The influence of urban heat island circulation in ...
Transcript of The influence of urban heat island circulation in ...
The influence of urban heat island
circulation in idealized city -- from building scale to city scale
Xiaoxue Wang, Yuguo Li
The University of Hong Kong
9th International Conference on Urban Climate July 20th-24th 2015 Toulouse, France
2
Why do we consider the urban heat island circulation ?
When UHIC happened, the worse pollution and warmer scenarios
generally happened compared with the windy day .
Preconditions for urban heat island circulation (UHIC)
Thermal inversion layer
Zero background wind
Horizontal temperature gradient Horizontal pressure gradient UHIC
ICUC9, Toulouse, France
(Falasca,2013)
ICUC9, Toulouse, France 3
Models
It is not easy to anlyze the air flow in the canopy or around a single
building by coupling meothd.
Coupling method?
The boundary condition changes with time during the UHIC
evolution.
CSCFD compared with traditional CFD
Porous model
Meso
-scale
(~ 105 m)
City
-scale
(~ 103 m)
Micro
-scale
(~ 101 m)
Porous model
Building details resolved
4
Modeling the mesoscale phenomena
Coordinate transformation
Modifying basic equations and turbulence by
comparing with mesoscale atmospheric
governing equation
Coriolis terms
Modified buoyancy term
Compressibility term for energy equation
Coordinate transformed terms
Modeling the overall city scale effect
Adding porous turbulence model
Modeling the 24-hour variation
Using daily surface temperature for ground
Improving numerical stability
Using an absorbing layer at the top
ICUC9, Toulouse, France
Modified CFD model -- City-Scal CFD (CSCFD)
eff 1l tJ
1/ 1 ξznJ
CSCFD - basic governing equations
0
eff 0 0
n
n n n n n n C n
d Vp V T T g F F F
dt
2 2
0 0 0
0
0
11
n
n c n n n
F
J T T g F w J p J z
Coordinate transformed terms
0 0
0
0
n
C
fv lw J
F fu
luJ
Coriolis terms
2 3
0
fl Fn n
CF V Q V
K K
Porous model terms
0 0nT T g
Modified buoyancy term
2
t
kC
0 p n
eff n T
d C Tk T S S
dt
0 ΓT p n s
p
gS JC w
C
A transformed compressibility term
not existing in the conventional CFD
0 sen
1S Q
H
Modified heat source
considering porosity 5
6
CSCFD - turbulence governing equations
0 0n n n
tn n n n k b n k k
k
dk k G G S S
dt
2
0 1 3 2 0n n
t n nn n n n k n b
n n
dC G C G C S S
dt k k
1 3n
n ts
n t p
gS C C g
k Pr C
2 3 3
0
82 2
3
l F Fk n n n k
C CS k Q k F
K K K
2 3 3
0
8 82
3 3
f
l nF Fn n n
r r
kC C QS Q
K x xK K
n
tk s
t p
gS g
Pr C
Coordinate transformed terms Porous turbulence model terms
22/07/2015 ICUC9, Toulouse, France
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.50
500
1000
1500
2000
2500
3000
x/D
Hei
ght
(m)
0 0.2 0.4 0.6 0.8 10
20
40
60
80
100
x/D
Hei
ght
(m)
7
Quasi-steady UHIC – domain
55 km×2.95km 5 km×100m
Not to scale
x
z
29
50
m
10 km
Symmetry
Wall
Pressure
outlet
Constant heat flux (city)
50 m
110 km
Table 1 Mesh and other model related setting
City
area
Other
area
Time step 1 second
Simulation
time
10 hours to achieve
quasi-steady state
Porosity 0.75 1.0
ICUC9, Toulouse, France
Constant heat flux (rural, 50W/m2)
8
Comparing with others’ data
-3 -2 -1 0 1
-3
-2
-1
0
Yoshikado (1992)
Richiardone and Brusasca (1989)
Kurbatskii (2001)
Kristof et al. (2009)
Hidalgo et al. (2009)
Our simulation
Godowitch et al. (1987)
Clarke and McElroy (1974)
Lu et al. (1997)
Cenedese and Monti (2003)
Catalano et al. (2012)
Falasca (2013)
zi/D=2.86Fr
z i/D
Fr
0.001 0.01 0.1 1 10
0.001
0.01
0.1
1
Simulation
Field experiment (St. Louis, Missouri)
Water tank
experiment
Our predicted mixing heights agree reasonably well with the field
data, scaled lab data and computations in literature.
