An investigation into the extent and impacts of hard surfacing of domestic gardens in an area of Leeds, United Kingdom


Perry,T.; Nawaz,R.

Landscape and Urban Planning 1: 1-13

2008


factors. In this paper, we show examples of forest management alternatives at macro- and meso-scales

Available
online
at
www.sciencedirect.com
qb
ScienceDirect
ELSEVIER
Landscape
and
Urban
Planning
86
(2008)
1-13
LANDSCAPE
AND
URBAN
PLANNING
www.elsevier.com/locate/landurbplan
An
investigation
into
the
extent
and
impacts
of
hard
surfacing
of
domestic
gardens
in
an
area
of
Leeds,
United
Kingdom
Thomas
Perry
a
,
Rizwan
Nawaz
b,
*
a
Environment
Agency,
Thames
Region,
South
East
Area
Office,
Swift
House,
Frimley
Business
Park,
Frimley,
Camberley,
Surrey,
GU16
75Q,
United
Kingdom
b
School
of
Geography,
University
of
Leeds, Leeds,
LS2
9J7;
United
Kingdom
Received
20
December
2006;
received
in
revised
form
19
November
2007;
accepted
5
December
2007
Available
online
1
February
2008
Abstract
There
is
limited
information
available
on
the
scale
and
potential
impacts
of
increasing
imperviousness
in
suburban
areas,
despite
significant
evidence
which
suggests
that
such
changes
could
have
severe
consequences,
particularly
for
urban
flooding.
In
this
investigation,
aerial
photographs
from
1971
to
2004
were
used
to
map
changes
in
the
impervious
cover
of
a
1.161=
2
suburban
area
of
Leeds
in
northern
England.
A
13%
increase
in
impervious
surfaces
was
observed
over
the
33-year
study
period.
Of
the
increase
in
impervious
surfaces,
75%
was
due
to
paving
of
residential
front
gardens.
To
allow
an
assessment
of
annual
runoff
sensitivity
to
the
increased
imperviousness
in
the
study
area,
an
empirically
based
model,
L-THIA
(long-term
hydrologic
impact
assessment
model)
was
applied.
The
L-THIA
hydrologic
model
predicts
that
average
annual
runoff
increased
by
12%
over
the
period
of
study.
The
L-THIA
model
is
also
used
to
create
a
graph
from
which
increases
in
annual
runoff
due
to
increased
imperviousness
can
be
predicted.
These
results
show
significant
increases
in
imperviousness
and
suggest
that
this
will
cause
an
increase
in
the
frequency
and
magnitude
of
flooding
in
the
area.
©
2007
Elsevier
B.V.
All
rights
reserved.
Keywords:
Impervious;
Land-use;
Urban;
Flooding;
GIS;
Green
space
1.
Introduction
The
extent
of
urban
areas
worldwide
continues
to
grow
from
around
471
million
ha,
accounting
for
4%
of
global
land
area
(UNDP,
2000),
due
to
population
and
housing
pres-
sures,
employment
and
other
socio-economic
factors.
In
the
UK,
urbanisation
has
substantially
increased
since
the
middle
of
the
last
century
and
recent
estimates
indicate
about
7%
of
England's
land
area
is
covered
by
cities
and
towns
with
popula-
tions
exceeding
10,000
people
(DETR,
2000).
In
the
UK
there
is
generally
a
lack
of
information
on
the
extent
of
urban
green
space.
This
may
in
part
be
due
to
the
fact
that
there
is
no
statu-
tory
requirement
for
local
authorities
to
provide
good
quality
urban
green
space,
and
the
issue
therefore
receives
less
attention
(UK
House
of
Commons,
2006).
This
should
concern
planning
authorities
given
that
numerous
studies
have
demonstrated
the
*
Corresponding
author.
Tel.:
+44
7894539052;
fax:
+44
113
3433308.
E-mail
address:
(R.
Nawaz).
0169-2046/$
see
front
matter
©
2007
Elsevier
B.V.
All
rights
reserved.
doi:10.1016/j.landurbplan.2007.12.004
high
degree
of
sensitivity
of
important
environmental
responses
(e.g.
surface
temperatures
and
stormwater
runoff)
to
the
extent
of
urban
green
space
and
building
density
(Pauleit
and
Duhme,
2000;
Whitford
et
al.,
2001;
Nowak
et
al.,
2003;
Gomez
et
al.,
2004).
There
is
even
less
information
on
the
extent
of
domestic
gar-
dens
in
urban
areas
(Mathieu
et
al.,
2007)
yet
they
comprise
a
significant
portion
of
many
urban
properties
and
must
there-
fore
constitute
a
sizeable
proportion
of
urban
areas
(Daniels
and
Kirkpatrick,
2006;
Gaston
et
al.,
2005).
The
few
studies
attempt-
ing
to
quantify
the
extent
of
domestic
garden
coverage
put
the
figure
at
19-27%
of
the
entire
urban
area
(Owen,
1991;
McCall
and
Doar,
1997;
London
Assembly
Environment
Committee,
2005;
Gaston
et
al.,
2005).
The
positive
effect
of
urban
green
space
(including
domestic
gardens)
on
the
health
and
well-being
of
city
dwellers
is
well
known
(Dunnett
and
Qasim,
2000)
but
less
well
studied
is
the
potential
role
of
domestic
gardens
in
alleviating
urban
flooding
and
aiding
urban
groundwater
recharge.
It
is
therefore
important
to
accurately
quantify
the
extent
of
green
space
within
such
envi-
2
T
Perry,
R.
Nawaz
1
Landscape
and
Urban
Planning
86
(2008)
1-13
ronments
(Gaston
et
al.,
2005)
and
assess
likely
environmental
impacts
due
to
any
changes.
In
the
UK,
the
high
proportion
of
urban
green
space
in
some
cities
is
now
threatened
by
a
relatively
new
phenomenon—the
growing
trend
of
paving
over
domestic
gardens.
Several
rea-
sons
including
increased
car
ownership,
difficulty
of
on
street
parking,
poor
public
transport
and
a
fashion
for
low
mainte-
nance
minimalist
gardens
have
been
identified
as
the
likely
causes
(London
Assembly
Environment
Committee,
2005).
The
practice
of
garden
paving
generally
attracts
little
attention
due
to
the
small
scale
of
each
change
but
could
have
far-reaching
consequences.
Garden
areas
are
beyond
the
scope
of
planning
laws
and
are
therefore
often
ignored
by
land
use
statistics
and
policies
(Thompson
et
al.,
2003;
Gaston
et
al.,
2004).
This
has
caused
widespread
hard
surfacing
of
gardens
to
remain
unde-
tected
in
many
areas.
Although
the
issue
has
been
largely
ignored
until
very
recently,
efforts
are
now
underway
to
highlight
the
possible
extent
and
impacts
of
domestic
garden
paving.
For
example
two
recent
reports
have
highlighted
the
large
increases
in
paving
that
have
been
occurring
in
London
for
many
years
(see
London
Assembly
Environment
Committee
Report,
2005;
Healey,
2004).
The
issue
also
received
extended
press
coverage
in
the
UK
recently
based
on
a
warning
from
the
British
Royal
Society
for
the
Protection
of
Birds
(RSPB,
2006).
The
first
published
report
on
the
subject
of
impervious
front
gardens
in
the
UK
prepared
for
the
London
Borough
of
Ealing
(Healey,
2004)
performed
an
extensive
review
of
published
research
and
found
almost
no
information
on
the
sub-
ject.
Since
the
publication
of
this
report,
the
London
Assembly
Environment
Committee
(2005)
produced
a
desk
study
on
paved
gardens,
sparking
an
increased
interest
in
the
subject.
Also,
Pauleit
et
al.
(2005)
provided
an
article
with
a
brief
discussion
on
urban
garden
paving
in
an
area
of
northern
England.
Pauleit
et
al.
(2005)
used
aerial
photographs
to
measure
land
cover
changes
in
the
urban
environment
of
eleven
residential
areas
in
Merseyside
between
1975
and
2000.
This
was
then
followed
by
applying
a
simple
hydrological
model
(adapted
from
Pandit
and
Gopalakrishnan,
1996),
to
determine
runoff
sensitivity
to
this
change.
Given
the
general
lack
of
studies
on
the
topic,
it
is
clear
that
further
investigations
on
urban
domestic
gardens
are
required.
The
present
study
aims
to
fill
this
research
gap
by
providing
an
assessment
of
the
extent
of
domestic
front
garden
paving
and
the
resulting
environmental
implications
for
a
small
area
of
Leeds
in
northern
England.
This
particular
case
study
area
was
selected
since
it
has
experienced
a
large
increase
in
the
percentage
of
impervious
surfaces
over
recent
years
and
the
investigation
will
attempt
to
explain
any
observed
land
use
changes
and
implica-
tions
for
flood
risk.
Over
100
urban
runoff
models
of
varying
degrees
of
com-
plexity
are
described
in
the
literature
(Mitchell
et
al.,
2001)
and
these
can
be
categorised
into
three
groups
(Nix,
1994);
(i)
sim-
ple
models,
(ii)
simple
routing
models
and
(iii)
complex
routing
models.
As
suggested
by
the
name,
the
first
type
of
model
offers
a
simplified
representation
of
the
urban
catchment
and
produce
long-term
averages
(e.g.
annual
runoff);
empirical
models
could
be
considered
as
simple
models.
Both
simple
and
complex
rout-
ing
models
are
based
on
physical
laws
describing
flow
within
a
catchment
(Zoppou,
1999).
Although
they
are
deterministic
models,
they
describe
catchment
behaviour
at
different
com-
plexities
(Zoppou,
1999).
Given
such
large
model
choice,
careful
consideration
needs
to
be
given
to
model
selection.
Nix
(1994)
stresses
that
in
urban
runoff
modelling,
the
key
issue
is
to
ensure
a
match
between
modelling
effort
and
study
objective—match
the
model
to
the
task
and
avoid
using
a
large
model
when
a
sim-
pler,
smaller
one
will
suffice.
