Flood probability analysis of the Huangpu barrier in Shanghai

Abstract

The Huangpu River meanders through downtown Shanghai City and links China's third largest Lake with the Yangtze River estuary. Typhoons passing Shanghai from June to October annually are the main trigger for flooding of the Huangpu River. When storm surges due to a tropical cyclone meet the local astronomical tide, the water levels in the river can easily exceed the warning water levels and cause flooding of the downtown area of Shanghai City. Moreover, during the typhoon season the water levels in the region are sustained higher because of the rain season earlier. As a result flooding of the urban area occurs frequently during the typhoon season. A storm surge barrier in the mouth of the river will effectively keep out high water levels caused by passing tropical cyclones. However, typhoons also bring heavy downpour to the area, which is able to temporarily increase the upstream discharge into the river substantially. The increased upstream discharge because of torrential rainfall triggered by passing tropical cyclones can be a cause for flooding of the river even though the barrier is closed. This study is aimed to determine the flood probability of the Huangpu River during closure of the storm surge barrier as a result of the upstream discharge for the three proposed barrier locations in the mouth of the river. This upstream discharge includes the base discharge of the Huangpu River and the torrential rainfall runoff triggered by the passage of a tropical cyclone. These discharges are investigated as well as their probabilities of exceedance. The situation in which the river is on the verge of flooding is termed the limit state condition. Hence, the flood probability of the river can be expressed as the probability that the upstream discharged water volume during barrier closure exceeds the storage capacity of the river. The storage capacity of the Huangpu River is considered to be deterministic, in contrast with the upstream discharge. The storage capacity is investigated with one-dimensional flow simulations of an enhanced model of the Huangpu River in SOBEK RIVER. The allowed upstream discharges per closure duration, i.e. critical discharge, per barrier location are computed to represent the storage capacity. Subsequently, the probabilities of occurrence of the critical discharges are studied with analyses of the probabilities of the components that build up the upstream discharge. Because of the limited availability of appropriate discharge records, the torrential rainfall runoff distribution is derived from the joint distribution of storm surge and torrential rainfall in the Shanghai area given storm tide levels in the mouth of the river equal to the barrier closure water level. The torrential rainfall probability is related to the storm surge level probability in the mouth of the river; since both are caused by passing typhoons. Therefore, the joint distribution of storm surge and torrential rainfall in the area is investigated with the known distribution of storm surge and torrential rainfall. These individual distribution functions are linked into their joint distribution with a copula. Copulas separate the dependence structure of multivariate distributions with the individual marginal distribution functions. The study of copulas and its applications is rather new but a rapid growing field in the literature of statistics. Eventually the runoff distribution is derived from the torrential rainfall probabilities with empirical runoff relations found from analysis of historical major rainfall events in the area. Prior to the analysis of the torrential rainfall probabilities, the required surge levels for barrier closure are computed by random combination of the tide and storm surge distribution in the mouth of the river. The present warning water level in the mouth of the river is regarded as the future barrier closure level. Subsequently, the required surge level for barrier closure is found by the conditional probability of the surge levels given the barrier closure water level.