气候变化和城市化进程导致沿海城市面临着日益严峻的台风洪涝风险，但减灾投入的财政资金相对有限，这构成了当前洪涝减灾工作的主要挑战。合理分配有限的减灾资金是缓解上述矛盾的关键，但需解决两大科学问题：首先，不同政府部门(如水利、市政、应急等)对于有限的减灾资金的申请存在竞争，而分属不同减灾部门的不同减灾措施在特定的区域和时间实现的边际效益不尽相同。因此，如何准确的量化不同减灾措施的经济效益，以科学合理的决定它们的优先顺序。其次，沿海地区台风次生洪涝灾害过程复杂，开展减灾措施经济评估面临高时空分辨率数据的获取、多尺度模型的集成等多重挑战，这些技术瓶颈限制了对减灾措施效益的准确量化。因此，如何准确模拟和重建沿海地区复杂的台风次生洪涝灾害过程，以更准确可靠的实现洪涝风险的量化评估。
针对上述问题，本研究的技术挑战聚焦于两方面：一是开发一个能够精细模拟台风降雨-洪涝-风暴潮灾害链过程的耦合模型，同时保证参数的准确性和模型的适用性。二是实现对历史洪涝灾害事件的精确重现，以及在风险评估中量化不确定性，以确保减灾措施的成本效益分析的准确性。具体而言，本文以海口市为研究对象，构建了台风次生洪涝水动力耦合模型，重建了海口市历史台风洪涝灾害事件，分析了台风次生洪涝灾害的时空格局变化及其驱动因素，并基于此开展了台风洪涝灾害风险分析和减灾措施经济效益评估。通过上述研究，本研究以期为沿海地区的台风洪涝灾害减灾工作开发一套实用的模型工具，并提供系统的理论分析框架。本研究的主要内容如下：
(1) 构建了面向沿海城区的台风次生洪涝灾害模拟及减灾效益评估方法框架。首先，建立了台风洪涝多过程耦合计算框架，包括一维河网、二维地表和风暴潮水动力模型的构建及耦合方法。并通过引入数据优选和动态参数设定等技术，改进了历史台风洪涝灾害事件的重建方法。其次，基于时空统计分析，建立了洪涝灾害时空格局变化及其驱动因素识别框架。然后，通过集成洪涝危险性建模、考虑不确定性的损失量化以及损失超越概率计算等方法，改进了洪涝灾害风险量化框架。最后，通过整合减灾措施情景设定、减灾措施成本估算以及基于风险量化的减灾措施效益评估方法，发展了沿海城市台风次生洪涝减灾措施经济评价框架。
(2) 构建了包括一维河网模型、二维地表模型和风暴潮模型的台风次生洪涝水动力耦合模型，开展了详尽的参数率定和历史场次验证，分析了模型的不确定性因素。首先，基于IFMS软件结合收集的海口地区19条河道断面、10m地表DEM、30m近岸地形数据，构建了一维河网、二维地表及风暴潮模型，并通过松散耦合策略实现了台风次生洪涝灾害链模拟。其次，结合土地利用等下垫面数据和模型参数参考表，预设了关键参数。然后，基于“威马逊”台风期间的61个淹没水深以及河道实测水位流量数据，以及22场数字化的历史台风风暴潮潮位，开展了模型参数率定与管网排水能力估算。同时，基于31场历史台风以及2011年 “尼格(Nalgae)”台风，开展了耦合模型的适用性验证。最后，开展了模型的数据、参数和简化假设的不确定性分析。
(3) 重建了海口市历史台风次生洪涝灾害事件，分析了其时空特征演变及驱动因素，拟合了洪涝危险性的典型重现期分布。首先，基于数据优选原则，筛选了历史台风路径、降雨以及河道径流等数据集，并模拟重建了数据完备的台风场次。其次，基于趋势检验等统计方法，分析了降雨、土地利用等台风洪涝驱动因素的变化趋势。然后，基于模拟的历史场次的最大淹没水深和最高潮位数据，开展了时空变化趋势以及驱动因素相关性分析。最后，通过组合3种采样方法和9个拟合分布函数，开展了降雨、地表洪涝以及风暴潮的重现期估算和结果优选，并计算了6种典型重现期下降雨、洪涝以及风暴潮的空间分布。
(4) 开展了海口市历史台风洪涝灾害损失量化及风险分析，分析了不同部门减灾措施情景的成本效益。首先，基于重建的台风洪涝危险性数据库，结合2000 ~ 2020年间的固定资本存量数据及率定调整后的脆弱性曲线，量化评估了历史台风洪涝灾害损失。其次，基于风险分析的结果，设计了包括管网升级改造、防洪堤、防潮堤强化加固以及灾前财产搬移4种减灾策略。通过相似案例估算了3类工程减灾措施的实施成本，并采用损失超越概率和年平均损失方法量化了各措施的效益。然后，基于净现值和内部收益率等成本效益分析评价指标，确定了各减灾措施的优先实施顺序。最后，基于3种参数敏感性情景，开展了3种减灾措施情景的不确定性分析。
本文的主要结论为：
(1) 构建的考虑台风降雨-洪涝-风暴潮灾害链过程的耦合水动力模型，能够更加准确的模拟沿海地区台风次生洪涝灾害过程，实现了研究区历史台风洪涝灾害事件的高时空分辨率的模拟重建。台风降雨-洪涝-风暴潮耦合模型在模拟精度和对中高风险区识别能力方面均优于仅考虑单一或双重洪水源的模型。具体来说，与仅考虑河流洪水和风暴潮的组合情景相比，耦合模型在R²指标上的提升幅度最大，从0.139增加至0.729。此外，相较于仅考虑降雨和河道洪水的组合情景，耦合模型的R2也实现了从0.671到0.729的提升。基于构建的耦合模型，实现了近40年台风洪涝和近70多年台风风暴潮的高时空分辨率的重建模拟。
(2) 研究区近40年的地表洪涝灾害整体上呈现增长趋势(p < 0.05)，近70多年的台风风暴潮灾害无变化趋势；自然驱动因素中，最大12小时降雨量与洪涝灾害低(< 0.6 m)、中(0.6 ~ 1.5 m)、高(> 1.5 m)危险等级的相关性最为高(mean(Corr) = 0.81, p < 0.05)；在人类活动驱动因素中，不透水地表面积与洪涝灾害各危险等级的相关性最高(mean(Corr) = 0.37, p < 0.01)。具体而言，1980 ~ 2019年期间，研究区内48.8%区域的台风地表洪涝最大淹没水深呈增长趋势，仅8.1%的区域出现减少趋势。地表洪涝的变化趋势与台风降雨量、不透水地表面积、河道径流峰值流量的变化趋势具有较好的一致性。而1949 ~ 2021期间台风风暴潮的最高潮位没有显示出变化趋势。驱动要素分析结果表明，自然因素中，研究区中、高风险区面积与最大12 h降雨量的相关性(Corr分别为0.83和0.73)更强，而低风险区面积与最大6 h降雨量的相关性更强(Corr = 0.88)，均呈现显著的正相关关系(p < 0.05)。人类活动要素中，研究区的各风险区面积均与不透水地表面积相关性最高，分别为0.41，0.40和0.31。
