733
River Flow 2014 – Schleiss et al. (Eds)
© 2014 Taylor & Francis Group, London, ISBN 978-1-138-02674-2
Representing skewed bridge crossing on 1-D and 2-D flood propagation
models: Compared analysis in practical studies
P. Costabile & F. Macchione
LAMPIT, Laboratorio di Modellistica Numerica per la Protezione Idraulica del Territorio,
Department of Environmental and Chemical Engineering, University of Calabria, Italy
G. Petaccia & L. Natale
Department of Civil Engineering and Architecture, University of Pavia, Italy
ABSTRACT: This paper deals with the application of flood propagation models in rivers with skewed
bridge crossings. In such situations, the application of 1-D models is not trivial because of the uncertainty
related to the narrowing induced by bridge piers within the cross-sections. In order to study this problem,
this paper investigates 1-D and 2-D state-of-the-arte unsteady flow models that have been applied to the
situations characterized by bridges or a sequence of skewed bridge crossing. The availability of terrestrial
and airborne LIDAR and the comparison between 1-D and 2-D results, allowed practical suggestions to
optimize the 1-D results in the above mentioned situations.
cross-sections with very small spacing if desired,
providing a more accurate depiction of the river
geometry (Pramanik et al. 2010). Recent advances
in LIDAR surveying techniques have reduced
the costs of obtaining Digital Elevation Model
(DEM). Moreover, the availability of LIDAR data
makes the use of 2-D modelling for flood inunda-
tion more convenient than before (Gallegos et al.
2009; Ernst et al. 2010). Despite studies addressing
the optimal cross-sectional spacing in 1-D models
of flood propagation (see, for example, Castel-
larin et al. 2009), particular care and experience
are required in identifying the most useful sections
for the hydraulic computations. This aspect has to
be managed carefully when applied to the correct
representation of the interaction between bridges
and varying flow regimes. Research related to the
representation of hydraulic structures (i.e. bridges)
in the hydraulic modelling, either using 1-D or 2-D
hydraulic models, is a topic of increasing interest
in the literature (see, for example, Pappenberger
et al. 2006, Cook & Merwade 2009, Fewtrell et al.
2011, Brandimarte & Woldeyes 2013, Hailemariam
et al. 2014). However, there are very few practical
studies that discuss, in practical situations, the
effects induced by bridges on the longitudinal
and transversal variation of the maximum water
surfaces and flow regimes aimed at flood hazard
assessments. For example, Costabile et al. (2014b)
discussed two situations in which bridges cross
rivers orthogonal to the principle flow direction,
located in almost rectilinear river reaches confined
1 INTRODUCTION
In producing flood inundation maps, especially
in urbanized areas, the use of one-dimensional
numerical models may be inadequate because they
may not represent all the features of the flow con-
ditions constrained by infrastructure that occur
in real cases. However 1-D models are still very
popular due to their reduced computational time,
their ease of implementation, the reduced need for
topographic data and lower level discretization
(Macchione & Viggiani 2004, Petaccia et al. 2013;
Costabile et al. 2014a). In order to take into account
two-dimensional features of flow within given
cross sections, terms representing the momentum
exchange between the main channel and the flood-
plain are often added to 1-D unsteady flow model-
ling investigations (Cao et al. 2006; Costabile and
Macchione 2012). Despite this, strong limitations
exist in representing water surface profile using 1-D
models where water surfaces are assumed horizon-
tal across section, and due to the inability in detect-
ing transverse changes of flow regime. Moreover,
the accuracy of flow simulations significantly
depends on the choice of the sequence of cross-
sections describing the river reach. In the past, the
river geometry was described using few surveyed
cross-sections, so numerical sections, obtained by
their interpolations, were added to complete the
computational domain. Nowadays, the use of
techniques for the automatic extraction of cross-
sections from DEMs allows the modeller to obtain