Estimating uncertainties for urban floods modelling

Abstract

Because stakes are concentrated in urban environment, flood modelling should be carried out there with high accuracy. However, uncertainties remain important because of:-the estimate of the inputs, these inputs being either rainfalls, or flows coming from the upstream basin or from the exchanges with the sewage network;-the complexity of the processes, which includes a frequent change of the flow regime-from subcritical to supercritical and oppositely-, the difficulty in predicting the flow distribution at the street junctions, the oscillations created by waves reflecting against , for instance, obstacles or walls of houses;-the parameters of the flooded area: variable roughness, very irregular topography, moving obstacles in the streets, water entering the built-up areas, etc. An assessment of these uncertainties is presented for a few examples coming from laboratory experiments or from the two-dimensional modelling of flood events in French towns. These latter results can be used in order to estimate the range of uncertainty during field studies.

Advances in experimental modelling of urban flooding

“…Ideally, flow rates should be accurate to 5%, however, this figure may differ consistently and errors may be much higher during flood events. Recent studies have also investigated more complex features such as the rain falling directly into streets (Paquier & Bazin, 2014). Rainfall has been considered for many years as a major uncertainty source, in the accurate prediction of urban flooding events (Niemczynowicz, 1999), but recent progress associated with the use of radar and microwave networks have reduced the uncertainties by facilitating the gathering of sufficient information on temporal and spatial variation of rainfall processes in urban catchments (Fletcher et al, 2013;Schellart et al, 2012).…”

Experimental modelling of urban flooding: A review

“…Sewer overflow (SO) is similar to river overflow except that in this case the water invading the surface comes from an exceeded capacity of the sewer within the urban domain; the upstream boundary condition is then a controlled discharge in the sewer inlet. The fourth origin of water is the rainfall over the studied domain (RA), which is the case of fully urban watersheds (Pons et al, 2005;Paquier and Bazin, 2014). Then the upstream boundary condition is a complex spatial (and temporal) distribution of water jets from a wellcontrolled rainfall simulator.…”

“…In particular, as the rain falls on the buildings roofs, part of the water reaches the sewer network (through the gutters) and the rest reaches the surrounding streets or gardens (with possible infiltration) with some surface runoff on the private slots. These processes, computed by Pons et al (2005) and Paquier and Bazin (2014), still require high quality data to enable deriving empirical parametrizations specific to urban catchments.…”

“…• additional processes were included in the models, such as the rain falling directly on the street network (Pons et al, 2005;Paquier and Bazin, 2014), short waves or tsunami long waves invading a city (Park et al, 2013), human evacuation during a flood (Bernardini et al, 2017), among others;…”

The influence of road network topology on street flooding in New York City—A social media data approach

“…Several studies over the last decade have investigated weather impacts on the road network, including topics of flooding disaster measures, road systems management strategies, and road system's adaptation to flooding disasters (Hammond et al, 2015;Kadaverugu et al, 2021;Kramer et al, 2016). However, existing research assessing the impact of flooding on the road did not capture the complexity of their interactions (Pregnolato et al, 2017), especially the possibility of two directions exchanges between the sewage network and the street, quantifying the influence between the road network and urban flooding is thus particularly complicated (Paquier & Bazin, 2014). Currently, research studying road networks' influence on flooding focused on (1) Small-scale studies of road junction morphology's influence on surface runoff using a hydrological model.…”