Based
on
this
recommendation,
it
was
decided
to
employ
a
simple
empirical
runoff
model
in
this
investigation.
2.
Methods
2.1.
Study
area
Research
was
undertaken
for
a
1.161=
2
area
of
subur-
ban
housing,
located
in
the
suburbs
of
Halton,
Austhorpe
and
Whitkirk
(the
areas
fall
within
the
Halton
Ward
according
to
the
2001
UK
census
classification)
on
the
urban
fringe
of
Leeds,
in
the
county
of
west
Yorkshire,
England
(Fig.
1).
The
majority
of
housing
in
the
study
area
consists
mostly
of
semi-detached
bungalows
with
large
gardens
and
space
between
bungalows
to
allow
driveways
to
be
constructed.
There
has
been
considerable
urban
densification
in
the
last
40
years
with
a
number
of
new
housing
and
commercial
developments
being
constructed.
The
Whitkirk
conservation
area
is
located
within
the
study
area.
The
study
area
was
selected
due
to
an
observed
increase
in
the
amount
of
impervious
gardens
in
recent
years,
noted
by
Leeds
City
Council
staff.
The
area
was
also
subjected
to
seri-
ous
flooding
in
August
2004,
triggered
by
3
h
of
heavy
rainfall
(Leeds
City
Council,
2004).
At
least
20
houses
were
inundated,
some
of
which
were
rendered
uninhabitable
in
the
short
term.
Floodwaters
are
believed
to
have
emanated
from
several
sources
including
overland
flow
(Leeds
City
Council,
2004)
but
the
main
cause
appears
to
be
the
inability
of
drains
to
cope
with
the
vol-
ume
of
water
travelling
down
the
highway.
In
three
parts
of
the
study
area
large
pools
of
water
built
up
in
2004
causing
inunda-
tion
of
nearby
properties
(see
Fig.
1).
A
possible
cause
identified
in
the
flood
report
by
Leeds
City
Council
(2004)
is
the
paved
driveways
draining
onto
the
road.
An
example
of
this
is
provided
in
Fig.
2
where
the
entire
front
garden
is
paved.
In
one
case
the
flooding
was
made
significantly
worse
by
the
fact
that
dropped
curbs
allowed
floodwater
to
escape
the
gutter
of
the
road
and
flow
onto
residential
property,
causing
internal
flooding
of
six
bungalows.
In
June
2007,
Leeds
was
again
affected
by
flooding,
and
it
was
the
Halton
Ward
once
more
which
became
one
of
the
most
badly
affected
areas.
Repeated
flooding
has
led
to
the
Ward
Councillor
to
call
for
urgent
action
to
solve
the
problems
of
the
flood-hit
neighbourhood.
2.2.
Calculation
of
impervious
area
The
total
impervious
area
was
calculated
by
mapping
aerial
photographs
of
the
study
area
into
the
ArcMap
GIS.
In
order
to
study
the
change
in
impervious
surfaces
over
time,
pho-
tographs
taken
in
1971
and
2004
were
studied.
The
1971
aerial
Cerny
th
)n
7
1
10K1130
I
OY
Whitkirk
as
v
ii
Xusth
ball
N
Hospl
Study
area
2004
2004
Flooding
sites
200
400
GOO
800
1000
in
T
Perry,
R.
Nawaz
Landscape
and
Urban
Planning
86
(2008)
1-13
3
(a)
(c)
(b)
r
BRADFORD
j::
a-11
r
Glic.
.
w
,
.4_
0
2
4
6
5
LEEDS
Leeds
SifiarKten
1.-1111,
Fig.
1.
The
study
area
showing
(a)
the
location
of
west
Yorkshire
region
in
England,
(b)
location
of
Leeds
City
and
(c)
location
of
the
study
area
in
Halton/Whitkirk
(source:
Ordnance
Survey/EDINA).
images
were
photographed
in
black
and
white
at
a
contact
scale
of
1/3000
on
the
20th
and
21st
of
May
1971
by
Meridian
Airmaps
Ltd.
whilst
images
for
2004
were
available
via
Google
Earth.
The
Google
Earth
images
supplied
by
the
Geoinforma-
Fig.
2.
A
paved
driveway
in
the
study
area:
runoff
is
likely
to
drain
to
the
street.
tion
Groupl
were
captured
by
aerial
photography
at
a
resolution
of
60
cm/pixel.
Although
a
standard
for
pixel
size
to
be
used
for
delineating
differences
in
land
use
has
not
yet
been
established
(Schuster
et
al.,
2005),
the
resolution
of
both
sets
of
images
is
well
above
the
minimum
standard
for
urban
investigations
indicated
by
Barnsley
and
Barr
(2000).
Both
the
aerial
photographs
and
the
Google
Earth
images
were
mapped
on
to
British
Ordnance
Survey
(OS)
Mastermap
data
from
2002.
This
was
done
manually
using
reference
points
that
are
present
on
the
aerial
images
and
on
the
Mastermap
data.
The
accuracy
of
transposition
from
the
images
to
the
GIS
was
tested
by
mapping
areas
of
land
that
are
present
as
poly-
gons
on
the
Mastermap
data.
Hand
drawn
polygons
were
found
to
be
on
average
98%
of
the
size
of
the
Mastermap
polygons
with
a
standard
deviation
of
4.4%.
The
calculated
impervious
area
is
therefore
believed
to
be
an
underestimate
of
the
true
size.
Mastermap
data
displays
a
variety
of
surfaces
from
roads
to
buildings
and
allows
the
total
area
occupied
by
these
surfaces
to
be
calculated.
With
Mastermap,
garden
areas
are
almost
exclu-
sively
classified
as
'general
surfaces'
giving
no
indication
of
4
T
Perry,
R.
Nawaz
1
Landscape
and
Urban
Planning
86
(2008)
1-13
whether
they
are
paved
or
not.
Polygons
for
impervious
surfaces
that
were
not
shown
on
the
Mastermap
data
were
drawn
on
top
of
the
Mastermap
'general
surface'
polygons.
The
total
area
cov-
ered
by
new
impervious
surfaces
could
then
be
added
to
the
total
obtained
from
the
Mastermap
data
to
give
a
figure
for
the
total
impervious
area.
The
newly
drawn
polygons
were
classified
into
a
number
of
categories
to
allow
further
analysis
of
the
data
to
take
place.
For
both
sets
of
data
(1971
and
2004)
all
significant
areas
of
paved
garden
were
mapped
and
classified
as
'impervious
gardens';
this
included
driveways
and
very
large
garden
paths.
Buildings
and
impervious
surfaces
from
other
sources—schools,
car
parks
and
roads
which
were
not
present
in
the
Mastermap
data
were
also
digitised
and
classified
as
'buildings'
and
'other
impervious',
respectively.
For
impervious
surfaces
present
on
the
Mastermap
data
but
not
on
aerial
photographs,
a
polygon
was
drawn
on
top
of
the
surface
and
classified
as
'unpaved'.
The
characteristic
colours
and
shapes
of
buildings
and
roads
make
them
easy
to
identify
as
impervious
surfaces.
However,
garden
areas
proved
more
difficult.
Pervious
gravel covered
gar-
dens
can
be
hard
to
differentiate
from
impervious
paved
gardens.
Where
identification
problems
occurred,
surfaces
were
differen-
tiated
by
looking
for
further
evidence.
For
example
patterns
in
the
material
(gravel
driveways
may
have
lines
where
a
car
has
driven
across),
the
colour
and
uniformity
of
the
surface
or
the
cleanness
of
the
line
defining
the
boundary
of
the
surface.
If
no
decision
could
be
reached,
a
cautious
approach
was
adopted
and
the
area
was
defined
as
pervious.
Shadows
from
houses
and
sections
of
ground
hidden
behind
buildings
or
trees
also
caused
problems
with
identification
of
the
type
of
surface.
In
these
cases,
unless
significant
evidence
pointed
towards
an
impervious
sur-
face
(for
example
a
road
disappearing
under
a
tree
and
appearing
on
the
other
side),
the
area
was
classified
as
pervious.
The
total
area
covered
by
impervious
gardens
is
therefore
likely
to
be
underestimated.
Bird
et
al.
(2000) highlight
a
further
problem
relating
to
areas
of
soil
which
are
heavily
used
by
vehicles
(for
example
builder's
yards
and
dirt
roads).
These
areas
can
be
so
compacted
that
they
are
functionally
impervious.
This
is
a
poten-
tial
source
of
error;
however,
it
is
not
believed
to
be
significant
because
most
of
the
runoff
from
these
areas
will
drain
to
less
compacted
sections
of
the
same
surface.
In
order
to
calculate
the
total
impervious
area
all
polygons
within
the
study
area
were
selected
and
total
areas
for
each
category
were
calculated.
The
impervious
areas
for
2004
were
also
studied
in
further
detail
by
mapping
additional
categories.
Garden
surfaces
were
categorised
as
either
'back
gardens'
or
`front
gardens'.
In
most
properties
this
distinction
is
easily
made,
however,
as
a
major
focus
of
the
study
is
on
driveways,
these
surfaces
were
categorised
as
front
gardens
regardless
of
their
position
relative
to
the
house
and
whether
they
exited
the
gar-
den
to
the
front
or
rear
of
the
house.
Areas
to
the
side
of
the
house
were
also
classified
as
front
gardens.
The
back
garden
area
should
therefore
be
interpreted
as
the
total
paved
garden
area
not
associated
with
transport
infrastructure
or
front
gar-
dens.
Gardens
that
are
associated
with
newly
developed
housing
were
also
differentiated
and
placed
in
two
further
categories
('new
development
back
gardens'
or
'new
development
front
gardens').
2.3.
Data
and
runoff
modelling
The
mean
annual
rainfall
for
the
rain
gauge
nearest
the
area
is
630
mm
(1970-2000).
Since
urbanisation
alters
runoff
so
drastically,
modelling
is
often
the
only
way
to
investigate
a
catch-
ment's
potential
to
flood.
