区域小麦赤霉病遥感动态预测方法研究

气候变化引起的病虫危害日益加剧，植保部门手工或半机械化调查方式和农气结合的时空粗放尺度预警模式，难以满足精准防控需求。本项目围绕精细尺度小麦赤霉病遥感动态预测机理和方法，在前期研究病害遥感特征参量提取方法的基础上，进一步深入研究：（1）开展小区控制实验和区域调查实验，解析病害的发生特征和流行状态与作物生长状况的光谱响应机理和与生境条件的时空响应机制，构建时空连续的病害遥感预测指标体系；（2）研究病害预测指标在病害发生发展过程模型中的输入和更新模式以及参数化方法，并与流行病学规律和机理解析相结合，实现病害动态预测建模；（3）融合多源数据评价病害生境适宜性，更新精细尺度下模型在不同地域的算法逻辑和联动关系，扩展模型的空间普适性。预期研究结果将突破现有基于气象信息建立的模型无法获得农田病害发生初始态和流行速率的现状，实现县域尺度病害动态预测高精度建模，为小麦赤霉病精准防控提供理论与技术支持。

Agricultural systems are facing a growing challenge of disease and pest threats by climate change. Ground monitoring of disease severity and pest population density, and meteorological-based forecasting models are not enough to support the scientific control. We urgently need spatial and temporal continuous hazard dynamic forecasting to promote decision making at regional scale. Taking wheat Fusarium head blight as the object, we will conduct the research of remote sensing mechanisms and methods for its early forecasting and epidemic dynamic forecasting, based on our previous research of disease characteristic parameters extraction. The main contents include: (1) conducting field observations and regional investigation experiments, analyzing the spectral response mechanisms of disease occurrence/epidemic characteristics with dynamic crop growth conditions and habitat, and then constructing spatiotemporal remote sensing forecasting indices for wheat Fusarium head blight; (2) exploring the effective input mode, update mode and parameterization method of disease forecasting indices with the disease occurrence and epidemic model, and then achieving the dynamic forecasting of disease by integrating with the disease epidemic mechanisms; (3) integrating multi-source data to evaluate the suitability of disease habitat, and updating the algorithm in different regions to expand the spatial universality of disease forecasting model. Our project outputs will construct the forecasting indices for wheat Fusarium head blight, and then take the initial state and its changing information of these indices as the input of epidemic model to achieve spatiotemporal disease forecasting. Furthermore, the results could be used to support decision making of plant protection for wheat Fusarium head blight at county scale.