Outline

  1. The role of green roofs in climate change mitigation. A case study in Seville (Spain)
  2. 1. Introduction
  3. 2. Methods
  4. 2.1. Regionalized scenarios of climate change for Spain
  5. 2.2. Relationship between land surface temperature and the normalized difference vegetation index (NDVI)
  6. 2.3. Location of the study area
  7. 2.4. Estimation of the necessary green roof area
  8. 3. Results
  9. 4. Discussion
  10. 5. Conclusions
  11. Acknowledgements
  12. Notation
  13. References
Previous PDF in this issueNext PDF in this issue
The
role
of
green
r
oofs
in
climate
change
mitigation.
A
case
study
in
Seville
(Spain)
Sergio
S.
Herrera-Gomez
a
,
Abel
Quevedo-Nolasco
a
,
Luis
P
erez-U
rrestarazu
b
,
*
a
Programa
de
Hidrociencias,
Colegio
de
Postgraduados,
Ctra.
M
exico-
T
excoco
km.36.5,
56230,
Montecillo,
T
excoco,
Mexico
b
Urban
Greening
&
Biosystems
Engineering
Research
Group,
Area
of
Agro-Forestry
Engineering,
Universidad
de
Sevilla,
ETSIA,
Ctra.
Utrera
km.
1
,
4
1
0
1
3,
Seville,
Spain
article
info
Article
history:
Received
30
May
201
7
Received
in
revised
form
24
July
20
1
7
Accepted
25
July
201
7
A
vailable
online
27
July
201
7
Keywords:
Urban
greening
Remot
e
sensing
Heat
island
Normalized
difference
vegetation
index
abstract
The
intense
anthropogenic
urban
activity
generates
a
phenomenon
known
as
heat
island,
which
is
related
with
high
temperatures
in
cities,
as
comp
ared
against
adjacent
suburban
and
rural
areas.
Due
to
this
effect,
the
comfort
conditions
of
the
citizens
deteriorate.
In
the
case
of
the
city
of
Seville
(Spa
in),
several
models
of
climate
change
forecast
increases
in
the
maximu
m
temperatures
ranging
from
1
.5
to
6
C
in
summer
.
This
article
explores
the
role
of
green
roofs
as
a
supplemen
t
to
the
green
spaces
of
the
city
,
in
order
to
buffer
the
nega
tive
effects
of
the
increase
of
the
maximu
m
temperatures
due
to
climate
change.
Images
from
the
Landsat
7
ETMþ
and
Sentinel-2
satellites
have
been
used
in
order
to
verify
the
inverse
relationship
between
land
surface
temperature
and
the
abundance
of
vegetation,
expressed
by
the
normalized
difference
vegetation
index.
For
Seville,
a
green
roof
surface
of
7
40
ha
shoul
d
be
implemented,
in
the
most
adverse
scenario,
which
means
cov
ering
40.6%
of
the
existing
buildings.
In
the
most
optimistic
scenario,
the
forecasted
green
roof
surface
required
is
207
ha
(1
1
.3%
of
the
roofs).
©
201
7
Elsevier
Ltd.
Al l
rights
reserved.
1
.
Introduction
Numerou
s
studies
on
climate
change
predict
a
global
rise
in
temperatur
es.
The
conseq
uences
of
this
increase
will
be
more
troublesome
in
urban
areas,
where
the
temperatures
are
already
higher
than
in
surrounding
rural
areas.
This
heating
phenomenon
is
mostly
due
to
anthropogenic
development
in
the
urban
area
[1]
and
the
increase
of
building
covered
areas
[2].
The
construction
materials
commonly
used
absorb
most
of
the
radiation
and
release
it
as
heat.
This
generates
the
urban
heat
island
phenomenon,
which
has
direct
and
indirect
impacts
on
the
health
and
life
quality
of
the
citizens
[3].
Urban
heat
islands
vary
in
magnitude
and
structure
according
to
two
main
groups
of
factors:
climatological
factors
(such
as
climatic
region,
season,
time
of
day
,
synoptic
conditions
and
wind
regime)
and
those
related
to
the
physical
and
human
nature
of
the
built
environmen
t,
such
as
geographic
location,
topograph
y
,
urban
landscape
geometry,
type
of
building
materials
and
intensity
of
human
activities
[4].
In
fact,
a
study
aiming
to
identity
heat
islands
at
different
height
levels
conducted
in
T
el-A
viv
(Israel)
showed
that
parks
and
open
areas
were
the
coldest
ele-
ments
within
the
city
during
day
and
night
[5].
There
is
a
clear
correlation
between
plant
cover
and
land
surface
temperature
[6,7],
and
consequentl
y
,
an
urban
increase
in
green
areas
would
contribute
to
mitigate
the
Heat
Island
[8].
Nevertheless,
in
many
modern
cities,
there
is
a
high
density
of
building
cover
ed
areas
which
does
not
allow
raising
the
number
of
green
areas.
Thus,
in
order
to
increase
the
presence
of
urban
vegetation,
it
is
necessary
to
dra
w
on
systems
implemented
on
existing
buildings.
Currently
,
the
sum
of
all
the
building
roofs
represents
a
high
percentage
of
exposition
in
urban
areas.
Estimations
for
dense
cities
prove
that
the
fraction
of
roof
area
varies
between
20
and
25%
of
the
total
area
[9].
Because
of
this,
the
use
of
these
surfaces
to
increase
urban
vegetation
is
an
interesting
option.
Green
roofs
are
urban
greening
systems
that
precisel
y
allow
installing
plant
life
in
the
roofs
of
buildings
through
more
or
less
complex
elements.
They
can
be
extensive,
lighter
,
and
with
less
substrate
when
establishing
smaller
species,
or
more
intensive
and
heavier
with
greater
amount
of
substrate
where
small
trees
and
shrubs
can
be
included
[2].
Green
roofs
have
existed
for
more
than
a
thousand
years,
although
their
use
has
become
more
relevant
in
modern
times
and
new
technical
solutions
that
fav
or
their
imple-
mentation
have
appeared.
This
development
has
come
about
since
*
Corresponding
author.
E-mail
addresses:
hserch@colpos.mx
(S.S.
Herrera-Gomez),
anolasco@colpos.mx
(A.
Quevedo-Nolasco),
lperez@us.es
(L.
P
erez-Urr
estarazu).
Contents
lists
available
at
ScienceDirect
Building
and
Environment
journal
homepage:
www.elsevier.com/locate/buildenv
http://dx.doi.org/1
0.
1
01
6/j.buildenv
.20
1
7
.07
.036
0360-1
323/©
201
7
Elsevier
Ltd.
All
rights
reserved.
Building
and
Environment
123
(201
7)
575e584
Previous PDF in this issueNext PDF in this issue
not
only
do
they
provide
a
nice
relaxing
space
or
scenery
,
but
also
ecosystems
services
such
as
microclimate
regulation,
rainw
ater
management,
improv
ed
building
insulation
(with
an
influence
on
inner
temperature),
noise
absorption,
decrease
of
air
pollution,
and
biodiversity
enhancing
[3,
1
0].
Moreover
,
they
contribute
to
increasing
the
albedo
of
urban
areas
[1
1].
Man
y
studies
on
green
roofs
are
oriented
to
their
capability
to
regulate
temperature.
However
,
depending
on
the
climate
and
the
type
of
green
roof
(different
plant
material,
substrate,
and
con-
struction
features),
their
efficiency
can
vary
[1
2].
The
thermal
effi-
cacy
of
a
green
roof
is
closely
related
with
the
climate,
and
it
becomes
more
significant
when
the
environmen
tal
temperature
rises
[3].
This
efficacy
is
measured
from
the
point
of
view
of
energy
savings
in
warm
areas
for
their
capacity
to
lower
temperatures
[2]
of
both
the
roof
surface
and
the
air
above
it
[1
3].
For
example,
an
analysis
of
the
surface
temperature
before
and
after
the
placement
of
a
green
roof
in
Singapore
showed
a
significant
decrease
once
the
green
roof
w
as
installed,
especially
for
high
plant
cover
,
making
the
maximum
temperature
difference
approximat
ely
1
8
C
[1
4].
Another
study
in
Hong
Kong
proved
that
the
heat
stored
in
a
bare
roof
was
75%
higher
than
that
of
a
green
roof
[1
5].
