Review of Urban Catchment Hydrological Modelling

Journal of Hydrology 529 (2015) 6281
Contents lists available at ScienceDirect
Journal of Hydrology
journal homepage: www.elsevier.com/locate/jhydrol
Review Paper
Hydrological modelling of urbanized catchments: A review and future

directions
Salvadore Elga a,b,, Bronders Jan b, Batelaan Okke a,c
a
Vrije Universiteit Brussel, Department of Hydrology and Hydraulic Engineering, Pleinlaan 2, 1050 Brussels, Belgium
VITO, Flemish Institute for Technological Research, Boeretang 20, 2400 Mol, Belgium
c
Flinders University, School of the Environment, GPO Box 2100, 5001 Adelaide, SA, Australia
b
a r t i c l e
i n f o
Article history:
Received 3 February 2015
Received in revised form 7 June 2015
Accepted 13 June 2015
Available online 19 June 2015
This manuscript was handled by
Konstantine P. Georgakakos, Editor-in-Chief,
with the assistance of Marco Borga,
Associate Editor
Keywords:
Urban hydrology
Hydrological modelling
Urban catchment
s u m m a r y
In recent years, the conceptual detail of hydrological models has dramatically increased as a result of
improved computational techniques and the availability of spatially-distributed digital data.
Nevertheless modelling spatially-distributed hydrological processes can be challenging, particularly in
strongly heterogeneous urbanized areas. Multiple interactions occur between urban structures and the
water system at various temporal and spatial scales. So far, no universal methodology exists for simulating the urban water system at catchment scale. This paper reviews the state of the art on the scientic
knowledge and practice of modelling the urban hydrological system at the catchment scale, with the
purpose of identifying current limitations and dening a blueprint for future modelling advances.
We compare conceptual descriptions of urban physical hydrological processes on basis of a selection of
43 modelling approaches. The complexity of the urban water system at the catchment scale results in an
incomplete understanding of the interaction between urban and natural hydrological systems, and in a
high degree of uncertainty. Data availability is still a strong limitation since current modelling practice
recognizes the need for high spatial and temporal resolution. Spatio-temporal gaps exist between the
physical scales of hydrological processes and the resolution of applied models. Therefore urban hydrology
is often simplied either as a study of surface runoff over impervious surfaces or hydraulics of piped
systems. Many approaches target very specic objectives and the level of detail in representing physical
processes is not consistent. Based on our analysis, we propose a blueprint for a highly complex integrated
urban hydrological model. We regard exibility, in terms of model structure and data assimilation, as the
key characteristic for overcoming these limitations. We advocate the use of modular, process-based
approaches, which are exible and adaptable to research needs. Higher complexity is inevitable, and
higher uncertainty is a major consequence. Remote sensing data, measurable model parameters
and new spatially-distributed calibration techniques might help to reduce uncertainties.
2015 Elsevier B.V. All rights reserved.
Contents
1.
2.
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hydrological processes in the urban environment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.
Precipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.
Evapotranspiration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.
Depression storage, overland flow and runoff . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.
Storm water drainage systems and combined sewer systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.
Retention basins and stormwater management systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.
Infiltration and subsurface processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.1.
Direct groundwater recharge: infiltration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.2.
Indirect groundwater recharge: leakage/drainage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.3.
Groundwater extraction, flow and discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Corresponding author at: Vrije Universiteit Brussel, Department of Hydrology and Hydraulic Engineering, Pleinlaan 2, 1050 Brussels, Belgium.
E-mail address: Elga.Salvadore@vub.ac.be (E. Salvadore).
http://dx.doi.org/10.1016/j.jhydrol.2015.06.028
0022-1694/ 2015 Elsevier B.V. All rights reserved.
63
63
65
66
66
67
67
67
67
68
68
E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281
3.
4.
Urban hydrological modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.1.
Spatial and temporal model resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.
GIS and remote sensing in urban hydrological modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.
Flow routing and hydraulic concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.
Process-based modelling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.
Parameter estimation and model evaluation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Blueprint for urban hydrological modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.
Urban modelling practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.
Future directions in urban modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Introduction
This paper aims to review the state of the art on the scientic
knowledge and practice of modelling hydrological processes in
urban environments at the scale of a catchment. Urban catchment
hydro(geo)logy, or the study of water uxes in urbanized catchments, is among the different sub-disciplines of the hydrological
sciences, clearly gaining in importance during the past few decades
(Niemczynowicz, 1999; Price and Vojinovic, 2011; Schirmer et al.,
2012; Braud et al., 2013; Fletcher et al., 2013). A major driver for
this trend is the growing concern about urban water sustainability
and human health protection. In fact, for the rst time in human
history since 2009, more than half of the world population is living
in urban and semi-urban zones (United Nations, 2010). Moreover,
population projections show that by the year 2030 urban population will exceed 80% of the total population, with growth/migration particularly concentrated in megacities and developing
countries. Growing population and urbanization augment the pressure on the environment and often challenge water resources sustainability (Lee and Heaney, 2003; Carle et al., 2005). Land use
modications, such as an increase in urbanization, can have a signicant impact on hydrological processes (DeFries and Eshleman,
2004). Governments are aware of these threats and demand solutions in terms of environmental monitoring and water-resources
protection (European Parliament, Water Framework Directive
2000/60/EC), (CWA/Clean Water Act, P.L. 92-500, 1972 and
amendments).
Urban development and urban water systems interact mutually: (1) on one hand urbanization is considered to be the major
cause of pollution/depletion of water resources; while (2) on the
other hand increased urban ooding threaten human security
and infrastructure integrity. Many researchers have investigated
the link between (ground)water pollution and urban growth
(Cronin et al., 2003; Carle et al., 2005; Rueedi et al., 2009;
Vizintin et al., 2009; Carey et al., 2013) and identied urbanization
as a major cause of contamination of groundwater and surface
water resources. This contamination is the consequence of accidental releases of toxic substances or more often by leakage of
sewers. The type of pollutants as well as their load can greatly vary
according to location and urbanization density (Beck, 2005;
Eiswirth et al., 2003). Urban areas have shown to be among the
most vulnerable systems to the adverse impact of heavy rainfalls.
Floods are becoming more frequent and more devastating than
ever before as urban areas are enlarging and becoming denser
(Kang et al., 1998; Mark et al., 2004; Schmitt et al., 2004; Chen
et al., 2009). Society suffers yearly from the consequences of (ash)
oods, with mortality nearly homogeneous over different continents (Jonkman and Vrijling, 2008). Economic losses are large,
overall losses in Europe due to weather disasters in the period
19802010 accounted for about 535 billion US$ per year (Kron
et al., 2012). The Floods Directive (2007/60/EC) was therefore
dened to mitigate the effects of oods by demanding oods risk
63
68
68
71
72
72
76
76
77
77
79
79
assessment and mapping (CEC, 2007). Reliable assessment of water

uxes is crucial for human lives, environmental protection and
infrastructures/goods safety. Water movement in urban areas is
however not well understood, and so are the physical principles
that regulate this movement as well as the interactions occurring
between the hydrological processes. Scientic understanding can
be supported by detailed and consistent measurements and by
hydrological modelling, therefore urban hydrology will be a major
issue in the decades ahead (Delleur, 2003; DeFries and Eshleman,
2004; Praskievicz and Chang, 2009; Fletcher et al., 2013).
Driven by developments in computer capability and the availability of remote sensing data, the use of distributed hydrological
models is becoming more and more a common practice.
Hydrological modelling of urban catchments is highly challenging
as urban catchments are strongly heterogeneous and have very
specic hydrological processes. The circulation of rainwater
within urban areas has not yet been described in a detailed manner, as studies on this topic often remain limited to the runoff on
impervious surfaces (Rodriguez et al., 2008). Developments in this
direction generally focus either on very specialized tools for a particular aspect of the urban hydrological cycle, or on generic software that combines and/or integrates several semi-specialized
components to describe the total water cycle in urban areas.
Bach et al. (2014) recognized the importance of integration in modelling the urban water system and proposed to classify models
based on their degree of integration. However there is little agreement so far on a universal concept or methodology for simulating
the urban water cycle at the catchment scale.
This review is articulated in four sequential sections with the
aim of answering 11 questions regarding urban hydrological modelling practice (Fig. 1). In this rst section we provided the motivations that make urban hydrological modelling a globally-relevant
subject. The second section gives a denition of urbanized catchment and discusses the main hydrological processes of the urban
water cycle. These processes are compared to natural hydrological
systems, while we highlight their characteristic spatial and temporal scales. As this topic was also addressed by the recent review of
Fletcher et al. (2013), the second section of this manuscripts therefore takes their work as a starting point for an extended discussion
on the spatial and temporal scales of the hydrological processes.
The purpose of the second section is to provide a basis for comparison for the following evaluation of modelling practice, and to
assess how consistent hydrological tools are with respect to the
complex interactions of physical processes in the urban environment. We particularly focus on the quantitative assessment of
the changes induced by urbanization on the water system with
brief references to the relevant qualitative aspects. It is however
not within the scope of this paper to review the chemical processes
of urban-water pollution. In the third section we describe the current practice in urban hydrological modelling, highlighting the
most important developments and the characteristics of such modelling approaches. To this end we compare and describe 43
64
Urban hydrology at the Catchment Scale
Table 1
Urban catchments are often dened on basis of impervious cover percentage,
however literature-based values can signicantly vary, ranging from 5% to 98%. The
values of impervious coverage are expressed as a percentage of the total catchment
area.
1) Is it RELEVANT?
Hydrology of Urban Environments
2) What is an URBAN CATCHMENT?

3) Which PHYSICAL PROCESSES play a role?
4) What are the SPATIAL and TEMPORAL scales?
Urban Hydrological Modelling : Current Pracce
5) What is the models SPATIO-TEMPORAL resoluon?

6) How do models deal with SPATIAL VARIABILITY?
7) How are HYDRAULIC PROCESSES simulated by the models?
8) How are PHYSICAL PROCESSES simulated by the models?
Authors
Urban catchment
size (km2)
Impervious cover
(%)
Zhou et al. (2010)

Chormanski et al. (2008)
Meja and Moglen (2010)
Carle et al. (2005)
Lhomme et al. (2004)
Mitchell et al. (2003)
Easton et al. (2007)
Ogden et al. (2011)
Rodriguez et al. (2000)
Xiao et al. (2007)
Tsihrintzis and Hamid (1998)
25.294.9
31
124
0.070.26
52
27
3.3
14.3
0.050.13
1.4 10 3
0.080.24
59.1
13
17
420.2
23
2024
24
34
3739
>50
3698
9) How do models deal with PARAMETER ESTIMATION

and UNCERTAINTY?
Blueprint for Urban Hydrological Modelling
10) How can we bridge the SPATIO-TEMPORAL GAPS?

11) What are the desirable CHARACTERISTICS
of the FUTURE models?
Fig. 1. This paper is structured in four sequential sections aiming at answering 11

questions on urban-catchment hydrological modelling practice and future. Clear
links exist between Sections 2 and 3: comparisons can be made between important
physical processes in urban catchments (question 3) and physical processes
simulated by models (question 8), and between spatio-temporal scales of processes
(question 4) and spatio-temporal resolution of models (question 5).
modelling approaches obtained from literature. This paper does

not aim at ranking models and selecting the most suitable
modelling approach for urban studies, but rather to draw a general
picture of the present situation by means of a representative set of
43 examples. The goal of the review is the identication of the
challenges urban hydrological modelling is facing in terms of
constantly growing data requirements and level of modelling
detail versus uncertainty. Based on the conclusions of this review,
we propose a blueprint for a highly complex integrated urban
hydrological model in the last section.
2. Hydrological processes in the urban environment
The conceptualization of the water uxes is a prerequisite for
understanding the hydrological behaviour of urbanized catchments. However, the identication of the hydrological components
and their interactions, as well as the water balance in these areas
present conceptual difculties (Van de Ven, 1990). To start with,
there is no simple denition of what constitutes an urban catchment. How large does the extent of human agglomerations within
a catchment have to be for it to be considered an urban catchment?
Should an urban catchment be dened upon the percentage of the
catchment area covered by housing and industries only?
Impervious coverage was used by many authors as a threshold
for catchments classication, however impervious thresholds
greatly differ from case to case, from very low percentage
(510%) to about the total surface of the catchment (Table 1).
Moreover, imperviousness estimations are prone to errors due to
the used methodologies, maps or algorithms, and the resolution
of the input satellite image used (Weng, 2012). According to Wu
and Murray (2003) errors of about 30% can be caused by
misclassication of urban vegetation and only in a few cases these
errors can be reduced below 10%. Because of these uncertainties,
Moglen and Kim (2007) and Jacobson (2011) propose to reject
the concept of an impervious cover threshold, as this assumption

over-simplies the imperviousness research, in favour of an
approach based upon environmental impact assessment of the
water system. This approach should be based on the identication
of how and where to locate development to minimize the impacts
on streams, groundwater, etc. Common sense however suggests
that the anthropogenic impact on the hydrological system is
directly proportional to the level of urbanization. This relationship
can be more or less strong according to the type of settlement; in
other words, a compact and very dense city might produce the
same impact at the catchment level as sparsely distributed villages.
As proved by Meja and Moglen (2010), imperviousness patterns
and distribution can result in major changes in the hydrological
response at the catchment outlet. A catchment should then be considered urban when the local anthropogenic impacts are such that
the hydrological uxes are signicantly disturbed, therefore a preliminary hydrological assessment of the area must be performed.
Factors that should be taken into account in this assessment are:
(1) housing and population density in relation to the catchment
size;
(2) comparison of pollutant concentrations and background values to assess river water quality status: high pollution is a
clear evidence of anthropogenic impact on the catchment;
(3) population water demand and main sources of water supply,
this information might indicate possible impacts on groundwater resources;
(4) modication of rivers: relocation, cross-sectional changes,
vaulted stretches;
(5) wastewater treatment facilities, whether or not the wastewater is treated within the catchment or it is transported elsewhere, if septic tanks or other systems of water treatment
are locally used.
Although no unique denition of what constitutes an urban
catchment is possible, screening the above elements would allow
identication of urban catchments. The hydrological system
differs very much from city to city, equal population and
impervious-surface area can in fact produce very different impacts
on the water resources. Differences result from government
policies, which might range from loose to strict water resources
protection, and from high to low population awareness.
Moreover, the concept of water cycle is complicated as water can
cross catchments boundaries by means of water distribution and
sewer networks. An urban catchment constitutes of a high heterogeneous mixture of natural and articial surface covers; natural
and anthropogenic-modied processes interact with each other
(Fig. 2). Hence there is neither a standard denition for an urban
65
B
A Precipitation
B Evapo(transpi)ration
C Overlandflow
D Runoff over sealed surfaces
E River flow
F Storm drainage and sewer flow
G Waste water treatment discharge
H Depression storage and open water
K
I
G
L
E
F
I Soil infiltration
J Interflow or lateral flow
K Groundwater recharge
L Pipe leackage
M Groundwater drainage
N Groundwater extraction
O Groundwater flow
P River/groundwater exchange
Fig. 2. In an urbanized catchment, the water cycle is modied by anthropogenic activity, leading to new pathways, and reduced or augmented natural pathways. Land cover
in urbanized catchment is extremely heterogeneous and multiple interactions occur within physical phenomena (adapted after http://www.miseagrant.umich.
edu/lessons/lessons/by-broad-concept/earth-science/exploring-watersheds/).
100 yrs
10-3
10-2
10-1
100
101
102
103
104
105
106 LENGTH [m]

