| With climate change and urbanization,urban floods caused by extreme weather events are becoming more frequent,especially for coastal cities.Floods have become an important obstacle for urban development.At present,there are still some problems to be solved and further improved in the study of coastal city floods,such as the correlation characteristic analysis of multi-dimensional flood hazard factors in coastal cities and their combined design methods,the zoning of flood-causing factors in coastal cities and the adaptability of drainage measures in different zones,flood risk classification methods based on intelligent algorithms and data mining techniques,and stage optimization models of urban drainage system considering climate change and hydrological model uncertainty,etc.In-depth study of these topics can further enrich the theory of flood control in coastal cities.Therefore,focusing on Haikou,this paper carries out research on risk analysis and disaster prevention decision-making from two aspects:flood risk analysis and risk management.The main research contents and conclusions include:First,the correlation characteristics and combination design of multi-dimensional hazard factors in coastal cities are investigated based on the integration correlation analysis method,trend test method,Copula function,multi-type return period and multi-variable hydrological event design method.It reveals the correlation characteristics and trends of rainfall and tide level in coastal cities,and comprehensively discusses the influences of parameter estimation methods,joint return period types and design methods on joint distribution characteristics and combing design values.The results indicate that there is a certain correlation between rainfall and tide level in coastal cities,and the correlation between heavy rainfall events and high tide events shows a significant increase trend.The disaster prevention standards for coastal cities should take the combined effects of rainfall and tide levels into account.When the selected design return period is large,it is recommended to calculate the design value by using the parameter estimation method to avoid the design value being too low.The design value of rainfall obtained by the same frequency method is a little larger than that of the maximum possible weight function method.The design value of tide level obtained by the maximum possible weight function method is a little larger than that of the same frequency method.The design rainfall and tide level values of the joint return period are greater than that of the Kendall return period,and both are greater than that of the co-occurrence return period.In the general case,if the dangerous event is not clearly defined,the corresponding rainfall and tide level design values of Kendall’s return period can be used as the engineering design standard.Second,this study presents an integrated method to quantify the hazard degree of disaster-causing factors,rainfall and tide level,and to investigate the optimal management of flooding risk in different disaster-causing factor areas.The results indicate that the hazard degree increases with the increasing distance between the drainage district and the Qiongzhou Strait or the Nandu River in the east of Haikou.Heavy rain is the main disaster-causing factor in inland areas,while high tide level is the main disaster-causing factor in island areas.For the area whose main disaster-causing factor is heavy rain,water storage projects could effectively reduce flooding.Meanwhile pump stations are economical choices for the area where tide level is the main disaster-causing factor.The results can provide reference for drainage planning in other coastal areas.For coastal cities,rainfall and tide level are two disaster-causing factors,which both have significant impact on flood events.The flood control planning should be given priority to drainage facilities for the main disaster-causing factor,and supplemented by other prevention measures.Third,to overcome the difficulty in determining the risk classification threshold in traditional flood risk analysis,an integrated methodology is proposed by incorporating urban flood inundation model,improved entropy weight method and k-means cluster algorithm to evaluate urban flood risk.The proposed approach is data driven without considering classification standard of different risk levels,and thus provides more reasonable and objective results.A region in Haikou,China is adopted to test the applicability of the proposed approach.The results indicate that high risk zones cover 13.7%of the total area,which generally exhibit higher inundation depth and lower elevations.The assessment result matches well with the historical data of flood events.The traditional cluster algorithm and the technique for order of preference by similarity to ideal solution(TOPSIS)methods are used for comparison with the improved entropy-cluster algorithm to validate the proposed approach for risk management.The result demonstrates that the proposed approach is feasible and exhibits the most reasonable classification result.The study outcomes provide a novel approach for flood risk assessment and can provide valuable information for urban flood management.Fourth,a more reliable optimization model(i.e.,SOCU model)for urban drainage design is proposed based on an integration of staged optimization policy,urban hydrological model,MC-based GLUE method,CCP model and heuristic algorithm.A case study of the Haidian island is used to demonstrate the proposed model.The minimum investment for the optimization of the Haidian Island disaster reduction project is 183 million yuan.The optimal solution shows that a pumping scale of 28.3 m~3/s should be constructed in the first time step(before 2020)and then extended to 38.4 m~3/s in next 10 years(before 2030).Compared with the implement-once design,the SOCU model is not only more flexible to adapt to climate change,but also is economically efficient in terms of net present value.The total present cost under climate change scenario would increase on average 1.4 times than that under present scenario.It reveals that climate change has a respectively significant influence on the investment of drainage system.Furthermore,the solutions under stationarity scenario would inevitably lead to the flood risk under future climate change conditions and the surcharging probability(i.e.,the total surcharging volume exceeds the corresponding threshold)was all higher than 50%.The total present cost of the deterministic optimization model(i.e.,without considering hydrological model uncertainty)is lower than the result of SOCU model.The deterministic solutions would lead to flooding problems due to the hydrological model uncertainty and the surcharging probability is 52%,which indicates that although the deterministic model saves some investment cost,the model results lack sufficient reliability.In addition,the total present cost would increase with an increase of the constraint satisfaction probability.When the constraint satisfaction probability increases from 0.75 to 0.95,the total present cost increases from 166 million RMB to 195 million RMB,which could help decision makers conduct a trade-off analysis between total system investment and the flooding risk.The research results can provide reference for urban flood management. |