In
the
city
of
Chicago,
the
temperatures
in
summer
of
the
surface
of
a
green
roof
and
a
neighboring
building
were
compared.
The
temperature
of
the
green
roof
varied
from
33
to
48
C,
while
in
the
conventional
dark
roof
of
the
adjacent
building
the
temperature
was
76
C.
The
air
temperature
near
the
surface
of
the
green
roof
was
4
C
lower
than
near
the
conventional
roof
[1
6].
This
decre
ase
in
temperature
happens
because,
in
a
green
roof,
the
flux
of
sensible
heat
is
low
due
to
the
high
latent
heat
flux
from
evaporation,
even
if
the
net
radi-
ation
is
high.
This
works
to
lower
the
tempera
ture
in
a
specific
area
[1
7].
Also,
some
simulation
studies
indicate
that
green
roofs
can
decrease
the
mean
environmental
temperature
from
0.3
to
3
Ca
ta
city
scale,
and
drastically
decrease
the
heat
island
effect
[2].
Now
adays,
in
many
cities
of
several
countries,
such
as
German
y
,
the
U.S.A.,
Denmark,
and
Canada,
their
gov
ernments
hav
e
devel-
oped
a
variety
of
norms,
incentives,
and
technical
services
to
pro-
mote
the
naturalizing
of
roofs
[1
8].
These
measures
will
foster
the
increase
of
the
area
covered
by
green
roofs,
which
will
ha
ve
favor
able
consequences
on
the
specific
climatic
conditions
in
the
urban
areas
where
they
are
installed.
In
fact,
the
mass
installment
of
green
roofs
might
work
as
a
mechanism
to
decrease
the
Heat
Island
effect
and
counteract
the
temperature
increase
due
to
climate
change.
In
order
to
do
so,
the
remaining
question
would
be
how
much
surface
would
be
needed
to
mitigate
the
effect
of
climate
change.
This
is,
precisely
,
the
main
objective
of
this
study
.
This
is
wh
y
the
effect
on
the
temperature
when
increasing
the
vegetation
areas
by
means
of
green
roofs
in
the
city
of
Seville
(Spain)
is
assessed.
2.
Methods
2.
1
.
Regionalized
scenarios
of
climate
change
for
Spain
The
National
Plan
of
Adap
tation
to
Climate
Change
(Plan
Nacional
de
Adaptaci
on
al
Cambio
Climatico
e
PN
ACC)
in
Spain
has
a
priority
work
line
for
developing
regionalized
climate
change
scenarios,
which
consists
of
generating
and
making
publicly
avail-
able
a
collection
of
scenarios
that
projects
how
climate
change
will
be
manifested
throughout
the
21st
century
in
Spain
[1
9].
For
this
study
case,
the
average
evolution
of
the
mean
annual
and
monthly
maximum
temperatur
es
in
the
Andalusia
region
is
used.
These
are
calculated
for
the
three
30-year
periods
comprised
between
20
01
and
21
0
0.
Herein
are
included
comparisons
with
the
different
global
models
used,
with
different
methods
of
regionalization,
and
according
to
the
SRES-IPCC
emission
scenarios
[1
9].
The
emission
scenarios
are
useful
in
the
analysis
of
climate
change,
particularly
to
create
climate
models,
to
evaluate
impact,
and
for
initiatives
on
adaptation
or
mitigation
[20].
The
A2
family
of
evolutionary
lines
and
scenarios
used
in
the
present
study
de-
scribes
a
very
heterogeneous
world.
Its
most
distinctive
features
are
self-sufficiency
and
preservation
of
local
identities.
The
fertility
patterns
in
the
group
of
regions
converge
very
slowly
,
thus
obtaining
an
ever
growing
world
population.
Economic
develop-
ment
is
basically
oriented
toward
regions,
while
per
capita
eco-
nomic
growth
and
technological
changes
are
more
fragmented
and
slower
than
other
evolutionary
lines
[20].
2.2.
R
elationship
between
land
surface
temper
ature
and
the
normalized
difference
vegetation
index
(NDVI)
The
NDVI
is
a
numerical
index
used
in
remote
sensing
analysis
to
ev
aluate
if
a
determined
objective
contains
live
vegetation.
Health
y
plant
life
absorbs
visible
light
(0.4e0.7
m
m)
and
reflects
near
infrared
light
(0.7e1.1
m
m).
Scarce
or
unhealth
y
vegetation
generally
reflects
more
visible
light
and
less
near
infrared
[2
1].
Therefore,
more
near
infrared
radiation
reflected
than
visible
light
wa
velengths
generally
indicates
the
presence
of
green
vegetation,
whereas
a
small
difference
in
the
intensity
between
these
two
wa
velengths
is
generally
an
indicator
of
scarce
vegetation
or
sur
-
faces
devoid
of
plant
life
[22].
Since
the
near
infrared
and
red
bands
of
satellites
are
the
most
sensitive
to
information
from
vegetation,
these
bands
can
be
used
to
quantify
the
growth
density
of
a
plant
in
a
given
pixel
[22].
This
index
has
been
used
in
some
works
to
determine
the
abundance
of
plant
life
in
a
determined
area
[23].
The
ND
VI
values
vary
from
0.0
to
1
.0,
depending
on
the
degree
of
plant
cover
and
the
ph
ysiological
state
o
health
of
the
plants.
An
ND
VI
over
0.2
indicates
the
presence
of
vegetation,
depending
on
the
amount
and
health
state
of
the
plant
cover
[24,25].
According
to
[26]
and
[27],
areas
with
high
ND
VI
values
can
lower
land
surface
temperatur
e
(T
s
).
This
correlation
is
due
to
the
influence
of
the
humidity
in
the
ground
and
evapotr
anspiration
of
plants
on
the
surface.
The
values
of
both
variables
can
be
fitted
to
a
linear
model
which
describes
an
inverse
dependence
between
T
s
and
the
NDVI
[23],
so
the
temperature
is
de fi ned
by
the
following
general
expression:
T
s
¼
-x
*
(ND
VI)
þ
y
(1)
Therefore,
the
increase
in
temperature
due
to
climate
change
(
D
T
max
)
could
be
mitigated
by
a
substantial
increase
in
the
mean
value
of
the
NDVI
(NDVI
mean
);
that
is
to
say,
increasing
the
current
plant
cover
(NDVI
CC
)(
Fig.
1).
Therefore,
ND
VI
CC
is
obtained
as
follows:
ND
VI
CC
¼
NDVImean
þ
[
D
Tmax*
(NDVImax-
ND
VImin)/(Tmax
-
Tmin)]
(2)
where
NDVI
mean
is
the
mean
of
the
NDVI
values
obtained
from
the
satellite
image
for
the
whole
study
area.
D
T
max
are
the
degrees
that
the
maximum
temperature
increases
due
to
climate
change;
these
values
are
provided
by
the
[1
9],
and
are
shown
in
T
able
2
.N
D
V
I
max
and
NDVI
min
,
are
the
maximum
and
minimum
NDVI
values
ob-
tained
from
processing
the
satellite
image
of
the
whole
study
area.
T
max
and
T
min
,
are
the
maximum
and
minimum
temperatur
e
values
obtained
from
processing
the
satellite
image
of
the
whole
study
area.
An
image
from
the
United
States
Geological
Survey
(USGS)
taken
by
the
Sentinel-2
satellite
on
August
3
1st,
20
1
6
was
used
to
calculate
the
T
s
and
NDVI
in
the
study
area.
It
was
acquired
at
around
1
4:00
local
time
with
atmospheric
conditions
of
0%
cloud
S.S.
Herrera-Gomez
et
al.
/
Building
and
Environment
1
23
(201
7)
575e584
576
cover
[28].
At
the
time
the
image
was
taken,
T
max
was
31
.
8
C
and
T
min
was
20.6
C.
The
calculation
of
NDVI
(Fig.
2)
was
done
from
the
bands
in
the
Sentinel-2
that
capture
red
(0.6
65
m
m)
and
near
infrared-NIR
(0.842
m
m)
colors
(bands
4
and
8,
respectively)
[29].
The
results
obtained
were
compared
against
those
from
another
image
acquired
by
the
Landsat
7
ETM
þ
on
May
1
8th,
20
03.