TIME [s]
108
1 yr
107
1 mon
106
105
1d
104
Transpiraon
1h
103
102
1 min
101
1 mm
1 cm
0.1 m
1m
10 m
100 m
1 km
10 km
100 km 1000 km 10000 km
Fig. 3. Spatial and temporal scales of hydrological processes in urban areas adapted after Blschl and Sivapalan (1995), Hall (1984), Berndtsson and Niemczynowicz (1988),
Maidment (1992), Ward and Robinson (1999), Dingman (2002), Brutsaert (2005), Bronstert et al. (2005), and Hendriks (2010) with additional information from the authors.
water cycle, although many texts agree in dividing the system in

two main networks of pathways: (1) a modied natural set of pathways; and (2) supply-sewerage pathways (Lerner, 1990). In the
modied natural network, (1) water follows the classic hydrological paths in the natural part of the catchment. For example canopy
interception will only take place on urban vegetation, inltration
on locations where the soil has not been sealed. Water movement
in the subsurface is also modied by the urban soils composition
which differs from natural ones, and groundwater discharge might
happen in different locations if the surface and the groundwater
system are articially disconnected. Part of supply-sewerage set
of pathways (2) are borehole groundwater extraction from urban
aquifers, ow of waste water in sewers, ow in mains for water
supply, stormwater ow, leakage from or inltration into pipes,
(over-)irrigation, inltration of water due to articial ponds and
septic tanks, wastewater release into surface water bodies after
or without treatment.
Water uxes and hydrological processes interact with each
other in multiple ways and at different temporal and spatial scales
(Fig. 3). We dene the characteristic spatial (temporal) scale of a
hydrological process as the region in space (or time) where the process takes place, variations of its uxes occur or the resolution in
space (or time) at which the process is best measured. All these
interactions are a major source of complexity and uncertainty for
urban catchment studies, and cause strong limitations on modelling the system (Bach et al., 2014). These interactions are very difcult to quantify and to take into account conceptually in the water
balance, as for example leakages from sewers and water mains.
The forthcoming sections briey discuss the major processes of
the urban water cycle with the focus of highlighting the urbanization impacts on water resources at different temporal and spatial
scales. We do not provide a comprehensive description of all
involved processes but rather give a general overview of the urban
hydrological cycle components at the catchment scale. This analysis is complementing the arguments of Fletcher et al. (2013), who
focussed on measurement techniques, forecasting urban precipitation and impacts of urban development on streamow, while we
discuss the impacts of urbanization in each water balance
component.
2.1. Precipitation
As stated by Fletcher et al. (2013), the study of urban-induced
changes in precipitation has been an active topic of research since
66
the 1970s when indications of these impacts were found for several large cities. Many studies tried to quantify these impacts by
comparing pre and post urbanization conditions and concluded
that seasonal changes locally increased the precipitation between
5% and 15% (Huff and Changnon, 1972; Taha, 1997; Shepherd
et al., 2002; Shepherd, 2006). However, these estimates have a high
degree of uncertainty due to the scarce data and the
non-quantiable impact of other atmospheric processes. Studies
on this matter were therefore considered as non-denitive by the
U.S. Weather Research Program (Dabberdt et al., 2000). It is however generally accepted that urbanization has an impact on atmospheric dynamics at a local scale, the main factors inuencing
precipitation are:
(1) The land cover change modies the energy balance and
together with anthropogenic heat release generates the
urban heat island effect, which might have an impact on precipitation occurrence and intensity. Moreover the local
increase in temperature can reduce snowfall in favour of
rainfall.
(2) The modied surface roughness patterns and the increased
land cover heterogeneity have an impact on wind circulation
and, as a consequence, might change precipitation patterns.
(3) Atmospheric pollution and ne particle release have a strong
inuence on the chemical composition of the precipitation
(e.g. acid rain), and can also affect the rain drops generation
process.
Since several studies suggest there might be a link between
urbanization and precipitation intensity, we advise to use carefully
historical precipitation data in analysis of strongly-urbanizing
catchments. Whenever possible, simulations with historical data
should be complemented with scenario analysis taking into
account future projections of climate variables. More research is
needed to incorporate the possible inuence of urbanization on
these projections. Urbanization seems to mainly affect precipitation intensity and patterns, while spatial and temporal scales of
rainfall are not strongly affected (Fig. 3). However, the high dynamics and heterogeneity of urban systems demand for very high spatialtemporal resolution of precipitation measurements (Fabry
et al., 1994; Tilford et al., 2002; Berne et al., 2004; Einfalt et al.,
2004).
2.2. Evapotranspiration
Urbanization can signicantly affect evapo(transpi)ration (ET),
and accurate estimation of ET is crucial as it represents a major
component of the global water balance (Cheng et al., 2011). ET
might be considerably less in urban catchments than in rural areas,
mainly due to the lack of vegetation (Taha, 1997; Chen et al., 2009).
Several studies however state that ET is not negligible in urban
zones, reaching up to about 40% annually and in extreme cases
up to 80% in summer of the total losses (Berthier et al., 2006;
Rodriguez et al., 2008; Grimmond and Oke, 1991). Many urban
green areas are (over-)irrigated in dry (and wet) periods
(Salvador et al., 2011). With adequate irrigation, root zone moisture conditions are as such that plants can transpire at potential
level, the total actual ET might therefore be higher than under natural (dry) conditions. As we discuss in Section 2.6.2, (waste)
water-networks leakage can also augment the soil water content
and so ET. The phenomenon of urban heat island (UHI), or the
increase in temperature of urban centres as compared to their surroundings, is perhaps the most known and studied effect of urbanization on local climatic conditions (Oke, 1982; Dixon and Mote,
2003; Arneld, 2003; Rizwan et al., 2008). The effects of UHI are
hydrologically relevant as higher temperatures increase the direct
evaporation of water stored in surface depressions, plant canopy,

articial and natural water reservoirs and can boost vegetation
growth.
The physical process of ET has a characteristic spatial scale that
ranges from <1 mm, size of leave stomata, to about 10 m, evaporation from homogeneous open water reservoirs (Fig. 3). Urban vegetation has a higher spatial variability than forested or agricultural
lands, which requires new techniques to asses ET (Nouri et al.,
2013a). In the recent review of Nouri et al. (2013b), several ET
measurements techniques for urban vegetation are described and
compared including standard approaches such as lysimeter and
soil water balances, and more recent ones such as remote sensing
(RS) based approaches. Shuttleworth (2007) stated that arguably,
the meteorological community is still ahead of the hydrological
community in exploiting the full potentials of RS techniques for
ET estimates, however in the past years this gap has been sensibly
reduced (Nagler, 2011). A detailed characterization of land cover is
essential for urban areas, as coarse spatial description might hide
urban heterogeneous ET and result in biased estimates. While
the high spatial variability demands an adequate representation,
implementing the true spatial scale of the ET processes is unfeasible for hydrological catchment models.
2.3. Depression storage, overland ow and runoff
In urban catchments, precipitation falls on (semi)natural surfaces or on sealed surfaces, runs off to receiving water bodies, pervious surfaces or ends in storm water collectors. Depression water
storage capacity of natural surfaces ranges from 0.5 mm to 15 mm,
and for impervious surfaces this range is reduced to 0.2 mm and
3.2 mm (Tholin and Keifer, 1960; Marsalek et al., 2007).
Depression storage is more inuential at low rainfall intensities,
whereas its impact is marginal during heavy storms.
Characteristic spatial scale of the depression storage process
ranges from about 10 1 mm, surface irregularity size, to about
1 m, according the spatial variability of urban surface covers
(Fig. 3). The process of lling depressions is relatively fast and is
a function of precipitation, evaporation and inltration rates and
in general is in the order of seconds to minutes.
Water, on (semi)natural surfaces, undergoes the classical hydrological processes of natural catchments. Sealed surfaces tend to
produce a higher amount of runoff because of their low permeability and smooth surface. Imperviousness is a key environmental
indicator (Arnold and Gibbons, 1996) and sealed surfaces are of
great importance in hydrological analysis, their spatial distribution
and connectivity is by far the most inuential factor in determining
ow velocity and volume of surface runoff (Shuster et al., 2005;
Zhou et al., 2010; Jacobson, 2011). Connectivity to storm water system has great impact on the amount and the dynamics of surface
runoff (Boyd et al., 1993; Brabec et al., 2002; Lee and Heaney,
2003; Roy and Shuster, 2009). High ow velocity decreases the
lag time, time interval between the centre of mass of the storm
and the centre of mass of the resultant hydrograph (Paul and
Meyer, 2001; Huang et al., 2008). This high velocity produces higher
ood peaks than in pre-urbanized conditions (Burns et al., 2005),
and augments erosion, which reduces slope stability and generates
more suspended sediments (Fletcher et al., 2013), resulting in an
overall higher ood risk and severity (Kang et al., 1998;
Niemczynowicz, 1999; Chen et al., 2009). Base ow can decrease
due to urbanization and surface sealing (Foster, 1990; Rumman
et al., 2005). Other studies however report no statisticallysignicant changes in base ow due to urbanization (Meyer,
2005). In natural conditions, if the ow regime changes, e.g. due
to an increase in velocity, the channel tends to adjust itself to
the new ow (Leopold, 1968). In urban catchments, water courses
are often articially modied, e.g. open channel converted
into a closed pipe system. These modications limit the natural

adaptability of water courses, as a consequence, ood frequency
and magnitude increase.
Temporal and spatial scales of the runoff process depend on the
geo-morphodynamic characteristics of the catchment. River runoff
is generally faster than overland ow and its temporal scale varies
mainly according to rainfall intensity and channel morphodynamic
properties. The characteristic time scales of river runoff and overland ow are in the order of minutes to days, for mountainous
creeks and lowland rivers respectively. The characteristic spatial
scales of these two processes is very wide, ranging from 10 1 to
103 m. Urbanization reduces the natural temporal and spatial scale
of the runoff phenomenon due to high spatial heterogeneity of
urban surfaces and by decreasing the catchment response time
to precipitation. A highly detailed spatio-temporal description of
this physical process is therefore needed to obtain an accurate
understanding of its dynamics.
2.4. Storm water drainage systems and combined sewer systems
Stormwater drainage systems date back several millennia BC
(Burian and Edwards, 2002), with the main purpose of protecting
urbanized zones from oods by moving surface runoff away as fast
as possible (Delleur, 2003). The amount of runoff into these pipes
or open drains depends on rainfall intensity, the drain or pipe
dimensions and whether the complete section is available for the
ux. In Europe, for example, such systems are designed to prevent
oods with return periods of 1050 years (European Standard EN
752, CEN, 1996, 1997). Pipes and drains deviate overland ow
and reduce its time scales.
Research on stormwater ow particularly focuses on (1) the
identication of the contributing areas for stormwater-system
design, and (2) the characterization of stormwater pollution,
streets washoff and pipe leakage. (1) Stormwater pathways are
complex and the ow regimes change as the water moves along
its path (Pan et al., 2012). The physical understanding is often
solely based on modelling results as data sets are limited.
Mackay and Last (2010) estimated that the stormwater system of
Birmingham collects 25% of the total surface runoff. Lee and
Heaney (2003) suggest to use imperviousness as proxy for
stormwater runoff production, and Vrebos et al. (2014) use, among
other data, effective impervious area for assessing the impact of
sewer infrastructure on the hydrological regime. The second point
receives by far the most attention in the literature as the assessment of pollution risks for receiving water bodies is eminent
(Barbee et al., 2002; Egodawatta et al., 2013).
Combined storm-drainage and sewer systems exist in many
cities worldwide. Mixed water has different ow characteristics
since density and quantity are higher than those of stormwater.
The description, given in Section 2.6.1, of the interaction between
the sewer system and the natural hydro(geo)logical system is
equally valid for combined storm-drainage and sewer systems.
2.5. Retention basins and stormwater management systems
Retention basins can store medium to large quantities of runoff
in particular locations and are articially created to protect urban
areas from oods and pollution. Several types of retention basins
are classied in Danso-Amoako et al. (2012) ranging from traditional ood retention to integrated sustainable wetland and ood
protection systems. Considerable amount of water can be stored
in such systems, Annunziato et al. (2009) analyse a hydraulic project of 29 reservoirs with a total volume of 20 106 m3. Unsteady
ow simulations show a reduction of peak discharge of about
50 m3/s and a time to peak delay of about 1 h. The Rhine River currently has a total retention volume of 121 106 m3 and in 2020
67
this should reach 294 106 m3 for a design discharge of

16,000 m3/s (Te Linde et al., 2010). Additional objectives of such
systems are nutrient and non-point source pollutant removal, like
in the case of Stormwater Treatment Areas (Labadie and Wan,
2010). Retention basins have therefore relevant impacts on both
quantity and quality aspects. Time scales of the lling of retention
basins are similar to the characteristic time of the overland ow
process, and spatial scales are in the order of m to km.
Living or green roofs can also contribute to stormwater management by intercepting part of the precipitation which evapotranspires during dry periods (Stovin, 2010). Fassman-Beck et al.
(2013) monitored four living roofs over periods of 8 month to
>2 years. They observed a 25 mm precipitation threshold for runoff
generation and a reduction of peak ow between 62% and 90%.
2.6. Inltration and subsurface processes
Understanding and quantifying the impact of urbanization on
groundwater systems is a key topic in urban hydrology (Sherlock,
1922; Lerner, 1990; Appleyard, 1995; Changming et al., 2001;
Schirmer et al., 2012). Human activity causes changes, in very short
time spans, of frequency, volume as well as quality of groundwater
recharge (Garcia-Fresca and Sharp, 2005). Due to urbanization, natural recharge mechanisms are modied and new mechanisms
appear (Foster, 1990): (1) impervious cover reduces the inltration
and so direct recharge, which produces an overall depletion of
groundwater resources; and (2) urbanization produces a general
increase in groundwater recharge mainly due to indirect recharge
generated by mains leakage (Lerner, 2002). Hence, the assessment
of anthropogenic effects on the groundwater system is extremely
challenging. To describe these phenomena we separately discuss:
(Section 2.6.1) inltration mechanisms leading to direct groundwater recharge, (Section 2.6.2) indirect recharge as a result of
mains leakage, and (Section 2.6.3) we give some consideration on
groundwater extraction, ow and discharge.
2.6.1. Direct groundwater recharge: inltration
Urban soil covers have generally a low permeability, often
assumed nearly zero. Hence, urbanization reduces (direct)
recharge and as a consequence groundwater table and storage drop
(Leopold, 1968; Rumman et al., 2005). Surface sealing in residential
areas can reach more than 50%, while in industrial areas can easily
go up to 7080% (Foster, 1990). Urban paved surfaces however are
not completely impervious, runoff losses can account for 3040% of
the total runoff (Ramier et al., 2011), and measured inltration on
roads in residential areas accounted for 69% of the total annual
rainfall (Ragab et al., 2003). The observation of most urban pavements moreover reveals abundant fractures that may permit inltration. Wiles and Sharp (2008) estimated, using a double-ring
inltrometer, (a) a non-negligible secondary permeability of urban
pavements, and (b) their equivalent hydraulic conductivity which
is comparable to that of ne-grained sands, sandstones, silts, and
loams. Sealed surfaces are generally smoother than vegetated surfaces, thus the residence time of surface runoff is shorter, and
water has little time to inltrate into the soil.
Other factors increasing direct groundwater recharge are: (1)
over-irrigation of green spaces; (2) reduced evapotranspiration;
(3) articial and near-natural stormwater inltration systems,
which can redirect large parts of the storm runoff to groundwater
recharge (Shuster et al., 2007). Gbel et al. (2004) tested several
congurations of such systems in a catchment in Germany, and
reported inltration rates between 60% and 80% of the annual precipitation (794 mm/yr), and thus a non-negligible increase of
area-averaged groundwater recharge of 70100 mm/yr.
Human modications of the natural environment do not end at
the soil surface. The development of most cities has a long history,
68
new cities have been constructed over old settlements. The rst
meters below a city are therefore a complex aggregate of soil, utilities trenches, buried structures and openings. Urban karstication, is an interesting concept that compares urban soils to karst
(Garcia-Fresca and Sharp, 2005). Permeability in karst as well as
in urban soils is highly variable, therefore the secondary permeability induced by such buried structures might create preferential
ow paths for water, enhancing inltration and direct recharge.
The ne scale of these processes demands a detailed spatial
description of urban land cover and subsurface preferably on the
scale of 1 m (Dujardin et al., 2011).
2.6.2. Indirect groundwater recharge: leakage/drainage
Although direct inltration might be reduced by anthropogenic
soil sealing, the global effect of urbanization on groundwater is in
many cases an increase of recharge (Appleyard, 1995). This
increase is the consequence of leakage. To avoid inltration of pollutants into drinking water networks, the water distribution is
pressurized and thus prone to leakage where damages/misconnections are present. Only few water utilities worldwide can claim
they were able to reduce leakage to less than 10%, while 2030%
in developed countries and 3060% in developing countries are
more common (Lerner, 2002; Garcia-Fresca and Sharp, 2005).
When leakage happens in permeable soils, potential groundwater
recharge raises up to 100300 mm/yr in normal cases and to about
3000 mm/yr in extreme cases (Lerner, 1990). In (semi-)arid
regions, leakage can be even higher than natural recharge.
Sewers are generally and fortunately less prone to leakage, nevertheless, sewer leakage belongs to the main sources of diffuse
groundwater contamination (Taylor et al., 2006; Wolf et al.,
2007; Vizintin et al., 2009). The European standard EN 752-2
(CEN, 1996) states that the structural integrity of urban sewer systems, including water tightness, should be guaranteed. Sewer leakage is not a major source of groundwater recharge; studies report
that 520% of the sewer ux is lost by leakage (Joss et al., 2008),
therefore leakage contributes to groundwater recharge only for
15% (Yang et al., 1999; Rutsch et al., 2006). Sewage pipes are generally positioned deeper than the water distribution network; in
case of shallow aquifers, sewage pipes might therefore act as
drains for the groundwater. Dilution of waste water concentrations
leads to a lower pollutant removal efciency and additional cost
for the wastewater treatment plants. A quantitative assessment
reveals that in 108 of the 194 investigated catchments in
Belgium, more than 50% of the dry-weather sewer ow is due to
groundwater inltration or parasitic water (Dirckx et al., 2009).
Similarly in Germany, Wei et al. (2002) state that on average
70% of treated water is originally non-polluted.
2.6.3. Groundwater extraction, ow and discharge
More than 70% of the global water consumption is provided by
groundwater extraction (Zektser and Everett, 2004). Unsustainable
groundwater extraction is not uncommon, natural or induced
recharge is generally not sufcient to allow groundwater systems
to recover from the overexploitation, allowing water tables to drop
up to several hundreds of meters (Foster and Chilton, 2003).
Consequences of this practice are degradation of groundwater
dependent ecosystems, reduced base ows, and land subsidence
with increased risk for the population and for the structural integrity of buildings. In the North China Plain, the total water demand
exceeds the natural groundwater recharge by a factor of 1.5, resulting in (1) 1.6 billion m3/yr water shortage for agriculture; (2) water
table decline in both shallow and deep aquifers of 0.4 m/yr; (3)
saltwater intrusion; (4) land subsidence; and (5) groundwater
quality deterioration (Changming et al., 2001). In other cases,
groundwater extraction is intended for lowering the water table
for large scale subsurface constructions (Kim et al., 2001).
Manmade changes of water courses, i.e. channelization,

