Use of impervious surface data obtained from remote sensing in distributed hydrological modeling of urban areas - Urban Remote Sensing: Monitoring, Synthesis and Modeling in the Urban Environment - page 243

18.1 Introduction

applications linking watershed imperviousness to environmental
quality is the Connecticut Nonpoint Education for Municipal
Officials (NEMO) project, which was initiated in 1991 to assist
communities in dealing with the complexities of polluted runoff
management (Arnold et al ., 1994). In this project remote sensing
data were used to estimate current impervious surface cover as
well as to determine zoning specific impervious surface cover
coefficients, which were then applied to estimate future levels
of imperviousness within watersheds. This gave town officials a
prediction of the future status of their town in terms of imper-
vious surface cover and health of their local water resources
(Arnold and Gibbons, 1996). Schueler, Fraley-McNeal and Cap-
piella (2009) performed a review of recent research on the
impervious cover model (ICM), which is based on the hypothesis
that the behaviour of urban stream indicators can be predicted
from the percentage of impervious cover in their contribut-
ing sub-watershed. Results of a meta-analysis of 65 new (since
2003) and 250 older studies showed that the majority of research
confirmed or reinforced the basic premise of the ICM. The
authors present a reformulated ICM model that can be readily
understood by watershed planners, storm-water engineers, water
quality regulators, economists, and policy makers. However, they
conclude also that more information is needed to extend the ICM
as a method to classify and manage small urban watersheds and
decide on the optimum combination of best management prac-
tices to protect or restore streams within each sub-watershed
classification (Schueler, Fraley-McNeal and Cappiella, 2009).
With the increasing awareness of the impact of impervi-
ous surface cover on the urban hydrological cycle there is also
an increase in environmentally sensitive design practices. This
requires that watershed professionals and storm water practi-
tioners not only have a need to better understand the role of
impervious cover types in urban settings, but also of pervi-
ous covers (Law, Cappiella and Novotney, 2009). Mejıa and
Moglen (2009) state that the relationship between flood condi-
tions and the spatial distribution of urban development has been
poorly studied, often because of limitations in streamflow data
availability or because of the common use of lumped watershed
models in urban hydrological modeling. They conclude that
measures for minimizing the impact of urban development on
water resources will have to take better account of spatial patterns
and morphology of urban areas.

More than half of the world's population lives in urbanized areas,
and in the future cities will house an even increasing number
of people in both absolute and relative terms (Martine, 2007).
Many cities worldwide also grow faster spatially than demo-
graphically. A study of the European Environment Agency states
that European cities have expanded on average by 78% since
the mid-1950s, while during the same period the population
increased by only 33% (EEA, 2006). One of the most obvious
impacts of urban expansion is a substantial increase of imper-
vious surfaces, i.e., man-made surface types that prevent direct
infiltration of surface water into the soil, forcing it to travel
downhill to sewers or to places where it can penetrate. Parking
lots, pavements, asphalted streets and rooftops are all examples
of urban surface types that make it impossible for precipitation
to directly seep into the ground. The increase of impervious
surfaces leads to more surface runoff, which in turn increases the
risk for water pollution (Peters, 2009) and floods in urbanized
watersheds, hampers the recharge of aquifers, and boosts erosion
(Schueler, 1994; Brun and Band, 2000). Incorporating informa-
tion on the spatial distribution of impervious surfaces is thus
important in hydrological modeling of urbanized areas. Imper-
vious surface cover is also increasingly used as a key indicator
for surface water quality (Carlson, 2008), aquatic fauna (Gillies
et al ., 2003) and for the overall ecological condition of watersheds
(Arnold and Gibbons, 1996; Sleavin et al ., 2000).
The importance of impervious surface cover for storm
drainage is well known to hydraulic engineers, who use since the
second half of the 19th century the ''rational method'' for the pre-
diction of peak discharge from small drainage catchments mostly
for the design of urban drainage systems (Mulvany, 1850; Pil-
grim and Cordery, 1992; TxDOT, 2009). In the rational method
a runoff coefficient is used, which is defined as the fraction of
rainfall that becomes runoff and which is assumed to be inde-
pendent of rainfall intensity or volume. The runoff coefficient
is typically selected from tables based on the type of land-use.
For a catchment mostly a spatially constant or an area-weighted
value based on the land-use distribution is determined. Since
the seminal paper of Horton (1933) on the role of infiltration
in the hydrological cycle much more awareness exists on the
importance of the infiltration capacity of soils for runoff genera-
tion. Horton describes how the infiltration capacity of a drainage
basin can be determined from runoff and rain-intensity data
and how conversely, if the infiltration capacity is known, the
surface runoff can be determined. He concludes that this is an
improvement over the rational method (Horton, 1933). Both the
rational method as well as the Hortonian runoff approach have
influenced hydrologists for decades in applying spatially constant
runoff concepts and models, although Beven (2006) remarks
that Horton did show how different parts of a catchment with
different infiltration capacities could be taken into account in
the modeling.
In the 1960s a clear impact was identified of urban devel-
opment on the hydrological regime in general and on peak dis-
charges in specific (Carter, 1961; Anderson, 1968; Leopold, 1968).
Since the 1980's various studies also pointed at the correlation
between imperviousness and the health of drainage basins, lead-
ing to the use of imperviousness as a simple, easily measurable
index of environmental disturbance (Schueler, 1994; Arnold
and Gibbons, 1996; Moglen, 2009). One of the best-known

18.2 Spatially distributed
hydrological modeling

Hydrological modeling has evolved enormously over the last four
decades and has taken advantage of the development of computa-
tional power. Traditionally, storm runoff response is determined
by the Rational Method and Soil Conservation Service (SCS)
Curve Number (CN) method (USDA, 1986). These, and other
simple empirical methods, have been widely used to predict the
effect of urbanization on precipitation runoff processes (Carlson,
2004, 2008; Melesse and Wang, 2008). Although these methods
can make use of area-weighted influence of different land-uses
on the runoff coefficient or curve number they are not able to
take spatial patterns into account.
An early example of the use of a simulation model for the esti-
mation of the effect of urban development on peak discharges was