Transcript Coutts et al 2007

An urban canopy model for Australian
regional climate and air quality modelling
www.cawcr.gov.au
Marcus Thatcher
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Overview
• Motivation and applications
• Basic features of Urban Canopy Models (UCM)
• Design of our UCM module (ateb.f90)
• Basic verification
• Future work
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Motivation
• There is currently growing interest in the dynamical modelling of urban climates
(TEB, UHSM, LUMPS, CM-BEM, VUCM, FVM, etc)
• Urban climate issues of interest include:
•
•
•
•
Response of urban climates under global warming scenarios
Modelling of aerosols and atmospheric chemistry in urban environments
Implications for public health
Implications for energy security
• To support the anticipated demand, we have upgraded our urban modelling
capabilities for regional climate modelling and air quality simulations
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Energy security - GENERSYS
2005 Load Duration Curve (SA)
2500
Demand (MW)
2000
Observed
1500
Model
1000
500
0
20
40
60
80
100
Percentage of tim e
+1oC climate change scenario (SA)
(no adaptation)
10.0%
Increased growth
8.0%
6.0%
4.0%
2.0%
0.0%
0
20
40
60
80
-2.0%
Percentage of tim e
•
GENERSYS National Electricity Market simulator (Grozev, et al). Electricity
demand from (Thatcher, 2007) based on Mk3.0 downscaled by CCAM
climate datasets
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
100
Air quality and health
CLIMATE RUN: 1996-2005. AVERAGE NUMBER OF DAYS CLIMATE RUN: 2021-2030. AVERAGE NUMBER OF DAYS
PER SEASON WITH 4-HR OZONE > 80 ppb
PER SEASON WITH 4-HR OZONE > 80 ppb
NEWCASTLE
NEWCASTLE
PENRITH
3
LITHGOW
TASMAN SEA
PENRITH
3
PENRITH
2
PICTON
PICTON
2
PICTON
1
1
WOLLONGONG
WOLLONGONG
3
SYDNEY
2
1
WOLLONGONG
0
0
1996-2005
TASMAN SEA
SYDNEY
SYDNEY
EASTING (m)
NORTH (m)
TASMAN SEA
NORTH (m)
NORTH (m)
NEWCASTLE
LITHGOW
LITHGOW
•
CLIMATE RUN: 2051-2060. AVERAGE NUMBER OF DAYS
PER SEASON WITH 4-HR OZONE > 80 ppb
EASTING (m)
0
EASTING (m)
2021-2030
2051-2060
Study by Martin Cope, et al, using the Chemical Transport Model (CTM),
after dynamically downscaling from CSIRO Mk3.0 using CCAM and TAPM.
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Estimated effect of climate change on Melbourne ozone
(Sean Walsh, 2009 – EPA Victoria)
Averages of Predicted Monthly Max 4-hr Ozone (ppb)
90
80
4-hour Ozone Air Quality Objective = 80 ppb
y = 0.15x - 238
R2 = 0.58
70
60
Nov-March Average of Monthly Max O 3
50
40
30
20
10
0
1990
2000
2010
2020
2030
2040
2050
2060
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
2070
Properties of Urban Canopy Models
• There are a few different approaches
to modelling the ‘building-averaged’
urban canopy within mesoscale
atmospheric models
• Slab models (e.g., Oke 1988)
• 2D canyon models (e.g., Masson,
2000)
• 3D building array models (e.g., Kanda
et al, 2005)
• Multi-level canopy models (e.g.,
Martilli et al, 2002)
• These models also attempt to solve
for the urban energy budget and
adapt MOST to the urban
environment
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of Urban Canopy Models
• There are many differences between urban and rural environments, e.g.,
•
•
•
•
•
•
•
•
•
Urban areas have increased roughness lengths (e.g., 1m)
The ratio of z0m/z0h is much larger in urban areas (e.g., ~1000 or greater)
Moisture availability is reduced due to impervious surfaces (high Bowen ratio)
Storage heat flux is significantly higher in urban areas (Q = Net – H – LE)
Anthropogenic contribution to heat flux in urban areas
Reduced emissivity and albedo
Increased thermal inertia compared to rural areas
Canyon shadowing effects
Buff bodies instead of a porous canopy
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of Urban Canopy Models
TAPM
slab
• Coutts et al 2007 used the urban slab
model in TAPM V3 to model the
partitioning of fluxes at different sites
in Melbourne
• After tunning the model, Coutts et al
were able to obtain a reasonable
partitioning of energy for January
2003
Obs
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of Urban Canopy Models
TAPM
slab
• However, the partitioning of the fluxes
in July 2003 is incorrect (note the
heat storage term)
Obs
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of TEB
• As part of the upgrading of our urban modelling capabilities for regional climate
simulations, we have coded a urban module based on an idealised 2D canyon
following Town Energy Budget (TEB) approach
• Note that the new urban model is designed around research requirements, not
operational requirements. Specifically, many parts of the model are designed
for future extensions.
