Transcript Numerical simulations of boundary layer structure during Air Quality Control

Numerical simulations of boundary layer structure during
the high pollution episode in Tehran region
Hossein Malakooti, Bruno Sportisse, Luc Musson Genon
Air Quality Control
Company (AQCC)
Motivation of the study:
Study of urban climate alteration in PBL structure and
meteorological fields in Tehran region.
Interaction between UHI circulation and mountain flows.
Challenges:
MM Models not designed to capture change in urban climate and
at low resolution fail to capture heat island mitigation.
Land use/land cover and physical properties of material and
surfaces becomes new critical issue on urban.
Anthropogenic heat fluxes form human activity.
How, in various situations, urbanization interacts with other local
forcing factors .
Topographic flows (mountain and valley winds)
Following the convention of Segal and Arritt
(1992) nonclassical mesoscale circulations
(NCMCs) can be generated by differential
sensible and latent heat fluxes to the
atmosphere produced by spatial gradients
thermal and radiative properties such as
albedo, thermal conductivity, differences in
snow cover or vegetation (Segal et al. 1989 ;
Segal et al. 1991 ; Rife et al. 2002,)
In valley cities and during calm periods,
condition may be lead to critical air quality
problem due to low horizontal ventilations
especially in transition between slope flows in
early morning and in the late afternoon
(Whiteman 1990, 2000) and also due to
combining of slop winds with urban heat
island, especially during winter and night
time, when vertical diffusion is small
(Atkinson, 1981).
Urban Heat island
What is causing :
Canyon effects (blocking of radiation,
and wind)
Anthropogenic heating (heat
generation by human activity)
Longwave re-emission from the
warmer air above the city (local
greenhouse effect and induced
vertical warm advection)
Changes in the thermal properties of
materials (thermal storage)
Lack of evapotranspiration (lack of
vegetation and standing water) ( Oke,
1998b).
Aerodynamic contrasts can lead to
enhanced or suppressed mechanical
mixing and therefore affect vertical and
horizontal diffusion and transport
(Grimmond and Oke 1999).
Local Meteorology



Average altitude from MSL: 1250 m
(1100-1700 meters above MSL)
Location: 51 E, 35 N
Average of annual rainfall : 230 mm/year
Synoptic condition during 4 – 6 th December 2005
500 mb chart
Sodar observation
5 December 2005
600
600
2:30 AM
8:30 AM
500
500
2:30 PM
Z (m)
Z (m)
400
400
7:30 PM
11:30 PM
2:30 AM
300
8:30 AM
200
2:30 PM
7:30 PM
100
11:30 PM
0
0
45
100
90200
135 300 180
400225
WindWind
speed
(Cm .Sec-1)
direction
270
500
315
600
360
700
DA-SM2-U urban canopy representation
Surface types for DA-SM2-U
nat
bare soil located between the sparse
vegetation elements
pav
Paved surface located between the sparse
vegetation elements
bare
Bare soil without vegetation
cova
Paved surface located under the
vegetation
vegan
Vegetation over the bare soil
bui
Building roofs
vega
Vegetation over paved surface
wat
Water surface
DA-SM2-U urban canopy representation
j : index for surface type
k : level above the ground
Surfj (k) : top area density of the surface type ‘j’ at the level k (m2 m-2)
fj : horizontal surface density of the surface type ‘j ’ (m2m-2)
Afj (k) and Apj (k): frontal area density (m2 m-3) and plan area density (m2 m-3) respectively
Surf tot (k )   f j Surf j (k )
: horizontal air density at the level k
j
k 1
Sair (k )  (1  f bui )   fbui Surfbui ( p)
: total top area density at the level k
p 0
Vair (k )  
z ( k )  0.5 z ( k )
z ( k )  0.5 z ( k )
S air ( z)dz
: volume air density (m3 m−2) at the level k,
DA-SM2-U modifications have been implemented inside two main parts:
MM5 GSPBL scheme by modifying of conservation equations and turbulence length.
Soil model SM2-U by modifying surface canopy heat fluxes equations and ground part equations.
GSPBL scheme modifications
A: Momentum equation:
  ui 
 Rui  Fuibui   Duij
t
j
Source due to the presence of horizontal surface
bui
ui
F
  (k )  fbuiSurf bui (k )u*bui (k )2  ui (k ) 