There
are
two
main
types
of
urban
runoff
model;
'planning'
models
are
used
to
give
an
overall
assessment
of
potential
urban
runoff
problems.
'Design'
mod-
els
on
the
other
hand
aim
to
simulate
a
single
storm
event
to
a
high
level
of
accuracy
(Huber
and
Dickinson,
1988,
cited
by
Bhaduri
et
al.,
2001).
Design
models
require
large
quantities
of
data
that
are
potentially
difficult
to
obtain,
relating
to
the
catch-
ment
and
its
drainage
system
Planning
models on
the
other
hand
are
generally
designed
with
low
mathematical
complexity
and
therefore
have
low
data
requirements.
Due
to
data
limitations
it
was
decided
to
apply
a
planning
model
to
determine
the
change
in
annual
surface
runoff
resulting
from
land-use
change
over
the
two
study
periods.
This
follows
the
approach
taken
by
Pauleit
et
al.
(2005)
and
Whitford
et
al.
(2001)
where
surface
runoff
was
used
as
an
indicator
of
the
flood-
ing
potential
of
the
catchment.
Urban
induced
flooding
is
caused
by
greater
runoff
levels
from
urban
surfaces
(Hollis,
1975);
the
increase
in
runoff
due
to
urbanisation
therefore
provides
a
good
indication
of
the
increased
flooding
potential
of
a
catchment.
The
empirically
based
long-term
hydrologic
impact
assess-
ment
(L-THIA)
model
was
used
to
calculate
the
average
annual
runoff
for
the
study
area.
L-THEA
estimates
runoff
based
on
the
relationship
between
the
nature
of
the
land
surface,
the
daily
rain-
fall
and
the
soil
type
(Lim
et
al.,
2006).
The
model
is
designed
to
simulate
the
variations
in
hydrology
generated
by
land
use
change.
Model
limitations
include
the
omission
of
groundwater
effects,
snowmelt
or
runoff
from
frozen
ground
in
calculations.
However,
the
percentage
increase
in
runoff
between
years
is
still
believed
to
be
accurate
for
planning
purposes,
as
errors
apply
equally
to
runoff
predictions
from
both
years.
L-THEA
can
be
easily
applied
to
any
small
catchment
and
is
freely
available
to
any
user
via
the
Internet
(Purdue
Research
Foundation,
2004).
The
simplicity
of
the
model
(e.g.
no
calibration
requirement)
makes
it
attractive
to
policymakers
wishing
to
identify
manage-
ment
issues
of
concern
and
to
target
resources
to
ameliorate
any
adverse
impacts
(Bhaduri
et
al.,
2001).
L-THIA
has
been
used
by
Bhaduri
et
al.
(1997),
Bhaduri
et
al.
(2001)
and
Tang
et
al.
(2005)
to
assess
the
impacts
of
urbanisation
in
the
USA.
The
model
was
also
successfully
applied
to
calculate
urban
runoff
for
an
area
in
Liverpool,
England,
by
Whitford
et
al.
(2001).
The
equation
for
runoff
depth
(Q)
is
expressed
as
(Harbor,
1994):
(P
0.2S)
2
Q
=
(P
+
0.8S)
where:
P
is
precipitation;
S
is
potential
maximum
storage
which
is
related
to
the
soil
and
land
cover
conditions
through
the
'curve
number'
(CN),
using
the
following
equation:
1000
S
=
10
CN
All
units
are
in
mm.
(1)
(2)
T
Perry,
R.
Nawaz
Landscape
and
Urban
Planning
86
(2008)
1-13
5
To
solve
Eqs.
(1)
and
(2)
requires
as
input,
daily
rainfall
data
and
a
CN
to
calculate
daily
runoff.
For
this
study,
30
years
(1969-1999)
of
daily
precipitation
data
was
available
(Leeds
Knostrop
Sewage
Works
rain
gauge
(SE
323315),
3
km
south
of
the
study
area.
The
fact
that
post-2000
data
was
unavailable
is
not
significant
as
we
are
interested
in
runoff
sensitivity
to
land-use
change
rather
than
climatic
(precipitation)
change.
The
CN
concept
was
developed
by
the
United
States
Depart-
ment
of
Agriculture,
Soil
Conservation
Service
(USDA,
1986)
and
the
CN
is
a
quantitative
description
of
an
area's
hydrologic
soil
group,
land
use,
treatment
and
hydrologic
condition—the
two
former
being
of
greatest
importance
(Weng,
2001).
USDA
(1986)
states
that
CNs
are
summary
values
ranging
from
0
to
100,
with
high
CN
values
corresponding
to
parking
and
paved
spaces
and
the
lower
values
indicative
of
water
and
wetlands
(see
also
Sanders,
1986).
It
should
however,
be
noted
that
the
value
of
zero
noted
by
USDA
(1986)
must
actually
refer
to
a
value
approaching
zero,
as
clearly
Eq.
(2)
is
invalidated
by
entering
a
zero
value.
Land
use
or
land
cover
types
that
are
found
in
urban
and
semi-urban
landscapes
can
be
approximated
to
CN
values
between
25
and
98,
depending
on
the
properties
of
the
surface
(or
surfaces),
the
hydrologic
soil
group,
and
the
antecedent
con-
ditions
(USDA,
1986).
The
properties
of
the
land
surface
are
particularly
affected
by
the
vegetative
cover
and
the
percentage
of
impervious
cover.
For
the
study
area,
the
total
impervious
and
total
pervious
area
were
assigned
individual
CNs
(95
and
80,
respectively)
in
order
to
assess
their
relative
contributions
to
the
runoff.
These
CNs
were
determined
by
characterising
the
land
surfaces
based
on
the
tables
in
the
USDA
(1986)
document
and
by
selecting
the
relevant
hydrologic
soil
group.
In
L-THEA,
soils
are
classified
into
four
hydrologic
soil
groups
based
on
the
runoff
potential,
as
defined
by
USDA
(1986)
(see
Table
1).
For
the
soils
under
imper-
vious
surfaces,
soil
type
D
was
used
as
this
is
the
recommended
value
for
compacted
urban
soils
(Purdue
Research
Foundation,
2004).
The
soils
types
were
not
differentiated
across
the
study
area
and
classified
as
type
D
due
to
the
low
permeability
of
the
soils
of
East
Leeds
(Crompton
and
Matthews,
1970).
The
CN
value
for
impervious
surfaces
is
lower
than
the
value
of
98
usually
associated
with
totally
impervious
ground
for
two
reasons.
Firstly,
paving
slabs
have
cracks
between
them
and
are
therefore
likely
to
let
a
significant
amount
of
water
through
(Hollis,
1988).
Secondly,
a
portion
of
the
measured
paved
garden
area
does
not
form
part
of
the
directly
contributing
impervious
area
(DCIA)
since
it
drains
onto
pervious
surfaces.
Therefore,
only
a
proportion
of
total
impervious
area
must
be
used
in
long-
term
runoff
calculations
(Lee
and
Heaney,
2003).
However,
the
DCIA
is
impossible
to
define
accurately
except
by
ground
sur-
veys
and
therefore
could
not
be
used
in
this
study.
Instead,
smaller
patches
of
impervious
ground
were
ignored
in
the
map-
ping
of
impervious
surfaces
and
larger
areas
were
accounted
for
by
reducing
the
CN
in
accordance
with
USDA
(1986).
For
pervious
surfaces
the
CN
was
calculated
as
80.
This
equates
to
open
space
with
established
vegetation
and
grass
cover
of
>75%
on
soil
type
D
(USDA,
1986).
This
CN
is
rep-
resentative
of
the
grassed
and
sparsely
vegetated
surfaces
that
are
generally
found
in
suburban
areas,
as
described
in
USDA
(1986).
After
supplying
L-THEA
with
daily
rainfall
and
CNs,
the
model
used
this
data
to
estimate
daily
runoff
for
each
CN
value.
The
model
then
aggregated
daily
values
to
annual
runoff
which
were
provided
as
model
output.
Use
of
annual
runoff
data
enables
a
linear
relationship
to
be
established
between
runoff
and
imperviousness
changes
(Shaw,
1994)
thus
enabling
simple
extrapolation
of
results
to
other
catchments.
Given
that
observed
runoff
data
was
unavailable
for
the
catchment,
model
validation
could
not
be
performed.
This
is
not
of
concern
since
we
are
pri-
marily
interested
in
relative
changes
in
runoff.
It
is
important
to
note
that
using
the
same
rainfall
record
in
the
model
applied
to
two
catchment
states of
permeability
(corresponding
to
1971
and
2004)
ensures
that
any
runoff
changes
observed
are
purely
due
to
land-use
change
excluding
changes
in
rainfall.
In
addition
to
the
runoff
modelling
for
the
Leeds
study
area,
a
sensitivity
study
was
performed
to
predict
runoff
for
different
levels
of
imperviousness.
The
L-THIA
model
was
run
using
the
same
rainfall
and
CN
data
as
was
used
for
the
study
area.
Runoff
values
were
calculated
for
a
theoretical
1
km
2
catchment
with
impervious
cover
values
ranging
from
0%
to
90%.
Soil
groups
B,
C
and
D,
were
tested
but
soil
group
A
had
to
be
omitted
from
calculation
as
the
predicted
CN
value
was
beyond
the
accuracy
threshold
of
L-THIA.
2.4.
Demographic
and
neighbourhood
statistics
To
gain
a
better
understanding
of
reasons
behind
any
changes
in
the
extent
of
garden
paving
in
the
study
area,
it
was
decided
Table
1
Classification
of
hydrologic
soil
groups
for
L-THIA
Hydrologic
soil
group
Description
Soil
types
A
B
C
D
Soils
with
high
infiltration
rates
even
when
thoroughly
wetted.
Generally
they
are
deep,
well
drained
sands
or
gravels.
Soils
with
a
moderate
infiltration
rate
when
thoroughly
wetted.
