川西地区森林野火风险预警遥感理论与方法

川西地区冬春季森林野火频发，多次造成人员伤亡，急需发展高精度的森林野火风险预警理论与方法。目前森林野火风险预警存在可燃物时空信息严重不足的理论缺陷，以及缺乏深入研究诱发因子多维特征及其与野火爆发之间的高度非线性关系等技术难题。这些关键科学与技术问题在面向川西复杂环境时更加凸显，严重影响野火预警效果。该项目拟通过多学科交叉研究，构建适合川西地区的森林野火风险预警理论与方法体系：一方面，研究不同类型可燃物的辐射传输、后向散射及能量与水分平衡过程等机理，发展复杂环境可燃物遥感反演理论与方法，实现空间连续、长时间序列的可燃物信息提取，构建川西森林可燃物数据集；另一方面，剖析川西历史森林野火事件的时空格局，建立动静态一体的野火诱发因子体系，耦合野火环境三角模型与多类别时空大数据挖掘方法，构建野火风险预警时空大数据挖掘混合模型，实现高精度的森林野火风险预警，服务于川西森林火灾风险预警与防控的重大需求。

Theory and methodology on high-precision forest wildfire risk early-warning for western Sichuan province are urgently needed, as a result of the frequent wildfires and consequent casualties in this area. However, spatial and temporal information of the fuel variables are not available at present due to their theoretical inversion issues, and the multidimensional characteristics of the fuel variables and other driving factors (weather and topography variables) were still unclear nowadays. These issues are more prominent when facing the case in the complex environment of western Sichuan province. Therefore, this project intends to build a wildfire risk early-warning framework for the complex environment of western Sichuan province. To this end, this project intends to develop the remote sensing-based theory and methodology for retrieving spatial and temporal fuel information in complex environments. Besides, by analyzing the spatiotemporal pattern of historical wildfire events in western Sichuan province, this project intends to establish wildfire database comprising dynamic and static driving factors, and then build a hybrid model for wildfire risk early-warning by coupling the wildfire environment triangle model and multi-category spatiotemporal big data mining methods. This project highlights the importance of theoretical and methodological support for the wildfire risks early-warning framework in the complex environment of western Sichuan province, allowing to serve this region in the field of wildfire risk warning and prevention.