 
2
2 0.5
 Vair (k )  ( ux (k )    u y (k )  )
Pressure and viscous drag force by vertical surfaces
Duij (k )    (k )Cdj Afj (k ) ui (k )
u x (k )  u y (k )
2
Cdbui = 0.4, Cdvegn = Cdvega = 0.2.
B: Thermal equation:
  L 
 R  D  A
t
 1  H sensmean (k )

D k   
cp
 Vair (k ) 
 1  Qurb (k )

A k   
 Vair (k )  c p
sensible heat from roofs &
vegetation
anthropogenic heat flux
C: Humidity equation:
  qw 
 Rq  Dq
t
Dq k  
Emean (k )
Vair (k )
humidity flux
2
D: Turbulent kinetic energy equation:
 ui E    u x
E

 K m 
t
xi
  z
1



2
  uy

 z

2 




2


 wE 

bui   g
 Sair  FE   wv  H E  1
   WEj   DEj
j
j

  4 z
  v
6 
7 

5 
3 
f Surf bui u*bui (k )
F (k )  bui
K m (k )
4
shear production by building horizontal surfaces
bui
E
buoyancy production by building and vegetation
sensible heat fluxes & anthropogenic heat fluxes
g
H E (k ) 
V (k )D (k )  A (k )
 v (k ) air
W (k )  Cdj Afi (k )( u x
j
E
2
2
 vy )
2
1.5
D (k )  4Cij Afj (k )( ux  u y ) E (k )
j
E
2
0.5
wake production of TKE due to interaction
between mean flow and canopy elements
TKE dissipation by canopy elements
E: Turbulent length scale (TLS)
The mixing length lBL is derived from the upward and downward
displacements (lup and ldown) that could be achieved by parcels
having kinetic energy equal to the mean TKE before being
stopped by buoyancy effects.
1
1
1


l (k ) lBL (k ) lcan (k )
1
lBL

1  1
1 



2  lup ldown 

g
z
z  l d o wn
j
 vs
( p)
1
1
 j
l can(k ) p  k 1  Surf tot (0) z ( p)
1/ lcan (k )  0
v ( z)  v ( z )dz  E ( z )
Bougeault and Lacarrère (1989)
and
 f Surf
j
g
 vs
z
above the canopy, l = lBL
k top
z  l up
for
for
Surf tot (0)  1
v ( z )  v ( z)dz  E ( z )
Surf tot (0)  1
Bélair et al. (1999)
Martilli et al. (2002),
Description of the SM2-U(3D) Model
Masson (2000) & Brown (2000) & Dupont (2001)
Mean heat flux inside the canopy
 mean (k )   f j Surf j k  j k   f pav*pav k 
j
Net radiation flux
k

 f A  p  z  p 

k
 ex p

 i pi 
 k 1  i
 exp 

cos
Z
e




top


Rnj k   1   j RG ktop    j Tsj k    aT k  1
4
4

Consideration: exponentially decay toward the ground ( kex = 1.5 the radiation extinction coefficient )
Latent heat flux from paved surfaces
The latent heat flux is considered as emitted at the floor level, thus,
LE *pav k   0
k to p
*
 p LEpav  p 
LE (0)   Surf pav
*
pav
p 0
where Surf*pav is the top area density of the street canyon.
Sensible and net rediation flux from paved surfaces

*
pav
k  
2z k ws  p, k 
p 1
 2z t   p, t   W  p   p,0
w s
t 1
*
0 
 pav
r s
W  p r s  p,0
p 1
 2z t   p, t   W  p   p,0
t 1
 pav  p 
w s
 pav  p 
at the ground level is
r s


*






2
Surf
p

z
k

p
,
k


pav
w s
*
k     p1
 pav
 pav  p 
p  k 1
  2z t ws  p, t   W  p r s  p,0

 t 1

ktop
The sensible and net radiative fluxes
assessed at the top of the street canyons
are distributed inside the street following the
sky view factor
To account for a vertical distribution of street
canyons, the contributions at the level k of the
higher street canyons having their tops between
the levels k + 1 and ktop are considered by