Generally
these
are
soils
with
a
moderately
fine
to
moderately
coarse
texture.
Soils
with
low
infiltration
rates
when
thoroughly
wetted.
Generally
these
soils
have
a
layer
that
impedes
downward
movement
of
water
or
soils
with
a
fine
texture.
Soils
with
very
low
infiltration
rates
when
thoroughly
wetted.
Generally
these
soils
have
a
high
swelling
potential,
a
permanently
high
water
table
or
a
clay
layer
near
or
at
the
surface.
Sand,
loamy
sand
and
sandy
loam
Silt
loam
or
loam
Sandy
clay
loam
Clay
loam,
silty
clay
loam,
sandy
clay,
silty
clay
or
clay
Based
on
US
Department
of
Agriculture
(2007)
and
Purdue
Research
Foundation
(2004).
6
T
Perry,
R.
Nawaz
Landscape
and
Urban
Planning
86
(2008)
1-13
Table
2
Changes
in
impervious
area
for
the
study
area
1971
2004
%Change
Total
impervious
area
0.371=
2
(31.7%)
0.52km
2
(44.3%)
12.6
Paved
garden
area
0.08
km
2
0.19
km
2
138
Average
paved
garden
size
39.4
m
2
53.4
m
2
36
Number
of
properties
with
paved
front
gardens
2130
(71%)
2700
(90%)
19
to
collect
2001
UK
census
data
from
the
Office
for
National
Statistics
(2001).
In
order
to
compare
the
results
of
the
current
study
with
those
of
Pauleit
et
al.
(2005),
data
on
the
National
Statistics
Index
of
Multiple
Deprivation
was
also
collected.
The
data
is
derived
from
multiple
variables
on
income,
employment,
health,
disability,
education,
skills
and
training,
housing
and
geographical
access
to
services.
Data
for
Leeds
was
collected
from
Leeds
Initiative
(2000)
and
relates
to
the
study
area
and
its
surrounding
suburbs.
The GIS
maps
constructed
as
part
of
this
study
also
enabled
the
number
of
houses
with
driveways/gardens
and
their
sizes
to
be
recorded.
Given
the
extensive
number
of
properties
in
the
study
area
(3001),
it
was
decided
to
obtain
the
data
by
randomly
sampling
100
properties.
3.
Results
3.1.
Land
cover
change
1971-2004
The
total
impervious
cover
increased
by
0.15
km
2
(12.6%)
over
the
33-year
study
period,
from
0.37
km
2
(31.7%)
in
1971
to
0.52
km
2
(44.3%)
in
2004
(see
Table
2).
This
increase
can
clearly
be
seen
by
comparing
Fig.
3a
and
b
which
shows
pervious
and
impervious
surfaces
in
the
study
area
for
the
2
years
in
question.
The
area
covered
by
paved
garden
surfaces
increased
by
a
staggering
138%
accounting
for
75%
of
the
total
increase
in
impervious
surfaces.
The
remaining
25%
of
the
increase
is
due
to
increases
in
the
'buildings'
and
'other
impervious'
categories.
Calculations
suggest
that
around
18%
is
accounted
for
by
an
increase
in
building
area
with
the
remaining
7%
attributed
to
an
increase
in
the
'other
impervious'
category.
The
vast
majority
of
the
increase
in
'other
impervious'
category
is
made
up
by
transport
infrastructure.
The
car
park
for
a
newly
constructed
large
shop
(see
Fig.
4c)
forms
a
large
portion
of
the
increase;
most
of
the
remainder
is
composed
of
car
parking
and
roads
for
new
housing
developments.
3.2.
Garden
paving
trends
Development
of
new
housing
during
the
study
period
con-
tributed
10%
of
the
increase
in
the
area
covered
by
paved
gardens,
of
that
total
6%
was
from
front
gardens.
A
number
of
these
new
housing
developments
can
be
seen
in
the
maps
in
Fig.
4c
and
d.
In
1971,
71%
of
houses
in
the
study
area
had
paved
gardens
(Table
2).
This
increased
by
671
houses
in
2004
such
that
90%
of
houses
had
paved
gardens
(Table
2).
Due
to
the
nature
of
housing
and
the
size
of
the
gardens,
the
vast
majority
of
this
paved
area
was
for
driveways.
Eighty
eight
percent
of
the
residential
properties
in
the
study
area
(excluding
blocks
of
flats)
are
semi-detached
and
(with
very
few
exceptions)
have
the
space
to
create
a
paved
driveway
in
their
garden.
Most
driveways
in
1971
were
generally
only
the
width
required
to
take
one
car
from
the
road
to
the
garage.
In
2004,
driveways
tended
to
take
up
a
larger
area
of
the
garden.
Many
houses
expanded
their
driveways
to
fill
the
space
between
the
house
and
boundary
instead
of
leaving
grass
or
flowerbed
on
one
or
both
sides,
as
was
generally
the
case
in
1971.
This
is
clearly
evident
in
Fig.
4b
and
d
from
where
it
can
be
seen
the
majority
of
houses
have
driveways
(during
both
1971
and
2004),
yet
there
is
clearly
a
large
increase
in
impervious
sur-
faces
over
the
study
period.
This
pattern
is
repeated
across
the
study
area
and
can
be
seen
in
Fig.
4
where
a
number
of
houses
have
paved
the
entire
area
in
front
of
and
beside
the
house
(see
also
Fig.
2).
(a)
Halton
1971
(b)
Halton
2004
..54Aet
v.
Legend
Impervious
Surfaces
Pervious
Surfaces
0
0.25
9.5
F<Jlometers
Legend
Impervious
Surfaces
Pervious
Sudases
0
0.25
0.5
Kilometers
Fig.
3.
Map
of
pervious
and
impervious
surfaces
in
the
study
area.
4
/
0
005
0.1
.
Kilometers
1971
2004
4
a
lyre,
.t•
yr
°v
,
4.
"..
q1
\
0
0.05
0.1
Kilometers
0.05
0.1
2004
Kilometers
Ale;
4
.1..fr
A...11.
la
a
'
41.
-
r
,.r
1.
-
irr
ses
4
l
e'
4
1
'
.46
1
\Nail>
S.
T
Perry,
R.
Nawaz
Landscape
and
Urban
Planning
86
(2008)
1-13
7
0
0.05
0.1
IGlomiaters
1971
1
-
oriiirr
(a)
South
-
west
Halton
(b)
North-east
Halton
Legend:
Pervious
Surfaces
=1
Impervious
Surfaces
Legend:
Pervious
Surfaces
Impervious
Surfaces
Fig.
4.
Comparison
of
two
sections
of
the
study
area
in
1971
and
2004.
In
2004,
footpaths
are
often
incorporated
into
the
drive
instead
of
the
driveway
past
the
house
and
separate
pathway
from
the
gate
to
the
front
door,
which
was
more
common
in
1971.
A
sample
of
100
paved
front
gardens
from
1971
had
an
average
size
of
39.4
m
2
with
a
standard
deviation
of
15.5
m
2
.
A
second
survey
of
100
gardens
for
the
2004
data
showed
that
the
average
paved
garden
size
had
increased
to
53.4
m
2
with
a
standard
deviation
of
28.1
m
2
(see
Table
2).
A
two
sample
T-test
showed
that
the
two
means
are
significantly
different
(P
=
0.001).
This
trend
for
larger
driveways
appears
to
have
accounted
for
much
of
the
increase
in
impervious
area.
Of
the
75%
increase
in
impervious
surfaces,
52%
was
from
paved
front
gardens
and
23%
from
paved
back
gardens
account-
ing
for
6.5%
and
3%
of
the
total
land
in
the
study
area,
respectively.
An
increase
in
paving
behind
many
houses
can
also
be
observed
in
Fig.
4.
The
majority
of
paved
back
garden
areas
in
2004
are
small
patio
areas
immediately behind
the
house.
In
1971,
back
gardens
tended
to
have
either
no
paving
at
all
or
a
path
running
around
the
walls
of
the
house
and
out
into
the
gar-
den.
These
small
paths
do
not
register
on
the
1971
maps
as
areas
of
this
size
were
omitted
due
to
their
minimal
impact
on
hydrol-
ogy
and
the
difficulty
of
mapping.
Whilst
some
patio
areas
may
drain
to
garden
surfaces,
many
have
a
drain
which
is
linked
to
the
household
drainage
system
and
therefore
contributes
to
urban
runoff.
It
was
not
possible
to
determine
hydrological
connectiv-
ity
from
aerial
photographs;
therefore
all
patios
were
included
in
impervious
area
calculations.
Driveways
in
the
study
area
are
generally
very
long
to
allow
garages
to
be
beside
or
behind
the
house,
where
they
do
not
block
light
from
the
windows
or
ruin
the
appearance
of
the
front
of
the
house.
This
means
that
they
account
for
a
very
large
area
of
paved
surfaces
relative
to
what
is
actually
needed.
The
impervious
area
for
a
single
garage
may
be
up
to
28
m
long
(24
m
of
driveway
and
4
m
of
garage),
enough
to
fit
seven
cars.
Twenty-eight
metres
is
the
longest
driveway
for
a
semi-detached
house
in
the
study
area.
However,
many
of
the
roads
in
the
study
area
have
driveways
that
are
typically
16-18
m
long.
The
vast
majority
(88%)
of
houses
in
the
study
area
are
semi-detached
and
only
12%
of
houses
are
terraced
(excluding
blocks
of
flats).
Many
terraced
houses
have
garages
but
these
tend
to
back
straight
onto
the
road
behind
the
house
and
therefore
account
for
only
a
small
impervious
area.
A
large
proportion
of
these
terraced
houses
have
paved
gardens.
In
Pervious
Impervious
L
To
ta
l
Annua
l
Flow
x
10
3
rn
3
7000
6000
5000
4000
3000
2000
1000
Soil
Type
B
Soil
Type
C
Soil
Type
D
%
inc
rease
in
ru
n
to
f
800
700
600
500
400
300
200
100
8
T
Perry,
R.
