David Carruthers, Jenny Stocker, Christina Hood

CERC

Jimmy Fung and Xie Bo

HKUST

BAQ 2012

ADMS-Urban nested within regional

models:

an air quality model for all scales

BAQ 2012

ADMS-Urban Model Capabilities

ADMS-Urban is designed to model dispersion from a wide range of urban sources

Gaussian type model with point, line area, road and grid sources; non-Gaussian vertical profile of concentration in convective conditions

Concentration calculated at high resolution (<10m)

Fully integrated street canyon model based on OSPM

Options for different chemical mechanisms ; considers effects of complex terrain

Integration with Geographical Information Systems (GIS) and an Emissions Inventory Database (EMIT)

Used in many major cities for air quality management: e.g: London, Dublin, Budapest, Rome, Beijing, Shanghai, Singapore

BAQ 2012

ADMS-Urban examples London: Modelled annual average NO2: 2001 & 2010 (DEFRA)

•N •O •2 • •( •µ •g •/ •m •³ •)

•> • •5 •6

•5 •2 • •- • •5 •6

•4 •8 • •- • •5 •2

•4 •4 • •- • •4 •8

•4 •0 • •- • •4 •4

•3 •6 • •- • •4 •0

•3 •2 • •- • •3 •6

•2 •8 • •- • •3 •2

•2 •4 • •- • •2 •8

•2 •0 • •- • •2 •4

•< • •2 •0

•0 •5 •1 •0 •K •i •l •o •m •e •t •e •r •s

•F •i •g •u •r •e • •6 •. •4

•L •o •n •d •o •n • •2 •0 •0 •1

•A •n •n •u •a •l • •a •v •e •r •a •g •e • •N •O •2 • •c •o •n •c

•N •O •2 • •( •µ •g •/ •m •³ •)

•> • •5 •6

•5 •2 • •- • •5 •6

•4 •8 • •- • •5 •2

•4 •4 • •- • •4 •8

•4 •0 • •- • •4 •4

•3 •6 • •- • •4 •0

•3 •2 • •- • •3 •6

•2 •8 • •- • •3 •2

•2 •4 • •- • •2 •8

•2 •0 • •- • •2 •4

•< • •2 •0

•F •i •g •u •r •e • •4 •. •1 •b

•L •o •n •d •o •n • •2 •0 •1 •0 • •- • •2 •0 •% • •N •O •2

•A •n •n •u •a •l • •a •v •e •r •a •g •e • •N •O •2

•onitoring network

CERC

Annual average NO2 concentration

National Environment AgencySingapore

NO2 (µg/m³)

>100

75 - 100

50 - 75

45 - 50

40 - 45

36 - 40

30 - 35

25 - 30

20 - 25

< 20

0 5 10 15 20 25 Kilometres

Beijing: AQ Forecasting Singapore: AQ Assessment (NEA)

BAQ 2012

Nesting local model (ADMS-Urban) within

regional model: motivation

• Why nest a local model within a regional model?

• What are the advantages of a nested model?

Model feature Model

Regional (eg grid based) Local (eg Gaussian plume)

Domain extent Country (few 1000 km) City (50km)

Meteorology Spatially and temporally

varying from meso scale

models

Usually spatially

homogeneous

Dispersion in low wind

speed conditions

Models stagnated flows

correctly

Limited modelling of

stagnated flows

Deposition and

chemical processes

Reactions over large spatial

and temporal scales

Simplified reactions over

short-time scales

Source resolution Low High

Validity Background receptors Background, roadside and

kerbside receptors

BAQ 2012

Regional model/ADMS-Urban nesting system

• Aim: to nest local model in regional model without double

counting emissions:

Concentration within

nested domain =

Regional

modelling of

emissions (all

time)

-

Gridded locally

modelled

emissions (ΔT)

+

Explicit locally

modelled

emissions (ΔT)

source

wind

speed, U

~U x ΔT

• ΔT is time for pollutants to be well mixed on grid scale

BAQ 2012

Regional model/ADMS-Urban nesting implementations

Regional model ADMS-

Urban

nested

domain

No. of

receptors

AQD

pollutants

modelled Name Met Domains

modelled

Resolution

(km x km)