*
2Surf pav  p W  p r s  p, k 


*
*
k     p1
0 pav 0
 pav
 pav  p   Surf pav
p 1
  2z t ws  p, t   W  p r s  p,0

 t 1

ktop
W  p 
*
 p S bui 
Surf pav
0.5
2 Surfbui  p 
it is assumed that (i) all buildings are parallelepipeds with the same square
horizontal section characterized by their area Sbui, (ii) surrounded by paved
surfaces, and (iii) each street canyon has a constant width with the height.
The calculations of the sky view factors are adapted from Masson (2000):
 z  p   2 
  1
 r s  p, k  0  


W
p



0.5

z p 
W  p
0.5
2
 z  p   z k 
  z  p   z k   1


 z  p   z k  


  1   

 r  s  p, k   
 1  






W
p
W
p
W
p








 
winter and summer workday city-scale anthropogenic heat flux profiles
Daily average workdays patterns released near surface in winter and summer
80
50
NG
NG
Traffic
60
Electricity
Qf (w/m2)
Metabolism
Qf (w/m2)
Traffic
40
Electricity
Total
40
20
Metabolism
30
Total
20
10
0
0
0
6
12
Local time
18
24
0
6
12
Local time
18
24
The diurnal profiles have morning and evening peaks, with summertime and wintertime
maxima up to 44Wm2 and 65Wm2 respectively. The foul consumption component is
the main one in winter with a 54% share and traffic with 44% in summer. Based on our
distribution analysis with 500 meter resolution of Tehran region we find that the urban
core region may have anthropogenic heating values 4-7 times the magnitudes of the
city-scale values presented in this paper, especially in morning and evening and with
higher spatial resolution, it is observed values more than this too.
Model application
One-way nested configuration for several days in December including five nested
computational domains with 81., 27., 9., 3., and 1. km horizontal grid spacing.
DA-SM2-U is used only on the 1. km domain.
To compare DA-SM2-U (Qf and non Qf) with RA MM5 versions, one other simulation
using RA is performed : a ‘standard’ version of MM5 (RA-SLAB).
The first four domains are run
with 40 vertical sigma layers.
The 1. km domain includes 130
- 130 grid points and 60 vertical
sigma layers covering the
Tehran metropolitan area.
10 urban Sub-categories for
land use
FDDA is used on the first three
domains (Stauffer and Seaman
,1994).
Analyses of the Vertical Profiles inside the RSL and PBL
At 2 pm
At 4 am
1200
50
50
1200
Sim-1
Sim-1
40
800
30
Z(m)
Z (m)
800
1000
Sim-2
Sim-2
Sim-3
Sim-3
(m)
ZZ(m)
1000
40
Sim-1
Sim-1
Sim-2
Sim-2
Sim-3
Sim-3
600
30
600
20
20
400
400
200
10
10
200
0
0 0
00
0
0.5 2
1
4
1.5
-1
U U(ms
(m s -1))
6
2
82.5
Sim-3
Sim-2
1000
40
Sim-2
6
8
2
2.5
Sim-3
Sim-1
Sim-2
Sim-2
Sin-3
Sim-3
Sim-3
800
30
600
600
20
400
20
10
200
10
0 0
292290
4
1 U (ms-1)1.5
U (ms-1)
Sim-3
Sim-1
ZZ(m)
(K)
Z (m)
(m)
Z
800
30
2
0.5
1200
50
50
1200
1000
40
0
0
10
3
400
200
292
293
294 294 296
298
295
Pot.tem
Tem
Pot.
p. p.
(K)(K)
300 296 302
304
297
0 0
297
297
298 298
299
300
299
Pot. Tem. (K)
Pott.temp.
(K)
301
300 302
303
301
Friction velocity & TKE
At 2 pm
At 4 am
40
40
Sim-1
Sim-1
30
30
Sim-2
Sim-2
Sim-3
Z (m)
Z (m)
Sim-3
20
10
20
10
0
0
0
0.2
0.4
0.6
0.8
1
1.2
0
0.2
0.4
0.6
0.8
1
u* / u*m ax
u* / u*max
400
1200
Sim-1
Sim -1
1000
Sim -2
300
Sim-2
Sim-3
Sim -3
800
Z (m)
Z (m)
1.2
200
600
400