ICUC9, Toulouse, France
Mixing height(zi): the height where the maximum difference between the plume
centerline and ambient density profiles occurs (Lu and Arya, 1997)
301
301
301
30
2
302
301.5
-1 -0.5 0 0.5 10
100
200
300
400300 301 302 303 304 305 306 307 308
302
30
2301
303
301
-1 -0.5 0 0.5 10
100
200
300
400300 301 302 303 304 305 306 307 308
Potential temperature
-2 -1.5 -1 -0.5 0 0.5 1 1.5 20
500
1000
1500
2000300 301 302 303 304 305 306 307 308 309 310 311 312
Potential temperature
-2 -1.5 -1 -0.5 0 0.5 1 1.5 20
500
1000
1500
2000300 301 302 303 304 305 306 307 308 309 310 311 312
Potential temperature (K)
Hei
ght
(m)
x/D
Quasi-steady UHIC– comparing a flat city and a porous city
(Porosity= 0.75 𝜆𝑝 = 0.25) Porous city Flat city
Wide neck Narrow neck
Hei
ght
(m)
x/D
Hei
ght
(m)
Hei
ght
(m)
x/D x/D
Potential temperature (K)
Potential temperature (K)
Potential temperature (K)
9
-2 -1.5 -1 -0.5 0 0.5 1 1.5 20
500
1000
1500
2000-5 -4 -3 -2 -1 0 1 2 3 4 5
Horizontal velocity
-2 -1.5 -1 -0.5 0 0.5 1 1.5 20
500
1000
1500
2000-5 -4 -3 -2 -1 0 1 2 3 4 5
Quasi-steady UHIC– comparing a flat city and a porous city
1 m/s 4 m/s
(Porosity= 0.75 𝜆𝑝 = 0.25) Porous city Flat city
Hei
gh
t (m
)
Horizontal velocity (m s-1)
x/D
10
-2 -1.5 -1 -0.5 0 0.5 1 1.5 20
500
1000
1500
2000-1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4
Hei
ght
(m)
x/D
Vertical velocity (m s-1)
-2 -1.5 -1 -0.5 0 0.5 1 1.5 20
500
1000
1500
2000-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Vertical velocity
x/D
Vertical velocity (m s-1)
Hei
ght
(m)
0.0 0 .0
0.0
0.0
0.0
0.0
0.4
0.4
0.80.8
0.8
1. 2
1.2
-1 -0.5 0 0.5 10
100
200
300
400-0.4 0.0 0.4 0.8 1.2
Hei
gh
t (m
)
x/D
Vertical velocity (m s-1)
0.0
0.00.0
0.0
0.00.0
0.0
0.0
0.0
0.0
0.00.0
0.00.0
0.0
0.4
0.40.4
0.4
0.8
1.2
1.2
-1 -0.5 0 0.5 10
100
200
300
400-0.4 0.0 0.4 0.8 1.2
Vertical velocity
Hei
ght
(m)
x/D
Vertical velocity (m s-1)
Quasi-steady UHIC– comparing a flat city and a porous city
Multiple upward flows Single upward flow
(Porosity= 0.75 𝜆𝑝 = 0.25) Porous city Flat city
11
• When the sensible heat flux is larger, the mixing height is higher,
vertical velocity is larger, the height of the maximum vertical velocity is
higher, and the UHIC is stronger.