(3) 在研究区在当前的减灾能力内，实施灾前财产搬移是4个减灾措施情景中经济效益最高的，3个工程减灾措施中城市排水管网升级改造经济效益更高，其次为河堤和海堤升级加固。具体而言，在灾前搬移30%财产的情景下，年期望减灾效益达到了33328.4万元，与升级改造管网相当(24947.4万元)，但财产搬移的实施成本明显低于后者。对比排水管网升级改造、河道堤防加固以及海岸堤防升级3类工程减灾措施的成本效益分析，前者的内部收益率(IRR)为8%，而河道及海岸堤防加固的IRR则分别为5%和6%。
本文的创新点为：
(1) 构建了一套基于危险性数值建模、灾害特征分析和概率风险量化的减灾措施经济效益评估方法框架，改进了现有评估框架面向台风洪涝灾害评估时存在的不足。首先，通过将灾害事件概率分布、损失评估的不确定性量化以及脆弱性曲线优调方法纳入考虑，优化了洪涝灾害的传统风险评估方法，提高了灾害损失和减灾效益量化的准确性。其次，基于灾害特征分析及上述风险评估方法，进一步改进了减灾情景选取和减灾措施效益量化方法。在此基础上，构建了一个针对沿海地区台风洪涝灾害的减灾措施经济效益评估框架，为减灾措施的经济效益评估提供了更为精确和全面的视角。
(2) 构建了考虑台风暴雨-洪涝-风暴潮灾害链过程的水动力学耦合模型，实现了海口地区近40年的高时空分辨率的历史台风洪涝事件重建。通过对台风暴雨-洪涝-风暴潮灾害链过程进行耦合建模，实现了在多尺度、多过程上对沿海地区台风次生洪涝灾害更准确的模拟，为更精细、准确地反映沿海区域洪涝灾害的时空分布特征提供了工具支撑。重建了海口地区近40年的历史台风洪涝事件集和近70多年的风暴潮事件集，弥补了以往灾害风险分析中样本数据不足的问题，为更合理可靠地评估海口市台风洪涝灾害特征及风险奠定了数据基础。

Climate change and urbanization have exposed coastal area to increasing risks of typhoon flooding. However, the limited financial resources invested in disaster mitigation have become a major challenge for flood mitigation efforts. Rational allocation of limited disaster risk reduction (DRR) funding is the key to mitigating these contradictions, but two major scientific issues need to be resolved. First, there is competition among different government departments (e.g., water conservancy, municipal, emergency, etc.) for applications of limited DRR funding. Moreover, different mitigation measures belonging to different disaster reduction departments do not achieve the same marginal benefits in specific regions and time. Therefore, the first scientific problem faced is how to accurately quantify the economic benefits of different disaster mitigation measures to decide on their priorities scientifically and rationally. Second, the complexity of the typhoon flooding process of typhoons in coastal areas and the multiple challenges of acquiring high spatial and temporal resolution data and integrating multi-scale models to conduct the economic assessment of disaster mitigation measures are technical bottlenecks that limit the accurate quantification of the benefits of disaster mitigation measures. Therefore, how to accurately simulate and reconstruct the complex typhoon flooding process in coastal areas, and then more accurately and reliably achieve the quantitative assessment of flood risk is the second scientific problem faced.