T
o
calculate
T
s
,
the
inverse
Planck
function
was
used
[30],
which,
for
satellite
images,
can
be
e
xpressed
as
[31]:
T
s
¼
K
2
/
{Ln
[(ε
NB
K
1
/R
c
)
þ1]}
(3)
where
T
s
is
land
surface
temperature
in
degrees,
R
c
is
the
corrected
thermal
radiance
of
the
surface,
and
K
1
and
K
2
are
constants
for
satellite
images
607
.7
6
and
1
260.56,
respectiv
ely
,
in
Wm
2
sr
1
m
m
1
.
ε
NB
is
the
emissivity
that
represents
the
performance
of
the
surface
for
thermal
emission
(1
0,4
to
1
2,5
m
m).
The
units
for
R
c
must
be
the
same
as
for
K
1
(Wm
2
sr
1
m
m
1
).
The
corrected
thermal
radiance
of
the
surface
R
c
was
used,
through
the
equation
developed
by
Ref.
[32]:
R
c
¼
[(L
6
-R
p
)/
t
NB
]
e
[(1-
ε
NB
)R
sky
]
(4)
Fig.
1
.
Relationship
between
T
s
and
NDVI
in
function
of
urban
green
surface
(S
ug
),
and
the
urban
green
areas
necessary
to
stabilize
the
effects
of
temperature
increases
due
to
climate
change
(S
ug
CC
).
(For
interpretation
of
the
references
to
colour
in
this
figure
legend,
the
reader
is
referred
to
the
web
version
of
this
article.)
Fig.
2.
NDVI
obtained
from
the
image
by
Sentinel-2
of
the
city
of
Seville.
S.S.
Herrera-Gomez
et
al.
/
Building
and
Environment
1
23
(201
7)
575e584
577
Where
L
6
is
the
spectral
radiance
of
band
6
(Wm
2
sr
1
m
m
1
),
R
p
is
the
radiance
register
ed
in
the
band,
1
0.4e
1
2.5
m
m,
R
sky
is
the
low
thermal
radiance
of
the
narrow
band
for
a
clear
sky
(Wm
2
sr
1
m
m
1
),
t
NB
is
the
transmissivity
of
air
in
the
narrow
band
(1
0.4e
12.5
m
m)
[33].
suggest
Rp
¼
0.91
,
t
NB
¼
0.866,
and
R
sky
¼
1
.32.
ε
0
is
the
emissivity
that
represents
the
performance
of
the
surface
for
the
thermal
emission
of
the
wide
thermal
spectrum
(6e14
m
m).
Emissivities
were
calculated
according
to
[34]:
ε
NB
¼
0.97
þ
0.033
LAI;
for
LAI3
ε
0
¼
0.95
þ
0.01
LAI;
for
LAI3
ε
NB
¼
0.98
and
ε
0
¼
0.981;
for
LAI>3
LAI
(Leaf
Area
Index)
is
the
total
one-sided
leaf
area
per
unit
ground
surface
area.
LAI
is
calculated
using
the
following
empirical
equation
developed
by
Ref.
[35].
LAI
¼
1
1*
SA
VI;
when
SA
VI
0.81
7
LAI
¼
6;
for
SA
VI
>0.81
7
SA
VI
is
an
index
that
tries
to
subtract
the
effect
of
the
soil
background
of
the
NDVI,
reducing
the
impact
of
soil
humidity
in
this
index.
SA
VI
¼
(1
þ
L)(
r
t,4
-
r
t,8
)/(
Lþ
r
t,4
þ
r
t,8
)
(5)
L
is
a
constant
for
SA
VI.
If
L
equals
zero,
SA
VI
is
equal
to
ND
VI.
For
this
research,
we
used
L
¼
0.5.
ND
VI
¼
(
r
t,4
-
r
t,8
)/(
r
t,4
þ
r
t,8
)
(6)
Where:
r
t,4
and
r
t,8
,
are
reflectivities
of
the
satellite
for
bands
4
and
8,
respectively
.
Equation
(1)
can
be
used
at
the
local
or
regional
level
and
the
slope
of
the
line
varies
depending
on
the
resolution
of
the
satellite
images
used
in
the
process
to
obtain
NDVI;
in
the
present
case,
Landsat
7
ETM
þ
obtains
images
with
a
resolution
of
30
m,
and
Sentinel-2
images
with
1
0
m.
2.3.
Location
of
the
study
area
Seville
is
a
city
located
in
the
region
of
Andalusia
in
southern
Spain.
In
the
last
20
years,
the
urban
area
of
Seville
has
experienced
an
important
growth
in
population,
becoming
an
expanding
metropolitan
area
[36].
Its
1
40
km
2
urban
area
at
the
moment
has
a
population
of
around
700,0
00
inhabitants,
with
a
population
density
of
about
50
0
0
citizens/km
2
.
According
to
Ramirez
et
al.
(201
3),
the
urban
design
of
the
city
of
Seville
represents
an
eco-design
with
a
central
urban
area
and
radiocentric
growth,
where
the
recently
built
disjointed
areas
hav
e
slowly
been
incorporated
(Fig.
3).
All
the
green
areas
in
Seville
add
up
to
890
ha
[37].
Seville
has
the
largest
total
green
surface
area
per
capita
in
Andalucía,
as
well
as
a
surrounding
green
ring,
although
it
is
not
close
enough
to
the
population
[38].
However
,
the
number
of
currently
existing
green
roofs
is
practically
null.
The
urban
center
of
the
city
of
Seville
has
an
area
of
61
90
ha
[39].
Qgis
2.
1
4
(Open
Source
Geospatial
Foundation)
software
is
used
to
estimate
the
building
covered
area,
as
well
as
the
geographic
and
urbanism
information
system
of
Seville
[40].
The
vectorial
shape-
files
of
buildings,
mass,
and
territorial
limits
of
the
city
of
Seville
are
over
laid
with
the
NDVI
shapefiles
obtained
from
processing
the
satellite
image
from
Sentinel-2,
discriminating
the
green
and
agricultural
areas
with
NDVI
greater
than
0.2.
In
Seville,
there
are
two
main
heat
islands,
in
the
east
and
the
southeast
parts,
both
made
up
of
polygons
of
buildings
with
“ in-
dustrial,
commercial,
public,
military,
and
private
unit”
infrastruc-
ture
and
scarce
vegetation.
This
type
of
polyg
on
presents
inverse
relationships
between
T
s
and
the
abundance
of
vegetation;
that
is,
the
higher
the
temperature,
the
less
the
established
vegetation
[39].
The
data
from
Ref.
[39],
regarding
the
polygons
of
urban
green
spaces
(S
ug
)
that
consider
small
agricultural,
park,
and
public
gar-
den
areas,
were
used
and
compared
against
temperature
and
NDVI
values
obtained
from
images
from
the
Sentinel-2
satellite.
2.4.
Estimation
of
the
necessary
green
roof
area
Every
representativ
e
ND
VI
value
corresponds
to
an
estimated
value
of
urban
green
spaces
(S
ug
)
in
the
city
of
Seville.
S
ug
now
represents
1
4%
of
the
urban
area
of
the
city
of
Seville
(Farina,
20
1
2)
[39].
Under
the
assumption
that
S
ug
does
not
decrease
in
the
future,
S
ug
CC
proportionally
depends
on
D
ND
VI.
Therefore,
at
the
end
of
each
of
the
periods
considered,
S
ug
CC
is
obtained
from
NDVI
CC
so
that:
S
ug
CC
¼
S
ug
*(
1
þ
D
ND
VI)
(7)
where:
D
ND
VI
is
the
relative
increment
of
NDVI
CC
with
respect
to
the
current
value
(NDVI
mean
).
For
this
case,
ND
VI
CC
is
alw
ays
higher
than
NDVI
mean
.
D
NDVI
is
expre
ssed
as:
D
ND
VI
¼
(NDVI
CC
-N
D
V
I
mean
)/NDVI
mean
(8)
At
the
same
time:
S
ug
CC
¼
S
ug
þ
D
S
ug
(9)
where
D
S
ug
is
the
vegetated
surface
that
would
need
to
be
created
to
buffer
the
temperatur
e
increase
produced
by
climate
change.
In
this
case,
no
change
is
contemplated
regarding
the
current
surface
of
the
conventional
green
areas,
so
this
increase
would
only
correspond
to
the
increase
of
the
surface
through
green
roofs.