Quasi-steady UHIC – effect of heat flux
12
301 302 303 304 305 306 307 308
0
200
400
600
800
1000
1200
1400
Heig
ht
(m)
Potential temperature (K)
Q =60 (W/m2)
Q =100(W/m2)
Q =150(W/m2)
Q =200(W/m2)
Small heat flux
Large heat flux
Sm
all
Lar
ge
-1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
0
200
400
600
800
1000
1200
1400
Heig
ht
(m)
Vertical velocity (m/s)
Q =60 (W/m2)
Q =100 (W/m2)
Q =150 (W/m2)
Q =200 (W/m2)
Large
heat flux
Small
heat flux
ICUC9, Toulouse, France
-2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0
0
1
2
3
z/z i
u/UD
Flat city
porosity=0.75
porosity=0.5
water tank experiment
(our group)
Theory
13
Comparing with water tank experiment
Case Relative reverse
height
ϕ=1.0
(Theory)
39.5%
ϕ=1.0
(Flat city)
42.6%
ϕ=0.75 50.3%
ϕ=0.5 51.3%
ϕ=0.8 ∗ 55.6% Flow reverse height
Reference height
Increase
* water tank data from Yan Yifan
Our water tank model produced a relatively lower scaled flow, which agreed with
other similar studies in the water tank, needs to be further studied.
Relative reverse height
=Reverse height
Reference height
ICUC9, Toulouse, France
(Lu, 1997)
14
The proposed model could simultaneously simulate the wind and
thermal environment around several specific buildings and the
basic dynamics in the whole meso-scale area.
In the future, the details around the building will be analyzed.
Preliminary result: simultaneously consider the
environment around buildings and city climate
ICUC9, Toulouse, France
Summary
• When the city is introduced by porous media, the
velocity in the city decreases, the neck of the plume
increases, and multi-upward flows are observed.
• Even when the background wind is zero, the velocity
in the city is not zero due to the UHIC.
• The mixing height simulated by CSCFD agrees well
with other data in literature.
• Dense cities also increase the relative reverse height
at the edge of the city.
• Further data will be needed, and validation of the
code for synoptic wind conditions is also needed.
22/07/2015 ICUC9, Toulouse, France 15
x (km)
Hei
ght(
m)
0 25 50 75 100 125 150 175 200 2250
500
1000
1500
2000
2500
Absorbing layer
x (km) x/D
Hei
ght(
m)
-2 -1.5 -1 -0.5 0 0.5 1 1.5 20
500
1000
1500298 299 300 301 302 303 304 305 306 307 308
x/D
Potential temperature
X
Heig
ht
(m
)
90000100000110000120000130000140000150000160000170000180000190000
0
500
1000
1500
2000
9 10 11 12 13 14 15 16 17 18 19
x (km)
520
The general plume features are well predicted.
Comparing with meso-scale model
Prediction of CSCFD Prediction of Yoshikado (1992)
17 kmKz
ICUC9, Toulouse, France 17
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 10
200
400
600
800
1000300 300.5 301 301.5 302 302.5 303 303.5 304
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 10
200
400
600
800
1000300 300.5 301 301.5 302 302.5 303 303.5 304
• The city height does not affect the
mixing height, but affects the
entrainment at the lower part.
Quasi-steady UHIC – effect of city
18
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 10
200
400
600
800
1000300 300.5 301 301.5 302 302.5 303 303.5 304
Hei
gh
t (m
)
Potential temperature (K)
x/D
Hei
gh
t (m
)
x/D
x/D
Hei
gh
t (m
)
302 303 304 305 306 307 308
0
200
400
600
800
1000
1200
1400
Heig
ht
(m)
Potential temperature (K)
H=100m
H=50m
H=10m
100 m
10 m
19
Application 2 – 12-hour evolution of daytime UHIC
The sensible heat flux
is partitioned with
25% in the porous air
and 75% at the ground
surface (assuming
porosity is 0.75).
7 8 9 10 11 12 13 14 15 16 17 18
-100
0
100
200
300
400
500
Sen
sibl
e he
at f
lux
(W m
-2)
Time (Hour)
city
rural
city - rural
anthropogenic heat
Anthropogenic
heat
Net urban-rural heat
Not to scale The same cities in Application
1 is used here.
