随着全球气候变化，极端天气事件增多，城市雨岛效应逐渐增强，城市化过程所带来的环境与灾害问题日益突出。对于人口密集、财产集中和文明中心的各大城市，内涝灾害带来的直接和间接损失尤为严重。而在突发的城市内涝灾害过程中暴露出来的城市内涝淹没过程模拟难、内涝隐患推演分析难等问题，成为城市防涝减灾与应急决策的瓶颈。当前对其研究的主要方法有结合水文学产汇流和水动力学的城区“双排水系统”法、一维二维水动力方程法、基于GIS的内涝淹没过程模拟法。针对当前研究中存在的栅格地表精细建模能力差、模型率定所需的实测数据要求高、精细排水管网数据可获性低、数值模拟耗时长、推演与隐患分析难等问题，本文从地理信息科学与城市内涝交叉研究的角度，采用约束特征集理论形成了一套城市内涝地表数据测量、组织和管理方法，进而采用CD-TIN构建精细城市地表。充分顾及影响地表径流和淹没的各关键约束特征，划分城区汇水单元以作为模拟的基本计算单元，对汇水单元内的产汇流进行建模并基于三棱柱的二分数值求解算法对城市内涝淹没时空过程进行模拟。利用本文模型和方法对试验区的内涝模拟进行了实例验证、对淹没风险进行了分析、动态推演了隐患区的淹没情景，为防灾预警提供了直接定量化的预案和平台。本文的主要的研究工作及成果总结如下：1）城市精细约束特征数据采集、组织、管理与精细建模方法城市内涝地表关键特征测绘、数据集成组织和精细建模，是进行城市内涝模拟的重要基础性空间数据。对影响城市地表汇水的各精细约束特征（FCFs）概化为约束点特征（如雨水篦、封闭空间出水口等）、约束线特征（马路牙、围墙等）、约束面特征（建筑物、草地等），并对其高程进行重点测量，形成城市内涝地表测量标准与入库技术流程。基于约束特征集理论组织和管理城市地表的各约束特征，将测绘的各类FCFs特征，统一组织编码入库并进行属性管理。进而，基于CD-TIN方法提出了一套顾及地表与道路、隔离带、马路牙、下穿式立交桥和地下空间等地表建（构）筑物的城市地表精细建模方法，实现了城市地表各属性特征的耦合和带拓扑结构的城市地表精细集成表达。2）顾及城市地表精细约束特征的面点和面边汇水单元划分模式依据三角形面片上汇流模式的不同，提出面点和面边汇流模式来划分城区汇水单元。（1）面点模式：根据CD-TIN构建的城市地表，进行面片顶点高程计算和各出水口的确定，即采用水往低处流的基本思想，设定三角形面片上水首先流向该面片的最低顶点，然后沿坡度最陡边流向邻域顶点，依次汇流追踪各三角形的顶点；根据该追踪的水流情况，标示各三角形的汇流终点，从而实现汇水单元的划分。（2）面边模式：先结合汇水边、分水边、过水边的定义，归类各三角形边的汇流类型，然后在汇水边中汇流追踪城区汇流路径，同时追踪过水边的汇流方向，标示各三角形汇流情况，从而实现汇水单元的划分，并随淹没积水量的增加给出了汇水单元动态划分方法；利用典型城区进行试验，表明了精细约束特征对水流的显著影响，分析验证了基于CD-TIN的方法能够更为精细地模拟城区汇流路径和汇水单元。3）基于水量平衡原理的汇水单元产汇流计算模型和淹没情景的时空序列模拟方法汇水单元内的净产流量为汇水单元内的产流量与流向邻域汇水单元的汇流量之差。各汇水单元内的产流量为降水量与下渗量、排水量之差。邻域汇水单元间的汇流量则依据水位是否到达汇水单元的出水口及其相邻汇水单元的水位高低进行定量计算。基于时间切片的时空序列快照模型，将连续内涝淹没过程划分为离散的时间切片，实现了各时间切片下的淹没场景的模拟。依据水量平衡原理，地表净积水量转换为淹没水量，建立了地表淹没情景的模型。通过以三棱柱为基本计算单元，采用数值二分求解某时间切片下的内涝淹没空间，进而基于序列快照模拟整个时空过程的内涝淹没情景。以北京师范大学为试验区对2012年“7.21”暴雨淹没进行模拟，并与实测数据对比最大误差为16.7%；与调研风险图对比验证了模拟范围的合理性。同时，对石景山金安桥区的“7.21”暴雨积水进行了模拟，与所获资料对比验证了金安桥下积水模拟的合理性。4）基于GIS的城市内涝淹没风险序列与隐患推演分析方法基于分层体系架构和面向对象的设计研发了城市暴雨内涝淹没模拟推演系统，可在工程尺度上对试验区内的重要街区、下穿式立交桥、地下空间等主要承灾体进行内涝淹没风险分析与隐患推演。以北京师范大学和石景山金安桥区为试验区，给出了不同暴雨重现期下的内涝淹没风险情景，形成了风险分析序列，可为城市隐患分析和防涝预警提供计量依据。从隐患推演的角度，本文顾及地下空间不同入口处的地表积水深度、地下空间室内地坪的坡度及排水能力等，利用双层CD-TIN的建模方法和地上下水量交互算法，研究提出了一种街区内涝并导致地下空间积水的动态仿真模型。以北京师范大学东门地下车库为例实现了地下车库在不同暴雨重现期下的淹没过程及其排水推演分析。以石景山区下穿式立交桥金安桥为例，模拟了其在不同暴雨重现期的内涝积水淹没过程，分析了不同的地表径流系数对积水量和淹没水深的影响程度，推演了50年一遇暴雨情形下穿桥添加排水泵站的排水情景、排水管网堵塞和下垫面改变的积水情景。本内容的推演方案，可为内涝积水排放和预演提供直接的定量化的防灾减灾方案。

With the changing of global climate, the increasing of extreme weather events and the enhancing of the urban rain island effect, the environment and disasters brought by urbanization has become a global concern. Urban served as center of population, property and resources, and one severe storm flood can result in dramatic property damage and financial losses. Due to the difficulty of inundation process simulation and weakness of incipient faults simulation analysis, the emergence urban storm flood has been the bottleneck for urban disaster reduction and emergency decisions. Current researches can be summarized as three types: dual urban drainage system method, one-dimensional and two-dimensional hydrodynamic method, GIS-based simulation method. However, there are defects for these methods, such as the limitation of raster to model accurate urban surface, the high requirement of real-time or historical data to calibrate parameters, the low possible of obtaining the detail drainage system data, the long time consuming for numerical simulation, and the difficulties to simulate the inundation scenarios and analysis the incipient faults. By employing the geographic information science, this paper can handle the above defects in certain extent. Based on the theory of constrained features set, this paper proposes the methodology to survey the Fine Constrained features (FCFs) on urban surface, to organize the data into geo-database, to manage their attributes, and to build the accurate urban surface by employing the Constrained Delaunay Triangular Irregular Network (CD-TIN). Then the paper employs the depression polygon division method that refers to Fine Constrained Features (FCFs) to construct computational urban water depressions. Storm-runoff yield is placed through mass conservation to calculate the volume of rainfall, runoff and drainage. The water confluences between neighboring depressions are provided when the water level exceeds the outlet of a certain depression. Numerical solutions achieved through a dichotomy are introduced to obtain the water level. Therefore, the continuous inundation process can be divided into different time intervals to obtain a series of inundation scenarios. The main campus of Beijing Normal University (BNU) and the Jinan area of Shijingshan district, Beijing were used as case studies to validate the storm inundation simulation, to analysis the inundation risks, and to simulate the inundation scenarios for incipient fault areas. This can provide the solutions and platforms in quantification to reduce the urban flood risk and influence. The main research work and achievements of this paper can be summarized as follows.1) Capturing, organizing, managing and modeling the FCFs on urban surface to build the accurate and seamless urban surface for flood modelingThe accurate and seamless urban surface including the FCFs is an essential spatial data for urban flood simulation. The FCFs that influence the water flow dramatically can be classified as constrained points (such as grated inlets, enclosing region outlet), constrained polylines (such as street curbs, walls) and constrained polygons (such as buildings, grass lands). This paper establishes the standards to survey the FCFs and their elevation data with fine survey. By the theory of constrained feature sets, this paper organizes the attribute and spatial data of FCFs into the geo-database with uniform code. Then the accurate and seamless urban surface can be modeled using the CD-TIN which integrates the FCFs, such as street curbs, buildings, underpass flyovers. This surface also couples all the attributes of FCFs and with strong topological structure. 2) The Facet to node (FN) and Facet to Edge (FE) confluence modes which considering the urban FCFs According to the characteristics of water flow on the triangular facet, the facet to node and facet to edge confluence mode are introduced to generate the depression polygons. (1) Facet to Node (FN) flow mode: Based on the CD-TIN built urban surface, the outlet and the lowest elevation node are first labeled. Referring to the natural law of water flowing from high to low place, this paper presents a FN flow mode for CD-TIN based urban surface. The FN flow mode assumes that: the water on each triangular facet flows to its lowest node initially; then the water on the lowest node flows along the steepest edge to its neighbor node; the flow continue until all the neighbor nodes’ elevations are higher than its own. Meanwhile, upstream triangles pass their flow flags to the downstream triangles along the flow path. After that we can generate depressions according to the flags. The case studies demonstrate that the proposed FN flow mode and depression division method for CD-TIN based surface are correct, which can provide more reasonable depression division results than raster, e.g., ArcHydro tools. (2) Facet to edge (FE) flow mode: With the definitions of co-fluent, di-fluent and trans-fluent edges, this paper develops a novel method for urban water depression fine division referring to constraint points, constraint polylines and constraint polygons. The data structures and algorithms for water flow path extraction and urban water depression division are also presented. And referring to the FCFs inundated by the water, the dynamic water depression method is introduced based on the hierarchy tree. The CD-TIN based method is validated by using typical urban terrain units and real case of the campus of Beijing Normal University. The