冬小麦旱灾致灾因子危险性高光谱遥感定量诊断研究

Hazard degree is the basic function for disaster risk assessment and loss assessment. Quantitative diagnosis of drought hazard degree is essential to decision-making of drought early warning and emergency relief in practice. The currently used diagnosis methods are based on disaster loss and drought indices. The response process and impacts of drought in different crop growth stages were ignored in these methods. So, the instructions were not dynamic and real-time. This study investigated the drought hazard degree diagnosing of winter wheat in Henan province based on continuous hyper-spectral remote sensing imagery and comprehensive ground-based observations. Considering the complicated association of drought driving factors, formation and effects and obvious regional disparity, the major hazard factors such as the meteorology factors, hydrology factors, geography factors, agriculture factors were determined and assessed quantitatively. Then, a functional and indicative drought hazard degree index of winter wheat was built based on the functional relationship between the various hazard factors. After that, spatial and temporal distribution of drought hazard degree index in different stages and different county-level administrative units was performed to identify the hazard diagnosing level and describe the signs of drought, combined with the quantitative disaster information. This study is important to improve the time effectiveness and precision of quantitative diagnosis of drought hazard degree. It would promote the hyperspectral remote sensing application in drought early warning and emergency relief response.

致灾因子危险性是旱灾风险和损失评估最基础的函数，定量诊断致灾因子危险性是旱灾预警和应急救助决策重要的现实需求。目前常用的灾损和干旱指数诊断方法，动态、实时指示特征不明显，未充分顾及不同作物发育期的旱灾响应过程和影响效应。本项目针对旱灾成因复杂、区域差异显著等特点，以河南省冬小麦为例，基于高光谱遥感连续成像和地面观测信息综合的技术优势，分析冬小麦旱灾驱动因素-形成过程-影响效应的关联关系，筛选并确定气象、水文、地理、农业等主要致灾要素及其定量提取方法，构建机理明晰、实时指示特征明显的致灾因子危险度量化指数，并根据不同的冬小麦生育期和县级行政单元，分析致灾因子危险度时空分布格局，结合灾情数据定量评价致灾因子危险度与危险性等级之间的关系，开展危险性分级诊断，并揭示其旱象特征和影响程度。本项目的实施可提高致灾因子危险性定量诊断的时效性和精度，促进高光谱遥感技术在旱灾预警和应急救助响应中的应用。

致灾因子危险性是旱灾风险和损失评估最基础的函数，定量诊断致灾因子危险性是旱灾预警和应急救助决策重要的现实需求。目前常用的灾损和干旱指数诊断方法，动态、实时指示特征不明显，未充分顾及不同作物发育期的旱灾响应过程和影响效应。本项目针对旱灾成因复杂、区域差异显著等特点，以河南省冬小麦为例，基于高光谱遥感连续成像和地面观测信息综合的技术优势，分析冬小麦旱灾驱动因素-形成过程-影响效应的关联关系，筛选并确定气象、水文、地理、农业等主要致灾要素及其定量提取方法，构建机理明晰、实时指示特征明显的致灾因子危险度量化指数，并根据不同的冬小麦生育期和县级行政单元，分析致灾因子危险度时空分布格局，结合灾情数据定量评价致灾因子危险度与危险性等级之间的关系，开展危险性分级诊断，并揭示其旱象特征和影响程度。本项目的实施可提高致灾因子危险性定量诊断的时效性和精度，促进高光谱遥感技术在旱灾预警和应急救助响应中的应用。主要取得研究成果包括：1）研发了一种基于区域生长法和自动阈值选择的全自动化的水体提取方法，可用于旱灾造成的水体变化信息自动提取；2）针对不同的冬小麦生育期构建基于高光谱遥感数据和其他致灾要素的冬小麦旱灾致灾因子危险度指数；3）基于国家标准和行业标准发展基于历史灾情数据的冬小麦旱灾致灾因子危险性分级诊断方法。本项目相关研究成果已经在防灾减灾救灾领域开展业务化应用，具有很好的适用性和推广价值。