• Before considering the new urban model, we first briefly review some of the
properties of TEB models
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of TEB
• In TEB, the urban energy budget is separated into three energy budgets for
roofs, roads and walls. The canyon is rotated through all possible orientations
and so a single wall energy budget is used.
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of TEB
• The TEB scheme recommends using
three layers when representing roofs,
walls and roads. Soil moisture and
in-canyon vegetation is usually
neglected (see later)
• Internal building heat source
(modifies behaviour of ‘heat storage’)
• A single layer snow scheme and a
simple model of surface water is
included for roof and road surfaces
(max 1 kg/m2)
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of TEB
• The inclusion of a surface water reservoir is to account for short term changes
to the fluxes (i.e., less than daily time scales)
Masson 2000
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of TEB
• TEB assumes isotropic reflections in shortwave and long wave radiation.
Usually an infinite number of reflections is assumed for shortwave and 1st order
reflections are modelled for longwave
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of TEB
• The canyon morphology (H/W)
has a significant influence on how
radiation is absorbed within the
canyon (i.e., shadowing effects)
• Higher canyon walls relative to
the canyon width results in less
radiation being reflected by the
canyon system
• The albedo also changes as a
function of solar zenith angle due
to differences in the albedo
between walls and roads
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of TEB
• TEB relies on MOST to describe the
turbulent heat exchange with the
atmosphere
• The urban sensible heat flux includes
contributions from roofs, roads, walls
and (potentially) snow
• Time dependent traffic fluxes are
included in the canyon sensible heat
flux budget
• ‘Industrial’ sensible heat is added to
the total urban sensible heat flux
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Properties of TEB
• Fairly simple relationships are used
by TEB for parameterising in-canyon
aerodynamical resistances
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90
• To improve the urban modelling capability for regional climate modelling and air
pollution, we have written an urban module called ateb.f90
• Although the new module has been coded from scratch, much of the science is
based on the TEB / 2D canyon approach
• However, we have taken the opportunity to structure ateb.f90 to allow for
extensions to various components including
•
•
•
•
Use of Harman et al (2004) scheme for in-canyon aerodynamical resistances
Prognostic equations for a second canyon wall
Vegetation and AC contributions to in-canyon fluxes
Use of Luhar et al (2008) parameterisation of turbulence for low winds
• Implicit numerical techniques are used to ensure stability with large time steps
(e.g., dt=20mins in CCAM 60 km resolution simulations)
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90
• ateb.f90 uses separate prognostic
equations for two canyon walls (i.e.,
the canyon is rotated through 180o
instead of the usual single wall
rotated through 360o)
road
buildings
canyon
buildings
westerly
wall
easterly
wall
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90
• In-canyon aerodynamical
resistances are based on
Harman et al (2004)
• This formulation accounts for
recirculation and venting
regimes in the canyon
• Some of the parameters have
been modified where necessary
• Parameterisations for
aerodynamical resistances used
by Masson (2000) and Kusaka
et al (2001) are also available
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90
• ateb.f90 includes an incanyon big-leaf vegetation
scheme based on Kowalczyk
(1994), assuming f=1 for
simplicity and therefore no
multi-layer soil model under
vegetation (see also Lee and
Park 2008)
• We have added a simple AC
heat flux into canyon (see
also Ohashi et al, 2007)
AC
• Time dependent traffic fluxes
are based on Coutts et al
(2007)
• Canyon air and vegetation
canopy temperatures are
solved simultaneously with
the Newton-Raphson method
traffic
Schematic representation of the aerodynamic resistances.