extracted flow paths are basically coincident with each other only in shape (coefficient of correspondence is 0.27), but the CD-TIN based method capable of providing more continuous flow paths. The CD-TIN based method can accurately simulate the water flow paths and depressions around the walls and buildings instead of the plume-shape and parallel-line-shape water flow paths and discrete-parallel water depressions via ArcHydro tools. The results also demonstrate the reasonability, efficiency and implementation of the CD-TIN based method for water flow path extraction and urban water depression division. 3) The accumulated water volume calculation and rainfall runoff modeling relying on water volume balance principle, and the inundation scenarios simulation based on spatio-temporal method There is an obvious mass balance principle in certain a water depression: the accumulated water is the volume of rainfall excluding water infiltrated into the underground and water conveyed by underground drainage system. The water confluences between neighboring depressions are provided when the water level exceeds the outlet of a certain depression. The spatio-temporal inundation process is cut into discrete time slice snapshots to simulate the inundation scenarios. Referring to the water volume balance, the accumulated water volume converts to the inundation volume. This paper proposes the triangular prism as the numerical calculation unit to obtain the water level through a dichotomy method. The main campus of Beijing Normal University (BNU) was used as a case study to simulate the “7.21” storm inundation event to validate the usability and suitability of the proposed methods. Compared the simulation results with in-situ observations, the maximum error is 16.7%. And the perception risk map generated through questionnaire survey also agrees with the simulation results. The experiments show that the proposed method is accurate and effective, with significantly lower drainage data requirements being obtained. Another case study is the Jinan underpass flyover in Shijingshan, Beijing. The simulated inundation scenarios agree with the reported data on “7.21” storm event. 4) GIS based urban storm flood inundation risk scenarios and the incipient faults deductive anlysis. This paper employs the layer architecture and object oriented software design mode to develop the urban storm flood inundation simulation and deducing system. It can simulate and deduce the inundation scenarios of high risk streets, underpass flyovers, and the underground spaces. Taking the campus of Beijing Normal University and Jinan flyover as experiment areas, this paper establishes the inundation risk scenarios at different return period of storm. This can provide the technical support for urban incipient faults analysis and disaster warning. To deduce the urban flood inundation, this paper employs the dual CD-TIN model to simulate the water logging referring to the inundation water depth in front of underground space inlet, the slope of terrace and the pump discharge capacity. This can simulate the inundation scenarios of underground space dynamically. This paper employs two cases studies: the underground garage of BNU and Jinan underpass flyover of Shijinshan to simulate their inundation scenarios at different return period of storm. Also, the influence of runoff coefficient on the accumulated water volume and water depth is studied at the Jinan underpass flyover experiments. The pump discharge and the inundation scenarios are also deduced under the 50 return period rain storm event. And this paper simulate the senarios of drain pipe blocking and underlying surfaces modification influence. This will be a practical solution for discharging the water logging and emergency preparation to reduce the disaster in direct quantitatively.