CMAQ MM5 SE England 3 x 3

9 x 9 km2,

centred on

central

London

17 NOx, NO2,

O3

EMEP4UK WRF

Europe 50 x 50 4 (5 x 5)

km2,

centred on

central

London

19

NOx, NO2,

O3, PM10,

PM2.5, SO2

UK 5 x 5

Scotland 1 x 1

CMAQ WRF

Europe 27 x 27 81 (1 x 1)

km2,

centred on

central

London

17

NOx, NO2,

O3, PM10,

PM2.5, SO2

UK 9 x 9

Southern

England

3 x 3

London 1 x 1

CMAQ

/CAMx

WRF/

MM5 Hong Kong

1 x 1

... .... ....

BAQ 2012

Preliminary modelling Model set up: NO2 emissions

初步模拟模型设置：NO2排放

NO2 t/year/km2

NO2_t_year

0 - 1

1 - 3

3 - 5

5 - 10

10 - 13

13 - 15

15 - 18

18 - 20

20 - 25

25 - 40

0 30 6015 Kilometers

Road

NO2_t_year

0 - 0.25

0.25 - 0.45

0.45 - 1.0

1.0 - 1.5

1.5 - 2.5

2.5 - 4.0

4.0 - 6.0

6.0- 8.50

Grid

NO2_t_year

6.852773 - 6.900000

6.900001 - 7.000000

7.000001 - 7.100000

7.100001 - 10.000000

10.000001 - 13.000000

13.000001 - 15.000000

15.000001 - 18.000000

18.000001 - 20.000000

20.000001 - 25.000000

25.000001 - 40.000000

•ADMS explicit

•9km

•CMAQ

BAQ 2012

CMAQ (Preliminary modelling)

Model results: receptors – NO2

(Summer)

Roadside and Kerbside Background

BAQ 2012

0

10

20

30

40

50

60

0 10 20 30 40 50 60M

od

elle

d P

M1

0 (

µg

/m³)

Measured PM10 (µg/m³)

ADMS nested

ADMS only

EMEP4UK

0

10

20

30

40

50

60

0 10 20 30 40 50 60

Mo

de

lle

d O

3 (

µg

/m³)

Measured O3 (µg/m³)

ADMS nested

ADMS only

EMEP4UK

O3 PM10

(All sites)

EMEP4UK (Full implementation)

Model results: receptors – O3 and PM10

BAQ 2012

CMAQ (Preliminary modelling)

Model results: NO2 average contours

CMAQ outer

domain

ADMS-Urban

nested domain

9km

Consistency of

background

concentrations

1km

NO2 µg/m³

BAQ 2012

•O

3

•NO2 •VOC

•NO

•13 Sep 99 •Level one •3:00p

m •O3

•NO2

•NO

•VOC

HKUST CMAQ/CAMx Modelling for Hong Kong;

ADMS-Urban to be nested within this

BAQ 2012

London Model Inter-comparison Study:

Annual Average NO2

• NO2

BAQ 2012

Pollution roses at Marylebone Road, London: NOx and

NO2

Marylebone Road 2008: road width 22m, canyon height 20m

ADMS/OSPM

Calculated NOx

Monitored

NOx

ADMS/OSPM

Calculated NO2

Monitored

NO2

BAQ 2012

New Canyon Model

• Well mixed recirculating region

• Flow/turbulence within canyon has

dependence on height and width; allows for

high aspect ratio

• Can consider pavements and road lanes

• One-sided/asymmetric canyons

• Channelling of flow along canyon

source ( )

plume centrelines

g

mth street

m+1th street

OSPM

Urban profiles of mean flow

and turbulence

BAQ 2012

Discussion

• Nesting a model with explicit source representation and street scale

spatial resolution (e.g.ADMS-Urban within a regional model

(e.g.CMAQ, CAMx, EMEP) allows the full range of relevant scales to

be considered within one modelling system

• This approach allows for modelling of:

– chemical reactions over large and small temporal and spatial

scales

– spatially varying meteorology

– low wind conditions

– local modelling (street canyons, noise barriers, cuttings etc)

• The system allows for model validation at all sites: background,

roadside and kerbside

• A revised street canyon model is being developed to allow for

asymmetric canyons and those of high aspect ratio; it will take

account of current knowledge of flow in urban areas.