基于作物生长模型和遥感数据同化的大尺度水稻重金属污染胁迫动态监测

This research takes the typical heavy metals contaminated paddy fields in Xiangjiang River Basin as the experimental area and aims at the large-scale dynamic monitoring of heavy metals pollution of rice using remote sensing. The main purpose of this research is exploring the dynamic spectral models for assessing heavy metal contamination stress of rice, revealing its spatial scale expansion mechanism and developing the assimilation methodology of multi-scale observational data and the local dynamic models. Based on the analysis of the spectral response mechanisms of heavy metals contaminated rice and our previous work, the parameter inversion theory will be explored, and the dynamic spectral models for reflecting the stress variations of contaminated rice in different growing stages will be constructed by combining the crop growth model and PROSPECT+SAIL model. The differences and relationships of the spectral effects of heavy metal contamination stressed rice in the different observed scales will be analyzed; thereafter the spectral conversion laws of pollution stress characteristics at different scales will be clarified in this project. Integrating experiments and the physical models, using Spatial and temporal interpolation, model simulation, climate zoning, raster conversion, normalization and other technologies, the impacting mechanism of the environmental spatial heterogeneity on end member and pixel spectrum will be explained, and the parameterization method of scale conversion model for evaluating the rice heavy metal stress will be established. In this research, the assimilation mechanisms and modalities for satellite remote sensing data at different scales and the local spectral dynamic models will be explored, and the spatial scale conversion of the dynamic inversion models for assessing heavy metals contamination stressed rice will be fulfilled. This project is to solve the major scientific issues in spatial scale conversion of point models, and help to improve the accuracy and application of the large-scale satellite remote sensing model.

以湘江流域典型重金属污染区为实验区，以大尺度水稻重金属污染遥感动态监测为目标，研究水稻重金属污染胁迫遥感反演动态模型建立及空间尺度扩展机理，发展多尺度遥感数据与动态模型同化方法。在作物生长模型中嵌入重金属胁迫因子，模拟重金属积累对水稻生长参数的影响及其光谱变化，并耦合辐射传输模型PROSPECT+SAIL建立水稻重金属污染胁迫动态光谱反演模型；建立区域尺度遥感-作物模拟同化框架模型，结合实验并运用时空插值、模型模拟、气候带划分和栅格化、归一化等技术对模型变量进行参数区域化，构建水稻重金属污染胁迫尺度转换模型参数化方法；探索卫星遥感数据与局地光谱动态模型的同化机制和模式，实现水稻重金属污染胁迫光谱反演模型空间尺度转换。通过本项目的研究，解决点模型空间尺度推演面临的主要科学问题，提高大尺度卫星遥感反演模型的精度与应用效果。

农田重金属污染是当今世界面临的重大生态环境问题之一，对全球环境质量、粮食安全和人类生存构成威胁。本项目针对大尺度水稻重金属污染胁迫遥感监测面临的关键科学问题，选择湘江流域典型重金属污染区作为试验区，以大尺度水稻重金属污染胁迫遥感动态评估为目标，开展了以下研究：分析了重金属胁迫对水稻生长及其光谱的影响机制，探索了卫星遥感观测数据与作物生长模型的耦合机制与模式，构建了基于遥感信息和作物生长模型同化的水稻重金属污染胁迫遥感监测模型。主要研究成果包括：从重金属胁迫对作物生长状况与生理功能影响的角度出发，构建了水稻重金属胁迫遥感监测与评估指标，动态模拟了重金属污染胁迫下水稻生长参数及其变化特征，揭示了不同浓度重金属污染胁迫下水稻生长状况变化及其差异性特征，实现了水稻重金属污染胁迫遥感监测；面向区域大尺度水稻重金属胁迫动态监测的应用需求，开展了同化框架尺度扩展研究，提升了水稻重金属胁迫遥感动态监测模型区域大尺度应用效果；通过揭示水稻重金属胁迫冠层辐射能量变化特征，建立了基于能量平衡理论的水稻重金属污染胁迫遥感监测方法；基于重金属污染对水稻生长周期持续性胁迫特征，探索并揭示了水稻遥感物候对重金属胁迫的响应规律。本课题的研究对水稻重金属污染胁迫遥感监测及其空间尺度扩展面临的主要科学问题进行了较为系统的探索，提高了水稻重金属胁迫遥感动态监测的可信度，突破了同化框架尺度外推的瓶颈，实现了大尺度水稻重金属污染胁迫动态监测，提升了作物重金属污染胁迫监测技术工程化应用的可行性，为农田生态系统变化遥感动态监测提供了新的思路和参考。