Flood risk analysis for metropolitan areas – a case study for Shanghai

In Shanghai, the main threat to the city’s safety is a typhoon induced storm surge in combination with a high astronomic tide in the Huangpu River. Historical flood events have shown that the weakness of the floodwall, with potential overtopping and breaching along the Huangpu River and its branches, has caused great economic damage and loss of life. In order to better understand flood risk in the city, flood risk analysis at the local city level is strongly required. With climate change, land subsidence and the rapid socio-economic development, flood risk is inevitably increasing if no measures are taken. Not only the current flood risk but also the flood risk in the future needs to be understood. Moreover, it is necessary to recommend effective risk-reduction measures to mitigate future flood risk. Therefore, the objective of this thesis is to quantify the current and future flood risk and to make recommendations on risk-reduction measures in a case study of Shanghai. It also aims to show and develop general methods for flood risk analysis in rapidly growing metropolitan areas. As a first step the Shanghai water system has been analysed. In terms of flood threats to the river, the water level is dominated by the storm tide at the mouth of the Huangpu River. The heavy precipitation mainly induces waterlogging due to an insufficient drainage capacity in the city, while not significantly increasing river runoff in the Huangpu River. Moreover, a control gate (between Tai Lake and the Huangpu River) is regulated to reduce drainage water from Tai Lake when a storm surge occurs. Therefore, a combination of a storm surge and a high astronomic tide will be the main flood threat. It is noticed that the current protection level of the floodwall is only based on the exceedance of the crest height of the floodwall by the water level and does not directly take other failure mechanisms into account. Failure mechanisms such as breaching of floodwalls and failure of the closure of floodgates would induce potential floods in Shanghai. In order to identify flood hazard in extreme events, a frequency analysis of annual maximum water levels was performed; the new frequency curves represent the relationships of water levels at three typical gauge stations along the Huangpu River with different return periods. The Generalized Extreme Value (GEV) distribution was suggested as a most suitable probability distribution for the datasets of annual maximum water levels at Wusongkou and Huangpu Park instead of a Pearson Type III distribution. With the aid of the 1D hydraulic model, water levels in each cross section of the Huangpu River were derived. The potential overtopping points were systematically identified by a comparison of the crest height of the floodwall and the water levels under different return periods (50yr, 100yr, 200yr, 500yr, 1,000yr and 10,000yr). It turned out that the current estimation of overtopping probability is 1/200 p.y. in the Huangpu River. In addition, potential breaching points and failure of floodgates were also hypothesized on both sides of the floodwall. Subsequently, inundation maps were produced by 1D2D hydrodynamic modelling (SOBEK) under different flood scenarios along the Huang-pu River. The results can be visualized on a map with information on maximum inundation depth and the extent of inundation. Firstly, a scenario without protection demonstrates the im-portant role of the floodwall and the infrastructures (e.g. floodgate) along the Huangpu River to protect Shanghai against river flooding. Secondly, various overtopping events at certain points along the floodwall were simulated. It was found that the inundation would merely oc-cur adjacent to the riverine area due to a limited flood volume under overtopping scenarios, since it only occurs during a limited period (e.g. 1 hour). Thirdly, as breach scenarios were developed to explore the worst–case flooding in Shanghai; it turned out that breaching would cause the most serious flooding along the Huangpu River, as parts of the city centre would get inundated with a maximum inundation depth of more than 3m. Lastly, the simulations of a failure of the floodgates were conducted at three selected locations. It showed that the inundation depth was a few decimeters in each scenario (40cm-80cm on average), which would pose threats to the buildings and the infrastructures adjacent to the floodgates. Ex-ante flood damage assessment plays an increasingly important role in flood risk management. Potential flood damage in cities like Shanghai can be massive due to the high rate of socio-economic development and the rapid urbanization in the near-future. Different flood scenarios result in different degrees of damage. New damage functions for various building categories were suggested in Shanghai; with the application of these functions, it was calculated that the potential damage under various breaching scenarios ranged from 88 to 440 million $USD in part of downtown area, which accounts for 1.5% - 7.6% of the maximum potential damage (5.77 billion $USD). In the estimation of indirect flood damage on the service interruption of the subway system, two typical subway stations were selected to estimate the revenue loss due to flooding. It was calculated that the service interruption at one subway station for one week would cause approximately 1 million $USD of revenue losses in Shanghai, which implied that it would cause a huge practical inconvenience for the inhabitants during such unexpected events. Furthermore, in the discussion of the effects of components on the flood damage, the damage function has the greatest influence on the final results, and this deserves priority for future study to reduce the uncertainty of flood damage estimation. Flood risk is calculated by the occurrence/exceedance probability and its associated potential damage. In this thesis, the total risk, which accounts for expected value and standard deviation of damages on the basis of risk aversion, represents the results of flood risk analysis. The results are represented below. The probability of flooding exists in events of overtopping, breaching and failure of floodgates. 26 scenarios were simulated along the Huangpu River based on various boundary conditions of the water level as a function of return periods of 200yr., 500yr., 1,000yr., and 10,000yr. at Wusongkou, in which 8 breaching points and 3 floodgates points were selected on the west and east side of the floodwall. The total (flood) risk is calculated between 40 million $USD/yr. and 112 million $USD/yr. along the Huangpu River of Shanghai, in which the point at ~45km away from the mouth is most likely to be over-topped, and the breaching point, ~26km away from the mouth at the west side of the Huangpu River in the city centre, leads to largest potential flood damage among all the scenarios. Furthermore, it is noticed that the economic damage due to breaching is a factor of 10 higher than the damage caused by overtopping scenarios. However, in terms of the contribution to the flood risk, the failure of floodgates accounts for ~41% of the overall flood risk due to its high-er probability of failure than breaching and parts of overtopping scenarios. Economic development appears to have the greatest effect on future flood risk, which could triple flood risk in 2030 and increase six fold in 2050 if no further measures are taken. Land subsidence is the second driver of future risk, and the ‘absolute’ sea level rise has the least effect on the future flood risk. The combination of all these affected factors would raise flood risk 4 times and 16 times in total in 2030 and 2050 respectively if no further measures are taken. In order to evaluate and recommend an effective (combination of) risk-reduction measure(s) to mitigate flood risk, a comparative study between Shanghai and Rotterdam was conducted to propose potential risk-reduction measures under the threats of future climate change and economic growth. It also showed that the metropolitan cities, under similar challenging flood threats, can learn from each other. Regarding the risk-reduction measures, the potential (structural and non-structural) measures have been proposed and evaluated by the methods of cost-benefit analysis and economic optimization. Preliminary results of the cost-benefit analysis show that the construction of a storm surge barrier has a somewhat larger benefit/cost ratio than the upgrading of the floodwall. Besides, since the Shanghai Municipal Government desires to upgrade the city to the level of an international metropolis with a high quality of life, the upgrading of the floodwall will largely hinder the view of rivers and lower the attractiveness of the city. Therefore, it is expected that the construction of the storm surge barrier is a better solution to protect Shanghai in the long run. Economic optimization led to a preliminary result of a safety level of 1/4,500p.y. for the Huangpu River in Shanghai due to fast economic growth in the future (2050). It is additionally noted that, given the current relatively low protec-tion level the flood barrier boards (to protect buildings for small floods) have advantages and it is also recommended to apply this measure at the entrance of all types of buildings in case of unexpected flood events. These results show how flood risk analysis can provide rational in-formation to support decisions for risk reduction for rapidly growing mega-cities, like Shanghai