3.
R
esults
Within
the
study
area
(Fig.
3),
the
to
tal
building
covered
surface
is
estimated
to
be
1
820
ha,
where
different
areas
were
established
in
function
of
their
characteristics
and
use:
housing,
commercial,
public
services
and
equipping,
and
green
areas.
Fig.
4
shows
an
example
of
this
categorization.
The
temperature
values
near
the
ground
and
the
NDVI,
obtained
by
Ref.
[39]
from
a
Landsat
7
ETM
þ
image,
were
digitalized
using
the
Graph
Expert
Professional
1
.3
software
into
poly
gons
classi
fied
for
green
areas.
These
data
were
complemented
with
those
ob-
tained
from
the
Sentinel-2
image,
and
confirm
an
inverse
rela-
tionship
between
the
calculated
temperature
and
the
ND
VI.
The
values
obtained
by
Ref.
[39]
present
a
better
fit
to
a
linear
model,
R
2
¼
0.4
4,
than
the
data
obtained
in
this
study
,
R
2
¼
0.33
(Fig.
5).
The
ND
VI
and
temperature
values
correspond
to
poly
gons
of
dense
urban
fabric
with
plant
cover
(park
s
and
gardens)
and
semi-
dense
urban
fabric
with
scarce
agricultural
or
forest
areas.
The
poly
gons
with
ND
VI
values
under
0.2
were
differentiated
as
they
are
areas
with
a
low
percentage
of
plant
cover
and
a
high
density
of
building
cover
.
T
o
verify
the
reliability
of
the
temperature
values
obtained,
the
temperature
readings
from
agroclimatic
stations
of
Seville
were
compared
against
those
calculated
(T
mean
/T
s
)
in
residential
S.S.
Herrera-Gomez
et
al.
/
Building
and
Environment
1
23
(201
7)
575e584
578
polyg
ons
and
those
with
vegetation
on
the
dates
that
the
satellite
images
were
taken
(Fig.
6).
The
T
mean
values
were
obtained
from
six
agroclimatic
stations
in
Seville,
which
are
distributed
in
different
places:
with
vegetation,
with
dense
urban
fabric,
and
semi-dense
urban
fabric.
For
each
of
the
polygons,
including
the
areas
with
urban
fabric,
housing,
factories,
and
plant
cover
of
different
heights
and
density
,
mean
temperatur
es
and
ND
VI
mean
values
we
re
obtained
from
Fig.
3.
Delimitation
of
the
study
area
(urban
area)
within
the
Municipality
of
Seville.
Fig.
4.
Example
of
the
determination
of
the
different
categories
within
the
study
area.
S.S.
Herrera-Gomez
et
al.
/
Building
and
Environment
1
23
(201
7)
575e584
579
processing
the
Sentinel-2
satellite
image.
As
an
example,
T
able
1
shows
some
representative
polygons
of
the
city
of
Seville
and
their
corresponding
values.
The
green
areas
conformed
by
bushes,
shrubs
and
trees
tend
to
present
higher
NDVI
values
(0.6e0.62),
and
therefore,
cooler
temperatures.
The
mean
NDVI
values
obtained
through
processing
the
Sentinel-2
images
range
between
0.2
and
0.62.
The
maximum
ND
VI
max
values
range
between
0.4
and
0.55
(Landsat
7
ETM
þ
image)
and
0.4
and
0.62
(Sentinel-2).
Using
the
Landsat
7
ETM
þ
image,
the
T
max
and
T
min
values
observed
are
40
C
and
34
C,
respectively
.
In
the
case
of
the
Sentinel-2
image,
these
values
are
39
C
and
30
C(
Fig.
7).
In
either
case,
the
tendency
to
decrease
the
temperature
as
the
NDVI
rises
is
confirmed.
T
able
2
shows
the
results
of
each
of
the
variables
used
to
esti-
mate
the
necessary
surface
of
green
roofs
to
cover
in
order
to
mitigate
the
temperature
increase
due
to
climate
change
for
different
SERES
A2
climatic
scenarios
and
global
models
for
three
periods
of
years
obtained
from
both
the
processed
satellite
images,
Landsat
7
ETMþ
and
Sentinel-2.
In
general,
less
vegetated
surface
is
needed
when
using
the
image
from
the
Landsat
(367e
682
ha)
than
from
the
Sentinel
(420e7
40
ha),
depending
on
the
model
and
cli-
matic
scenario.
For
the
three
global
climatic
models
analyzed
in
the
20
1
1e
2040
period,
in
the
most
optimistic
scenario,
the
forecasted
increase
in
temperatur
e
is
1
.5
C,
which
makes
it
necessary
to
implement
a
green
roof
surface
of
207
ha.
Therefore,
1
1
.3%
of
the
current
roofs
would
have
to
be
cover
ed
with
vegetation.
At
the
other
extreme,
for
the
most
adverse
scenario,
the
ECHAM4-A2_INM
global
model
for
the
2071
e2
1
0
0
period,
the
increase
in
tempera-
ture
would
be
6.5
C.
In
this
case,
the
percentage
of
green
roofs
considering
the
current
built
surface
in
the
city
of
Seville
would
have
to
reach
37
.4%
(682
ha)
according
to
the
Landsat7
ETM
þ
image,
and
40.6%
(7
40
ha)
in
the
case
of
Sentinel-2.
4.
Discussion
In
this
paper
,
the
surface
of
green
roofs
needed
necessary
to
Fig.
5.
Relationship
between
T
s
and
NDVI
for
(a)
Landsat
7
ETMþ
and
(b)
Sentinel-2.
Fig.
6.
Comparison
of
mean
temperatures
in
the
city
of
Seville
(T
mean
/T
s
)
measured
in
agroclimatic
stations
against
those
obtained
from
the
Landsat
7
ETMþ
and
Sentinel-2
satellite
images.
S.S.
Herrera-Gomez
et
al.
/
Building
and
Environment
1
23
(201
7)
575e584
580
mitigate
the
impact
of
climate
change
on
the
temperature
in
the
case
of
the
city
of
Seville
is
estimated.
T
o
do
so,
the
increase
in
green
areas
necessary
for
this
aim
is
calculated
taking
into
account
the
relationship
between
the
T
s
and
ND
VI
values
obtained
from
the
Sentinel-2
and
Landsat
7
ETM
þ
images.
The
differences
between
the
temperature
values
observed
in
each
point
with
each
of
the
images
mentioned
are
because
they
were
taken
on
different
dates,
May
1
8th,
2003,
and
August
31st,
201
5,
respectively
.
Also,
the
variation
in
the
maximum
NDVI
values
obtained
with
the
Landsat
7
ETM
þ
image
and
those
from
the
Sentinel-2
image
could
be
due
to
the
difference
in
resolution
of
the
images
used.
The
Landsat
images
have
a
resolution
of
30
m,
while
those
of
the
Sentinel-2
satellite,
T
able
1
Examples
of
T
s
,
NDVI,
area
and
type
of
vegetation
for
representative
polygons
in
Seville.
Representative
polygons
with
plant
cover
T
s
(
C)
NDVI
mean
Area
(ha)
a
Type
of
vegetation
a
Agricultural
plot
in
the
northern
area
38
0.4
1.4
Crops
María
Luisa
park
31
0.60
34.0
Grass,
bushes,
shrubs,
and
trees
Príncipes
park
31
0.61
10.8
Grass,
bushes,
shrubs,
and
trees
Jardines
del
Prado
San
Sebasti
an
31
0.62
5.8
Bushes,
shrubs,
and
trees
Gardens
of
the
Real
Alc
azar
31
0.61
8.0
Grass,
bushes,
shrubs,
and
trees
Santa
Justa
roundabout
37
0.62
0.07
Grass,
shrubs,
and
trees
Alameda
de
H
ercules
(tree
walkway)
37
0.6
2.0
Shrubs
and
trees
S
anchez
Pizjuan
Stadium
football
field
33
0.5
0.8
Grass
a
Provided
by
the
Parks
and
Gardens
Services
of
the
Seville
City
Hall.
T
able
2
Estimation
of
the
green
roof
surface
necessary
to
mitigate
D
T
due
to
climatic
change.