Note snow and water have been omitted in this diagram
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90
• The stability functions have
been modified for low winds
following Luhar et al (2008)
• However, there are
complications with the
aerodynamical resistance
between the canyon and the
atmosphere
AC
traffic
Schematic representation of the aerodynamic resistances.
Note snow and water have been omitted in this diagram
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90
• Based on the results of Kanda et al
(2008), we have experimented with
changes to the roughness length for
heat when calculating the sensible
heat flux between the atmosphere
and the canyon air temperature
• Basically zh depends on the
definition of the ‘urban’ temperature
Kanda et al 2008
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90
zh=z0m
Solve for z0h
(z0h <<z0m)
approximated
by
TS=(L/)1/4
Kanda et al 2008
• Note that zoh for the urban tile is
still much smaller than zom
• z0h=10-3z0m is fed back to the
host atmospheric model for
diagnosing screen level
variables (ateb.f90 also
supports the Brutsaert (1982)
parameterisation)
ateb.f90
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90
• We have included a single layer snow
scheme based on Douville et al
(1995)
• Note that changing snow albedo can
effect the albedo of the canyon
system
• Therefore, we have simply expanded
the number of reflections to Nth order
for both longwave and shortwave (3rd
order by default), rather than solve for
an infinite number of reflections
(slightly complicated by the 2nd
canyon wall)
• The time dependent urban albedo is
fed back to the radiation scheme in
the host atmospheric model
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90
• ateb.f90 is a single layer model for
consistency with the vegetation
canopy scheme (e.g., CABLE)
• For regional climate modelling, this is
acceptable since the results of
Kusaka et al (2001) suggest that the
fluxes calculated by a single level
model are close to that obtained from
multi-level models
• Nevertheless, air pollution modelling
would probably benefit from a multilevel model and this may be a future
extension
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Basic evaluation of ateb.f90
• Our study is based on Melbourne
after downscaling from 1o resolution
NCEP analyses to 3 km resolution
using a multiple nesting technique
MEL
ALP
• TAPM is run for 1 year (2003)
FOO
• The results presented here are based
on the Alphington site (ALP)
• Observations are from the
Environmental Protection Authority
(EPA) monitoring stations and from
Coutts et al (2007)
BRI
Port Phillip Bay
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Simulated temperature climatology
No urban
SLAB urban
0.3
0.3
0.2
0.2
pdf_mod
0.1
pdf_obs
pdf
pdf
pdf_obs
pdf_mod
0.1
0.0
0.0
0
5
10
15
20
25
Tem perature ( oC)
30
35
40
0
5
10
15
20
25
Tem perature ( oC)
30
35
40
• ‘No urban’ shows a cold bias due to
absence of urban heat island
aTEB urban
0.3
0.2
pdf
pdf_obs
pdf_mod
• ‘aTEB urban’ predicts low temperatures
better than the SLAB approach
0.1
0.0
0
5
10
15
20
25
Tem perature ( oC)
30
35
40
• The correlation between modelled and
observed temperatures is also improved
with aTEB
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Simulated wind climatology
No urban
SLAB urban
0.3
0.3
0.2
pdf
pdf_obs
pdf_mod
0.1
pdf_obs
pdf
0.2
pdf_mod
0.1
0.0
0.0
0
5
10
Wind Speed (m s -1)
15
20
0
5
10
Wind Speed (m s -1)
15
20
aTEB urban
0.3
0.2
pdf
pdf_obs
• aTEB and SLAB predict similar wind
speed climatologies, but both are better
than for ‘No urban’
pdf_mod
0.1
• Wind direction errors are essentially the
same for all three approaches
0.0
0
5
10
Wind Speed (m s -1)
15
20
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
March energy partitioning
SLAB urban
500
500
400
400
300
300
NETR
200
SENS
Flux (W/m2)
Flux (W/m2)
No urban
200
100
0
0
6
12
18
