基于不确定情景构建和挖掘的区域突发环境事件风险分析与应对策略研究

Regional sudden environmental incident risk assessment is the scientific basis for risk prevention at the planning stage. In general, Integrated multi-objective risk assessments focus on identification of environmental risk factors such as sources, receptors and emergency prevention and control ability in a static perspective, which is set by limited planning scenarios. Without scenario uncertainty analysis, it is unlikely to reach the reliability of results and declare the future risks in advance. To address these issues, this project aims to support the regional area planning by proposing a creative risk analysis method based on construction and mining for uncertain environmental risk scenarios. On the basis of the global sensitivity analysis of the risk assessment model，key uncertain elements in planning will be gathered, then the scenario discovery algorithm and the risk assessment model will be employed to generate a large database of environmental risk scenarios. Subsequently, statistical analysis and visualization technologies will be conducted for tracing the responses of environmental risk value under various uncertainty scenarios. Key factors and risk characteristics of unacceptable risk scenarios will be determined, and coping strategies will be made. This project is innovative in terms of methodology and management strategy, will help improve the scientific nature of risk prevention, and provide a scientific method and tools for sudden environmental incident risk management in China.

区域突发环境事件风险评估是从规划层面加强区域风险防控的重要科学依据。目前，综合化多目标的风险分析方法多基于有限规划情景对环境风险源、风险受体和防控应急能力等关键要素进行识别和表征，缺少对情景不确定性的解析或预测，影响结果的可靠性和对风险的应对能力。鉴于此，本研究尝试探讨一种基于不确定情景构建和挖掘的区域风险逆向分析策略：在风险评估模型全局敏感性分析的基础上，识别影响模型结果的关键不确定因子；针对性地识别规划中的不确定要素，利用情景发掘算法和风险表征模型生成完备的环境风险情景数据库；利用统计分析和可视化技术来研究不确定情景下的风险响应；分析高风险情景的关键因子与生成机制；制定不确定情景的控制策略。本研究从方法和管理策略上进行创新，将有助于提高风险管控的科学性，为我国区域突发环境风险管理提供方法和工具支撑。

区域环境风险防控是我国现阶段生态文明建设的重要任务，也是区域环境生态安全的重要保障，研究从四个方面开展研究：（1）面向环境风险受体的脆弱性，基于生态敏感性-恢复力-压力度框架，构建脆弱性指标体系评估模型，并评估模型的不确定性和结果空间变异特征；（2）面向风险受体的价值核算，提出耦合谷歌地球引擎、随机森林算法和时空调节因子的生态系统服务价值评估模型；并应用于区域土地利用变化和价值核算，识别环境生态价值变化的关键驱动因素和功能衰减的热点区域；（3）面向区域风险应急防范能力评估，基于两步移动搜索法研究了环境风险防范核心应急资源（应急避难场所、应急站点和消防站等）在不同时间范围内的可达性特征；（4）面向环境风险信息化管控需求，基于 C/S 结构，.NET 等技术构建基于风险严重度-环境脆弱性-应急管理能力的区域环境风险决策支持系统和面向污染损害的生态服务价值支持系统，实现风险源管理、风险分区、风险与价值评估等功能，为区域风险管理提供工具及策略。相关研究成果为区域环境风险管理实践提供了支持。