Urban sewer flooding:analysis of the behavior of drainage systems during extreme rain events

Abstract

Currently cities and communities are experiencing ever growing problems related to urban pluvial flooding. This is due primarily to inefficient drainage inlets and overloaded sewer systems. In fact, existing drainage systems rapidly reach their maximum capacity and tend to work pressurized even in the case of medium-entity storms. Damage and losses caused by flood events in urban areas, primarily life and economic losses and traffic disruption, can be significant. Moreover, this situation is destined to worsen in the immediate future due to the fervent urbanization process and the ongoing climate changes. This research is therefore aimed at investigating this type of event, because to guarantee an efficient working of the drainage systems is a prerequisite in modern societies. Specifically the broader objective of the study is to contribute to an improvement of urban flood management by enhancing urban drainage modeling and storm motion forecasting. In order to achieve such scope the following detailed tasks were performed: 1. Investigation of the various LiDAR Digital Terrain Models (DTMs) available for the drainage modeling of a study area. From literature review it is evident that a great effort has been made to improve existing hydraulic models and to develop new ones. Nevertheless, little interest has been devoted to evaluate the effects of the use of different available LiDAR DTMs on hydraulic modeling. The research is therefore motivated by the need to know how LiDAR DTMs with different detail scale (LiDAR DSM first, LiDAR DSM last and LiDAR DTM bare earth with overlapped building) can affect the hydraulic modeling of drainage networks. Every DTM is in fact characterized by a variable presence of non-ground surface features, such as cars, buildings or vegetation, that will influence surely the hydraulic response of the urban catchment differently. Consequently every data set was studied by GIS-based analysis methods, such as calculation of surface depressions, in order to evaluate whether the consideration of all the non-ground features is necessary for hydraulic modeling purposes, or whether the use of a less detailed LiDAR DTM, adequately improved, could be an approachable solution. 2. Analysis of improvements brought by a dual drainage approach in simulating the behavior of a drainage network during extreme rain events, compared to the use of a conventional methodology. Another question that justifies the work carried out by the author and presented in the thesis is related to the need of improving available urban drainage modeling. Most of these models are in fact based on process simplifications that are far removed from reality, such as assuming that when water leaves the sewer it is stored in a virtual reservoir and does not follow the natural flow paths, i.e. the effect of local topography is neglected. This approach provides a very biased image of flooding process. Consequently the research was aimed at quantifying capabilities and limits of two urban drainage modeling with diverse sophistication level. The first one was based on the classical hypothesis according to which the drainage system is composed only of the sewer system, that is to consider that stormwater, once entered the sewer system, can no longer leave this system coming back to the surface. Instead the second one was based on the dual drainage approach, i.e. it was assumed that the urban drainage system was composed of a surface network and the sewer network. The evaluation of the best approach was performed by comparing the water volume distributions in the sewer network and the number of surcharged sewer trunks resulting from hydraulic simulations. Specifically the issues relative to the development of the most complicated model, that is the dual drainage one, were studied in more detail: the influence of buildings and DTM resolution on the surface network definition, and the introduction of criteria to be taken into account for pond filtering parameters were the topics deepened through the use of an innovative methodology, the AOFD tool (Automatic Overland Flow Delineation).3. Study of the potentials of a dense network of rain gauges in forecasting storm movements for flood prevention purposes. This research was performed because, currently, methods for rainfall prediction are mainly based on radar measurements. However rain gauge data are often available whereas radar data are not. Furthermore radar instruments enable the investigation of convective cells motion, whereas rain gauges data allow the analysis of the movement of rainfall patterns recorded on the ground, that is more important for hydraulic modeling. Consequently storm movement parameters, velocity and direction, were derived by analyzing rainfall data trough available storm tracking procedures. The method proposed by Diskin was tested and, in particular, the extent to which the choice of the reference feature in the hyetograph and the location of the recording stations inside the catchment can affect the results of the methodology was studied in detail. The quality of the elaborations was estimated by comparing the results obtained with other physical phenomena which are related to storm movement, such as wind movement data. In particular statistical analysis, based on the computation of the correlation coefficient and root mean square deviation between storm and wind data sets, were performed. With the results from the research presented herein, it is expected that: 1. DTM enhancement methods generate hydraulically corrected DTMs that can potentially lead to improvements in urban pluvial flood modeling. 2. more realistic simulations of the drainage system are performed by developing dual drainage models. In this way engineers could aim at minimizing both the costs of construction of new works and maintenance of existing structures by evaluating systematically the effectiveness of all the possible design solutions. Actually, the use of such a modeling will have to push them to optimize the working conditions of both the surface and sewer networks when evaluating flood control and mitigation measures. 3. rain gauges are considered as valid alternatives in rainfall movement prediction, to be taken into account in areas where radar measurements cannot be obtained yet. In fact the results of the elaborations will demonstrate how such instruments, that are more approachable than radar ones for economical and practical reasons, are very useful in forecasting the movements that future storm events can make in a monitored area. Similar information could be also used in connection with hydraulic models, previously calibrated for the same study area,in order to evaluate in advance the possible flood-prone areas. In addition the analysis of the results, obtained by considering an ever decreasing number of recording stations, will give interesting information to municipalities having limited budget for equipping themselves with an adequate number of such instruments.