24
-100
-200
EVAP
100
STORE
0
0
-100
NETR
SENS
EVAP
STORE
6
12
18
24
-200
Time (hrs)
Time (hrs)
Coutts et al 2007
aTEB urban
500
400
Flux (W/m2)
300
NETR
200
SENS
100
EVAP
STORE
0
-100
0
6
12
18
24
-200
Time (hrs)
• aTEB shows improved partitioning of energy (particularly ‘storage’). Note the
The Centre for Australian Weather and Climate Research
slight lag in fluxes
A partnership between CSIRO and the Bureau of Meteorology
May energy partitioning
SLAB urban
500
500
400
400
300
300
NETR
200
SENS
Flux (W/m2)
Flux (W/m2)
No urban
200
100
0
0
6
12
18
24
-100
-200
EVAP
100
STORE
0
0
-100
NETR
SENS
EVAP
STORE
6
12
-200
Time
(hrs)
aTEB
urban
Time (hrs)
18
24
Coutts et al 2007
500
400
Flux (W/m2)
300
NETR
200
SENS
100
EVAP
STORE
0
-100
0
6
12
18
24
-200
Time (hrs)
• aTEB also partially captures changes in energy partitioning between different
The Centre for Australian Weather and Climate Research
months
A partnership between CSIRO and the Bureau of Meteorology
Nesting in CCAM
Modelled vs Observed Temperature
Nested in NCEP analyses
40
0.3
35
y = 0.9415x + 0.7508
R2 = 0.9007
30
25
0.2
pdf
pdf_mod
0.1
T_MOD
20
pdf_obs
15
10
5
0
-5
0.0
0
5
10
15
20
25
Tem perature ( oC)
30
35
-10
40
-15
-15 -10 -5 0 5 10 15 20 25 30 35 40
Modelled vs Observed
Temperature
T_OBS
Nested in CCAM
(forced by NCEP analyses)
40
35
0.3
y = 0.9087x + 2.2917
R2 = 0.759
30
25
0.2
pdf
pdf_mod
0.1
T_MOD
20
pdf_obs
15
10
5
0
-5
0.0
0
5
10
15
20
25
Tem perature ( oC)
30
35
40
-10
-15
-15 -10 -5
0
5
10 15 20 25 30 35 40
• Urban temperature climatology is still retained when theT_OBS
correlation for timeThe Centre for Australian Weather and Climate Research
paired temperatures is (intentionally) degraded
A partnership between CSIRO and the Bureau of Meteorology
Nesting in CCAM
Modelled vs Observed Wind Speed
Nested in NCEP analyses
20
y = 0.7306x + 0.7397
R2 = 0.6567
0.3
0.2
pdf
pdf_obs
pdf_mod
0.1
WS_MOD
15
10
5
0.0
0
5
10
Wind Speed (m s -1)
15
20
0
0
Nested in CCAM
(forced by NCEP analyses)
5
10
15
Modelled vs Observed Wind Speed
WS_OBS
20
20
y = 0.5255x + 1.1164
R2 = 0.3634
0.3
0.2
pdf
pdf_obs
pdf_mod
0.1
WS_MOD
15
10
5
0.0
0
5
10
Wind Speed (m s -1)
15
20
0
0
5
10
15
20
WS_OBS
• The wind climatology is also preserved when TAPM is nested in CCAM
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
ateb.f90 and CABLE
• ateb.f90 has also been used with
a mosaic/tiled version of CABLE
in CCAM
Oklahoma city, JU2003 study – Screen temperature
• In the Oklahoma city JUL2003
study we dynamically downscale
to 1 km resolution
• Urban prognostic variables are
initialised from an urban tile in the
CCAM 60 km resolution
simulation that has been spun-up
Urban heat island
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Future work
• Parameter tunning
• The simulated climatology is independent of may parameters (e.g., industrial and
traffic heat fluxes seem to have a minor effect)
• Canyon geometry is estimated from Coutts et al (2008) study for suburban
Melbourne
• Thermal properties of building materials is taken from the ECOCLIM dataset
• Issues with heat storage in roofs and methods for predicting screen level
observations within the canyon
• More realistic treatment of turbulent heat fluxes between canyon and
atmosphere
• Planned experiments include
• Simulation of the 20thC Australian urban climate downscaled from various GCMs
• Simulation of 2050 and 2070 Australian urban climate
• Application to air quality and energy models
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
Marcus Thatcher
Research Scientist
Phone: 03 9239 3530
Email: [email protected]
Web: www.cawcr.gov.au
Thank you
www.cawcr.gov.au
Thank you
Acknowledgements to Peter Hurley
March energy partitioning
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology
May energy partitioning
The Centre for Australian Weather and Climate Research
A partnership between CSIRO and the Bureau of Meteorology