Assessment of sewer network data quality on urban pluvial flood modeling with a 2D/1D dual drainage model

Urban pluvial floods are one of the most common water-related hazards and are going to become more frequent and severe looking at the upsetting climate projections. These events mainly occur due to intense and short precipitation events, leading to the overload of the sewer network and resulting in physical, economic, and even human losses. To address this hazard, effective methods are needed to estimate the scale and impact of pluvial flood events and to develop mitigation strategies. In this context, 2D/1D dual drainage models have become one of the most useful tools for these purposes, being able to simulate all hydraulic phenomena occurring on and beneath the surface. However, these models require detailed information about the topography and geometrical specifications of the sewer network, which are not always readily accessible or, when available, are often incomplete or of poor quality, particularly in large urban environments.

In this work, considering that pluvial flood studies are becoming more popular and several numerical tools are available, we wanted to address a recurrent question raised by the flood modeler community: is the effort/level of complexity of implementing a detailed dual drainage model worth it? To answer this question, we assess the influence of sewer network data quality on the results of water depth and velocity obtained with a 2D/1D dual drainage model applied to urban flood modelling. For this purpose, an ad-hoc 2D/1D hydraulic model was implemented to simulate the complex network system of the city of Differdange (LU) exploiting the recently developed Iber-SWMM. This city was chosen as study case because it has experienced several flooding events in recent years, such as those recorded in 2021, and it has an extensive dataset of detailed geospatial data available, enabling the setup of a high-resolution resolution and fully coupled 2D/1D dual drainage model.

Sewer network links were classified based on their physical properties, such as diameter and length. The sewer network layout was gradually simplified, starting from the minor links to the more complex segments of the network, obtaining new simplified versions of the network that could represent incomplete or poor-quality scenarios. These simplified versions were successively implemented in the 2D/1D model. The comparison between the results of the complete and comprehensive model and the simplified scenarios reveals the impacts of the quality of the sewer network information on pluvial flood modeling.