straightening, conversion of open rivers into closed pipe systems,
slope modication, river bed composition, have repercussions on
the groundwater system. The ecologically-vulnerable hyporheic
exchange zone between surface and groundwater is altered by
these modications, and natural water uxes are disturbed
(Hancock, 2002). In extreme cases, water courses and groundwater
systems can become partially disconnected, river ow regimes can
drastically change, and groundwater discharge can be localized in
different zones.
3. Urban hydrological modelling
Urban hydrological models are primarily developed (1) to evaluate the effect of urbanization on the natural water system and to
augment the knowledge of this complex system; (2) to compensate
for the lack of reliable data, as measurements in the heterogeneous
urban environment are even more challenging than in the natural
environment; and (3) to make predictions towards the future, i.e.
ood forecasting, land use and climate change scenario impact
assessment and ecosystem protection. The primary needs of urban
population are the supply of clean fresh water and the evacuation
of the waste(water). These needs guided the initial development
of models to support the design of water supply and sewer networks. In a later stage, scientists addressed the problems of safety,
i.e. ood and pollution risk assessment, with the development of
hydraulic and transport models (Price and Vojinovic, 2011).
Hydraulic models are capable of numerically solving the governing
De Saint Venant ow equations at a very high resolution. However
the spatial scales covered by these models are generally limited to
the build-up area, and are therefore incapable of evaluating the
hydrological conditions of urban catchments. Moreover these types
of models generally focus on the assessment of particular hydrological/hydrochemical variables, like sewer ow or pollutant load, and
thus they lack interactions with other relevant processes of the
urban water cycle. They also require strong boundary assumptions,
such as incoming surface runoff, as they do not simulate the entire
hydrological cycle. As outlined by OLoughlin et al. (1996) and
Mitchell et al. (2007) the consistency of model integration at several
scales is a major problem for todays models when attempting to
simulate the complete water system at the catchment scale.
This section assesses the status of urban hydrological modelling
by analyzing: the spatial and temporal scale of a selection of modelling tools (Section 3.1), the use of GIS and Remote Sensing
(Section 3.2), the ow routing options (Section 3.3), the advances
in process-based description of the urban water system
(Section 3.4), and how parameters are estimated and models
evaluated (Section 3.5). To do so, we compared 43 models and
modelling strategies described in 58 publications in the period
19872014 for a total of 66 case studies (Table 2). Only a few of
the analyzed tools are specially developed for urban studies, most
of them are generic hydrologic/hydraulic rainfall-runoff models
applied to urban areas or catchments. However, the comparison
of these studies can provide insights in the modelling advances in
and limitations of particular parts of the urban water system.
When available, additional information derived from user manuals
and websites is used, however the purpose of this section is not to
support model selection but to identify the current practice in
urban hydrological modelling at catchment scale.
3.1. Spatial and temporal model resolution
Hydrological models are often classied based on their spatial
and temporal resolution. The selected modelling approaches deal
with urbanization at the catchment scale or at city scale. We dene
69
Table 2
43 modelling approaches for small city to large catchment scale applications, temporal and spatial resolution and references. We also mention if the modelling approaches use
opens source (OS), free (F) or commercial software (C) for the cases in which this information was available to the authors.
Model
References
Study area (km2)
Spatial scale
Temporal scale
(1) Urban residential model

(2) SURF
(3) SWMM, and
SWMM + inundation
model (OS)
(4) MUSIC (C)
Xiao et al. (2007)

Bellal et al. (1996), and Rodriguez et al. (2000)
Lee and Heaney (2003), Girons et al. (2010), Hsu et al. (2000),
Petrucci and Bonhomme (2014), and Krebs et al. (2014)
1.4 10 3
0.047, 0.134, 0.16
0.0581, 0.6, 0.7,
1.3, 2.3, 10.5, 250
UHE
5m
Semi-distr, 120 m
1h
6 min
2 min, 1 min, 5 min,
10 min, 15 min?, 1 h
Dotto et al. (2011) and Hamel and Fletcher, 2013
Semi-distr, lumped
6 min
(5) KAREN
Kleidorfer et al. (2009) and Dotto et al. (2011)
Lumped
6 min
(6) Conceptual model

(7) URBS-UH, URBS-MO
Aronica and Cannarozzo (2000)

Rodriguez et al. (2003) and Rodriguez et al. (2008)
Semi-distr
UHE
1 min 1 h
1 min, 2 min, 5 min
(8) SUES (F)

(9) UHE model
(10) Hyperbolic model
(11) MOUSE + FORM (C)
(12) GUFIM
(13) U-McIUH (OS)
(14) Distributed Urban
Drainage model
(15) SMDR (F)
(16) InfoWORKS CS (C)
(17) RORB (F)
(18) UVQ + NEIMO (F)
(19) TOPURBAN
(20) GwNeu + HYDRU2D +
SPRING (C)
Grimmond and Oke (1991) and Mitchell et al. (2008)

Berthier et al. (2004)
Aronica and Lanza (2005)
Thorndahl and Willems (2008)
Chen et al. (2009)
Girons et al. (2009)
Pan et al. (2012)
0.105, 0.28, 0.38,

0.89, 1.06, 40
0.105, 0.28, 0.38,
0.89, 1.06
0.128
0.18, 0.6, 0.6, 1.8,
1.8
0.21, 30
0.47
0.82
0.87
1.5
1.64, 10.9
2, 100
UHE
UHE, cadastral parcels
5m
Semi-distr
10 m
20 m
10 m, 200 m
P1 h, 5 min
1h
3 min
1 min
15 min
5 min
1 s 1 min
Easton et al. (2007)

Hurford et al. (2010)
Selvalingam et al. (1987)
Morris et al. (2007) and Mitchell and Diaper (2005)
Valeo and Moin (2000)
Gbel et al. (2004)
3.32
5.41
6.11
6.3
8
11.5
10 min 1 d
1h
1h
1d
1h
Steady state, 15 min
(21)
(22)
(23)
(24)
(25)
(26)
(27)
(28)
(29)
Greene and Cruise (1995)

Ogden et al. (2011)
Faulkner and Barber (2009) and Kjeldsen et al. (2005)
Cuo et al. (2008)
Fewtrell et al. (2011)
Zhou et al. (2010)
Mitchell et al. (2003)
Jia et al. (2001)
Chormanski et al. (2008), Berezowski et al. (2012), and
Verbeiren et al. (2013)
Thielen and Creutin (1997)
Meja and Moglen (2010)
Hamdi et al. (2010)
Brun and Band (2000)
Franczyk and Chang (2009), and Dixon and Earls (2012)
12
14.3
15997
21.9
25
25.15, 94.88
27
27
31, 105.9, 142
10 m
Semi-distr
Semi-distr
UHE + Lumped
Semi-distr
Homogeneous subarea, 2m 4 m, 1
25 m
HRU-UHE, 30.48 m
30 m
Lumped
30 m
25 m
UHE (object), 1030 m
UHE
50 m
30 m
52
52
116
124
162
170
194.8, 855
Semi-distr
25 m
Semi-distr
30 m
10 km
Lumped
HRU
30 s
10 min?
5 min
1h
1h
1d
1 d, 1 mon
Huang et al. (2008)

Im et al. (2009) and Vrebos et al. (2014)
Vieux and Bedient (2004)
McColl and Aggett (2007) and Suriya and Mudgal (2011)
204
258, 350
260
300, >500
Lumped
200 m, 250 m
30 m
Semi-distr
1d
1 d, 1 h
15 min
1h
Singh et al. (2009)

Mackay and Last (2010)
Te Linde et al. (2010)
625
670
18.5 104
Lumped
UHE
Semi-distr, 500 m
1 d, 1 mon, 1 yr
1d
1d
GIS + hydr. modelling

GSSHA + SUPERLINK (F)
ReFH (C)
DHSVM (OS)
ESTRY-TUFLOW (C)
DORS
Aquacycle (F)
WEP
WetSpa (OS)
(30) CANOE
(31) MERCEDES
(32) CAREDAS
(33) Event model
(34) SURFEX (OS)
(35) BASINS (HSPF + GIS) (F)
(36) AVSWAT-X model, and
SWAT + land use change
(OS)
(37) SCS, NLP, etc.
(38) MIKE SHE, MIKE 11 (C)
(39) Vow (C)
(40) HEC-(Geo)HMS, HEC(Geo)RAS (F)
(41) Peri-Urban SIMHYD
(42) CWB (Aquacycle) (OS)
(43) HBV + SOBEK (OS)
models that are applied to an area equal or larger than 10 km2 as

catchment scale applications; and models applied to smaller areas
are city scale applications. 30 out of the 66 case studies are catchment scale applications, the remaining city scale applications are
discussed because they often have more advanced algorithms or
modelling concepts for urban applications. Table 2 gives full details
on (1) the spatialtemporal resolution, (2) coverage of the study
area, and (3) references of these modelling studies. We dene
model implementation as the application of one of the 43 modelling approaches to a particular case study with a specic combination of dt (time step) and dx (spatial resolution). In our sample,
we identied 84 model implementations (Fig. 4).
Based on the spatial discretization, we grouped the 84
model implementations into ve categories: (1) lumped; (2)
5 min
1 min?
1h 1 d
1h
15 min
1h
1d
1h
1 h, >1 h
semi-distributed using sub-catchments; (3) Hydrologic Response

Unit (HRU) based, where an HRU represents a combination of the
same soil and land use type, typical of SWAT (Soil and Water
Assessment Tool) models (Arnold et al., 1998; Arnold and Fohrer,
2005); (4) grid-based spatially distributed; (5) Urban
Hydrological Element (UHE) based. Lumped models are generally
simpler than other approaches and particularly useful when limited data is available, they can provide a rst assessment of the
urban water system at the catchment outlet but they lack spatial
information. Models considering sub-catchments or HRU as basic
unit of calculation are similar in the sense that they consider
sub-regions in the catchment as uniform with respect to the
hydrological processes. They both require more spatial data compared to lumped approaches and can potentially provide more
70
Dynamic
Spaal Resoluon
4+1
Lumped
Semi-distributed
HRU
UHE
10 km
5
3
Small-size
applicaons
Non-grid
based
2+5
2+3
2+3
Steady
state
2+1
Grid based
500 m
250 m
200 m
Catchment-size
applicaons
120 m
50 m
30 m
25 m
20 m
10 m
3
5m
Small-size
applicaons
3m
1s
30 s
1 min 2 min 3 min
5 min
6 min 10 min
Temporal Resoluon
15 min 1 hr
>1 hr
1 day
1 mon
1 yr
steady
state
Fig. 4. The 84 identied model implementations of 43 analyzed modelling approaches are separated into: (a) blue dots or models applied at the urban catchment scale
(P10 km2) and (b) red crosses or models applied to small urban areas (<10 km2). The models simulating the catchment scale cover the full range of spatial and temporal
resolution, while models applied to smaller study sites are concentrated in two regions: (1) Non-spatially explicit zone with temporal resolution ranging from 1 min to 1 day;
and (2) very ne spatial resolution with temporal resolution ranging from 1 s to 1 day. (For interpretation of the references to colour in this gure legend, the reader is
referred to the web version of this article.)
information if the size of the sub-catchments or HRUs is reasonably

small. These two categories use a sort of hybrid modelling
approach, since they take into account a certain degree of spatial
variability, sub-catchments, but the sub-catchment internal
description is lumped. The main difference between HRU and
sub-catchment approaches is the way homogeneous sub-areas
are delineated and calculation executed: semi-distributed models
use topography-derived sub-catchments while SWAT models generate HRUs based on geo-morphological considerations as combination of soil and land cover types. This approach takes into
account a higher number of spatially-distributed hydrological
parameters but, once HRUs are identied, the exact geographical
location is lost as identical combinations of soil and land cover
types are grouped together and considered as one element for
the calculations. Hence, spatial variability provided by GIS maps
and remote sensing can better be exploited by distributed models,
which perform calculations for each grid cell. The level of complexity varies with respect to the spatial resolution of the input data
and with the internal description of the processes. Nearly 36% of
the analyzed modelling approaches are spatially distributed with
spatial resolution ranging from 3 m to 10 km. Finally UHE-based
models are primarily developed for analyzing urban water uxes,
9 out of the 43 analyzed approaches belong to this category. The
concept of UHE is slightly different for each model but the common
idea is the identication of an object or a unit of calculation small
enough to represent the urban heterogeneity and to be homogeneous with respect to the urban hydrological processes. UHEs
can represent (a) urban blocks, buildings and surrounding areas
such as cadastral parcels, e.g. URBS-MO (Urban Runoff Branching
Structure Model) (Rodriguez et al., 2008), Aquacycle (Mitchell
et al., 2003), and UVQ (Urban Volume Quality) (Morris et al.,
2007); (b) single neighbourhoods, like in SUES (Single-source
Urban Evapotranspiration-interception Scheme) (Grimmond and
Oke, 1991) or CWD (City Water Balance) (Mackay and Last,
2010); or (c) objects derived from combinations of spatially distributed catchment characteristics as in DORS (Distributed
Object-oriented Rainfall-runoff Simulation Model) (Zhou et al.,

