变化环境下考虑不确定性的流域洪水演变及适应性调控研究

Coupled effects of climate change and intensive anthropogenic activities have induced frequent occurrences of extreme flood events, altered evolvement rules of floods, and increased flood risk. Therefore, carrying out study on adaptive regulation system for flood control at basin scale should be attached great importance, which is extremely beneficial for exploring the safety of flood control system and reducing flood risk. To achieve this purpose, this study takes the upper Hanjiang basin, one flood-prone region, as research object, which is also one significant strategic region of water resources allocation to our country. Meanwhile, we proposed one gradual and deep-going three-step research strategy, i.e. "recognition, reconstruction and adaptation", to systematically disclose the evolvement rules of floods and related adaptive regulation mechanism in the context of multiple uncertainties under changing environment. Here, recognition is the basis, focusing on uncovering evolvement rules of floods at basin scale under changing environment; reconstruction is the key, aiming to construct flood events in the past and future through incorporating climate model and distributed hydrological model; adaptation is the target, devoting to explore the adaptability of flood control system to changing environment, propose scientific adaptive regulation rules to flood control for reservoir(s) and identify flood control capability of river basin through establishing reservoir(s) operation model and constructing relevant evaluation index system. Related results are of great importance for flood risk management at the basin scale under changing environment.

气候变暖和剧烈的人类活动共同作用导致极端洪水事件频发，改变了流域洪水规律，增加了洪涝灾害风险，因此，亟需开展流域适应性调控的系统研究，以期判断流域的防洪安全并减少洪水灾害。为实现这一目标，本次研究以国家水资源配置战略水源地和洪灾频发的汉江上游流域为研究对象，提出了“识别—重构—应对”的三层次逐步递进研究思路，考虑多重不确定性因素的影响，系统研究变化环境下的洪水演变规律和适应性调控机制。“识别”是基础，重在定量分析变化环境下流域洪水时空变化规律；“重构”是关键，通过耦合气候模型与分布式水文模型等手段，旨在构建非一致性条件下过去与未来洪水序列；“应对”是目标，建立水库（群）调度模型，构建评价指标体系，揭示水库（群）防洪调控对环境变化的适应能力，提出合理的水库（群）适应性防洪调控规则，判断变化环境下的流域防洪能力。研究成果对变化环境下流域洪水灾害风险管理具有重要意义。

本研究以揭示变化环境下的洪水演变规律及适应性洪水调控策略为目标，以国家水资源配置战略水源地和洪灾频发的汉江上游流域为研究对象，从“识别-重构-应对”三个层面出发开展研究。诊断了流域石泉、安康水文站极端水文序列变异情况，针对趋势变异、跳跃变异等变异类型，提出了相应的洪水序列重构方法，重构了流域历史洪水序列；采用SOM、BP等神经网络方法，对流域洪水进行了聚类分析，建立了洪水分类预报Topmodel模型及实时校正模型，计算了CMIP5气候模式下RCP4.5、RCP8.5情景下的设计洪水；考虑概率分布线型、参数及样本多重不确定性因素，提出了基于Bayes理论的设计洪水计算方法；基于信息熵理论，量化了基于Copula函数的联合分布模型参数不确定性对设计洪水计算结果的影响，发现边缘分布的形状参数与尺度参数对设计洪水不确定性的影响程度最大，Copula函数参数影响次之；针对设计洪水不确定性，分析了其产生的水库防洪调度风险情况，并推荐了在此环境下安康水库汛限水位的适宜动态控制域为325m-327.5m；考虑洪水实时预报不确定性，建立了水库多目标防洪调度模型，采用随机动态规划算法求解，推荐了特定预报精度下的水库防洪调度规则。研究成果可为变化环境下流域防洪规划、调度提供借鉴。