100
200
0
0
0
0.05
0.12 -2
TKE (m S )
0.15
0.2
0
0.2
0.4
0.6
0.8
2
1
-2
TKE (m s )
1.2
1.4
1.6
th 2005
Wind
Windfield
fieldat
at44pm
pmat
at25
25m
mAGL
AGL December
December
December
55thth 5
2005
2005
DA-SM2-U
DA-SM2-U
RA-SLAB
&&QQf:f:on
off
th2005DA-SM2-U
Temperature
Temperature
Temperature
field
field
field
at
atat
44 pm
4pm
pm
at
atat
22 2
m
mm
AGL
AGL
AGLDecember
December
December
55thth
52005
2005
DA-SM2-U
RA-SLAB
&
&Q
Qff:: off
on
Circulation
&&Pot.
Tem.
Distribution
across
BB,
55thth2005
DA-SM2-U
&
Circulation
& Pot.
Tem.
Distribution
across
BB,
5th 2005
RA-SLAB
Circulation
Pot.
Tem.
Distribution
across
BB,Dec
DecDec
2005DA-SM2-U
&Q
Qff:: off
on
10-6m2s-1Kkg-1 = 1 PVU
at 4 pm
Wind
Wind field
field at
at 44 am
am at
at 25
25 m
m AGL
AGL
th 2005
th 5
December
December
5th
2005
DA-SM2-U
RA-SLAB
& Qff: off
on
th
Temperature
Temperature
field
field
at at
4 am
4 am
at at
2m
2m
AGL
AGLDecember
December
December
55thth52005
2005DA-SM2-U
2005
DA-SM2-U
RA-SLAB
& Qf: on
off
th 2005
Circulation
Circulation
Circulation
&&Pot.
Pot.
& Pot.
Tem.
Tem.
Tem.
Distribution
Distribution
Distribution
across
across
across
BB,
BB,
BB,
Dec
Dec
Dec
55thth52005
2005DA-SM2-U
DA-SM2-U
RA-SLAB
& Qf: on
off
4 am
Comparison between observation and simulations
360
290
3.5
Obs
Obs Obs
RA
RA RA
DA-SM2-U
Qf : off
DA-SM2-U
Qf DA-SM2-U
: off Qf : off
DA-SM2-U
Da-SM2-U
Qf :DA-SM2-U
on Qf : onQf : on
289
315
2883
Wind direction (k)
-1
Temperature
Wind
speed (ms )
270
287
2.5
286
225
2852
180
284
1.5
283
135
2821
90
281
0.5
280
45
279
00
0
0
0
4
66
8
12
12
12
Local
Localtime
time
Local
time
16
1818
20
24
24 24
Conclusions:
Tehran PBl structure is extremely under effect of topographic and heat island flows and
circulations and their interaction and show a multi layer structure.
The roughness approach (RA) are unsatisfactory at neighbourhood scales in this area.
Urban Roughness-Induced Convergence and UHI Thermal Circulation.
within the canopies, the DA-SM2-U meteorological fields seem well simulated following
the canopy morphology: decrease of the wind speed inside the dense canopies, skirting of
the flow around the canopy blocks, warmer air inside the vegetation canopy than above
open areas during the night and conversely during the day, and constantly warmer air
inside the urban canopy .
A discontinuity in the eddy diffusivity is simulated by DA-SM2-U between the canopies
and the upper atmosphere, with small values of the eddy diffusivity inside the canopies,
inducing a limitation of the turbulent exchanges between the inside and the outside of the
canopies.
UHI Displacement/De-stabalized Boundary Layer.
The vertical profiles of the potential air temperature in urban areas have shown that
simulations using SM2-U reduce the tendency toward stable stratification. They even yield
a neutral layer during the night because of the anthropogenic heat fluxes and also the heat
released by urban surfaces.
These results indicate that using a UCP may have significant ramifications for air-quality
modeling at this scale because the dynamical characteristics of the volume in which
pollutants are injected has been altered.
Thank you
for
listening