2010). The last type of UHE can be in part comparable with an
HRU, but UHEs preserve their geographical location and have generally smaller dimensions.
For each approach we identied the time steps used for the simulations regardless the length of the simulation period. Some modelling approaches solve the differential De Saint Venant ow
equations with numerical schemes that allow variable time steps
according to ow conditions, in this case we considered either an
average condition, when the smallest and the largest time steps
were very close to each other, or we included the two extremes
as separate implementations of a particular model. The identied
time steps range from 30 s to 1 year, however more than 90% of
our sample use a time step of 1 day or less.
To compare the models we further divided them in grid based,
i.e. spatially distributed models, and non-grid based, which include
all the other categories since sub-catchments or urban hydrological
elements can have different spatial extents even within the same
application. Non-grid based approaches are popular in urban
hydrological modelling as more than 60% of the analyzed cases
are of this type, the reason is twofold: (a) this approach is easily
applicable when data is scarcely available by considering the
catchment as one big element, e.g. lumped approaches, or increasing the size of sub-elements; (b) UHEs are specically designed to
describe urban systems, they are exible and can include highly
detailed spatial information.
In Fig. 4, we identify two distinct practices in applying hydrological models in urban areas: (1) model applications at the catchment level (blue dots) cover the full range of spatial and temporal
scales, both grid and non-grid based. The current practice in urban
hydrological modelling at catchment scale is therefore highly
diverse in terms of spatialtemporal discretization. The general
trend is such that spatial and temporal resolution are directly proportional; a very high spatial resolution approach, i.e. <10 m, will
result in a ne time step discretization, in the order of minutes,
while a lumped approach is more likely to be described with a
71
Non-spaal-explicit
model implementaons
25
60%
40%
20
Catchment-size
applicaons
Small-size
applicaons
15
10
6 min
seconds
minutes
5m
30 m
< 10 m
< 50 m
Number of model
implementaons
Cumulave % of
model implementaons
Spaal-explicit model implementaons
100
90
80
70
60
50
40
30
20
10
0
1 hr
hours 1d 1 mth1yrS.S.
Temporal Resoluon
< 500 m
0
1-10 km UHE HRUsemi Lumped
distr.
Spaal Resoluon
Fig. 5. The temporal and spatial resolution of the 84 analyzed urban model implementations are displayed in panel A and B respectively. The model implementations are
applied either to small-sized areas (<10 km2) or to catchment-sized areas (>10 km2). Urban hydrological modelling of small-sized areas (red lines and bars) are generally
applied at a ner temporal and spatial resolution than catchment-sized applications (blue lines and bars). (For interpretation of the references to colour in this gure legend,
the reader is referred to the web version of this article.)
daily or larger time step. (2) The city scale application cases (red
dots) are localized into two distinct spatio-temporal zones: in the
very high spatial resolution zone, below 10 m, and in the
non-grid based zone. Modelling approaches in this last zone have
a slightly coarser spatial resolution, between 23 m (Berthier
et al., 2004) and 80 m (Lee and Heaney, 2003) when the study area
is small and the number of identied elements is large. In both
regions, high resolution and non-grid based, the time steps vary
between 1 s and 1 min to 1 day. This behaviour is clearly explained
by the small-scale study areas, which do not allow a too coarse
spatial resolution but that still leaves some degree of freedom in
the choice of the time step.
Urban hydrological models applied at the city scale have generally a smaller temporal and spatial resolution than urban hydrological applications at the catchment scale (Fig. 5). More than 50% of
the small size applications (<10 km2) have a temporal resolution
equal or lower than 6 min and a spatial resolution equal or lower
than 5 m (grid-based group), while for catchment applications
(>10 km2) these values are respectively 1 h and 30 m. Spatial and
temporal resolutions are relatively high when compared to traditional rainfall-runoff applications of non-urbanized catchments.
The current modelling practice, therefore recognizes the need for
a ne spatialtemporal resolution to account for the high urban
heterogeneity and for the fast dynamics of hydrological processes
in urban areas. The most used time steps are however 1 h (25%)
and 1 day (20%), which hinders models from accurately characterizing the fast components of the urban uxes. Higher temporal resolution comes at high computational cost, for this reason most of
the approaches which use a ne time step are event-based.
3.2. GIS and remote sensing in urban hydrological modelling
Since more than two decades hydrological simulation tools
have taken advantages of GIS functionalities. The majority of the
currently available hydrological models are either loosely or tightly
coupled with a GIS. In urban hydrological modelling, this practice
is even more common as 80% of the analyzed model approaches
have a GIS link. The spread of GIS tools certainly was one of the
main drivers in the development of spatially-distributed hydrological models, but nowadays nearly all types of models take benet
from spatial data management to derive for example average
parameter values. GIS is well suited for urban hydrology because
of the strong spatial variability and fast dynamics of urban lands.
Urban databases, with detailed and up to date information, are a
precious source of information for hydrological models. GIS platforms are used for: (a) pre- and post-processing of model
parameters and inputs; (b) managing and displaying spatial data;

(c) watershed surface representation, which includes topographic
aspects and (sub-)watershed delineation; (d) remote sensing data
assimilation; (e) identication of hydrologic unit response or unit
hydrological elements; and (f) as framework for model coupling.
More than 80% of the analyzed modelling approaches uses GIS
functionalities from ArcView, ArcGIS, ArcInfo, ArcGIS-Python, and
GRASS. The more appreciated functionalities are data management
and map algebra calculations to estimate model parameters
(Fig. 6). GIS graphical display tools are used by more than 30% of
the models and nearly 50% of the models delineate watershed from
digital elevation models (DEMs) or digital terrain models (DTMs).
In this regard, Girons et al. (2010) show that the delineation of
urbanized watershed, with varying level of details, is highly sensitive to the selected method. A group of models use routing
approaches based on DEM and/or on other GIS layers by deriving
Geomorphological Instantaneous Unit Hydrographs (GIUHs), e.g.
the model proposed by Meja and Moglen (2010), and the
WetSpa model (Chormanski et al., 2008). In the WetSpa model
for example overland and channel ow are routed to the outlet
using the diffusive wave approximation, a unit response function
is derived for each grid cell of the catchment based on
spatially-distributed morphological characteristics such as ow
direction, slope (DEM derivatives) and roughness. This method,
developed with GIS functions is computationally effective but
assumes the linearity of the system and is not specically developed for urban applications. The Urban Unit Hydrograph
(Rodriguez et al., 2003) is specically designed to describe the
water movement in urban areas by using geometric information
76.5%
Data management,
parameter esmaon
45.5%
Watershed delineaon
36%
RS integraon
32%
Graphical display
Flow roung,
UHE extracon
28%
20.5%
GIS-based tools
0
10
20
30
40
50
60
Number of models
Fig. 6. Urban applications make extensive use of GIS functionalities, mainly for data
management, parameter estimation and graphical display. One out of four of the 43
selected urban approaches is a GIS-based model.
72
from urban databanks for the denition of parcel cells as urban

hydrological element (UHE) and to derive transport functions
based on street location and connectivity as a vector map of ow
paths. Non-linearity in the rainfallrunoff response of urban catchment can be taken into account with the Urban Morpho-climatic
Instantaneous Hydrograph (U-McIUH) developed by Girons
et al. (2009).
Remote sensing data usage is also supported by GIS platforms,
nearly half of the models use remote sensing derived information.
Imperviousness is the most commonly used RS product: airborne
technologies provide very high resolution impervious cover maps
(Lee and Heaney, 2003; Easton et al., 2007; Zhou et al., 2010),
spaceborne imagery can also be used for this purpose (Dotto
et al., 2011; Berezoswki et al., 2012). Radar precipitation data is
slowly making its way into the urban hydrological modelling practices. Five cases use spatially-distributed rainfall input data
(Thielen and Creutin, 1997; Vieux and Bedient, 2004; McColl and
Aggett, 2007; Meja and Moglen, 2010; Pan et al., 2012).
Classication of space(air)borne imagery produces land cover
maps from which important hydrological parameters can be
derived. Most of the urban hydrological models use land cover
information, in the form of maps or averaged values, however only
a very small group takes advantage of the spatial and temporal
coverage remote sensing can offer (Im et al., 2009). The methodology proposed by Verbeiren et al. (2013) tries to exploit both spatial
and temporal benets of remote sensing for land cover mapping. In
their work a time-series of satellite imageries is used to analyse
urban development and to assess the relative impact of urbanization (sealed surface proportions) on the hydrological system.
Finally RS is used for producing DEM from Lidar technology
(Chen et al., 2009; Fewtrell et al., 2011), for vegetation density
and indirect evapotranspiration parameter estimation, such as
Leaf Area Index (LAI) (Zhou et al., 2010), for roughness direct estimation (Vieux and Bedient, 2004), and to estimate channel bottom
width (Ogden et al., 2011).
approaches (Table 2) use a large variety of (hydraulic) methods to

simulate overland ow, storm drainage, river and sewer routing as
well as the presence of hydraulic structures. We therefore grouped
these methods into four categories: (1) (non-)linear reservoirs/conceptual; (2) kinematic/diffusive/dynamic wave approximation,
shallow water equations, 1D and 2D De Saint Venant equations,
further on called hydrodynamic routing; (3) IUH and GIUH; and
(4) no or unknown routing methods.
Roughly all the modelling approaches take into account overland ow routing, nearly 28% also consider river routing, while
only a small minority simulates the hydraulics of storm drainage
and sewer systems (15%). Urban hydrological modelling practice
therefore recognizes the importance of overland ow routing,
however around half of the approaches (20) describe this physical
phenomenon with conceptual methods, such as (non-)linear reservoirs or empirical transfer functions, and only a quarter uses
hydraulic/hydrodynamic conceptualizations (Fig. 7). The remaining cases use the instantaneous unit hydrograph approach or do
not simulate overland ow routing. The routing of the other water
uxes displays a different trend: even though the total number of
approaches is less than in the previous case, these approaches are
mostly highly detailed and physically based. Surprisingly, typical
urban features, such as storm drainage and sewer systems, are
not often simulated by urban hydrological models.
Urban-catchment hydrology diverges from the classical urban
hydraulic concepts, mainly due to scale issues. Detailed description
of sewers and drainage networks appears unfeasible for hydrological models of large catchments. Hydrological models therefore opt
for more integrated paradigms where the level of complexity is
uneven throughout the different ow components. For example,
overland ow is often simulated with conceptual approaches and
sewer ow with detailed hydrodynamic concepts. River ow,
storm drainage and sewer routing require hydraulic simulations,
as when these components are integrated in a model, they are
mostly simulated with rigorous hydraulic formulations.
3.3. Flow routing and hydraulic concepts
3.4. Process-based modelling
Traditionally, a key aspect of urban hydrology is the hydraulic

modelling of sewers, rivers and structures. The analyzed modelling
Models and their applications are developed to answer specic

research and/or practical questions. The way physical processes are
Number of
modelling
approaches
20
16
Overland ow
Channel ow
Storm drainage
Sewer
12
8
4
0
20
16
12
8
4
0
(non-)linear
reservoir/
conceptual
kinemac/
diusive/
dynamic wave,
shallow water,
1D 2D DSV
IUH, GIUH
no or unknown
roung
(non-)linear
reservoir/
conceptual
kinemac/
diusive/
dynamic wave,
shallow water,
1D 2D DSV
IUH, GIUH
no or unknown
roung
Fig. 7. The ow routing concepts, used by the 43 analyzed modelling approaches, are here classied in four group: (non-)linear reservoir, hydrodynamic routing, IUH, GIUH,
and no or unknown routing. Different ow routing concepts are applied to different ow processes.
73
described by models should reect the modelling purpose, which

means developers should select (1) which hydrological processes
to simulate; and (2) level of model complexity in accordance to
the objective. Modelling purposes can be general when they
address the urban water problems at a comprehensive scale.
These general purposes can be stand-alone approaches or can be
a combination of a series of specic-purpose approaches. Both general and specic model purposes in urban hydrological studies are
directly connected to urban hydrological processes (Fig. 8). General
purpose models tend to include a larger variety of hydrological
processes in a rather simplied way while specic/process-based
purpose models have a more detailed description of fewer hydrological processes and consider secondary processes in a more
lumped/empirical way. The modelled surface area extent is also
changing according to the purpose of the study. Specic purpose
models are usually applied at a small city scale, i.e.
storm-drainage network design and pollution assessment, while
general purpose models work on catchment to regional scale for
water resources allocations or health and safety risk management.
The 43 analyzed modelling approaches can be divided into
seven categories according to their main purpose and methodology
used to simulate urbanization and water resources (Fig. 9). In the
next paragraphs, we will address each of these functional categories separately, giving some examples from the 43 selected
approaches. At the end of the section we provide a summary of
model characteristics and limitations for each functional category
(Table 3).
(1) Impervious cover and land-use change impact analysis. The
most common approach (nearly 30% of cases) is to characterize
urbanization exclusively as a phenomenon that alters inltration
and runoff-production patterns by modifying surface characteristics. The approach is based on the different hydrological behaviour
of sealed and natural surfaces and is established by means of classication of urban land covers, each with a certain degree of sealing (Valeo and Moin, 2000; Dixon and Earls, 2012), or by using
spatially-distributed imperviousness (Zhou et al., 2010; Verbeiren
et al., 2013). This type of approach can assess the impact of land
Water
Resources
Sustainability
12 %
16 %
5 models
Integrated
Physicallybased
7 models
16 %
Conceptual,
Integrated
Water
Balance
7 models
21 %
5%
9 models
2 models Urban Soil &
28 %
Groundwater
Urban Climate
12 models
Imperviousness & Land use change

Fig. 9. The 43 analyzed modelling approaches belong to seven functional groups
according to the approaches used for characterizing the urban environment.
cover change on the surface water system and so it can simulate

and evaluate urban planning strategies. The impacts of urbanization on the hydrological system are however more physically complex and the effects of other typically-urban processes, such as
storm water, sewers, mains network and leakage, are highly relevant for assessment of water uxes. Moreover, in land-cover
change studies, the focus is often limited to surface processes
and little or no attention is given to sub-surfaces urban groundwater processes.
(2) Conceptual Integrated Water Balance. Models of this kind
(21%) aim at an integrated water balance assessment in urban
areas with conceptual and mostly empirical methods. These
models, applied at small to medium sized catchments, introduce
concepts like the Urban Hydrological Elements (UHE) for
their calculations. Generally, they require detailed data on
domestic and industrial water use that is estimated on basis of
population data. Urban-specic simulated processes are: storm
drainage, sewer and mains, ow and leakage with various
empirically-based approaches, wastewater treatment plants,
Change
Impact
Evapo
Precipitaon (transpi)
Temperature
raon
Climate
Storm Drainage
Design
Inundaon,
Floods
Protecon
Ecology &
Biodiversity
Preservaon
Groundwater
Floods
Control
Water
Demand
Overland ow
Storm
Drainage
Design
River ow
Storm drainage
Hydrological
Processes
Flood
Retenon
Leakage
Inltraon
Polluon
Recreaon
Scienc
Theories
Land Use Change

Impact
Health &
Safety
Fig. 8. Urban hydrological modelling purposes can be classied into general (pink) and specic purpose (blue). We identied four general purposes: water resources
sustainability, ecology and biodiversity preservation, health and safety, and scientic theory validation. These general purposes refer to the complete set of urban hydrological
processes while specic modelling purposes take into account only a subsection of the physical processes of urban catchments. Some of the images were downloaded from
www.freerangestock.com. (For interpretation of the references to colour in this gure legend, the reader is referred to the web version of this article.)
74
pumping stations, groundwater extraction, domestic and industrial

water use, retention basins functioning and pollutant load and fate.
Examples of this kind of models are: Peri-Urban SYMHYD (Singh
et al., 2009), CWB (City Water Balance) (Mackay and Last, 2010)
and its predecessors Aquacycle (Mitchell et al., 2003), and UVQ
(Urban Volume Quantity) (Morris et al., 2007). Conceptual models
can also be used to simulate Best Management Practices (BMPs) or
source control devices affect the inltration process (Hamel and
Fletcher, 2013). All these models, in general, provide great support
for water resources management and are used to evaluate water
policies. They simplify the physical processes involved and derive
therefore a general picture of the urban water status. Hence, they
lack a full spatial characterization of the hydrological system and
cannot be used for applications like ood or groundwater level
mapping and estimation of interactions between surface and
sub-surface uxes.
(3) and (4) Inundation, Flood Protection and Storm Drainage
Design. A wide range of commercial and open source packages
are available for ood modelling and for the design of protective
infrastructure, including stormwater drainage systems. Many
examples of these two types of purpose models are included in
our sample of models (28%, sum of 16% and 12%, Fig. 9), sharing
generally a more rigorous formulation of pure hydraulic principles
in their calculation of water ow. Floods and inundations, frequently recognized as major negative consequence of urbanization,
have a highly complex dynamics. Inundation is generally considered as a local phenomenon and it is quantied for small areas,
in the analyzed sample ranging from a few hectares to 2 km2.
The main physical processes involved in ood events are surface
runoff, inltration, and water ow in storm drainage systems and
retention basins. In modelling practice, the surface runoff for ood
assessment is generally simplied to a two-dimensional phenomenon with water uxes simulated by one of the various forms
of the De Saint Venant equations, i.e. kinematic, diffusive or
dynamic wave. Examples are the hyperbolic model proposed by
Aronica and Lanza (2005), the models of Ogden et al. (2011) and
Fewtrell et al. (2011). Other less rigorous approaches take full
advantage of high spatial resolution information to characterize
the ow direction and avoid ow through buildings by articially
increasing the elevation of high resolution DEM at building locations, like in the case of the Semi-Urbanized Runoff Model (SURF)
model (Bellal et al., 1996; Rodriguez et al., 2000), or use the
at-water concept when detailed topographic data is not available
(Chen et al., 2009). Most of these examples also take into account
the impact of storm water drainage, which can either decrease
the intensity of the oods or augment it when the maximum drainage capacity is reached and surcharge starts. The storm drainage
network can be simulated as a combination of virtual hydraulic
elements with inlets considered as internal sinks (Aronica and
Cannarozzo, 2000) or water ux in storm drainage pipes can be
simulated assuming uniform ow condition and deriving transfer
time curves, like in the case of the SURF model (Bellal et al.,
1996). More detailed, physically-based descriptions of stormwater
conduits drainage use the 1D form of the De Saint Venant equation
or couple an inundation model with storm drainage design software, like in the case of Hsu et al. (2000) where the Storm Water
Management Model SWMM (Huber and Dickinson, 1988) was coupled to an inundation model. Beside the SWMM model, other generally accepted software for stormwater drainage (and sewers)
ow simulations, present in our sample of models, are MUSIC
(Dotto et al., 2011), InfoWORKS CS (Hurford et al., 2010), MOUSE
(Thorndahl and Willems, 2008), and CANOE (Lhomme et al.,
2004). These models are capable of accurately estimating water
ow in articial networks thereby providing solutions for most
of the practical demands regarding design and implementation.
They also often include modules to assess pollutant load and fate.
Generally, they are referred to as urban hydrological models,