The 12-hour profiles for the sensible heat flux
(calculated from the new 24-hour boundary condition)
295 300 305 310 315
0
500
1000
1500
2000
2500
295 300 305 310 315
Heig
ht
(m)
Potential temperature (K)
at x/D=0 (city center)
in clear case
8:00
10:00
12:00
14:00
16:00
18:00
Potential temperature (K)
at x/D=0 (city center)
in porous case
8:00
10:00
12:00
14:00
16:00
18:00
8 9 10 11 12 13 14 15 16 17 18
0
500
1000
1500
2000
2500
Hig
ht
(m)
Time (Hour)
city center (porous)
city center (clear)
city edge (porous)
city edge (clear)
rural area (porous)
rural area (clear)
20
Application 2 – 12-hour evolution of daytime UHIC
– Comparing mixing height for a flat city and a porous city.
The porous city has a lower mixing
height due to flow resistance, which
leads to multiple smaller plumes.
x/D
z 1
2
0-0.5-1 0.5 1
Flat city Porous city
Flat city
Porous city
22/07/2015 ICUC9, Toulouse, France
21
The mixed porous-resolved approach
0
Wind
Wind
x
y
W B
B
Other “city” area Area of interest
(fully resolved)
Other “city” area
(as porous in case P&B,
as flat in case N&B,
and resolved in case B&B.)
B=W=H
x/H
z/H
7H 6.6H 20.3H
Wind
22/07/2015 ICUC9, Toulouse, France
Porosity = 0.75
Similar
Different
Only interested area is modelled
All areas are CFD modelled
Mixed porous-resolved approach
The mixed porous-
resolved approach can
predict well the airflow
pattern in the other
“city” areas and its
impact on micro-scale
flows in the area of
interest
The mixed porous-resolved approach
Other
“city” area
Other
“city” area Area of interest
22
Verification - Comparing the mixed
porous-resolved approach with full
resolution in an ideal city
-1 0 1 20
1
2
3
4
5
6
7
8
u (m s-1
) at line1
z/H
P&B
B&B
-1 0 1 20
1
2
3
4
5
6
7
8
u (m s-1
) at line2
P&B
B&B
-1 0 1 20
1
2
3
4
5
6
7
8
u (m s-1
) at line3
P&B
B&B
-1 0 1 20
1
2
3
4
5
6
7
8
u (m s-1
) at line4
P&B
B&B
-1 0 1 20
1
2
3
4
5
6
7
8
u (m s-1
) at line5
P&B
B&B
-1 0 1 20
1
2
3
4
5
6
7
8
u (m s-1
) at line6
P&B
B&B
Not to scale
The mixed approach
produces nearly
identical u-velocities
with the fully
resolved approach.
23
22/07/2015 ICUC9, Toulouse, France
is the temperature lapse rate
24
CSCFD Coordinate transformation
0
11 ln 1
g z
z R T
Based on the pressure equation,
The transformation coefficient is
Kristóf et al. (2009)
1
ln 1 nz z
z
nw e w
z
p np e p
After the transformation:
;nT T z
0 1 z
n e ;nu u
;nv v
22/07/2015 ICUC9, Toulouse, France
Including all buildings in a CFD model is not possible
25
Methodology -- City-Scale CFD (CSCFD)
Resolved
Only modelling neighborhood (such
as AVA) is not accurate I propose -
Treat the other urban area using a porous
turbulence model approach
Porous media
22/07/2015 ICUC9, Toulouse, France
How to analyze the UHIC’s effect in the city/around the
buildings ?
by modeling the environment around several buildings in isolation?
or by modeling the whole city with simple parameterizations?
or by a modified CFD model -- City-Scal CFD (CSCFD)?
The urban heat island
circulation is a multi-scale
flow problem, which includes
ICUC9, Toulouse, France 26
Introduction
the mesoscale (~103 km)
the city scale (101~103 km)
the local scale (10-3~101 km)
22/07/2015