Period
SERES
climatic
scenario
T
CC
max
(
C)
D
T
max
(
C)
Landsat
7
ETMþ
Sentinel-2
NDVI
CC
(dim)
S
ug
CC
(ha)
D
A
gr
(ha)
Percentage
of
roofs
to
vegetate
(%)
NDVI
CC
(dim)
S
ug
CC
(ha)
A
gr
(ha)
Percentage
of
roofs
to
vegetate
(%)
2011e2040
A2
a/
35.5
3.5
0.47
1257
367
20.1
0.56
1310
420
23.1
2041e2070
A2
a/
34.0
5.0
0.52
1414
524
28.8
0.63
1470
580
31.8
2071e2100
A2
a/
33.0
6.0
0.56
1519
629
34.5
0.67
1577
687
37.7
Global
models
2071e2100
HadAM3H_A2-FIC
a/
34.5
4.5
0.51
1362
472
25.9
0.61
1417
527
28.9
2071e2100
HadCM3_A2-SDSM
a/
34.0
5.0
0.52
1414
524
28.8
0.63
1470
580
31.8
2071e2100
HadAM3H_A2-INM
a/
34.0
5.0
0.52
1414
524
28.8
0.63
1470
580
31.8
2071e2100
HadAM3H_A2-RCM
a/
34.0
5.0
0.52
1414
524
28.8
0.63
1470
580
31.8
2011e2040
CGCM2_A2_FIC
b/
33.5
1.5
0.39
1047
157
8.6
0.47
1097
207
11.3
ECHAM4-A2_INM
b/
33.5
1.5
0.39
1047
157
8.6
0.47
1097
207
11.3
HadCM3_A2_SDSM/INM
b/
33.5
1.5
0.39
1047
157
8.6
0.47
1097
207
11.3
2041e2070
CGCM2_A2_FIC
b/
32.0
3.0
0.45
1205
315
17.3
0.54
1257
367
20.1
ECHAM4-A2_INM
b/
31.0
4.0
0.49
1310
420
23.0
0.58
1364
474
26.0
HadCM3_A2_SDSM/INM
b/
31.5
3.5
0.47
1257
367
20.1
0.56
1310
420
23.1
2071e2100
CGCM2_A2_FIC
b/
31.0
4.0
0.49
1310
420
23.0
0.58
1364
474
26.0
ECHAM4-A2_INM
b/
28.5
6.5
0.58
1572
682
37.4
0.70
1630
740
40.6
HadCM3_A2_SDSM/INM
b/
31.0
4.0
0.49
1310
420
23.0
0.58
1364
474
26.0
a/
Mean
maximum
temperature
for
the
warmest
month.
b/
Mean
maximum
annual
temperature.
S
ug
CC
Green
area
necessary
to
stabilize
the
temperature
increase
effects
of
climate
change.
A
gr
Area
of
green
roofs
to
establish
to
stabilize
the
temperature
increase
due
to
climate
change.
Fig.
7
.
Mean
temperature
values
near
the
ground
and
NDVI
index
for
(a)
Landsat7
ETMþ
and
(b)
Sentinel-2.
The
graph
corresponds
to
the
polygons
with
agricultural
and
green
area
fabric
in
the
city
of
Seville.
S.S.
Herrera-Gomez
et
al.
/
Building
and
Environment
1
23
(201
7)
575e584
581
according
to
ESA
(201
6),
hav
e
a
spatial
resolution
of
1
0
m
(bands
4
and
8),
and,
in
the
latter
case,
the
smaller
green
areas,
such
as
roundabouts
and
parterres
are
also
analyzed.
Although
these
small
green
areas
show
high
NDVI
values,
given
that
the
vegetable
cover
in
this
case
is
of
little
significance,
there
is
no
decrease
in
the
environmental
temperature
observed
in
these
sites.
The
NDVI
values
obtained
by
Ref.
[39]
vary
between
0.2
and
0.55,
while
in
this
study
they
fluctuate
between
0.2
and
0.62
(Fig.
5),
possibly
due
to
the
higher
resolution
of
satellite
images.
But
in
both
cases
they
show
the
inverse
relationship
between
the
NDVI
and
the
temper
-
ature;
the
higher
the
NDVI,
the
lower
the
temperature.
However
,
T
able
1
shows
that
the
types
of
vegetation,
and
especially
the
area
of
the
polygon
considered,
also
influence
the
decrease
in
temper
-
ature.
For
example,
for
the
representative
sites
chosen
in
T
able
1
,
the
lowest
temperatures
(31
C)
are
in
three
parks
and
gardens:
Prado
San
Sebasti
an,
Gardens
of
the
Real
Alc
azar
,
and
María
Luisa
Park.
These
sites,
besides
having
a
greater
area,
are
characterized
for
ha
ving
grass
,
shrubs,
and
mostly
trees.
This
confirms
the
recommendation
by
Ref.
[41],
that
in
order
to
decrease
the
tem-
perature
at
1
.6
m
from
the
ground,
trees
and
shrubs
have
the
greatest
impact.
In
this
study,
it
is
assumed
that
there
will
be
a
D
NDVI
in
each
of
the
poly
gons
considered,
according
to
the
increase
of
the
vegetat
ed
surface,
thank
s
to
the
incorporation
of
green
roofs
in
the
building
covered
area.
Nevertheless,
this
D
ND
VI
would
also
depend
on
the
types
of
vegetation
used
in
the
green
roofs,
as
well
as
the
green
roof
system
used,
the
depth
of
the
substrate,
and
the
water
status,
all
of
which
were
not
considered.
Therefore,
the
appropriate
selection
of
plant
species
when
designing
the
green
roof
is
essential
in
order
to
obtain
the
desired
effect,
as
well
as
a
proper
maintenance
to
keep
them
health
y
.
For
example,
according
to
[42],
an
extensi
ve
green
roof
with
Zoysia
matrella
has
a
mean
NDVI
of
0.6.
This
type
of
green
roof
would
be
sufficient
to
stabilize
the
increase
in
maximum
temperature
with
just
the
implementation
of
1
1
.3
and
26%
(207e
47
4
ha)
green
roofs
(rows
where
ND
VI<0.6),
under
the
global
models
with
A2
evolutionary
lines.
That
is
to
say
,
where
economy
and
technology
grow
slowly
and
production
is
oriented
at
self-
sufficiency
and
preserv
ation.
Contrariwise,
in
the
more
adverse
forecasts,
where
the
surface
of
green
roofs
needed
is
such
that
ND
VI>0.6,
from
28.9%
to
40.6%
(527e7
40
ha),
semi-intensive
or
intensiv
e
green
roofs
would
be
more
effective.
Other
variables,
such
as
the
state
of
the
green
roof
(with
health
y
and
well-w
atered
plants)
or
the
characteristics
and
location
of
the
building,
would
also
affect
the
effectiveness
of
this
strategy
.
For
example,
the
level
of
water
content
in
the
substrate
used
in
the
green
roof
is
a
variable
that
significantly
influences
the
surface
temperature
of
green
roofs,
being
an
effective
mean
for
tempera-
ture
regulation
[43,4
4].
Moreover
,
the
height
of
the
buildings
where
the
green
roofs
are
to
be
installed
would
also
change
the
decrease
in
temperature.
Actually
[45],
state
that
green
roofs
can
be
more
efficient
when
the
height
of
the
building
is
less
than
1
0
m.
Howev
er
,
even
with
taller
buildings,
the
fact
that
the
land
surface
temperature
(be
it
the
ground
or
the
roof)
decreases
is
a
fact.
The
only
difference
is
that
for
higher
buildings,
the
effect
of
the
tem-
perature
reduction
will
obviousl
y
not
be
noticed
at
street
level
but
on
the
roof.
It
will
also
alter
the
temperature
inside
the
building
and
thus
the
need
to
use
air
conditioning
systems,
which
would
also
increase
the
urban
heat
island
effect.
In
fact,
a
distinction
can
be
made
between
the
Canopy
layer
(which
occurs
in
the
layer
of
air
where
people
live,
from
the
ground
to
below
the
tops
of
trees
and
roofs)
and
the
Boundary
layer
(starting
from
the
rooftop
level
and
ext
ending
up
to
the
point
where
urban
landscapes
no
longer
in-
fluence
the
atmosphere,
around
1
.5
km
from
the
surface)
urban
heat
islands
[1
6,46].
Then,
green
roofs
in
higher
buildings
would
not
affect
the
former
but
would
influence
the
latter
.