however (1) they are not designed to analyse catchment sized
problems; (2) their description of typical natural hydrological processes is rather simplied, e.g. considering sub-catchment runoff
production as input for the storm water drainage system; and (3)
rarely take into account sub-surface and leakage processes.
(5) Urban Soils and Groundwater. Soil water conditions in urban
areas are of great importance in the simulation of surface and subsurface catchment hydrological processes. Urban soil properties
are heavily modied by construction and piped (waste)water ow
can interact with the unsaturated as well as saturated zone.
Notwithstanding the hydrological importance of urban soils,
hydrological models for urban catchments tend to limit model
complexity by neglecting most of the sub-surface processes, justifying the assumptions on basis of temporal scales as
soil-groundwater processes are generally slow compared to ood
generation. In the cases where sub-surface processes are simulated, there is no particular emphasis on characterizing urban soils.
The level of complexity of this type of approaches goes from relatively simple, using for example the concept of linear reservoirs
(Easton et al., 2007; Faulkner and Barber, 2009), to a more
physically-based two or three dimensional approach comparable
to traditional groundwater ow models (Jia et al., 2001; Gbel
et al., 2004). The concepts of urban porosity and urban transmissivity appear in Aronica and Lanza (2005) but are not resolved by
their model, while Berthier et al. (2004) model the role of urban
soils in the runoff generation. Their modelling approach takes into
account most of the anthropogenic modications of soils: storm
drainage networks can drain groundwater when pipes are not
watertight, soils below streets are composed of several layers,
which may create preferential pathways for water, and buildings
are considered as fully sealed surfaces but the presence of foundation drains water or modies the general ow paths. Soil water
uxes are calculated using Richards equation with a nite element
approach and pipes draining groundwater are considered as ideal
drains. This unconventional approach is however applied to a very
small area, 5 10 2 km2, and other important urban processes are
neglected, such as evapotranspiration and inltration on paved
surfaces. Another attempt of detailed characterization of some soil
water processes, limited to inltration and exltration from drainage and sewer network, is the Network Exltration and
Inltration MOdel (NEIMO) (Morris et al., 2007). NEIMO is based
on a deterministic approach for a quantitative and qualitative
assessment of leakage, using Darcys law and a detailed description
of conduits defects and surrounding soil physical properties.
(6) Urban Climate. Including atmospheric variables in urban
hydrological modelling allows taking into account both local urban
climatic effects on the water balance as well as the impact of
urbanization on the hydro-climatology. Nearly all 43 models consider precipitation processes, most of them include evapotranspiration analysis, some also include temperature for simulating
snow accumulation and snowmelt phenomena (Cuo et al., 2008),
and a few take into account other energy balance variables, such
as long- and short-wave radiation, sensible and latent heat uxes
(i.e. Jia et al., 2001). None of the 43 models tries to evaluate the
impact of urbanization on rainfall patterns, which means that this
topic remains a niche exercise not included in hydrological modelling practice at the catchment scale. Only the work of Thielen
and Creutin (1997) implies that their methodology could potentially be applied to evaluate the consequences of urbanization on
urban climate as they couple a climate model with a
rainfall-runoff model. Their approach is modular and composed
of physically-based and lumped approaches; the atmospheric
module simulates rainfall events in urban areas, and the hydrological module is composed of two sub-models: (a) overland ow and
inltration; and (b) a storm drainage system. The level of detail is
not consistent because the hydrological simulations are lumped

and cannot fully exploit the detailed characterization of atmosphere processes. The analyzed models treat evapotranspiration
with a varying degree of detail, from neglecting ET contribution
in the urban water balance for event-type simulation (Chen
et al., 2009), or using empirical formulations (Mitchell et al.,
2003; Chormanski et al., 2008) up to more physically-based
approaches supported by the PenmanMonteith equation
(Franczyk and Chang, 2009; Zhou et al., 2010). The SUES model
(Grimmond and Oke, 1991; Mitchell et al., 2008) uses the more
sophisticated single-source urban evapotranspiration-interception
scheme to represent the energy balance in urban areas in a
fully-physical way, considering the anthropogenic heat uxes,
the change in heat storage in the urban canopy (buildings and
streets), and aerodynamic and surface resistance. The SUES model
is a purpose-specic tool that requires detailed input and many
parameters, which are physically-based and can therefore be measured. The model is capable of reproducing ET at a high spatial and
temporal resolution but it does not take into account other important hydrological processes which can affect evapotranspiration by
modifying the soil water availability, such as overland ow, sewers
and mains leakage. The detailed description of ET requires very
high resolution spatially distributed data and was only applied to
catchments with a very small (0.21 km2) to medium size (30 km2).
(7) Integrated Physically-based models. Finally, there is a group of
modelling approaches that addresses the problem of water in the
urban environment in a more comprehensive and holistic way,
mostly at catchment scale, with an authentic hydrological perspective. These models are physically-based and consider both the
water and the energy balance of urban catchments. To this end,
they aim at describing most hydrological processes in a detailed
way, causing the models to become complex, computationally
75
heavy, data demanding and potentially more uncertain. However,

the models belonging to this category do not cover the full spectrum of urban hydrological processes with a consistent level of
detail, but they can be considered a step forward towards the
recognition that urban catchments intrinsically differ from natural
systems. The Urban Runoff Branching Structure MOdel (URBS-MO) is
perhaps one of the clearest examples of this typology (Rodriguez
et al., 2008). This innovative approach models the hydrological
processes at the small Urban Hydrological Element (UHE) scale
and considers: interception and evapotranspiration with
PenmanMonteith approach, saturated soil processes with a mixture of 3D equations and simplied approaches, drainage of
groundwater by means of storm water pipes with the ideal drain
concept, overland ow and street ow calculated with Urban
Geomorphological Instantaneous Unit Hydrograph (Urban GIUH)
based on detailed GIS urban databanks, and routing of stormwater
drainage with the MuskingumCunge scheme. Weak points of this
approach are the fact that pipe leakage is neglected and the use of
the Urban GIUH method that implies the linearity of the hydrological response of the catchment, which might not be realistic in
cases like ash oods. Hence, the model does not allow the rigorous assessment of spatially-distributed water uxes, making the
model not suitable for simulating ood zones. Gbel et al. (2004)
is a good example of increased model capabilities by combining
three purpose-specic models: (a) a water budget model
(GwNeu) for the assessment of groundwater recharge for
non-urban soils; (b) the deterministic 2D vertical model
HYDRUS_2D to model the inltration from urban inltration locations; and (c) the groundwater ow model SPRING. These methodologies are process based, which is an advantage because physical
processes can be simulated with a targeted approach, for example
one can select the most appropriate form of the ow equation to
Table 3
We identied 7 functional categories for the 43 analyzed modelling approaches. The main model characteristics (processes, purposes and application scale) and limitations are
summarized here.
Model category
Characteristics
Limitations
(1) Impervious cover, Land

use change
MAIN PROCESSES: Detailed description of the urban inltration process

and the urban runoff production
PURPOSE: Evaluation land-cover change impacts, general hydrological
simulations
APPLICATION: Catchment scale
Limited treatment of articial pathways (sewers, stormwater

system, water supply, etc.)
Groundwater description and the urban soils characteristics
are often simplied
(2) Conceptual, Integrated

Water Balance
MAIN PROCESSES: Complete set of articial pathways, including

leakage.
PURPOSE: Water resources management, pollution assessment
APPLICATION: Small city-scale
Conceptual-empirical formulations; limited applicability for

ood assessment and sub-surface water simulation
Urban catchment scale?
(3) and (4) Inundation, Flood

Protection and Storm
Drainage Design
MAIN PROCESSES: Rigorous formulation of hydraulic principles for

surface runoff, storm drainage systems, sewers. They might include
pollutant simulation.
PURPOSE: Flood simulation, storm drainage/sewers design
APPLICATION: Small-city scale, localized phenomena
Natural hydrological phenomena are less rigorously treated;

pipe leakage and water movement in the sub-soil is rarely
simulated
(5) Urban Soils and

Groundwater
MAIN PROCESSES: Highly detailed description of sub-surface urban

processes, urban porosity, urban transmissivity, physically-based
leakage
PURPOSE: Urban groundwater system assessment, pollutants
APPLICATION: Small city-area
Other processes of urban catchments are not evaluated in

detail; very specic purpose models; limited applicability
MAIN PROCESSES: Urban precipitation and urban evapotranspiration.

Detailed energy balance. Physically-based equations. Urban vegetation
PURPOSE: energy balance, urban climate-evapotranspiration
No impact of urbanization on precipitation patterns and

intensity
Level of detail is not consistent: very detailed description of
atmospheric processes and lumped description for others
hydrological processes
(6) Urban climate
APPLICATION: small to medium size catchments

(7) Integrated Physicallybased
MAIN PROCESSES: Comprehensive description of urban hydrological

processes, energy and water balance, physically based equations, subsurface processes are included
PURPOSE: hydrological assessment of urban catchments, general
purpose models
APPLICATION: Catchment scale, spatially-distributed
High complexity/uncertainty; data demanding;

computational heavy; urban hydrological processes are not
treated with a consistent level of detail
76
describe the river ow or overland ow, or adapt the spatial and

temporal resolution. Pan et al. (2012) for example, uses layers
and combines the one-dimensional Richardss equation for the vertical soil water movement, the 2D De Saint Venant equation in its
diffusive form for overland ow, and the 1D dynamic form of De
Saint Venant equation for the street, sewers, and ditch/river layers.
The modelling approach in Xiao et al., 2007, also represents an
advance in the eld as (1) it takes into account ow processes in
both the soil and at the soil-surface interface (interception and
evaptranspiration), and (2) it allows to better consider irrigation
inputs and the analysis of Best Managements Practices (BMPs).
3.5. Parameter estimation and model evaluation
Hydrological models for urban applications are considerably
more complex than classical hydrological models as they try to
simulate many more processes at a generally ner
spatio-temporal scale. In this type of modelling approach, the
number of parameters involved and data requirement signicantly
increase (Hamel and Fletcher, 2013; Girons et al., 2009, 2010;
Hamdi et al., 2010; Mackay and Last, 2010; Im et al., 2009; Singh
et al., 2009; Chormanski et al., 2008; Xiao et al., 2007; Morris
et al., 2007; Rodriguez et al., 2000; Grimmond and Oke, 1991). As
a consequence, there is a growing issue with the reliability of the
results of these kind of models (Vrebos et al., 2014; Petrucci and
Bonhomme, 2014). Conceptual lumped approaches might not be
sufciently detailed in terms of spatial coverage but, in case only
discharge at the outlet is required, they might be more accurate
as they pose less over-parameterization problems (Vrebos et al.,
2014). In many cases, the high data demand is not fullled and
proper calibration and validation is unfeasible (Chormanski et al.,
2008, Aronica and Lanza, 2005). Moreover, it is recognized that
more accurate and detailed spatio-temporal measurements would
have improved modelling results (Dotto et al., 2011; Hurford et al.,
2010; Morris et al., 2007). The preferred choice for increasing
model performances is often manual calibration (Vrebos et al.,
2014; Hamel and Fletcher, 2013; Pan et al., 2012; Ogden et al.,
2011; Morris et al., 2007; Aronica and Cannarozzo, 2000).
However some approaches apply sequentially manual and automated calibration (Berezowski et al., 2012; Franczyk and Chang,
2009) or explore various automatic calibration techniques
(Petrucci and Bonhomme, 2014; Dotto et al., 2011; Meja and
Moglen, 2010; Kleidorfer et al., 2009; Berthier et al., 2004;
Rodriguez et al., 2000).
Urban integrated models are intrinsically complex and tend to
have many parameters. Hence a number of approaches are used
for dealing with parameter complexity.
(1) Parameter reduction: sensitivity analysis can reduce the
number of parameters for calibration and identify the most
inuential and potentially most uncertain parameters
(Krebs et al., 2014; Vrebos et al., 2014; Dotto et al., 2011;
Meja and Moglen, 2010; Kleidorfer et al., 2009; Rodriguez
et al., 2003, 2000). In many cases the most sensitive parameters are the roughness coefcient and imperviousness
(Ogden et al., 2011; Girons et al., 2010; Cuo et al., 2008;
Lee and Heaney, 2003; Rodriguez et al., 2003). Spatial resolution is also recognized as having an impact in modelling
results and uncertainty, it should therefore be considered
in the sensitivity/uncertainty analysis (Pan et al., 2012;
Fewtrell et al., 2011; Valeo and Moin, 2000). According to
Petrucci and Bonhomme (2014) a trade-off should be found
as excessive spatial distribution can lead to severe
over-parameterization. Other authors however consider RS
data very benecial as ne spatio-temporal resolution can
improve simulation results and reduce the number of
parameters or x their values (Krebs et al., 2014;

Verbeiren et al., 2013; Girons et al., 2010; Zhou et al.,
2010; Lee and Heaney, 2003; Bellal et al., 1996).
(2) Measurable, physically meaningful parameters: Many of the
modelling approaches are physically-based. Corresponding
publications state that such models should perform better
(Te Linde et al., 2010), are more numerically robust
(Fewtrell et al., 2011), have identiable parameters with
optimal values (Vieux and Bedient, 2004), and their parameters can provide a more detailed and potentially more correct description of the hydrological processes (Jia et al.,
2001). Physically measurable parameters (a) can increase
the understanding of typical urban processes, such as
BMPs (Xiao et al., 2007); (b) can support manual calibration
techniques as each model parameter is sensitive to a particular observable process (Morris et al., 2007); and (c) have
one major advantage: physical limits on their values
(Ogden et al., 2011). Moreover, physical parameters can be
retrieved from literature or eld measurements (Rodriguez
et al., 2008), eld investigations for example can be used
for identifying hydraulic properties, including connectivity
of impervious surfaces (Lee and Heaney, 2003) so that
parameters in fact become data.
(3) Calibration/validation: Standard techniques for calibration/validation of integrated and spatially-distributed urban
models are minimization/maximization of a single objective
function, which is a simple or complex form of the residuals
between observed and measured river discharge at the
catchment outlet (Verbeiren et al., 2013; Berezowski et al.,
2012; Hamdi et al., 2010; Meja and Moglen, 2010; Zhou
et al., 2010; Franczyk and Chang, 2009; Im et al., 2009;
Cuo et al., 2008; Lhomme et al., 2004; Mitchell et al., 2003;
Valeo and Moin, 2000). Alternative methods are also investigated: (a) different calibration algorithms for the same study
(Rodriguez et al., 2000); (b) independent quantity and quality calibrations (Petrucci and Bonhomme, 2014; Kleidorfer
et al., 2009) including cross-validation over stream ow
and waste water discharge (Singh et al., 2009); (c) sequential
calibration, i.e. rst water supply, then stormwater and
nally wastewater (Mackay and Last, 2010); (d) calibration
and or validation on different observations or different
statistics of the same observations. Examples belonging to
this last category are: various ow statistics (Hamel and
Fletcher, 2013); soil moisture (Ogden et al., 2011; Xiao
et al., 2007), soil moisture and surface runoff (Easton et al.,
2007), surface runoff and groundwater levels (Gbel et al.,
2004; Rodriguez et al., 2008), surface runoff, groundwater
levels, net radiation and surface temperature (Jia et al.,
2001); soil water pressure head in the upper subsoil,
groundwater drainage runoff rate, total runoff rate and volume (Berthier et al., 2004); surface water depth for ood
modelling (Chen et al., 2009); spatially distributed water
levels and runoff (Fewtrell et al., 2011); indirect calibration
for leakage and measured effective contribution of precipitation from green spaces (Morris et al., 2007).
4. Blueprint for urban hydrological modelling