To address these problems, this study focuses on two technical challenges. The first is to develop a coupled model that can accurately simulate the typhoon rainfall-flood-storm surge disaster chain while ensuring parameter accuracy and model applicability. The second is to achieve accurate reproduction of historical flooding events and quantify uncertainty in risk assessment to ensure the accuracy of cost-benefit analyses of mitigation measures. Specifically, this paper takes Haikou City areas as the object of study and constructs a typhoon flood hydrodynamic coupling model. The study reconstructs historical typhoon flood disaster events in Haikou City areas and analyzes the temporal and spatial pattern changes of the typhoon flood disaster and its driving factors. Based on these findings, the study carries out a risk analysis of the typhoon flood disaster and assesses the economic benefits of disaster mitigation measures. And the main contents of this study are as follows:
(1) A methodological framework for assessing changes in spatial and temporal patterns of typhoon flooding in coastal area and the benefits of disaster reduction was constructed. Firstly, a computational framework for multi-process coupling of typhoon flooding was established, incorporating a one-dimensional river network model, a two-dimensional hydrodynamic model, and a storm surge hydrodynamic model. On this basis, the reconstruction method of long historical typhoon flood events was developed, incorporating techniques such as data optimization and dynamic parameter setting. Secondly, based on spatio-temporal statistical analysis, a framework for identifying the changes in the spatio-temporal pattern of flood hazards and their driving factors is established. Then, the flood risk quantification framework is improved by integrating flood hazard modeling, loss quantification considering uncertainty, and loss exceeding probability calculation. Finally, the economic evaluation framework of typhoon flood mitigation measures in coastal area was developed by integrating scenario setting for mitigation measures, cost estimation of mitigation measures, and risk quantification-based methods for assessing the benefits of mitigation measures.