Therefore,
in
our
study
,
the
height
of
the
buildings
has
not
been
considered.
In
fact,
there
are
other
analyses
that
have
attempted
to
quantify
the
potential
temperature
reductions
ov
er
a
br
oad
area
from
the
widespread
adoption
of
green
roof
technology
without
taking
into
account
the
height
of
buildings.
For
example,
a
modeling
study
in
T
oronto,
Canada,
predicted
that
adding
green
roofs
to
50%
of
the
av
ailable
surfaces
downtown
would
cool
the
entire
city
by
up
to
1
C
if
they
have
sufficient
moisture
for
evaporative
cooling,
playing
a
role
in
reducing
atmospheric
urban
heat
island
[4
7].
A
similar
study
in
New
Y
ork
City
assumed
the
full
conversion
of
all
available
roofs
area
to
green
roofs,
obtaining
air
temperature
reductions
2
m
above
the
roof
surface
of
about
0.2
C
for
the
city
as
a
whole
[48].
In
the
present
study
it
has
been
assumed
that
green
roofs
could
be
installed
on
all
the
buildings.
Nevertheless,
this
is
not
completely
true,
as
not
every
building
is
structurally
prepared
to
support
a
green
roof.
Also,
not
all
the
roof
surface
can
be
dedicated
for
that
use,
given
that
additional
space
is
required
for
building
facilities
such
as
solar
panels,
water
tank
s,
etc.
However
,
the
vegeta
ted
surface
area
on
buildings
can
also
be
increased
by
installing
living
walls
on
the
facades,
as
they
represent
an
even
larger
percentage
of
the
building
surface
than
the
roofs
themselves.
For
example,
ac-
cording
to
a
model
proposed
by
Ref.
[4
9],
a
living
wall
system
with
LAI
¼
2
(medium
density
foliage)
can
reduce
the
intensity
of
the
urban
heat
island
by
0.55
C
in
a
dry
climate.
There
are
other
solutions
that
might
have
the
same
effects
or
even
improv
e
the
performance
of
green
roofs,
such
as
cool
roofs
[50].
Nevertheless,
green
roofs
provide
many
other
benefits
that
posit
them
as
a
great
solution
against
other
effects
of
climate
change,
such
as
the
possible
increase
in
heavy
rainfall
[5
1].
In
these
cases,
the
green
roofs
would
act
as
a
buffer
to
avoid
the
fast
discharge
of
water
into
the
sewe
r
system,
decreasing
the
amount
of
runoff
per
building
by
approximat
ely
50%,
depending
on
the
type
of
green
roof.
For
example,
according
to
a
model
proposed
by
Ref.
[52],
if
1
0%
of
the
buildings
in
Brussels
had
an
extensi
ve
green
roof,
there
would
be
a
decrease
of
runoff
of
about
2.7%.
Moreov
er
,
the
use
of
green
roofs
would
also
influence
the
decrease
of
the
energy
de-
mand
of
buildings
[53]
[54],
or
the
amount
of
heat
generated
by
air
conditioning
[50].
5.
Conclusions
By
processing
the
images
from
the
Landsat
7
ETMþ
and
Sentinal-2
satellites,
the
inv
erse
relationship
between
land
surface
temperature
and
the
abundance
of
vegetation
was
verified
through
ND
VI,
in
other
words,
the
higher
the
temperature,
the
less
estab-
lished
veg
etation.
This
inverse
relationship
shows
a
linear
behavior
in
the
NDVI
range
from
0.2
to
0.6.
Therefore,
a
methodology
based
on
the
inverse
temperature-
NDVI
relationship
was
used
to
estimate
the
surface
of
green
roofs
needed
in
the
city
of
Seville,
Spain,
to
mitigate
the
increase
in
maximum
temperatures
in
summer
,
attributable
to
climate
change.
The
precision
and
number
of
temperature
and
NDVI
values
is
limited
by
the
degree
of
resolution
of
the
satellite
images
used.
It
is
indispensable
to
supervise
the
data
obtained
with
real
data
extr
acted
from
the
local
meteorological
stations.
In
general,
the
changes
in
maximum
temperature
in
the
city
of
Seville
are
expected
to
be
between
1
.5
and
6.5
C
according
to
the
climate
change
scenarios
and
models
for
the
periods
analyzed.
T
o
mitigate
the
negative
effects
of
climate
change
in
Seville,
it
would
be
necessary
to
cover
an
area
of
semi-intensive
or
intensive
green
roofs
of
up
to
7
40
ha,
for
the
most
adverse
scenario,
which
means
40.6%
of
the
total
building
covered
area.
Semi-intensive
and
intensive
green
roofs
play
an
important
role
in
improving
urban
microclimates,
thus
public
policies
to
foster
their
use
in
cites
should
be
implemented.
S.S.
Herrera-Gomez
et
al.
/
Building
and
Environment
1
23
(201
7)
575e584
582
The
results
of
these
estimations
can
be
considered
as
pre-
liminary
,
as
their
validation
req
uires
the
establishing
of
significant
areas
of
green
roofs
in
the
city
of
Seville
and
obtaining
the
exper
-
imental
data
measured
in
the
study
area.
Further
studies
are
rec-
ommended
using
a
greater
number
of
images
with
higher
resolution
in
different
seasons,
in
order
to
register
the
annual
variability
of
the
NDVI.
Ackno
wledgements
W
e
would
like
to
thank
CONA
CYT
(M
exico)
for
their
support
with
a
postgraduate
scholarship
for
the
first
author
of
this
work.
Notation
ND
VI
normalized
difference
vegetation
index
ND
VI
mean
mean
NDVI
values
obtained
from
the
satellite
image
for
the
whole
study
area.
NDVI
max
and
NDVI
min
:
maximum
and
minimum
NDVI
values
obtained
from
processing
the
satellite
images
in
the
whole
study
area
ND
VI
cc
ND
VI
necessary
to
counteract
the
effects
of
the
increase
in
maximum
temperature
due
to
climate
change
S
ug
urban
green
surface
S
ug
cc
urban
green
surface
necessary
to
stabilize
the
effects
of
the
increase
in
temperature
due
to
climate
change
A
gr
Area
of
green
roofs
to
be
deployed
to
stabilize
the
increase
in
temperature
due
to
climate
change
T
max
and
T
min
maximum
and
minimum
temperature
values
obtained
from
processing
the
satellite
images
in
the
whole
study
area
D
T
max
increase
in
the
maximum
tempera
ture
due
to
climate
change
T
s
land
surface
temperature
T
CC
max
Current
mean
monthly
temperature
plus
the
increase
in
temperature
produced
by
climate
change
Refere
nces
[1]
J.S.
Lee,
J.T.
Kim,
M.G.
Lee,
Mitigation
of
urban
heat
island
effect
and
green-
roofs,
Indoor
Built
Environ.
23
(2013)
62e69,
http://dx.doi.org/10.1177/
1420326X12474483.
[2]
M.
Santamouris,
Cooling
the
cities
-
a
review
of
reflective
and
green
roof
mitigation
technologies
to
fight
heat
island
and
improve
comfort
in
urban
environments,
Sol.
Energy
103
(2014)
682e703,
http://dx.doi.org/10.1016/
j.solener.2012.07.003.
[3]
B.-S.
Lin,
C.-C.
Yu,
A.-T.
Su,
Y.-J.
Lin,
Impact
of
climatic
conditions
on
the
thermal
effectiveness
of
an
extensive
green
roof,
Build.
Environ.
67
(2013)
26e33,
http://dx.doi.org/10.1016/j.buildenv.2013.04.026
.
[4]
M.
Sofer,
O.
Potchter,
The
urban
heat
island
of
a
city
in
an
arid
zone:
the
case
of
Eilat,
Israel,
Theor.
Appl.
Climatol.
85
(2006)
81e88,
http://dx.doi.org/
10.1007/s00704-005-0181-9.
[5]
H.
Saaroni,
E.
Ben-Dor,
A.
Bitan,
O.
Potchter,
Spatial
distribution
and
micro-
scale
characteristics
of
the
urban
heat
island
in
Tel-Aviv,
Israel,
Landsc.
Urban
Plan.
48
(2000)
1e
18,
http://dx.doi.org/10.1016/S0169-2046(99)00075-4.
[6]
T.
Susca,
S.R.