The main objectives of this paper were: (1) the identication of
urban hydrological modelling practice; and (2) the denition of a
blueprint for the future development of urban-catchment hydrological modelling. Section 4.1 summarizes the main conclusions
on the practice of urban hydrological modelling on which basis
we list a set of desirable model characteristics for future progress
in urban catchment hydrology in Section 4.2.
4.1. Urban modelling practice

Urban-catchment hydrology is strongly gaining in importance
due to the need to better manage water with increasing urban population and area. Technological advances have boosted the development of hydrological models for simulating and understanding
physical processes in the urban environment. Although the modelling panorama offers ample choice, no standard methodology
has yet been identied. The complexity and thus the incomplete
knowledge of the urban hydrological system is the major cause
of the lack of standardization. Hydrological processes in urban
environments, as we described in Section 2, differ from those in
natural catchments because urbanization modies the physical
environment and hence the water quantity, quality and its dynamics. Natural hydrological processes, such as inltration, overland
ow, etc. are altered and new processes are introduced, like the
articial stormwater drainage ow. These processes strongly interact with each other and are characterized by specic spatial and
temporal scales. These complex spatio-temporal interactions make
urban hydrology, by denition, a multidisciplinary subject.
Modelling practice however tends to neglect certain processes
and their interaction, i.e. stormwater drainage and ow through
sewer networks. The property that above all characterizes urban
environments is the high spatial variability of the land (sub)surface, a variability which is difcult to take into account with the
currently scarce data, hence it introduces high uncertainty in modelling approaches. Moreover, urban catchments have their own
dynamics as urban population and urban land cover change over
time. How can hydrological models take into account all these
characteristics and accurately estimate water uxes in urban
catchments? We have selected and analyzed 43 urban modelling
approaches to assess the current practice and shortcomings, our
ndings are summarized as follow:
(1) Spatial and temporal scales of physical processes do not usually correspond to spatial and temporal discretization of
hydrological models. This modelling practice deals with high
uncertainty and creates conceptual problems in the reliability of the simulation results. As stated by Kleme (1983):
Can we expect a greater similarity in behaviour between
a 5 km2 basin of a creek and the Orinoco Basin
[880,000 km2] than between a kitchen sink and the Aswan
Dam? Our approaches often suggest that we believe so.
Therefore we cannot expect to accurately describe ash
oods occurring in small urban catchments with a temporal
scale of 1 day and a spatial resolution of 50 m. Current urban
hydrological models recognize the need for ner spatial and
temporal resolution, when compared to traditional hydrological models. However the average urban hydrological
simulations have a spatial resolution of 30 m and a time step
of 1 h, which greatly limits the applicability elds. Moreover,
many applications still prefer lumped or semi-distributed
approaches, and when distributed approaches are used they
mainly serve to derive punctual outputs, i.e. river discharge
at the catchment outlet, and none or little analysis is done on
spatially distributed outputs.
(2) In the past, the main reasons for not simulating physical processes at their characteristic scales were the incomplete
knowledge of the natural phenomena, limited computational power and lack of consistent data. While the rst
two points were partially solved in the last decades, limited
data still represent an important issue for todays models.
Measurement techniques have advanced but not as fast as
numerical models require. The main data advancement
relate to remote sensing techniques, which can provide spatial and temporal coverage of precipitation, soil moisture,
77
evapotranspiration, land cover, imperviousness, ood zones,

vegetation indices, etc. However, only a small minority of
the analyzed approaches make use of RS products.
(3) Geographical Information Systems are nowadays commonly
used to support hydrological models in data management
and visualization, and in some cases GIS are an integral part
of the simulation tool or used as a platform for model coupling. Spatially distributed modelling approaches fully
exploit the power of GIS, while lumped or semi-distributed
models only marginally benet. The availability of urban
GIS databases promoted the development of physical
object-based approaches. Detailed urban features allow the
identication of Urban Hydrological Elements that are used
as physically-meaningful spatial units for calculations.
(4) Sewers, drinking water networks, reservoirs, etc. are only
partially incorporated into modelling practice. Standard
hydrological tools are used to analyse urban catchments
with none or limited modications. The most commonly recognized difference between urban and non-urban catchments relates to sealed surfaces versus vegetated areas.
The impact of soil sealing, reduced inltration and increased
surface runoff, is the most commonly modelled urban
aspect. To do so, models use urban land cover classications
and imperviousness maps. A group of models address the
problem of urban water resource management, incorporating in the calculations the articial water networks and possible interactions. Models of this group however are
conceptual and empirically-based and are generally used
for small-sized urban applications. Well-established urban
hydraulic (hydrodynamic) models are available for rigorous
and detailed calculations of pipe ow and ood risk assessment. These models are also used as hydrological tools from
small-sized to catchment scale applications, but the natural
hydrological processes are generally more simplied.
Purpose-specic models aim at simulating particular processes of the urban environment, such as evapotranspiration
and sub-surface urban processes. Only a small minority of
models belong to this category, and these models have a less
consistent level of process detail. Recently, few attempts
have been made to combine several purpose-specic models
into more general urban hydrological models.
4.2. Future directions in urban modelling
How do we expect to overcome current model limitations?
Based on the analyses of the current literature and extrapolating
towards the future we list some model characteristics that could
represent a step forward in urban-catchment hydrology together
with future technological advances and research needs (Fig. 10).
FLEXIBILITY in terms of spatio-temporal discretization, process
parameterization and input requirements is perhaps the most
desirable model characteristic. A exible model structure allows
the selection of physical processes relevant for the particular case
study and the adaptation to the available data. This characteristic is
of importance for urban hydrological models as urban catchments
are extremely diverse and a xed, t-for-all model structure is not
appropriate. Model structure selection should be based upon
objective criteria, and hence methods for assessment of MODEL
STRUCTURE UNCERTAINTY should be developed. Flexible models
can also be easily adapted to take advantage of data from new
measurement techniques.
Flexibility can be achieved with a MODULAR approach, when
process-based independent hydrological components are identied. Modules should be able to interact and can be INTEGRATED
into one comprehensive model to describe the urban hydrological
system. Conceptually, this approach is similar to the real system
78
Shortcomings
Possible Soluons/Approaches
Urban Catchments:
Complex physical systems
Theorecal processes interacons
Incomplete knowledge
Data acquision Remote Sensing
Mulple processes interact

at various spaal and temporal scales
High spaal variability
Hydrological Models of Urban Catchments:
Spao-temporal gaps
between physical scales
and models resoluon
Limited data
Limited number of physical
processes described
Inconsistent level of details
GIS, spaal explicit distributed,

urban object-oriented
Flexibility
enhanced coupling opportunies
Modular,
process-based structure
Physically-based equaons
New spaally-distributed
validaon & calibraon techniques
Measurable parameters
High complexity High uncertainty

Fig. 10. The conclusions of this paper are summarized in a graphical way, by connecting shortcomings of urban catchment modelling with possible identied solutions.
where hydrological processes interact and so all urban water

problems, whether runoff quantity or quality, or water supply
and wastewater treatment, can no longer be evaluated in isolation
but will have to be looked at in an integrated way (Delleur, 2003).
Urban hydrological models could either be developed in a modular
way or they might take advantage of existing and well-established
specic-purpose tools. There is in fact no hydrological process in
the urban environment that is not, at least partially, covered by a
modelling approach, however a more coherent and organic
approach is still missing. COUPLING FRAMEWORKS might support
the integration by providing standardization and utilities for dening interconnection rules (Leniak et al., 2013; David et al., 2013).
The level of detail in process representation should be in function of the PHYSICAL hydrological characteristics. Under the
assumption that our knowledge of the laws of nature is sufciently
deep, physically-based models are the most suited approximation
of reality. Since computation capabilities are hardly a limitation,
more rigorous physically-based formulations can be solved for
large areas. Model parameters should be kept to a necessary minimum to avoid increasing uncertainty and should refer to measurable physical properties. Future urban hydrological models can no
longer neglect (1) evapotranspiration, which was proven to be a
signicant component in urban areas and new techniques for ET
measurements in heterogeneous urban environments need further
investigation; (2) sub-surface processes in the (un)saturated zone
as the physical behaviour of water in the complex urban subsoil
considerably distances itself from that of natural environments;
and (3) pipes inltration and exltration since sewer leakage is a
major source of groundwater contamination and leakage from
mains a major source of groundwater recharge.
PROCESS INTERACTIONS are the weakest theoretical aspects in
the physical description of the urban hydrological system.
Examples are surface and sub-surface interactions, sewer-storm
drainage system inltration and exltration. More research is
therefore needed to correctly parameterize these process interactions, new experiments should be conducted and results should
be incorporated into modelling practice.
The fast dynamics of urban-catchment processes requires a
high temporal resolution to correctly simulate hydrological processes and their interactions. The high TEMPORAL VARIABILITY
poses the important question whether conducting long-term simulations is feasible or only applicable to single event or design
storms. High temporal resolution, using for instance time interval

disaggregation, can increase numerical accuracy but also computational burden (Xiao et al., 2007). Current practice solves this shortcoming by nding compromises between detail and computational
efciency (De Vleeschauwer et al., 2014; Wolfs et al., 2013;
Willems, 2010). However, for most applications, computational
costs are not the bottle neck of urban hydrological models, current
technological development will support the research needs for
high resolution and fast computation. Cloud computing and clusters are slowly making their way into modelling practice and will
continue to do so in the future.
SPATIAL VARIABILITY is a key element in urban catchment
dynamics. Models should take into account the process spatial
variation and select a coherent resolution for the simulations.
Urban catchments are characterized by a highly heterogeneous
land cover, a complex topography which affects the physical processes. Lumped approaches are not capable of describing
spatially-variable processes, therefore the future of urban hydrological models will be employing spatially distributed urban
hydrological elements at a high spatial resolution. Promising
methodologies use object-oriented concepts to describe
hydrological-meaningful urban elements.
Spatial variability description can be supported by spatial measurements. REMOTE SENSING is the major source of spatial information for hydrological models, this fact is well recognized in
theory but in the current practice RS data are not extensively used.
We expect that the use of RS derived parameters will increase in
the future as such data are getting more and more available with
increasing accuracy. Urban hydrological models should incorporate both standard remote sensing data, like precipitation, Leaf
Area Index (LAI), as well as urban-specic measurements, i.e.
imperviousness, land cover, roughness, ooding zones. Hydrology
and remote sensing should work in synergy with each other on
new methods and techniques to measure and evaluate data to support hydrological model development.
A better use of remote sensing data might help in reducing
uncertainty. Spatially-distributed models should be veried with
spatially-distributed measurements. NEW CALIBRATION and
VALIDATION TECHNIQUES are therefore needed beside traditional
ones. The validation of model performance has well-developed
methods, but generally focussed on curve tting in singular locations. New spatially-distributed indices should be developed and
used, using both statistical measures and pattern comparisons, as