(2) A coupled typhoon flooding hydrodynamic model including a one-dimensional river network model, a two-dimensional hydrodynamic model and a storm surge model is constructed, the detailed parameter calibration and historical field validation are conducted, and model uncertainties are analyzed. Firstly, based on the IFMS model software and the collected data of 19 major river cross sections, 10m resolution DEM, and 30m resolution near-shore topography in Haikou area, a one-dimensional river network model, a two-dimensional hydrodynamic model and a storm surge model were constructed, and the simulation of the typhoon flooding chain was realized through the loose coupling strategy. Secondly, the key parameters were preset by the land use and other subsurface data with the model parameter reference table. Then, based on 61 inundation depths and measured water level and flow data of rivers during Typhoon Rammasun, and 22 digitized historical typhoon storm surge levels, the model parameter calibration and municipal network drainage capacity estimation were carried out. At the same time, based on 31 historical typhoons and Typhoon Nalgae in 2011, the applicability validation of the coupled model was carried out. Finally, an uncertainty analysis of the model's data, parameters and simplifying assumptions was conducted.
(3) Historical typhoon flooding events in Haikou City areas were reconstructed, and their spatial and temporal feature evolution and driving factors were analyzed. Additionally, the distribution of flooding hazard during typical return periods was determined. Firstly, historical typhoon tracks, rainfall, river runoff, and impervious surface datasets were screened based on the principle of data preference. Using these screened datasets, simulation reconstruction was conducted for historical typhoon events with complete data. Secondly, based on statistical methods such as trend tests, the trends of typhoon flooding drivers such as rainfall and land use were analyzed. Then, based on the maximum inundation water depth and maximum tide level data of the simulated historical events, the temporal and spatial trends as well as the correlation analysis of the drivers were carried out. Finally, the estimation and optimization of return periods and six typical return periods for rainfall, land flooding, and storm surges were carried out by combining three sampling methods and nine fitting distribution functions.
(4) A quantification of historical typhoon and flood losses and risk analysis of Haikou City was conducted to analyze the cost-benefits of scenarios of disaster mitigation measures in different sectors. Firstly, based on the reconstructed typhoon flood hazard database, the historical typhoon flood losses were quantitatively assessed by used the fixed capital stock data and rate-adjusted vulnerability curves for the period 2000 ~ 2020. Secondly, based on the results of the risk analysis, 4 types of mitigation strategies including pipeline network upgrading, levee and tide embankment reinforcement, and pre-disaster property relocation were designed. The implementation costs of the three types of engineering mitigation measures were estimated through similar cases, and the benefits of each measure were quantified using the loss exceedance probability and average annual loss methods. Then, based on the cost-benefit analysis evaluation indexes such as net present value and internal rate of return, the prioritization of the implementation order of each mitigation measure was determined. Finally, uncertainty analyses of three scenarios of disaster reduction measures were conducted based on three-parameter sensitivity scenarios.
The main conclusions of this paper are:
(1) The coupled hydrodynamic model, which considers the typhoon rainfall-flooding-storm surge disaster chain, can more accurately simulate the typhoon flooding process of typhoons in the coastal area of Haikou City. The model can achieve the simulation and reconstruction of historical typhoon flooding events with high spatial and temporal resolution in the study area. The coupled typhoon rainfall-flood-storm surge model outperforms the model considering only a single or dual flooding source in terms of simulation accuracy and the ability to identify medium and high-risk areas. Specifically, the coupled model achieves the greatest improvement in the R² metric from 0.139 to 0.729 compared to the combined scenario that only considers river flooding and storm surge and achieves an improvement in R2 from 0.671 to 0.729 compared to the combined scenario that only considers rainfall and river flooding. And the reconstruction simulation with high temporal and spatial resolution of typhoon flooding in the last 40 years and typhoon storm surge in the last 70 years was achieved based on the coupled model.