Nawaz
1
Landscape
and
Urban
Planning
86
(2008)
1-13
1971
2004
Fig.
5.
Simulated
total
annual
flow
volume
for
the
study
area
showing
the
relative
contributions
of
the
pervious
and
impervious
surfaces
for
1971
and
2004.
terraced
houses,
paved
areas
are
commonly
back
gardens
which
exit
to
a
lane
behind
the
terrace.
Terraced
houses
generally
have
garden
sizes
between
40
m
2
and
100
m
2
compared
to
the
range
of
140-500
m
2
which
is
typical
for
semi-detached
houses
in
the
study
area.
However,
it
appears
to
be
more
usual
for
these
small
gardens
to
be
entirely
paved
and
they
can
therefore
significantly
impact
the
total
imperviousness.
People
in
semi-detached
houses
tend
to
have
a
lot
more
garden
space.
However,
it
is
very
rare
to
see
all
of
this
area
paved.
The
vast
majority
of
semi-detached
houses
tend
to
have
retained
a
large
portion
of
grass
in
their
back
garden
with
only
a
small
paved
patio
area.
There
appears
to
be
far
less
desire
to
retain
front
garden
lawns
in
semi-detached
houses.
3.3.
Runoff
sensitivity
Based
on
the
observed
increase
in
the
impervious
area
(12.6%),
the
L-THIA
model
predicts
a
similar
change
(a
12%
increase)
in
the
average
annual
runoff
from
the
study
area
(the
similarity
being
coincidental).
Fig.
5
shows
the
relative
con-
tributions
of
the
pervious
and
impervious
surfaces.
In
1971,
the
impervious
area
accounted
for
56%
of
the
total
runoff,
increasing
to
66%
in
2004.
The
modelled
runoff
volumes
equate
to
average
annual
runoff
depths
of
490
mm
for 1971
and
560
mm
for
2004.
Fig.
6
displays
the
results
of
a
sensitivity
study
using
the
L-
THIA
model.
A
number
of
model
runs
have
been
combined
to
display
increases
in
runoff
relative
to
increases
in
the
percentage
of
impervious
cover,
for
three
of
the
hydrologic
soil
groups.
Soil
type
A
has
been
excluded
because
the
calculated
CN
is
below
the
accuracy
threshold
of
the
model.
It
is
clear
that
as
percent-
age
imperviousness
increases,
there
is
an
increase
in
percentage
runoff;
the
relationship
being
linear
as
expected.
The
slope
of
the
relationship
varies
with
soil
type
and
in
some
cases
(soil
B)
is
very
high
resulting
in
huge
changes
in
runoff
from
rela-
tively
small
changes
in
percentage
imperviousness
(e.g.
a
runoff
change
as
high
as
360%
can
result
from
a
40%
runoff
increase
in
pristine
catchment).
3.4.
Demographic
and
neighbourhood
statistics
Data
from
the
2001
UK
census
was
collected
on
(a)
vehicle
ownership,
(b)
mode
of
transport
used
to
get
to
work
and
(c)
distance
of
residence
from
workplace.
Fig.
7
compares
the
number
of
vehicles
per
household
in
the
study
area
with
those
of
Leeds
City
and
England.
It
can
be
seen
that
levels
of
car
ownership
are
high
although
most
households
have
only
one
car.
The
percentage
of
two
car
households
is
close
to
the
national
average
but
higher
than
the
citywide
figure,
whilst
the
number
of
households
with
three
or
more
cars
is
lower
than
the
average
for
England.
An
interesting
finding
is
that
90%
of
households
had
a
paved
front
garden
in
2004
yet
24%
of
households
did
not
own
a
car
(in
2001).
On
average,
57%
of
people
in
the
study
area
drive
to
work
in
a
car
or
van.
This
is
slightly
above
the
national
average
of
55%.
Seventy
eight
percent
of
the
employed
people
in
the
study
area
work
over
2
km
away
from
their
home
which
is
well
above
the
national
average
of
69%.
Employment
levels
are
high
in
the
0
10
20
30
40
50
60
70
80
90
100
%
increase
in
imperviousness
Fig.
6.
Modelled
runoff
increases
in
response
to
increased
catchment
imperviousness
for
three
soil
types.
V;
1 ,
Leeds
-
:
-
ANIL.
tak;
Southport
r.
AIN
T
Perry,
R.
Nawaz
Landscape
and
Urban
Planning
86
(2008)
1-13
9
o
0
cars
1
car
2
cars
3
cars
4
or
more
cars
in
Study
area
Leeds
England
Fig.
7.
Vehicle
ownership
per
household
for
the
study
area,
the
city
of
Leeds
and
England.
Data
is
from
the
2001 UK
census
(Office
for
National
Statistics,
2001).
Employed
Unemployed
[II
Student
0
Retired
E
Economically
Inactive
_
=
Halton
Study
area
Leeds
England
Fig.
8.
Percentage
of
people
in
each
economic
group
for
the
study
area,
the
City
of
Leeds,
and
England.
Data
is
from
the
2001
census
(Office
for
National
Statistics,
2001).
Pe
rce
n
tag
e
o
f
house
ho
lds
50
45
40
35
30
25
20
15
10
5
0
70
60
50
co
c
40
0
30
20
10
0
study
area,
as
are
the
number
of
retired
people
(see
Fig.
8).
It
would
also
be
useful
to
obtain
some
statistics
on
a
nearby
coun-
cil
estate (Halton
Moor)
council
estate
where
there
is
significant
deprivation.
Sixty
six
percent
of
residents
have
no
qualifications
and
61%
are
either
economically
inactive
or
unemployed
com-
pared
to
national
figures
of
29%
and
36%,
respectively.
Between
April
2003
and
March
2004
there
were
144
crimes
against
vehi-
cles
in
the
Halton
Moor
area.
This
is
12%
of
the
number
of
cars
owned
in
the
area.
The
Leeds
average
was
5%
for
the
same
period
(Office
for
National
Statistics
(2004))
(Fig.
9).
4.
Discussion
4.1.
Garden
paving
By
directly
mapping
land
cover
measurement
of
the
total
impervious
area
was
achieved
to
a
high
level
of
accuracy
relative
to
the
size
of
the
study
area.
This
method
of
mapping
provides
Fig.
9.
Arial
photos
of
small
sections
of
the
Leeds
study
area
and
the
Southport
suburb
studied
by
Pauleit
et
al.
(2005).
Both
pictures
obtained
from
Google
Earth
and
were
captured
at
an
eye
level
of
256
m,
the
Leeds
picture
was
taken
in
August
2004
and
the
Southport
picture
between
July
2003
and
July
2006.
10
T
Perry,
R.
Nawaz
1
Landscape
and
Urban
Planning
86
(2008)
1-13
a
more
accurate
measurement
than
estimates
from
similar
stud-
ies
(Bird
et
al.,
2000;
Pauleit
et
al.,
2005).
However,
mapping
impervious
areas
manually
is
time-intensive
and
the
1.16
km
2
study
area
is
far
smaller
than
the
coverage
achieved
by
the
above
studies.
Therefore
the
method
employed
in
the
current
study
is
only
able
to
offer
an
in
depth
investigation
of
a
single
study
area,
which
is
believed
to
be
exceptional
in
terms
of
the
amount
of
paved
garden
surfaces.
The
results
show
a
significant
increase
in
the
paved
areas
and
demonstrate
that
the
vast
majority
of
this
change
is
due
to
the
recent
trend
of
paving
garden
surfaces.
These
results
are
believed
to
be
representative
of
similarly
designed
housing
areas
which
have
space
provided
for
garden
paving
and
are
of
high
socio-economic
status.
There
have
been
a
variety
of
reasons
suggested
to
explain
the
trend
of
impervious
gardens.
These
include
increased
car
ownership,
difficulty
of
on
street
parking,
poor
public
trans-
port
and
a
fashion
for
low
maintenance
minimalist
gardens
(London
Assembly
Environment
Committee,
2005).
Fear
of
car
crime
or
robbery
whilst
walking
from
the
car
to
the
house
has
also
been
cited
as
reasons
to
park
on
front
gardens
(Healey,
2004).
More
parking
is
also
required
in
plots
where
older
family
sized
properties
are
converted
into
flats
(Healey,
2004).
Many
of
these
suggestions
reflect
significant
societal
changes
that
have
occurred
over
the
last
30
years.
Perhaps
the
chance
to
display
a
car
in
front
of
the
house
is
taking
precedence
over
the
more
traditional
household
status
symbol;
a
well
kept
front
garden.
Developers
are
also
trying
to
provide
only
the
minimum
garden
space
required
meaning
that
the
majority
of
newly
developed
residential
properties
have
a
low
percentage
of
pervious
surfaces
(Kellet,
1982).
In
addition
to
these
reasons,
there
could
be
several
reasons
for
the
large
proportion
of
the
observed
12.6%
increase
in
the
percentage
of
impervious
surfaces
attributed
to
garden
paving
in
the
study
area.
The
majority
of
the
minor
residential
roads
are
only
the
width
of
two
and
a
half
cars.
Driving
around
the
study
area
can
therefore
be
difficult
when
there
are
cars
parked
in
the
street.
Thin
streets
give
less
room
for
other
drivers
to
pass
cars
and
may
also
lead
to
an
increased
likelihood
of
accidental
damage
whilst
passing
cars
in
the
street.
This
may
have
caused
residents
to
keep
their
cars
off
the
road.
Socio-economic
statistics
of
the
area
also
provide
clues
to
the
increasing
trend
in
garden
paving.
The
relatively
high
numbers
of
employed
and
retired
people
in
the
study
area
(see
Fig.
8)
might
be
more
likely
to
pave
their
garden.
This
could
be
for
a
number
of
reasons;
retired
people
are
generally
older
and
are
therefore
more
likely
to
have
mobility
problems
which
would
require
the
car
to
be
parked
very
near
to
the
house,
people
with
mobility
issues
are
also
more
likely
to
need
a
car
as
they
are
unable
to
walk,
or
cycle
to
public
transport
and
other
services.