e.g. for the disaggregation of spatial rainfall (Gires et al., 2014;
Perica and Foufoula-Georgiou, 1996).
Finally, uncertainty reduction is the key topic when it comes to
developing complex integrated models (Bach et al., 2014). The aim
of new development is progress, in terms of understanding and of
functionalities, and it is not the creation of integronsters (Voinov
and Shugart, 2013). A sustainable urban hydrological model
development is required, where the modelling tools and measurement techniques are in balance, such that model complexity can be
justied by measurement advances, and uncertainty can be properly handled (Niemczynowicz, 1999; Beven, 2008). Efforts in the
eld of soil water measurements has high priority, as only few
research papers treat this topic and better observations of soil
moisture, groundwater levels and base ow in sewers could be
helpful for a better understanding of the urban water system.
Acknowledgments
This research was funded by a FWO (Research Foundation
Flanders) and VITO (Flemish Institute for Technological Research)
PhD grant. We wish to thank the three anonymous reviewers, their
inspiring comments substantially improved this manuscript. We
would also like to thank Mrs. Tuy To for her review of the
manuscript.
References
Annunziato, S., Stocchino, A., M., Colombini, 2009. Case study: design of ood
control systems on the vara river by numerical and physical modeling. J.
Hydraul. Eng. ASCE 135 (12), 10631072.
Appleyard, S., 1995. The impact of urban development on recharge and
groundwater quality in a coastal aquifer near Perth, Western Australia.
Hydrogeol. J. 3 (2), 6575.
Arneld, A.J., 2003. Two decades of urban climate research: a review of turbulence,
exchanges of energy and water, and the urban heat island. Int. J. Climatol. 23
(1), 126.
Arnold, J.G., Fohrer, N., 2005. SWAT2000: current capabilities and research
opportunities in applied watershed modeling. Hydrol. Process. 19 (3), 563572.
Arnold, C.L., Gibbons, C.J., 1996. Impervious surface coverage the emergence of a
key environmental indicator. J. Am. Plan. Assoc. 62 (2), 243258.
Arnold, J.G., Srinivasan, R., Muttiah, R.S., Williams, J.R., 1998. Large area hydrologic
modeling and assessment Part 1: Model development. J. Am. Water Resour.
Assoc. 34 (1), 7389.
Aronica, G., Cannarozzo, M., 2000. Studying the hydrological response of urban
catchments using a semi-distributed linear non-linear model. J. Hydrol. 238 (1
2), 3543.
Aronica, G.T., Lanza, L.G., 2005. Drainage efciency in urban areas: a case study.
Hydrol. Process. 19 (5), 11051119.
Bach, P.M., Rauch, W., Mikkelsen, P.S., McCarty, D.T., Deletic, A., 2014. A critical
review of integrated urban water modelling urban drainage and beyond.
Environ. Modell. Softw. 54, 88107.
Barbee, R., Vermij, P., van de Laak, W.J., 2002. Policy development for reduction of
pollution caused by trafc experiences from the Netherlands. Water Sci.
Technol. 49 (3), 183188.
Beck, M., 2005. Vulnerability of water quality in intensively developing urban
watersheds. Environ. Modell. Softw. 20 (4), 381400.
Bellal, M., Sillen, X., Zech, Y., 1996. Coupling GIS with a Distributed Hydrological
Model for Studying the Effect of Various Urban Planning Options on Rainfallrunoff Relationship in Urbanized watersheds, vol. 235. IAHS-AISH Publication
(pp. 99106).
Berezowski, T., Chormanski, J., Batelaan, O., Canters, F., Van de Voorde, T., 2012.
Impact of remotely sensed land-cover proportions on urban runoff production.
Int. J. Appl. Earth Obs. Geoinf. 16, 5465.
Berndtsson, R., Niemczynowicz, J., 1988. Spatial and temporal scales in rainfall
analysis some aspects and future perspectives. J. Hydrol. 100 (13), 293313.
Berne, A., Delrieu, G., Creutin, J.D., Obled, C., 2004. Temporal and spatial resolution
of rainfall measurements required for urban hydrology. J. Hydrol. 299 (34),
166179.
Berthier, E., Andrieu, H., Creutin, J.D., 2004. The role of soil in the generation of
urban runoff: development and evaluation of a 2D model. J. Hydrol. 299 (34),
252266.
Berthier, E., Dupont, S., Mestayer, P.G., Andrieu, H., 2006. Comparison of two
evapotranspiration schemes on a sub-urban site. J. Hydrol. 328 (34), 635646.
Beven, K.J., 2008. RainfallRunoff Modelling. John Wiley & Sons Ltd., Chichester,
England (February 2008 received).
79
Blschl, G., Sivapalan, M., 1995. Scale issues in hydrological modelling: a review.
Hydrol. Process. 9 (34), 251290.
Boyd, M.J., Bull, M.C., Knee, R.M., 1993. Pervious and impervious runoff in urban
catchments. Hydrol. Sci. J. 38 (6), 463478.
Brabec, E., Schulte, S., Richards, P.L., 2002. Impervious surface and water quality: a
review of current literature and its implications for watershed planning. J. Plan.
Lit. 16 (4), 499514.
Braud, I., Fletcher, T.D., Andrieu, H., 2013. Hydrology of per-urban catchments:
process and modelling. J. Hydrol. 485, 14.
Bronstert, A., Carrera, J., Kabat, P., Ltkemeier, S., 2005. Coupled models for the
Hydrological Cycle. Integrating Atmosphere, Biosphere, and Pedosphere.
Springer.
Brun, S.E., Band, L.E., 2000. Simulating runoff behavior in an urbanizing watershed.
Comput. Environ. Urban Syst. 24 (1), 522.
Brutsaert, W. (Ed.), 2005. Hydrology an Introduction, rst ed. Cambridge University
Press, Cambridge, United Kingdom.
Burian, S.J., Edwards, F.G., 2002. Historical Perspectives of Urban Drainage. Portland,
Oregon, sn.
Burns, D., Vitvar, T., McDonnell, J., Hassett, J., Duncan, J., Kendall, C., 2005. Effects of
suburban development on runoff generation in the Croton River basin, New
York, USA. J. Hydrol. 311 (14), 266281.
Carey, R.O., Hochmuth, G.J., Martinez, D.J., Boyer, T.H., Dukes, M.D., Toor, G.S., Cisar,
J.L., 2013. Evaluating nutrient impacts in urban watersheds: challenges and
research opportunities. Environ. Pollut. 173, 138149.
Carle, M.V., Halpin, P.N., Stow, C.A., 2005. Patterns of watershed urbanization and
impacts on water quality. J. Am. Water Resour. Assoc. 41 (3), 693708.
CEC, 2007. Directive Concerning EU Floods. 2007/60/EEC. CEC, Luxemburg.
CEN, 1996. Drain and Sewers Systems Outside Buildings Part 2: Performance
Requirements, European Standards, European Committee for Standardization
CEN, Brussels, Belgium.
CEN, 1997. Drain and Sewer Systems Outside Buildings Part 4: Hydraulic Design
and Environmental Considerations, European Standard, European Committee
for Standardization CEN, Brussels, Belgium.
Changming, L., Jingjie, Y., Kendy, E., 2001. Groundwater exploitation and its impact
on the environment in the north China plain. Water Int. 26 (2), 265272.
Chen, J., Arleen, A.H., Lensyl, D.U., 2009. A GIS-based model for ood inundation. J.
Hydrol. 373 (12), 184192.
Cheng, L., Zongxue, X., Wang, D., Cai, X., 2011. Assessing interannual variability of
evapotranspiration
at
the
catchment
scale
using
satellite-based
evapotranspiration data sets. Water Resour. Res. 47 (9), W09509.
Chormanski, J., Van de Voorde, T., De Roeck, T., Batelaan, O., Canters, F., 2008.
Improving distributed runoff prediction in urbanized catchments with remote
sensing based estimates of impervious surface cover. Sensors 8 (2), 910932.
Cronin, A.A., Taylor, R.G., Powell, K.L., Barrett, M.H., Trowsdale, S.A., Lerner, D.N.,
2003. Temporal variations in the depth-specic hydrochemistry and sewagerelated microbiology of an urban sandstone aquifer, Nottingham, United
Kingdom. Hydrogeol. J. 11 (2), 205216.
Cuo, L., Lettenmaier, D.P., Mattheussen, B.V., Storck, P., Wiley, M., 2008. Hydrologic
prediction for urban watersheds with the distributed hydrology soil
vegetation model. Hydrol. Process. 22 (21), 42054213.
Dabberdt, W.F., Hales, J., Zubrick, S., Crook, A., Krajewski, W., Doran, K.C., Mueller, C.,
King, C., Keener, R.N., Bornstein, R., Rodenhuis, D., Kocin, P., Rossetti, M.A.,
Sharrocks, F., Stanley, E.M., 2000. Forecast issues in the urban zone: report of
the 10th prospectus development team of the U.S. weather research program.
Bull. Am. Meteorol. Soc. 81 (9), 20472064.
Danso-Amoako, E., Scholz, M., Kalimeris, N., Yang, Q., Shao, J., 2012. Predicting dam
failure risk for sustainable ood retention basins: a generic case study for the
wider Great Manchester area. Comput. Environ. Urban Syst. 36 (5), 423433.
David, O., Ascough II, J.C., Lloyd, W., Green, T.R., Rojas, K.W., Leavesley, G.H., Ahuja,
L.R., 2013. A software engineering perspective on environmental modeling
framework design: the object modeling system. Environ. Modell. Softw. 39,
201213.
De Vleeschauwer, K., Weustenraad, J., Nolf, C., Wolfs, V., De Meulder, B., Shannon, K.,
Willems, P., 2014. Greenblue water in the city: quantication of impact of
source control versus end-of-pipe solutions on sewer and river oods. Water
Sci. Technol. 70 (11), 18251837.
DeFries, R., Eshleman, N.K., 2004. Land-use change and hydrologic processes: a
major focus of the future. Hydrol. Process. 18 (11), 21832186.
Delleur, J.W., 2003. The evolution of urban hydrology: past, present, future. J.
Hydraul. Eng. ASCE 129 (8), 563573.
Dingman, S.L., 2002. Physical Hydrology, second ed. Prentice Hall, Inc., Upper
Saddle River, New Jersey, USA.
Dirckx, G., Bixio, D., Thoeye, C., De Gueldre, G., Van De Steene, B., 2009. Dilution of
sewage in Flanders mapped with mathematical and tracer methods. Urban
Water J. 6 (2), 8192.
Dixon, B., Earls, J., 2012. Effects of urbanization on streamow using SWAT with real
and simulated meteorological data. Appl. Geogr. 35 (12), 174190.
Dixon, P.G., Mote, T.L., 2003. Patterns and causes of Atlantas urban heat islandinitiated precipitation. J. Appl. Meteorol. 42 (9), 12731284.
Dotto, C.B.S., Kleidorfer, L., Deletic, A., Rauch, W., McCarthy, D.T., Fletcher, T.D.,
2011. Performance and sensitivity of stormwater models using a Bayesian
approach and long-term high resolution data. Environ. Modell. Softw. 26 (10),
12251239.
Dujardin, J., Batelaan, O., Canters, F., Boel, S., Anibas, C., Bronders, J., 2011. Improving
surface-subsurface water budgeting using high resolution satellite imagery
applied on a browneld. Sci. Total Environ. 409 (4), 800809.
80
Easton, Z.M., Grard-Merchant, P., Walter, M.T., Petrovic, A.M., Steenhuis, T.S., 2007.
Hydrologic assessment of an urban variable source watershed in the northeast
United States. Water Resour. Res. 43 (3), W03413.
Egodawatta, P., Ziyath, A., Goonetilleke, A., 2013. Characterizing metal build-up on
urban road surfaces. Environ. Pollut. 176, 8791.
Einfalt, T., Arnbjerg-Nielsen, K., Golz, C., Jensen, N.-E., Quirmbach, M., Vaes, G.,
Vieux, B., 2004. Towards a roadmap for use of radar rainfall data in urban
drainage. J. Hydrol. 299 (34), 186202.
Eiswirth, M., Wolf, L., Htzl, H., 2003. Balancing the contaminant input into urban
water resources. Environ. Geol. 46 (2), 246256.
European Parliament, Directive 2000/60/EC of the European Parliament and of the
Council of 23 October 2000 Establishing a Framework for Community Action in
the Field of Water Policy.
Fabry, F., Bellon, A., Ducan, M., Austin, G., 1994. High resolution rainfall
measurements by radar for very small basins: the sampling problem
reexamined. J. Hydrol. 161 (14), 415428.
Fassman-Beck, E., Voyde, E., Simcock, R., Hong, Y.S., 2013. 4 Living roofs in 3
locations: does conguration affect runoff mitigation? J. Hydrol. 490, 1120.
Faulkner, D.S., Barber, S., 2009. Performance of the revitalised ood hydrograph
method. J. Flood Risk Manage. 2 (4), 254261.
Fewtrell, T.J., Neal, J.C., Bates, P.D., Harrison, P.J., 2011. Geometric and structural
river channel complexity and the prediction of urban inundation. Hydrol.
Process. 25 (20), 31733186.
Fletcher, T.D., Andrieu, H., Hamel, P., 2013. Understanding, management and
modelling of urban hydrology and its consequences for receiving waters: a state
of the art. Adv. Water Resour. 51, 261279.
Foster, S.S.D., 1990. Impacts of Urbanization on Groundwater. Hydrological
Processes and Water Management in Urban Areas, vol. 198. IAHS Publication
(pp. 187207).
Foster, S.S.D., Chilton, P.J., 2003. Groundwater: the processes and global signicance
of aquifer degradation. Philos. Trans. Roy. Soc. London Ser. B Biol. 358 (1440),
19571972.
Franczyk, J., Chang, H., 2009. The effect of climate change and urbanization on the
runoff of the Rock Creek basin in the Portland metropolitan area, Oregon, USA.
Hydrol. Process. 23 (6), 805815.
Garcia-Fresca, B., Sharp, J.M., 2005. Hydrogeological considerations of urban
development: urban-induced recharge. Rev. Eng. Geol. 16, 123136.
Gires, A., Tchiguirinskaia, I., Schertzer, D., Schellart, A., Berne, A., Lovejoy, S., 2014.
Inuence of small rainfall variability on standard comparison tools between
radar and rain gauge data. Atmos. Res. 138, 125138.
Girons, J., Niemann, J.D., Roesner, L.A., Rodriguez, F., Andrieu, H., 2009. A morphoclimatic instantaneous unit hydrograph model for urban catchments based on
the kinematic wave approximation. J. Hydrol. 377 (34), 317334.
Girons, J., Niemann, J.D., Roesner, L.A., Rodriguez, F., Andrieu, H., 2010. Evaluation
of methods for representing urban terrain in storm-water modeling. J. Hydrol.
Eng. 15 (1), 114.
Gbel, P., Stubbe, H., Weinert, M., Zimmermann, J., Fach, S., Dierkes, C., Kories, H.,
Messer, J., Mertsch, V., Geiger, W.F., Coldewey, W.G., 2004. Near-natural
stormwater management and its effects on the water budget and
groundwater surface in urban areas taking account of the hydrogeological
conditions. J. Hydrol. 299 (34), 267283.
Greene, R.G., Cruise, J.F., 1995. Urban watershed modeling using geographic
information-system. J. Water Res. Plan. ASCE 121 (4), 318325.
Grimmond, C.S.B., Oke, T.R., 1991. An evapotranspiration-interception model for
urban areas. Water Resour. Res. 27 (7), 17391755.
Hall, M.J., 1984. Urban Hydrology. Elsevier Applied Science Publishers.
Hamdi, R., Termonia, P., Baguis, P., 2010. Effects of urbanization and climate change
on surface runoff of the Brussels Capital Region: a case study using and urban
soil-vegetation-atmosphere-transfer model. Int. J. Climatol. 31 (13), 1959
1974.
Hamel, P., Fletcher, T.D., 2013. Modelling the impact of stormwater source control
inltration techniques on catchment baseow. Hydrol. Process. 28 (24), 5817
5831.
Hancock, P.J., 2002. Human impacts on the stream-groundwater exchange zone.
Environ. Manage. 29 (6), 763781.
Hendriks, M.R., 2010. Introduction to Physical Hydrology. Oxford University Press.
Hsu, M.H., Chen, S.H., Chang, T.J., 2000. Inundation simulation for urban drainage
basin with storm sewer system. J. Hydrol. 234 (12), 2137.
Huang, S., Cheng, S., Wen, J., Lee, J., 2008. Identifying peak-imperviousnessrecurrence relationships on a growing-impervious watershed, Taiwan. J. Hydrol.
362 (34), 320336.
Huber, W.C., Dickinson, R.E., 1988. Storm Water Management Model. Users
Manual. Ver. IV, sl: US EPA Environmental Protection Agency.
Huff, F.A., Changnon, S.A.J., 1972. Climatological assessment of urban effects on
precipitation at St. Louis. J. Appl. Meteorol. 11 (5), 823842.
Hurford, A.P., Maksimovic, C., Leitao, J.P., 2010. Urban pluvial ooding in Jakarta:
applying state-of-the-art technology in a data scarce environment. Water Sci.
Technol. 62 (10), 22462255.
Im, S., Kim, L., Kim, C., Jang, C., 2009. Assessing the impacts of land use changes on
watershed hydrology using MIKE SHE. Environ. Geol. 57 (1), 231239.
Jacobson, C.R., 2011. Identication and quantication of the hydrological impacts of
imperviousness in urban catchments: a review. J. Environ. Manage. 92 (6),
14381448.
Jia, Y., Ni, G., Kawahara, Y., Suetsugi, T., 2001. Development of WEP model and its
application to an urban watershed. Hydrol. Process. 15 (11), 21752194.
Jonkman, S.N., Vrijling, J.K., 2008. Loss of life due to oods. J. Flood Risk Manage. 1
(1), 4356.
Joss, A., Siegrist, H., Ternes, T.A., 2008. Are we about to upgrade wastewater
treatment for removing organic micropollutants? Water Sci. Technol. 57 (2),
251255.
Kang, I.S., Park, J.I., Singh, V.P., 1998. Effect of urbanization on runoff characteristics
of the On-Cheon stream watershed in Pusan, Korea. Hydrol. Process. 12 (2),
351363.
Kim, Y.-Y., Lee, K.-K., Sung, I.H., 2001. Urbanization and the groundwater budget,
metropolitan Seoul area, Korea. Hydrogeol. J. 9 (4), 401412.
Kjeldsen, T.R., Stewart, E.J., Packman, J.C., Folwell, S.S., Bayliss, A.C., 2005.
Revitalisation of the FSR/FEH rainfall-runoff method., CEH Wallingford, UK:
Defra R&D Technical Report FD1913/TR.
Kleidorfer, M., Deletic, A., Fletcher, T.D., Rauch, W., 2009. Impact of input data
uncertainties on urban stormwater model parameters. Water Sci. Technol. 60
(6), 15451554.
Kleme, V., 1983. Conceptualization and scale in hydrology. J. Hydrol. 65 (13), 1
23.
Krebs, G., Kokkonen, T., Setl, H., Koivusalo, H., 2014. Spatial resolution
considerations for urban hydrological modelling. J. Hydrol. 512, 482497.
Kron, W., Steuer, M., Lw, P., Wirtz, A., 2012. How to deal properly with a natural
catastrophe database analysis of ood losses. Nat. Hazards Earth Sys. 12 (3),
535550.
Labadie, J.W., Wan, Y., 2010. Fuzzy optimal control of reservoir-assisted stormwater
treatment areas for aquatic ecosystem restoration. Environ. Modell. Softw. 25,
16921701.
Lee, J.G., Heaney, J.P., 2003. Estimation of urban imperviousness and its impacts on
storm water systems. J. Water Res. Plan. ASCE 129 (5), 419426.
Leniak, G.F., Olchin, G., Goodall, J., Voinov, A., Hill, M., Glynn, P., Whelan, G., Geller,
G., Quinn, N., Blind, M., Peckham, S., Reaney, S., Gaber, N., Kennedy, R., Hughes,
A., 2013. Integrated environmental modeling: a vision and roadmap for the
future. Environ. Modell. Softw. 39, 323.
Leopold, L.B., 1968. Hydrology for Urban Land Planning A guidebook on the
Hydrologic Effects of urban Land Use. U.S. Government Printing Ofce,
Washington.
Lerner, D.N., 1990. Groundwater recharge in urban areas. Atmos. Environ. Part B
Urban Atmos. 24 (1), 2933.
Lerner, D.N., 2002. Identifying and quantifying urban recharge: a review. Hydrogeol.
J. 10 (1), 143152.
Lhomme, J., Bouvier, C., Perrin, J.-L., 2004. Applying a GIS-based geomorphological
routing model in urban catchments. J. Hydrol. 299 (34), 203216.
Mackay, R., Last, E., 2010. SWITCH city water balance: a scoping model for
integrated urban water management. Rev. Environ. Sci. Biotechnol. 9 (4), 291
296.
Maidment, D.R., 1992. Handbook of Hydrology. Mc-Graw-Hill Inc., New York, USA.
Mark, O., Weesakul, S., Apirumanekul, C., Aroonnet, S.B., Djordjevic, S., 2004.
Potential and limitations of 1D modelling of urban ooding. J. Hydrol. 299 (34),
284299.
Marsalek, J., Jimnez-Cisneros, B., Karamouz, M., Malmquist, P.-A., Goldenfum, J.,
Chocat, B., 2007. Urban Water Cycle Processes and Interactions: Urban Water
Series UNESCO-IHP. Springer, Berlin.
McColl, C., Aggett, G., 2007. Land-use forecasting and hydrologic model integration
for improved land-use decision support. J. Environ. Manage. 84 (4), 494512.
Meja, A.I., Moglen, G.E., 2010. Impact of the spatial distribution of imperviousness
on the hydrologic response of an urbanizing basin. Hydrol. Process. 24 (23),
33593373.
Meyer, S.C., 2005. Analysis of base ow trends in urban streams, northeastern
Illinois, USA. Hydrogeol. J. 13 (56), 871885.
Mitchell, V.G., Diaper, C., 2005. UVQ: a tool for assessing the water and
contamination balance impacts of urban development scenarios. Water Sci.
Technol. 52 (12), 9198.
Mitchell, V.G., McMahon, T.A., Mein, R.G., 2003. Components of the total water
balance of an urban catchment. Environ. Manage. 32 (6), 735746.
Mitchell, V.G., Duncan, H., Inman, M., Rahilly, M., Stewart, J., Vieritz, A., Holt, P.,
Grant, A., Fletcher, T.D., Coleman, J., Maheepala, S., Sharma, A., Deletic, A., Breen,
P., 2007. State of the Art Review of Integrated Urban Water Models. Lyon,
France, sn, pp. 507514.
Mitchell, V.G., Cleugh, H.A., Grimmond, C.S.B., Xu, J., 2008. Linked urban water
balance and energy balance models to analyse urban design options. Hydrol.
Process. 22 (16), 28912900.
Moglen, G.E., Kim, S., 2007. Limiting imperviousness are threshold-based policies
a good idea? J. Am. Plan. Assoc. 73 (2), 161171.
Morris, B., Rueedi, J., Cronin, A.A., Diaper, C., DeSilva, D., 2007. Using linked process
models to improve urban groundwater management: an example from
Doncaster England. Water Environ. J. 21 (4), 229240.
Nagler, P., 2011. The role of remote sensing observations and models in hydrology:
the science of evapotranspiration. Hydrol. Process. 25 (26), 39773978.
Niemczynowicz, J., 1999. Urban hydrology and water management present and
future challenges. Urban Water 1 (1), 114.
Nouri, H., Beecham, S., Hassanli, A.M., Kazemi, F., 2013a. Water requirements of
urban landscape plants: a comparison of three factor-based approaches. Ecol.
Eng. 57, 276284.
Nouri, H., Beecham, S., Kazemi, F., Hassanli, A.M., 2013b. A review of ET
measurement techniques for estimating the water requirements of urban
landscape vegetation. Urban Water J. 10 (4), 247259.