Gaffin,
G.R.
Dell’osso,
Positive
effects
of
vegetation:
urban
heat
island
and
green
roofs,
Environ.
Pollut.
159
(2011)
2119e2126,
http://
dx.doi.org/10.1016/j.envpol.2011.03.007.
[7]
Q.
Weng,
D.
Lu,
J.
Schubring,
Estimation
of
land
surface
temperature-
vegetation
abundance
relationship
for
urban
heat
island
studies,
Remote
Sens.
Environ.
89
(2004)
467e483,
http://dx.doi.org/10.1016/
j.rse.2003.11.005.
[8]
M.A.
Hart,
D.J.
Sailor,
Quantifying
the
influence
of
land-use
and
surface
characteristics
on
spatial
variability
in
the
urban
heat
island,
Theor.
Appl.
Climatol.
95
(2009)
397e406,
http://dx.doi.org/10.1007/s00704-008-0017-5.
[9]
H.
Akbari,
L.S.
Rose,
Urban
surfaces
and
heat
island
mitigation
potentials,
J.
Human-Environment
Syst.
11
(2008)
85e101,
http://dx.doi.org/10.1618/
jhes.11.85.
[10]
J.
Yang,
Q.
Yu,
P.
Gong,
Quantifying
air
pollution
removal
by
green
roofs
in
Chicago,
Atmos.
Environ.
42
(2008)
7266e7273,
http://dx.doi.org/10.1016/
j.atmosenv.2008.07.003.
[11]
O.
Saadatian,
K.
Sopian,
E.
Salleh,
C.H.
Lim,
S.
Riffat,
E.
Saadatian,
A.
Toudeshki,
M.Y.
Sulaiman,
A
review
of
energy
aspects
of
green
roofs,
Renew.
Sustain.
Energy
Rev.
23
(2013)
155e168,
http://dx.doi.org/10.1016/j.rser.2013.02.022.
[12]
S.-E.
Ouldboukhitine,
R.
Belarbi,
R.
Djedjig,
Characterization
of
green
roof
components:
measurements
of
thermal
and
hydrological
properties,
Build.
Environ.
56
(2012)
78e85,
http://dx.doi.org/10.1016/j.buildenv.2012.02.024.
[13]
J.S.
MacIvor,
L.
Margolis,
M.
Perotto,
J.A.P.
Drake,
Air
temperature
cooling
by
extensive
green
roofs
in
Toronto
Canada,
Ecol.
Eng.
95
(2016)
36
e42,
http://
dx.doi.org/10.1016/j.ecoleng.2016.06.050.
[14]
N.H.
Wong,
P.Y.
Tan,
Y.
Chen,
Study
of
thermal
performance
of
extensive
rooftop
greenery
systems
in
the
tropical
climate,
Build.
Environ.
42
(2007)
25e54,
http://dx.doi.org/10.1016/j.buildenv.2005.07.030
.
[15]
S.W.
Tsang,
C.Y.
Jim,
Theoretical
evaluation
of
thermal
and
energy
perfor-
mance
of
tropical
green
roofs,
Energy
36
(2011)
3590e3598,
http://dx.doi.org/
10.1016/j.energy.2011.03.072.
[16]
U.S.
Environmental
Protection
Agency,
Reducing
Urban
Heat
Islands:
Com-
pendium
of
Strategies,
Washington,
DC
20460
USA,
2008,
http://dx.doi.org/
10.1175/1520-0450(2002)041<0792:THFIUA>2.0.CO;2.
[17]
H.
Takebayashi,
M.
Moriyama,
Surface
heat
budget
on
green
roof
and
high
reflection
roof
for
mitigation
of
urban
heat
island,
Build.
Environ.
42
(2007)
2971e2979,
http://dx.doi.org/10.1016/j.buildenv.2006.06.017
.
[18]
U.
Berardi,
A.
GhaffarianHoseini,
A.
GhaffarianHoseini,
State-of-the-art
anal-
ysis
of
the
environmental
benefits
of
green
roofs,
Appl.
Energy
115
(2014)
411e428,
http://dx.doi.org/10.1016/j.apenergy.2013.10.047
.
[19]
AEMET,
Generaci
on
de
Escenarios
Regionalizados
de
Cambio
Clim
atico
para
Espa
~
na.
Agencia
Estatal
de
Meteorología.
Ministerio
de
Medio
Ambiente
y
Medio
Rural
y
Marino.
Gobierno
de
Espa
~
na,
2009.
http://www.aemet.es/es/
idi/clima/escenarios_CC.
[20]
IPCC,
Informe
Especial
del
IPCC.
Escenarios
de
Emisiones.
Resumen
para
responsables
de
Políticas.
Informe
especial
del
Grupo
de
trabajo
III
del
IPCC.
Grupo
Intergubernamental
sobre
el
Cambio
Clim
atico.,
Grupo
Inter-
gubernamental
de
Expertos
sobre
el
Cambio
Clim
atico,
2000.
[21]
FSNAU,
Understanding
the
Normalized
Diference
Vegetation
Index
(NDVI),
2010,
p.
2.
http://www.fsnau.org/downloads/Understanding_the_
Normalized_Vegetation_Index_NDVI.pdf.
(Accessed
1
January
2016).
[22]
J.
Weirer,
D.
Herring,
Measuring
Vegetation
(NDVI
&
EVI).
(NASA
Earth
Ob-
servatory),
2010.
http://earthobservatory.nasa.gov/Features/
MeasuringVegetation/measuring_vegetation_2.php.
(Accessed
1
January
2016).
[23]
C.
Huang,
X.
Ye,
Spatial
modeling
of
urban
vegetation
and
land
surface
tem-
perature:
a
case
study
of
Beijing,
Sustainability
7
(2015)
9478e9504,
http://
dx.doi.org/10.3390/su7079478.
[24]
L.
Liao,
L.
Zhang,
L.
Bengtsson,
Analyzing
dynamic
change
of
vegetation
cover
of
desert
oasis
based
on
remote
sensing
data
in
hexi
region,
in:
Proc.
Int.
Symp.
Sustain.
Water
Resour.
Manag.
Oasis-hydrosphere-desert
Interact.
Arid
Reg,
2005,
pp.
279e295.
Beijing,
PR
China.
[25]
K.C.
Tan,
H.S.
Lim,
M.Z.
MatJafri,
K.
Abdullah,
Landsat
data
to
evaluate
urban
expansion
and
determine
land
use/land
cover
change
in
Penang
Island,
En-
viron.
Earth
Sci.
60
(2010)
1509
e1521.
[26]
F.
Yuan,
M.E.
Bauer,
Comparison
of
impervious
surface
area
and
normalized
difference
vegetation
index
as
indicators
of
surface
urban
heat
island
effects
in
Landsat
imagery,
Remote
Sens.
Environ.
106
(2007)
375e
386,
http://
dx.doi.org/10.1016/j.rse.2006.09.003.
[27]
W.
Su,
C.
Gu,
G.
Yang,
Assessing
the
impact
of
land
use/land
cover
on
urban
heat
island
pattern
in
nanjing
city,
China,
J.
Urban
Plan.
Dev.
136
(2010)
365e372,
http://dx.doi.org/10.1061/(ASCE)UP.1943-5444.0000033
.
[28]
USGC,
United
States
Geological
Survey.
Science
for
a
Changing
World,
2016.
http://earthexplorer.usgs.gov/.
(Accessed
20
August
2016).
[29]
ESA,
User
Guides,
SENTINEL
2-MSI,
European
Space
Agency.
User
Services
and
Mission
Planning
Office
Via
Galileo
Galilei
00044
Frascati,
(Rome)
Italy,
2016.
https://sentinel.esa.int/web/sentinel/user-guides/sentinel-2-msi/resolutions/
spatial.
(Accessed
1
January
2016).
[30]
V.M.
Salinas,
Estimaci
on
de
la
evapotranspiraci
on
de
un
cultivo
de
vid
con
apoyo
de
im
agen
satelital
y
validaci
on
utilizando
Eddy
covariance,
Colegio
de
Postgraduados,
2013,
http://www.biblio.colpos.mx.
Master
thesis
10521/
1910.
[31]
F.
Li,
T.
Jackson,
W.
Kustas,
T.
Schmugge,
A.
French,
M.
Cosh,
R.