They
may
also
wish
to
have
a
garden
that
is
easy
to
maintain.
Similarly,
residents
who
are
employed
may
have
little
time
or
inclination
to
do
any
gardening,
they may
also
rely
on
and
prize
their
car
highly,
increasing
their
desire
to
protect
it.
Also
linked
to
the
mobility
issue
is
the
fact
that
it
was
noted
a
large
proportion
of
people
work
more
than
2
km
from
home.
This
is
probably
due
in
part
to
the
proximity
of
the
study
area
to
a
major
motorway
(the
M1).
This
suggests
that
a
large
number
of
people
in
the
study
area
rely
heavily
on
their
cars.
This
may
lead
them
to
want
to
park on
their
driveways
as
the
garden
area
is
perceived
to
be
safer
than
leaving
a
car
on
the
street
(London
Assembly
Environment
Committee,
2005).
The
safety
issue
further
brings
to
attention
a
nearby
council
estate—in
Halton
Moor.
This
is
an
area
of
significant
deprivation
and
notorious
for
crime
and
social
problems
which
may
lead
residents
to
park
on
their
driveways
or
in
garages
in
order
to
keep
their
cars
safe
from
vandals
venturing
into
the
study
area.
In
general,
there
appeared
to
be
less
desire
to
retain
front
gar-
den
lawns
in
semi-detached
houses
compared
to
back
gardens.
This
may
be
due
to
householder
aiming
for
better
functionality
from
their
garden.
A
lawn
in
front
of
the
house
is
not
a
comfort-
able
place
for
many
garden
activities;
for
example
kids
playing,
reading
the
papers
or
sunbathing.
The
front
garden
may
therefore
be
viewed
as
an
area
requiring
significant
maintenance
but
which
gives
little
benefit
in
terms
of
function.
By
paving
the
area
an
extra
car
parking
space
is
provided
increasing
the
functionality
of
the
area.
4.2.
Runoff
sensitivity
The
predicted
increase
in
annual
runoff
(12%)
is
likely
to
have
a
significant
effect
on
the
frequency
and
magnitude
of
flooding
in
the
study
area.
The
amount
of
runoff
from
a
catchment
pro-
vides
a
good
indicator
of
the
likely
flooding
potential
(Hollis,
1975).
The
circumstances
of
the
2004
flood
event
in
the
study
area
(Leeds
City
Council,
2004)
suggest
that
these
predictions
are
well
founded.
The
increase
in
runoff
will
have
altered
the
catchment's
response
to
storm
events
and
the
drainage
system
in
the
study
area
will
no
longer
be
able
to
cope
with
the
maximum
rainfall
event
it
was
designed
to
convey.
The
modelled
increase
in
runoff
will
also
cause
an
increase
in
the
pollution
loads
transported
into
the
river
system.
This
is
partly
due
to
pollutants
being
washed
off
the
increased
amount
of
impervious
surfaces
(Tang
et
al.,
2005)
and
also
due
to
the
increase
in
combined
sewerage
overflow
incidents
that
occur
in
response
to
more
frequent
flood
events
(Old
et
al.,
2006).
The
increased
runoff
predicted
by
modelling
reflects
the
higher
total
runoff
that
can
be
expected
not
just
from
a
12.6%
increase
in
paved
surfaces
but
also
a
corresponding
12.6%
decrease
in
veg-
etated
surfaces.
The
increase
in
runoff
is
explained
not
only
by
the
lower
levels
of
infiltration
into
the
soil
but
also
the
reduced
evapotranspiration
and
storage
from
vegetation.
It
should
be
noted
that
the
reported
runoff
changes
are
not
sensitive
to
rain-
fall
changes
since
the
same
rainfall
data
was
routed
through
the
L-THIA
model
but
for
the
two
different
levels
of
catchment
imperviousness.
The
sensitivity
study
reveals
that
the
runoff
change
is
depen-
dent
on
the
soil
type.
The
infiltration
rates
of
soil
type
D
are
much
closer
to
that
of
an
impervious
surface
and
therefore
have
far
lower
predicted
increase
in
runoff
levels
due
to
increased
imperviousness.
How-
ever,
for
soil
type
B,
with
relatively
high
infiltration
rates
(see
Table
1),
any
paving
will
result
in
very
large
runoff
increase.
The
findings
from
the
sensitivity
study
are
important
as
they
suggest
that
areas
with
soils
of
greater
permeability
than
in
the
study
T
Perry,
R.
Nawaz
1
Landscape
and
Urban
Planning
86
(2008)
1-13
11
area
would
experience
much
greater
runoff
sensitivity
to
garden
paving.
4.3.
Comparison
with
other
studies
The
observed
12.6%
increase
in
the
percentage
of
imper-
vious
surfaces
and
its
impact
on
runoff
compares
favourably
with
the
results
of
the
study
by
Pauleit
et
al.
(2005)
on
land-
use
change
in
Liverpool,
UK.
The
National
Index
of
Multiple
Deprivation
(Leeds
Initiative,
2000)
places
the
area
in
that
par-
ticular
study
within
the
high
socio-economic
status
category,
where
the
average
increase
in
imperviousness
was
8%.
These
areas
experienced
an
increase
of
4.5%
in
the
building
area
and
a
5.5%
increase
in
surfaces
associated
with
transport
over
the
25-year
study
period.
This
compares
to
an
increase
of
around
18%
in
buildings
in
the
Leeds
study
area
and
a
75%
increase
in
transport
area
(Impervious
gardens'
+
'other
impervious').
In
both
studies
a
high
proportion
of
the
increase
in
transport
area
was
made
up
of
increases
in
paved
garden
surfaces.
However,
this
increase
in
the
Leeds
study
area
is
of
a
far
higher
magnitude
for
reasons
provided
earlier.
An
important
difference
between
the
two
investigations
is
the
fact
that
the
period
2000-2004
was
not
considered
by
the
study
of
Pauleit
et
al.
(2005);
obser-
vations
by
Leeds
City
Council
employees
suggest
that
this
is
the
period
where
the
greatest
increases
in
imperviousness
were
experienced.
The
modelled
effects
of
land-use
change
on
runoff
were
expectedly
higher
for
the
Leeds
study
area
compared
to
results
reported
by
Pauleit
et
al.
(2005)
for
the
high
socio-economic
status
areas
of
Liverpool
with
projected
increases
of
12%
and
5%,
respectively.
Apart
from
the
investigation
of
Pauleit
et
al.
(2005)
there
are
no
other
published
studies
that
allow
a
basis
for
comparison
with
the
findings
reported
in
the
present
investigation.
However,
the
work
of
Hollis
(1975)
is
of
some
relevance
and
is
briefly
dis-
cussed
here.
Hollis
(1975)
predicted
that
urbanisation
increased
flood
peaks
by
between
1.5
and
6
times
those
experienced
in
the
pristine
catchment.
These
figures
were
reached
by
looking
at
a
variety
of
studies,
for
which
total
imperviousness
averaged
around
30%,
a
figure
that
was
fairly
typical
of
urban
areas
in
the
UK
at
the
time
(this
is
exemplified
by
the
measured
32%
imperviousness
for
the
Leeds
study
area
in
1971).
Hollis
(1975)
predicted
that
a
30%
paving
of
the
urban
area
led
to
an
increase
in
flood
peaks
of
at
least
100%.
This
is
signifi-
cantly
higher
than
the
60%
increase
predicted
in
the
current
study
for
the
same
degree
of
land
use
change.
This
is
because
average
annual
runoff
does
not
increase
to
the
same
extent
as
flood
peaks
in
response
to
urbanisation.
Extreme
events
are
evened
out
by
lower
flows
over
the
course
of
the
period
studied.
These
smaller
events
do
not
experience
such
a
large
increase
in
runoff,
as
a
greater
proportion
of
the
total
rainfall
will
be
taken
up
filling
sur-
face
voids.
The
increase
in
peak
flow
from
the
Leeds
study
area
could
potentially
be
far
higher
than
the
12%
predicted
increase
in
average
runoff.
It
is
however,
important
to
note
that
imper-
vious
effects
are
more
pronounced
where
there
is
a
low
initial
level
of
imperviousness
in
a
catchment
(Schuster
et
al.,
2005)
which
is
not
the
case
in
the
Leeds
study
area.
4.4.
Future
outlook
A
significant
proportion
of
the
land
area
across
cities
in
the
developed
world
is
occupied
by
domestic
gardens
(Daniels
and
Kirkpatrick,
2006)
yet
gardens
remain
the
least
studied
habitat
in
urban
areas
(Mathieu
et
al.,
2007).
In
the
UK,
media
and
political
attention
has
only
begun
to
focus
on
the
issue
of
paved
gardens
in
the
last
4
years.
This
is
likely
to
be
because
very
large
increases
in
garden
paving
have
only
been
occurring
for
a
relatively
short
time
period.
Current
social
and
planning
trends
suggest
that
the
period
of
growth
is
set
to
continue.
Possible
reasons
for
this
include
increasing
per-household
car
ownership,
higher
numbers
of
commuters,
poor
public
trans-
port,
and
conversion
of
urban
housing
into
flats;
factors
which
continue
to
increase
the
demand
for
parking
on
city
streets.
There
is
evidence
that
the
problem
could
become
further
exac-
erbated
due
to
feedback
effects:
the
'neighbour
mimicry
effect'
observed
by
Zmyslony
and
Gagnon
(1998)
and
the
reduction
in
on
street
parking
that
generally
occurs
due
to
creation
of
off-
street
parking.
The
combined
effect
of
these
trends
may
cause
rate
of
growth
in
paved
garden
surfaces
to
far
exceed
the
trends
seen
over
the
last
33
years.
However,
it
must
be
noted
that
the
level
of
growth
observed
in
the
present
study
is
not
thought
to
be
typical,
and
exists
in
a
suburb
of
Leeds
where
conditions
are
particularly
favourable
to
paving
driveways.