Ogden, F.L., Pradhan, N.R., Downer, C.W., Zahner, J.A., 2011. Relative importance of
impervious area, drainage density, width function, and subsurface storm
drainage on ood runoff from an urbanized catchment. Water Resour. Res. 47
(12), W12503.
Oke, T.R., 1982. The energetic basis of urban heat island. Quart. J. Roy. Meteor. Soc.
108 (455), 124.
OLoughlin, G., Huber, W., Chocat, B., 1996. Rainfall-runoff processes and modelling.
J. Hydraul. Res. 34 (6), 733751.
Pan, A., Hou, A., Tian, F., Ni, G., Hu, H., 2012. Hydrologically enhanced distributed
urban drainage model and its application in Beijing city. J. Hydrol. Eng. 17 (6),
667678.
Paul, M.J., Meyer, J.L., 2001. Streams in the urban landscape. Annu. Rev. Ecol. Syst.
32, 333365.
Perica, S., Foufoula-Georgiou, E., 1996. Model for multiscale disaggregation of
spatial rainfall based on coupling meteorological and scaling descriptions. J.
Geophys. Res. 101 (D21), 2634726361.
Petrucci, G., Bonhomme, C., 2014. The dilemma of spatial representation for urban
hydrology semi-distributed modelling: trade-offs among complexity,
calibration and geographical data. J. Hydrol. 517, 9971007.
Praskievicz, S., Chang, H., 2009. A review of hydrological modelling of basin-scale
climate change and urban development impacts. Prog. Phys. Geog. 33 (5),
650671.
Price, R.K., Vojinovic, Z., 2011. Urban Hydroinformatics: Data, Models and Decision
support for Integrated Urban Water Management. IWA Publishing, London, UK.
Ragab, R., Rosier, P., Dixon, A., Bromley, J., Cooper, J.D., 2003. Experimental study of
water uxes in a residential area: 2. Road inltration, runoff and evaporation.
Hydrol. Process. 17 (12), 24232437.
Ramier, D., Berthier, E., Andrieu, H., 2011. The hydrological behaviour of urban
streets: long-term observations and modelling of runoff losses and rainfall
runoff transformation. Hydrol. Process. 25 (14), 21612178.
Rizwan, A.M., Dennis, Y.C., Liu, C., 2008. A review on the generation, determination
and mitigation of Urban Heat Island. J. Environ. Sci. 20 (1), 120128.
Rodriguez, F., Andrieu, H., Zech, Y., 2000. Evaluation of a distributed model for urban
catchments using a 7-years continuous data series. Hydrol. Process. 14 (5),
899914.
Rodriguez, F., Andrieu, H., Creutin, J.-D., 2003. Surface runoff in urban catchments:
morphological identication of unit hydrographs from urban databanks. J.
Hydrol. 283 (14), 146168.
Rodriguez, F., Andrieu, H., Morena, F., 2008. A distributed hydrological model for
urbanized areas model development and application case studies. J. Hydrol.
351 (34), 268287.
Roy, A.H., Shuster, W.D., 2009. Assessing impervious surface connectivity and
applications for watershed management. J. Am. Water Resour. Assoc. 45 (1),
198209.
Rueedi, J., Cronin, A.A., Morris, B.L., 2009. Estimation of sewer leakage to urban
groundwater using depth-specic hydrochemistry. Water Environ. J. 23 (2),
134144.
Rumman, N., Lin, G., Li, J., 2005. Investigation of GIS-based surface hydrological
modelling for identifying inltration zones in an urban watershed. Int. Soc.
Environ. Inform. Sci. 1 (5), 315322.
Rutsch, M., Rieckermann, J., Krebs, P., 2006. Quantication of sewer leakage: a
review. Water Sci. Technol. 54 (67), 135144.
Salvador, R., Bautista-Capetillo, C., Playn, E., 2011. Irrigation performance in
private urban landscapes: a case study in Zaragoza (Spain). Landscape Urban
Plan. 100 (3), 302311.
Schirmer, M., Leschik, S., Musolff, A., 2012. Current research in urban hydrogeology
a review. Adv. Water Resour. 51, 280291.
Schmitt, T.G., Thomas, M., Ettrich, N., 2004. Analysis and modeling of ooding in
urban drainage systems. J. Hydrol. 299 (34), 300311.
Selvalingam, S., Liong, P.C., Manoharan, P.C., 1987. Use of RORB and SWMM models
to an urban catchment in Singapore. Adv. Water Resour. 10 (2), 7886.
Shepherd, J.M., 2006. Evidence of urban-induced precipitation variability in arid
climate regimes. J. Arid Environ. 67 (4), 607628.
Shepherd, J.M., Pierce, H., Negri, A.J., 2002. Rainfall modication by major urban
areas: observations from spaceborne rain radar on the TRMM satellite. J. Appl.
Meteorol. 41 (7), 689701.
Sherlock, R.L., 1922. Man as a Geological Agent: An Account of his Action on
Inanimate Nature. H.F. & G. Witherby, London.
Shuster, W.D., Bonta, J., Thurston, H., Warnemuende, E., Smith, D.R., 2005. Impacts
of impervious surface on watershed hydrology: a review. Urban Water J. 2 (4),
263275.
Shuster, W.D., Gehring, R., Gerken, J., 2007. Prospects for enhanced groundwater
recharge via inltration of urban storm water runoff: a case study. J. Soil Water
Conserv. 62 (3), 129137.
81
Shuttleworth, W.J., 2007. Putting the vap into evaporation. Hydrol. Earth Syst. Sc.
11 (1), 210244.
Singh, R., Maheshwari, B., Malano, H.M., 2009. Developing a Conceptual Model for
Water Accounting in Peri-urban Catchments. Cairns, Australia, sn, pp. 3654
3660.
Stovin, V., 2010. The potential of green roofs to manage Urban Stormwater. Water
Environ. J. 24, 192199.
Suriya, S., Mudgal, B.V., 2011. Impact of urbanization on ooding: the Thirusoolam
sub watershed a case study. J. Hydrol. 412413, 210219.
Taha, H., 1997. Urban climates and heat islands: albedo, evapotranspiration, and
anthropogenic heat. Energy Build. 25 (2), 99103.
Taylor, R.G., Cronin, A.A., Lerner, D.N., Tellam, J.H., Bottrell, S.H., Rueedi, J., Barrett,
M.H., 2006. Hydrochemical evidence of the depth of penetration of
anthropogenic recharge in sandstone aquifers underlying two mature cities in
the UK. Appl. Geochem. 21 (9), 15701592.
Te Linde, A.H., Aerts, J.C.J.H., Kwadijk, J.C.J., 2010. Effectiveness of ood management
measures on peak discharges in the Rhine basin under climate change. J. Flood
Risk Manage. 3 (4), 248269.
Thielen, J., Creutin, J.-D., 1997. An urban hydrological model with high spatial
resolution rainfall from a meteorological model. J. Hydrol. 200 (14), 5883.
Tholin, A.L., Keifer, C.J., 1960. The hydrology of urban runoff. In: Trans. ASCE 125. sl:
sn, pp. 13081379.
Thorndahl, S., Willems, P., 2008. Probabilistic modelling of overow, surcharge and
ooding in urban drainage using the rst-order reliability method and
parameterization of local rain series. Water Res. 42 (12), 455466.
Tilford, K.A., Fox, N.I., Collier, C.G., 2002. Application of weather radar data for urban
hydrology. Meteorol. Appl. 9 (1), 95104.
Tsihrintzis, V.A., Hamid, R., 1998. Runoff quality prediction from small urban
catchments using SWMM. Hydrol. Process. 12 (2), 311329.
United Nations, 2010. World Population Prospects: The 2009 Revision, New York, sn.
Valeo, C., Moin, S.M.A., 2000. Grid-resolution effects on a model for integrating
urban and rural areas. Hydrol. Process. 14 (14), 25052525.
Van de Ven, F.H.M., 1990. Water Balances of Urban Areas. Hydrological Processes
and Water Management in Urban Areas, vol. 198. IAHS Publication (pp. 2132).
Verbeiren, B., Van De Voorde, T., Canters, F., Binard, M., Cornet, Y., Batelaan, O., 2013.
Assessing urbanisation effects on rainfallrunoff using a remote sensing
supported modelling strategy. Int. J. Appl. Earth Obs. Geoinf. 21, 92102.
Vieux, B.E., Bedient, P.B., 2004. Assessing urban hydrologic prediction accuracy
through event reconstruction. J. Hydrol. 299 (34), 217236.
Vizintin, G., Souvent, P., Veselic, M., Curk, C., 2009. Determination of urban
groundwater pollution in alluvial aquifer using linked process models
considering urban water cycle. J. Hydrol. 377 (34), 261273.
Voinov, A., Shugart, H., 2013. Integronsters, integral and integrated modelling.
Environ. Modell. Softw. 39, 149158.
Vrebos, D., Vansteenkiste, T., Staes, J., Willems, P., Meire, P., 2014. Water
displacement by sewer infrastructure in the Grote Nete catchment, Belgium,
and its hydrological regime effects. Hydrol. Earth Syst. Sci. 18 (3), 11191136.
Ward, R., Robinson, M., 1999. Principles of Hydrology, fourth ed. McGraw-Hill
Publishing Company, Berkshire, England.
Wei, G., Brombach, H., Haller, B., 2002. Inltration and inow in combined sewer
systems: long-term analysis. Water Sci. Technol. 45 (7), 1119.
Weng, Q., 2012. Remote sensing of impervious surfaces in the urban area:
requirements, methods, and trends. Remote Sens. Environ. 117, 3449.
Wiles, T.J., Sharp, J.M., 2008. The secondary permeability of impervious cover.
Environ. Eng. Geosci. 14 (4), 251265.
Willems, P., 2010. Parsimonious model for combined sewer overow pollution. J.
Environ. Eng. ASCE 136 (3), 316325.
Wolf, L., Klinger, J., Hoetzl, H., 2007. Quantifying mass uxes from urban drainage
systems to the urban soil-aquifer system. J. Soil. Sediment. 7 (2), 8595.
Wolfs, V., Villazon Gomez, M.F., Willems, P., 2013. Development of a semidistributed model identication and calibration tool for conceptual modeling of
sewer systems. Water Sci. Technol. 68 (1), 167175.
Wu, C., Murray, A.T., 2003. Estimating impervious surface distribution by spectral
mixture analysis. Remote Sens. Environ. 84, 493505.
Xiao, Q., McPherson, E.G., Simpson, J.R., Ustin, S.L., 2007. Hydrologic processes at the
urban residential scale. Hydrol. Process. 21 (16), 21742188.
Yang, Y., Lerner, D.N., Barrett, M.H., Tellam, J.H., 1999. Quantication of
groundwater recharge in the city of Nottingham, UK. Environ. Geol. 38 (3),
183198.
Zektser, I.S., Everett, L.G., 2004. Groundwater Resources of the World and their Use.
United Nations Educational, Scientic and Cultural Organization, Paris.
Zhou, Y., Wang, Y., Gold, A.J., August, P.V., 2010. Modeling watershed rainfall-runoff
relations using impervious surface-area data with high spatial resolution.
Hydrogeol. J. 18 (6), 14131423.

Review of Urban Catchment Hydrological Modelling

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

Review of Urban Catchment Hydrological Modelling

Încărcat de

Drepturi de autor:

Formate disponibile

Journal of Hydrology 529 (2015) 6281

Contents lists available at ScienceDirect

Hydrological modelling of urbanized catchments: A review and future

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

Urban hydrological modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

assessment and mapping (CEC, 2007). Reliable assessment of water

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

Urban hydrology at the Catchment Scale

2) What is an URBAN CATCHMENT?

5) What is the models SPATIO-TEMPORAL resoluon?

Zhou et al. (2010)

9) How do models deal with PARAMETER ESTIMATION

10) How can we bridge the SPATIO-TEMPORAL GAPS?

Fig. 1. This paper is structured in four sequential sections aiming at answering 11

modelling approaches obtained from literature. This paper does

the concept of an impervious cover threshold, as this assumption

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

D Runoff over sealed surfaces

G Waste water treatment discharge

H Depression storage and open water

106 LENGTH [m]

100 km 1000 km 10000 km

water cycle, although many texts agree in dividing the system in

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

evaporation of water stored in surface depressions, plant canopy,

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

into a closed pipe system. These modications limit the natural

this should reach 294 106 m3 for a design discharge of

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

Manmade changes of water courses, i.e. channelization,

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

Study area (km2)

(1) Urban residential model

Xiao et al. (2007)

Dotto et al. (2011) and Hamel and Fletcher, 2013

Kleidorfer et al. (2009) and Dotto et al. (2011)

(6) Conceptual model

Aronica and Cannarozzo (2000)

(8) SUES (F)

Grimmond and Oke (1991) and Mitchell et al. (2008)

0.105, 0.28, 0.38,

Easton et al. (2007)

Greene and Cruise (1995)

Huang et al. (2008)

Singh et al. (2009)

GIS + hydr. modelling

models that are applied to an area equal or larger than 10 km2 as

semi-distributed using sub-catchments; (3) Hydrologic Response

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

1 min 2 min 3 min

information if the size of the sub-catchments or HRUs is reasonably

Object-oriented Rainfall-runoff Simulation Model) (Zhou et al.,

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

Spaal-explicit model implementaons

parameters and inputs; (b) managing and displaying spatial data;

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281

from urban databanks for the denition of parcel cells as urban

approaches (Table 2) use a large variety of (hydraulic) methods to

3.3. Flow routing and hydraulic concepts

3.4. Process-based modelling

Traditionally, a key aspect of urban hydrology is the hydraulic

Models and their applications are developed to answer specic

E. Salvadore et al. / Journal of Hydrology 529 (2015) 6281