Bindlish,
Deriving
land
surface
temp
from
Landsat
5
and
7,
Remote
Sens.
Environ.
92
(2004)
521e534.
[32]
G.E.
Wukelic,
D.E.
Gibbons,
L.M.
Martucci,
Radiometric
calibration
of
Landsat
thematic
mapper
thermal
band,
Remote
Sens.
Environ.
28
(1989)
339e347.
[33]
R.G.
Allen,
M.
Tasumi,
A.T.
Morse,
Satellite-based
energy
balance
for
mapping
Evapotranspiration
with
internalized
calibration
(METRIC)-Applications,
J.
Irrig.
Drain.
Eng.
133
(4)
(2007)
395e406.
[34]
M.
Tasumi,
R.G.
Allen,
R.
Trezza,
Soil
Heat
Flux
Estimation
Method.
Appendix
12
in
M.
Tasumi,
Progress
in
Operational
Estimation
of
Regional
Evapo-
transpiration
Using
Satellite
Imagery,
2003.
[35]
R.
Trezza,
Evapotranspiration
Using
a
Satellite-based
Surface
Energy
Balance
with
Standardized
Ground
Control,
Utah
State
University,
2002.
[36]
A.
L
opez,
J.
Reca
~
no,
The
comeback
of
the
central
city
in
Southern
Europe:
population
growth
and
sociodemographic
change
in
the
Spanish
urban
cores,
in:
Ext.
Abstr.
Prep.
Eur.
Popul.
Conf
vol.
2010,
2010.
http://epc2010.princeton.
edu/papers/100918.
[37]
E.
Ayuntamiento
de
Sevilla,
Ayuntamiento
de
Sevilla.
Servicio
de
parques
y
jardines.
Area
de
H
abitat
Urbano,
Cultura
y
Turismo.,
(2016).
http://www.
sevilla.org/ayuntamiento/competencias-areas/area-de-habitat-urbano-
cultura-y-turismo/a-servicio-de-parques-y-jardines/parques/parques-
jardines-y-zonas-verdes.
S.S.
Herrera-Gomez
et
al.
/
Building
and
Environment
1
23
(201
7)
575e584
583
[38]
MAGRAMA,
Informe
de
Medio
Ambiente
en
Andalucía.
Secretaría
General
de
Medio
Ambiente
y
Agua.
Consejería
de
Agricultura,
Pesca
y
Medio
Ambiente
de
Andalucía,
2011.
http://www.juntadeandalucia.es/medioambiente/site/
rediam/menuitem.04dc44281e5d53cf8ca78ca731525ea0/?vgnextoid¼eedde
9762819d310VgnVCM2000000624e50aRCRD&vgnextchannel¼c2770219f56
0f210VgnVCM1000001325e50aRCRD&vgnextfmt¼rediam&lr¼lang_es.
[39]
A.
Farina,
Exploring
the
Relationship
between
Land
Surface
Temperature
and
Vegetation
Abundance
for
Urban
Heat
Island
Mitigation
in
Seville,
Lund
University,
Spain,
2012.
[40]
IDE,
IDE.Sevilla.
Datos
abiertos
espaciales
de
Sevilla.
Gerencia
de
Urbanismo
Ayuntamiento
de
Sevilla,
2016.
http://sig.urbanismosevilla.org/sevilla.art/
datosabiertos/index.html.
(Accessed
1
August
2016).
[41]
K.
Perini,
A.
Magliocco,
Effects
of
vegetation,
urban
density,
building
height,-
and
atmospheric
conditions
on
local
temperatures
and
thermal
comfort,
UrbanForestry&UrbanGreening
13
(2014)
495e506,
http://dx.doi.org/
10.1016/j.ufug.2014.03.003.
[42]
N.
Ntoulas,
P.
Nektarios,
E.
Charalambous,
A.
Psaroulis,
Zoysia
matrella
cover
rate
and
drought
tolerance
in
adaptive
extensive
green
roof
systems,
UrbanForestry&UrbanGreening
12
(2013)
522e531.
[43]
A.P.
Bevilacqua,
D.
Mazzeo,
N.
Arcuri,
Surface
temperature
analysis
of
an
extensive
green
roof
for
the
mitigation
of
urban
heat
island
in
southern
mediterranean
climate,
Energy
Build.
150
(2017)
318e327,
http://dx.doi.org/
10.1016/j.enbuild.2017.05.081.
[44]
C.L.
Tan,
P.Y.
Tan,
N.H.
Wong,
H.
Takasuna,
T.
Kudo,
Y.
Takemasa,
C.V.J.
Lim,
H.X.V.
Chua,
Impact
of
soil
and
water
retention
characteristics
on
green
roof
thermal
performance,
Energy
Build.
(2017),
http://dx.doi.org/10.1016/
j.enbuild.2017.01.011.
[45]
N.
Kabisch,
J.
Stadler,
H.
Korn,
A.
Bonn,
Nature-based
Solutions
to
Climate
Change
Mitigation
and
Adaptation
in
Urban
Areas,
Bundesamt
für
Natur-
schutz
(BfN),
Bonn,
Germany,
2016.
http://www.bfn.de/0502_skripten.html.
[46]
T.
Oke,
The
energetic
basis
of
the
urban
heat
island,
Q.
J.
R.
Meteorol.
Soc.
108
(1982)
1e24.
[47]
K.
Liu,
B.
Bass,
Performance
of
green
roof
systems,
Atlanta,
GA.,
May
12-13,
in:
Cool
Roofing
Symposium,
2005,
pp.
1e18,
https://www.researchgate.net/
profile/Brad_Bass/publication/44077726_Performance_of_green_roof_
systems/links/0c96052b4deed36266000000.pdf.
[48]
C.
Rosenzweig,
W.
Solecki,
L.
Parshall,
S.
Gaffin,
B.
Lynn,
R.
Goldberg,
J.
Cox,
S.
Hodges,
Mitigating
New
York
city's
heat
island
with
urban
forestry,
living
roofs,
and
light
surfaces,
Atlanta
GA,
in:
Sixth
Symp.
Urban
Environ.
Forum
Manag.
Our
Phys.
Nat.
Resour.
Am.
Meteorol.
Soc,
2006,
https://pdfs.
semanticscholar.org/6a84/7e63248369137027ced29c7808343f30255d.pdf.
[49]
A.
Afshari,
A
new
model
of
urban
cooling
demand
and
heat
islanddapplication
to
vertical
greenery
systems
(VGS),
Energy
Build.
(2017).
https://doi.org/10.1016/j.enbuild.2017.01.008
.
in
press.
[50]
V.
Costanzo,
G.
Evola,
L.
Marletta,
Energy
savings
in
buildings
or
UHI
mitiga-
tion?
Comparison
between
green
roofs
and
cool
roofs,
Energy
Build.
114
(2016)
247e255,
http://dx.doi.org/10.1016/j.enbuild.2015.04.053.
[51]
R.
Fioretti,
A.
Palla,
L.G.
Lanza,
P.
Principi,
Green
roof
energy
and
water
related
performance
in
the
Mediterranean
climate,
Build.
Environ.
45
(2010)
1890e1904,
http://dx.doi.org/10.1016/j.buildenv.2010.03.001
.
[52]
J.
Mentens,
D.
Raes,
M.
Hermy,
Green
roofs
as
a
tool
for
solving
the
rainwater
runoff
problem
in
the
urbanized
21st
century?
Landsc.
Urban
Plan.
77
(2006)
217e226,
http://dx.doi.org/10.1016/j.landurbplan.2005.02.010
.
[53]
U.
Berardi,
The
outdoor
microclimate
benefits
and
energy
saving
resulting
from
green
roofs
retrofits,
Energy
Build.
121
(2016)
217e229,
http://
dx.doi.org/10.1016/j.enbuild.2016.03.021.
[54]
J.
Coma,
G.
P
erez,
C.
Sol
e,
A.
Castell,
L.F.
Cabeza,
Thermal
assessment
of
extensive
green
roofs
as
passive
tool
for
energy
savings
in
buildings,
Renew.
Energy
85
(2016)
1106e1115,
http://dx.doi.org/10.1016/
j.renene.2015.07.074.
S.S.
Herrera-Gomez
et
al.
/
Building
and
Environment
1
23
(201
7)
575e584
584
Previous PDF in this issueNext PDF in this issue