Nonetheless,
planners
should
be
concerned
that
there
is
cur-
rently
hardly
any
legislation
in
the
UK
aimed
at
specifically
managing
the
problem.
The
40-year
old
Civic
Amenities
Act
(1967)
is
the
only
piece
of
legislation
worthy
of
note.
The
Act
allows
local
planning
authorities
to
define
areas
of
spe-
cial
architectural
or
historic
interest
in
order
to
preserve
their
character
and
appearance
(Thomas,
1983).
This
provides
the
Council
with
powers
to
prevent
garden
walls
being
removed
to
create
driveways
in
order
to
preserve
the
character
of
the
area.
However,
the
powers
provided
by
the
act
are
often
not
exercised
due
to
the
potential
costs
of
prosecution
(Healey,
2004).
Indeed,
Article
4
of
the
UK
Town
and
Country
Planning
(General
Permitted
Development)
Order
(1995)
permits
hard
surfacing
within
the
curtilage
Vbr
any
purpose
incidental
to
the
enjoyment
of
the
dwelling".
In
order
to
create
a
driveway
a
'vehicular
crossing'
application
must
be
made
to
the
local
Council
which
takes
a
variety
of
factors
into
account
before
reaching
a
decision.
Councils
do
not
yet
take
the
potential
envi-
ronmental
impacts
of
a
driveway
into
account.
Moreover,
the
kerb
lowering
order
is
not
strictly
enforced
and
prosecutions
for
illegally
'dropped
kerbs'
are
very
rare
due
to
the
cost
of
court
proceedings.
A
survey
of
pavement
crossovers
in
the
London
Borough
of
Ealing
found
that
9%
of
properties
with
parking
on
their
front
gardens
had
not
submitted
a
crossover
application
(Healey,
2004).
A
number
of
solutions
are
available
to
better
manage
the
problem
of
urban
garden
paving
and
its
environmental
impacts.
Firstly,
the
public
and
policymakers
need
to
be
made
more
aware
of
the
environmental
implications
of
domestic
garden
paving.
Media
campaigns
such
as
that
undertaken
recently
undertaken
by
the
Royal
Horticultural
Society
(2005)
are
a
good
starting
point.
Such
campaigns
need
to
highlight
the
value
of
domestic
12
T
Perry,
R.
Nawaz
Landscape
and
Urban
Planning
86
(2008)
1-13
gardens
to
societal
well-being
and
suggest
alternatives
to
hard
surfacing.
Secondly,
both
the
Civic
Amenities
Act
(1967)
and
the
UK
Town
and
Country
Planning
Order
(1995)
could
be
updated
and
enforced
more
rigorously.
Further
research
is
clearly
required
and
this
should
build
on
the
work
of
Gaston
et
al.
(2005)
carried
out
under
the
UK
Natu-
ral
Environmental
Research
Council
project
URGENT.
Special
consideration
needs
to
be
given
to
the
impacts
of
garden
paving
on
urban
drainage
systems.
The
annual
average
runoff
sensi-
tivity
to
garden
paving
reported
in
the
current
study
should
be
more
fully
explored
by
considering
seasonal
changes
and
peak
flows,
especially
in
areas
underlain
by
clay
soils.
This
is
because
urbanisation
can
affect
summer
flows
much
more
than
winter
flows
due
to
the
fact
that
clay
soils,
when
saturated
in
the
win-
ter,
behave
much
like
an
urban
paved
surface
(Taylor
and
Roth,
1979).
In
the
summer
months
the
dry
soil
responds
very
dif-
ferently
to
rainfall.
Clearly,
remedies
are
also
needed
for
many
of
the
social
problems
discussed
earlier
that
might
be
linked
to
garden
paving.
5.
Conclusion
The
issue
of
impervious
garden
surfaces
has
attracted
the
attention
of
council
employees
who
are
involved
in
flooding
and
drainage
management
(Leeds
City
Council,
2004).
However,
this
interest
has
not
yet
resulted
in
any
efforts
to
tackle
the
prob-
lem.
The
fact
that
changes
occur
gradually
allows
the
issue
to
slip
under
the
carpet.
The
observed
0.111=
2
increase
in
impervious
gardens
and
would
have
attracted
significant
attention
were
it
paved
as
a
single
plot.
Modelled
increases
in
runoff
show
that
increases
in
imper-
viousness
have
the
potential
to
cause
serious
urban
flooding.
Combined
with
the
other
effects
of
impervious
paving,
the
observed
increases
are
likely
to
have
a
longstanding
effect
on
the
social
and
environmental
quality
of
the
study
area.
The
geo-
graphical
and
social
conditions
in
the
study
area
are
favourable
to
increases
in
garden
paving,
particularly
for
the
creation
of
off-street
parking.
Similar
increases
would
be
expected
in
other
areas
where
these
conditions
exist.
However,
the
con-
ditions
existing
in
the
study
area
are
not
believed
to
be
the
optimum
for
inducing
paving
of
garden
surfaces.
There
is
ample
on
street
parking
in
the
study
area
and
in
areas
where
this
is
limited
more
extensive
increases
in
garden
paving
may
occur.
Future
land
use
policies
for
the
study
area
should
give
significant
consideration
to
the
issue
of
impervious
gardens.
Further
increases
in
impervious
surfaces
could
cause
signifi-
cant
problems
either
through
increased
flooding
or
by
forcing
expensive
improvements
in
the
drainage
network.
The
policy
for
the
Whitkirk
conservation
area
may
provide
a
platform
for
the
issue
to
be
considered
in
the
study
area.
Nationally,
further
study
is
required
both
to
make
a
better
assessment
of
the
prob-
lem
in
a
variety
of
different
areas.
The
UK
government
recently
proposed
the
accelerated
development
of
the
new
green
space
database
(in
conjunction
with
local
authorities)
to
establish
a
consistent
baseline
on
the
total
amount
of
public
green
space
in
the
UK
and
its
distribution,
particularly
in
deprived
areas
(UK
House
of
Commons,
2006).
The
database
therefore
offers
scope
for
integration
of
information
on
domestic
garden
space
extent.
Accurate
measurement
is
also
required
to
ascertain
the
true
extent
of
the
problems
caused
in
order
to
validate
modelling
results
presented
here
and
by
Pauleit
et
al.
(2005).
Further
research
is
also
needed
into
methods
of
controlling
the
effects,
not
just
on
increased
runoff
but
on
the
wide
variety
of
other
issues
that
are
highlighted
by
this
study.
The
issue
of
impervi-
ous
garden
surfaces
does
not
appear
to
be
widely
recognised.
With
no
plans
for
changes
in
planning
laws
and
no
significant
public
awareness
campaign,
the
problem
seems
likely
to
con-
tinue
to
increase.
Rising
urban
imperviousness
appears
to
be
a
potentially
significant
cause
of
present
and
future
environmen-
tal
and
social
degradation
and
must
therefore
be
investigated
further.
Acknowledgements
A
postgraduate
study
award
by
the
UK
National
Environ-
mental
Research
Council
to
the
first
author
enabled
successful
completion
of
the
research
at
the
University
of
Leeds
and
this
is
gratefully
acknowledged.
The
authors
would
like
to
thank
the
British
Atmospheric
Data
Centre
and
the
UK
Met
Office
for
pro-
viding
access
to
the
Land
Surface
Observation
Stations
Rainfall
Data.
Finally,
the
authors
would
like
to
extend
their
thanks
to
two
anonymous
reviewers
for
their
valuable
comments
which
helped
to
improve
the
manuscript.
References
Bamsley,
M.J.,
Barr,
S.L.,
2000.
Monitoring
urban
land
use
by
earth
observation.
Surv.
Geophys.
21,
269-289.
Bhaduri,
B.,
Grove,
M.,
Lowry,
C.,
Harbor,
J.,
1997.
Assessing
long
term
hydro-
logic
effects
of
land
use
change.
J.
Am.
Water
Works
Ass.
89,
94-106.
Bhaduri,
B.,
Minner,
M.,
Tatalovich,
S.,
Harbor,
J.,
2001.
Long
term
hydrologic
impact
of
urbanization:
a
tale
of
two
models.
J.
Wat.
Resour.
Urban
Plan.
127,
13-19.
Bird,
S.L.,
Exum,
L.R.,
Alberty,
S.W.,
2000.
High
quality
impervious
cover
data.
Qual.
Assur.
8,
91-103.
Civic
Amenities
Act,
1967.
c.
69.
Crompton,
A.,
Matthews,
B.,
1970.
Soils
of
the
Leeds
District.
Adlard
and
Son,
Dorking.
Daniels,
G.D.,
Kirkpatrick,
J.B.,
2006.
Comparing
the
characteristics
of
front
and
back
domestic
gardens
in
Hobart,
Tasmania,
Australia.
Landscape
Urban
Plan.
78,
344-352.
DETR
Department
of
Environment,
Transport
and
the
Regions,
2000.
Our
towns
and
cities:
the
future
delivering
an
urban
renaissance,
avail-
able
at
the
following
website:
http://www.urban.odpm.gov.uk/whitepaper/
ourtowns/index.htm.
Dunnett,
N.,
Qasim,
M.,
2000.
Perceived
benefits
to
human
well-being
of
urban
gardens.
Horticulture
Tech.
10,
40-45.
Gaston,
K.J.,
Smith,
R.M.,
Thompson,
K.,
Warren,
P.H.,
2004.
Gardens
and
wildlife—the
BUGS
project.
Br.
Wildl.
16,
1-9.
Gaston,
K.J.,
Warren,
P.H.,
Thompson,
K.,
Smith,
R.M.,
2005.
Urban
domestic
gardens
(IV):
the
extent
of
the
resource
and
its
associated
features.
Biodivers.
Conserv.
14,
3327-3349.
Gomez,
K.J.,
Jabaloyes,
J.,
Vano,
E.,
2004.
Green
zones
in
the
future
of
urban
planning.
J.
Urban
Plan.
D.—ASCE
130,
94-100.
