，实现作物氮素状况的连续动态监测与预测，提高其实际应用价值。

作物氮素状况是进行田间管理和产量品质预报的重要指标，而传统的田间调查取样分析方法和遥感反演方法均难以实现作物氮素状况的连续动态监测与预测。本项目借助作物生长模型和遥感数据同化手段，开展叶片尺度氮含量（LNC）和群体尺度冠层氮密度（CND）的连续动态与预测研究，为实现作物氮素诊断从“遥感瞬时”向“时相连续”扩展提供有效手段。主要研究内容包括：(1) 本地化 DSSAT-CERES 作物模型参数，特别是氮素模块参数的校正，利用作物生长模型开展作物双尺度氮素水平的精确模拟研究；(2) 综合利用光谱信息和农学知识模型，构建作物双尺度氮素指标的遥感精确反演模型；(3) 分别采用以叶片尺度氮含量、群体尺度氮素累积量及双尺度氮素指标共同作为状态变量等3种同化方式，开展作物模型和遥感氮素同化研究；(4) 对所构建的模型方法进行多站点验证，定量评价其稳定性和实际效用。按照项目研究计划和目标，本课题完成了上述各项研究任务并开展了广泛深入的国内外合作交流；基于研究成果，已发表学术论文26篇（其中SCI论文16篇，EI论文8篇）；撰写学术专著1部；申请授权发明专利2项；晋升中国工程院院士1人，高级职称人才1人，副高级职称2人，中级职称1人，培养博士研究生6名，硕士研究生5名。综上，本项目圆满完成了各项预期指标，具体信息详见后述。