(2) There was an overall increasing trend in land flooding in the study area over the last 40 years (P < 0.05), and no trend in typhoon storm surge hazard over the last 70 years; among the natural drivers, the correlation between the maximum 12-hour rainfall and the degree of flood hazard was the most significant (mean(Corr) = 0.81, p < 0.05); and among the anthropogenic drivers, the correlation between the area of impermeable surfaces and the degree of flood hazard (mean(Corr) = 0.37, p < 0.01). Specifically, the maximum inundation depth of typhoon flooding in 48.8% of the areas in the study area showed an increasing trend from 1980 to 2019, while only 8.1% of the areas showed a decreasing trend. The trend of land flooding is in good agreement with the trend of typhoon rainfall, impervious surface area, and peak flow of river runoff. The maximum tidal levels of typhoon storm surges do not show any significant trend changes during 1949–2021. The results of the driving factor analysis showed that the correlation between the medium and high-risk zones area and the maximum 12-h rainfall (Corr = 0.83 and 0.73) was the highest, while the area of the low-risk zone area had the highest correlation with the maximum 6-h rainfall (Corr = 0.88), which all showed significant positive correlations (p < 0.05). Among the anthropogenic driving factors, the area of each risk area has the highest correlation with the impervious surface area, which is 0.41, 0.40, and 0.31, respectively.
(3) In the current mitigation capacity framework in the study area, the implementation of pre-disaster property removal is the most economically efficient of the four mitigation measure scenarios, while the upgrading of the urban drainage network is more economically efficient among the three engineering mitigation measures, followed by the upgrading of riverbanks and seawalls. Specifically, under the scenario of 30% property removal before disaster, the annual expected hazard mitigation benefit reaches RMB 333.284 million, which is comparable to the upgrading of the urban drainage network (RMB 249.474 million), but the implementation cost of property removal is significantly lower than the latter. Comparing the cost-benefit analyses of the three types of engineering mitigation measures, namely upgrading of the drainage network, river embankment reinforcement, and upgrading of coastal embankments, the IRR of the former is 8%, while that of the river and coastal embankment reinforcement is 5% and 6%, respectively.
The innovations in this paper are:
(1) A methodological framework for assessing the economic benefits of disaster reduction measures based on hazard characterization and risk quantification is constructed, which improves the deficiencies of the existing framework for assessing the economic benefits of disaster reduction measures when oriented towards typhoon floods. Firstly, the traditional risk assessment methods for flood disasters are optimized by considering the probability distribution of disaster events, the quantification of uncertainty in loss assessment, and the method of optimal adjustment of vulnerability curves, which improves the accuracy of the quantification of disaster losses and mitigation benefits. Second, based on the disaster characterization and the risk assessment methods mentioned above, the methods for selecting mitigation scenarios and quantifying the benefits of mitigation measures are improved, and an economic benefit assessment framework for typhoon and flood mitigation measures in coastal areas is constructed, which fills in the deficiencies and gaps in the existing economic assessment framework for mitigation measures.
(2) A hydrodynamic model coupling the typhoon rainfall-flooding-storm surge disaster chain process was constructed, and the reconstruction of the historical typhoon flood event set of the Haikou area in the past 40 years with high spatial and temporal resolution was achieved. Through coupled modeling of typhoon rainstorm-flood-storm surge hazard chain processes, a more accurate simulation of typhoon flooding in coastal areas at multiple scales and processes is achieved, which provides tool support to reflect the spatial and temporal distribution characteristics of flooding in coastal areas in a more refined and accurate way. The reconstructed historical typhoon and flooding event set of the Haikou area in the past 40 years and the storm surge event set in the past 70 years, which makes up for the problem of insufficient sample data in the previous disaster risk analysis and lays a data foundation for a more reasonable and reliable assessment of typhoon and flooding characteristics and risks in Haikou city area.