Harbor,
J.M.,
1994.
A
practical
method
for
estimating
the
impact
of
land-use
change
on
surface
runoff,
groundwater
recharge
and
wetland
hydrology.
J.
Am.
Plan.
Assoc.
60,
91-104.
T
Perry,
R.
Nawaz
Landscape
and
Urban
Planning
86
(2008)
1-13
13
Healey,
P.,
2004.
Hard
surfacing
of
front
gardens:
report
on
desk
research
for
London
Borough
of
Ealing,
Pene
Healey
Associates,
available
at
the
following
website:
http://www.la21.org/subjects/frontgardens/fgfinal.pdf.
Hollis,
G.E.,
1975.
The
effect
of
urbanisation
on
floods
of
different
recurrence
interval.
Water
Resour.
Res.
11,431-435.
Hollis,
G.E.,
1988.
Rain,
roads,
roofs
and
runoff:
hydrology
in
cities.
Geography
73,9-18.
Huber,
W.C.,
Dickinson,
R.E.,
1988.
Storm
water
management
model,
ver-
sion
4:
User's
manual,
Report.
No.
EPA-600/3-88-001a,
U.S.
Environmental
Protection
Agency,
Washington,
D.C.
Kellet,
J.E.,
1982.
The
private
garden
in
England
and
Wales.
Landscape
Plan.
9,105-123.
Lee,
J.G.,
Heaney,
J.P.,
2003.
Estimation
of
urban
imperviousness
and
its
impacts
on
storm
water
systems.
J.
Water
Res.
Plan.-ASCE
129,419-426.
Leeds
City
Council,
2004.
Flooding
in
East
Leeds:
12th
August
2004,
Leeds
City
Council
and
Yorkshire
Water,
available
at
the
follow-
ing
website:
http://www.leeds.gov.uk/page.aspx?pagelD=9f91647a-7025-
4dbb-bb8d-aa352c1e2f0d.
Leeds
Initiative,
2000.
National
Index
of
Multiple
Deprivation:
Leeds,
available
at
the
following
website:
http://www.leedsinitiative.org/
initiativeDocuments/2004213_46029299.pdf#search=%22%22national
%20index%20of%20multiple%20deprivation%22%22.
Lim,
K.J.,
Engel,
B.A.,
Tang,
Z.,
Muthukrishnan,
S.,
Chio,
J.,
Kim,
K.,
2006.
Effects
of
calibration
on
L-THIA
GIS
runoff
and
pollutant
estimation.
J.
Environ.
Manage.
78,35-43.
London
Assembly
Environment
Committee,
2005.
Crazy
Paving:
The
Environ-
mental
Importance
of
London's
Front
Gardens.
Greater
London
Authority,
London.
Mathieu,
R.,
Freeman,
C.,
Aryal,
J.,
2007.
Mapping
private
gardens
in
urban
areas
using
object-oriented
techniques
and
very
high
resolution
satellite
imagery.
Landscape
Urban
Plan.
81,179-192.
McCall,
A.,
Doar,
N.,
1997.
The
state
of
Scottish
greenspace.
Scot.
Natl.
Heritage
Rev.,
88.
Mitchell,
G.,
McDonald,
A.,
Lockyer,
J.,
2001.
Pollution
Hazard
from
Urban
Non-Point
Sources:
A
GIS
Model
to
Support
Strategic
Environmental
Plan-
ning
in
the
UK.
University
of
Leeds,
available
at
the
following
website:
http://www.geog.leeds.ac.uk/projects/nps/.
Nix,
S.J.,
1994.
Urban
Storm
Water
Modelling
and
Simulation.
CRC
Press,
Lewis.
Nowak,
D.J.,
Crane,
D.,
Walton,
J.T.,
Stevens,
J.,
Escobedo,
F.,
Heisler,
G.,
Hoehn,
R.,
Luley,
C.,
Bond,
J.,
Ina,
G.,
Binkley,
M.,
Tian,
L.,
Thou,
J.,
Endreney,
T.,
Wang,
J.,
Post,
C.,
2003.
A
tool
for
assessing
urban forest
structure,
functions
and
benefits.
In:
'Collin,
C.
(Ed.),
Conference
Pro-
ceedings,
2003
National
Urban
Forest
Conference. American
Forests,
pp.
69-71.
Office
for
National
Statistics,
2001.
UK
2001
Census,
available
at
the
following
website:
http://www.statistics.gov.uk/census/.
Office
for
National
Statistics,
2004.
Neighbourhood
statistics,
available
at
the
following
website:
http://neighbourhood.statistics.gov.uk/dissemination/.
Old,
G.H.,
Leeks,
G.J.L.,
Packman,
J.C.,
Smith,
B.P.G.,
Lewis,
S.,
Hewitt,
E.J.,
2006.
River
flow
and
associated
transport
of
sediments
and
solutes
through
a
highly
urbanised
catchment,
Bradford,
West
Yorkshire.
Sci.
Total
Environ.
360,98-108.
Owen,
J.,
1991.
The
Ecology
of
a
Garden.
Cambridge
University
Press,
pp.
343-369.
Pandit,
A.,
Gopalakrishnan,
G.,
1996.
Estimation
of
annual
storm
runoff
coeffi-
cients
by
continuous
simulation.
J.
Irrigat.
Drain.
Eng.
122,211-220.
Pauleit,
S.,
Duhme,
F.,
2000.
Assessing
the
environmental
performance
of
land
cover
types
for
urban
planning.
Landscape
Urban
Plan.
52,1-20.
Pauleit,
S.,
Ennos,
R.,
Golding,
Y.,
2005.
Modelling
the
environmental
impacts
of
urban
land
use
and
land
cover
change:
a
study
in
Merseyside,
UK.
Landscape
Urban
Plan.
71
(2-4),
295-310.
Purdue
Research
Foundation,
2004.
Long-term
Impact
Assessments,
Perdue
University,
Indiana,
available
at
the
following
website:
http://www.ecn.
purdue.edukunoffilthia/basic_input_file_upload.htm.
Royal
Horticultural
Society,
2005.
Front
gardens:
are
we
parking
on
our
gardens?
Gardening
Matters
Series,
Royal
Horticultural
Society,
avail-
able
at
the
following
website:
http://www.rhs.org.uk/Leaming/research/
gardeningmatters/documents/FrontGardens.pdf.
RSPB-Royal
Society
for
the
Protection
of
Birds,
2006.
Britain's
best
loved
birds
starving
to
death,
available
at
the
following
website:
http://www.rspb.org.uk/feedthebirds/birdsstarving.asp.
Sanders,
R.A.,
1986.
Urban
vegetation
impacts
on
the
hydrology
of
Dayton
Ohio.
Urban
Ecol.
9,361-376.
Schuster,
W.D.,
Bonta,
J.,
Thurston,
H.,
Wamemuende,
E.,
Smith,
D.R.,
2005.
Impacts
of
impervious
surface
on
watershed
hydrology:
a
review.
Urban
Water
J.
2,263-275.
Shaw,
E.M.,
1994.
Hydrology
in
Practice,
third
ed.
Stanley
Thomes,
Chel-
tenham,
pp.
315-341.
Tang,
Z.,
Engel,
B.A.,
Pijanowski,
B.C.,
Lim,
K.J.,
2005.
Forecasting
land
use
change
and
its
environmental
impact
at
a
watershed
scale.
J.
Environ.
Manage.
76,35-45.
Taylor,
C.H.,
Roth,
D.M.,
1979.
Effects
of
suburban
construction
on
runoff
contributing
zones
in
a
small
southern
Ontario
drainage
basin.
Hydro.
Sci.
J.
24
(3),
289-301.
Thomas,
A.D.,
1983.
Planning
in
residential
conservation
areas.
Prog.
Plan.
20,
173-256.
Thompson,
K.,
Austin,
K.C.,
Smith,
R.M.,
Warren,
P.H.,
Angold,
P.
G.,
Gaston,
K.J.,
2003.
Urban
domestic
gardens:
putting
small-scale
plant
diversity
in
context.
J.
Veg.
Sci.
14,71-78.
UK
House
of
Commons,
2006.
Public
accounts-58th
Report,
available
at
the
following
website:
http://www.publications.parliament.uk
.
UK
Town
and
Country
Planning
Order,
1995.
General
Permitted
Development
Order,
Staturtory
Instrument
No.
418,
Article
4.
UNDP-United
Nations
Development
Programme,
2000.
World
Resources
2000-2001:
People
and
Ecosystems-The
Fraying
web
of
Life.
Elsevier
Science,
Amsterdam,
The
Netherlands.
USDA
(United
States
Department
of
Agriculture
Natural
Resources
Conser-
vation
Service),
1986.
Urban
Hydrology
for
Small
Watersheds,
Technical
Release
55,
second
ed.,
NTIS
PB87-101580,
Springfield,
Virginia,
also
available
on
the
web
in
PDF
format
at
ftp://ftp.wcc.nrcs.usda.
gov/downloads/hydrology_hydraulics/tr55/tr55.pdf.
US
Department
of
Agriculture,
Natural
Resources
Conservation
Service,
2007.
National
Soil
Survey
Handbook,
title
430-VI,
available
online:
http://soils.usda.gov/technicaVhandbook/.
Weng,
Q.,
2001.
Modeling urban
growth
effects
on
surface
runoff
with
the
integration
of
remote
sensing
and
GIS.
Environ.
Manage.
28,737-748.
Whitford,
V.,
Handley,
J.,
Ennos,
R.,
2001.
City
form
and
natural
process:
indi-
cators
for
the
ecological
performance
of
urban
areas.
Landscape
Urban
Plan.
57,91-103.
Zmyslony,
J.,
Gagnon,
D.,
1998.
Residential
management
of
urban
front-yard
landscape:
a
random
process?
Landscape
Urban
Plan.
40,295-307.
Zoppou,
C.,
1999.
Review
of
stormwater
models,
CSIRO
Land
and
Water,
Canberra,
Technical
Report
52/99.