EXPERIMENTAL AND MODELLING STUDIES OF ULTRAFINE PARTICLECONCENTRATIONS IN URBAN STREET AND BACKGROUND ENVIRONMENT

M. KETZEL1,2,#, P. WÅHLIN1, R. BERKOWICZ1 and F. PALMGREN1

1National Environmental Research Institute, Roskilde, Denmark

2Division of Nuclear Physics, Physics Dept., Lund University, Lund, Sweden

#Corresponding author. mke@dmu.dk (M.Ketzel)

SummaryThe presentation has two main parts focussing on a) results from ambient particle measurements and b)application of different air pollution models to estimate the particle concentration at kerbside and in theurban background.Measurements

We reported previously on long term measurements of particle size distribution (size range 10-700nm)and PM10 at street and urban background location in Copenhagen (Ketzel et al., 2003a). Thesemeasurements enable us to estimate vehicle emission factors. In addition we present here seasonalvariations and correlation with meteorological parameters (e.g. temperature, wind speed, rain events)(Palmgren et al., 2003).We show that we are able to fit the size distributions averaged separately for each hour of the week with 3lognormal modes. The variation of these 3 modes together with CO and NOx measurements gives thebasis for a factor analysis. As one possible solution we obtain 3 factors resembling the following profiles:

Factor 1: Mode 1+2; NOxFactor 2: Mode 1+3; NOxFactor 3: - NOx; CO

An interesting aspect is that the Factor 3 explains all the variation in CO (tracer for gasoline vehicles) butdoes not contain any ultrafine particles. That means that we can interpret our field measurements withoutgasoline particle emissions and only two different types of 'diesel factors'.We also report first results of a short term measuring campaign (6 weeks) conducted simultaneously at anurban background station in Copenhagen and two rural background locations ca. 30 km west and 60 kmnortheast of Copenhagen (Ketzel et al., 2003b). These measurements allow for estimating the contributionfrom the city to the elevation of particle number concentration and PM10.Modelling

In this study we are applying these emission factors together with a street pollution model OSPM(Berkowicz, 2000b) to predict time series of particle number and NOx at street level. We show that anassumed temperature dependence of the particle emissions can improve the model results.The time scales for particle dynamic processes (e.g. coagulation, deposition, dilution) are estimated forurban street and background environment and discussed in their relevance for influencing the particle sizedistribution. In agreement with the literature (Pohjola et al., 2003; Vignati et al., 1999) we conclude thatcoagulation is too slow to alter the size distribution in the exhaust plume and dilution is the dominatingprocess. A similar conclusion can be drawn for the street scale. In a more confined environment as e.g. aroad tunnel the removal processes coagulation and deposition might play a important role (Gidhagen etal., 2003).First model results for the urban scale using a simple plume approach for the pollutantdispersion(Berkowicz, 2000a) in connection with the aerosol model AERO3 (Vignati, 1999) show that 1)the coagulation and deposition account for ca. 20% loss of particles in the urban background compared tostreet level and 2) the size dependent removal lead to a slight increase in ‘max’ - diameter of the particlesize distriution.

REFERENCES

Berkowicz, R. (2000a): A simple model for urban background pollution. Environmental Monitoring and Assessment65, 259-267.

Berkowicz, R. (2000b): OSPM - A parameterised street pollution model. Environmental Monitoring and Assessment65, 323-331.

Gidhagen, L., Johansson, C., Ström, J., Kristensson, A., Swietlicki, E., Pirjola, L., and Hansson, H.-C. (2003): Modelsimulation of ultrafine particles inside a road tunnel. Atmospheric Environment 37, 2023-2036.

Ketzel, M., Wåhlin, P., Berkowicz, R., and Palmgren, F. (2003a): Particle and trace gas emission factors under urbandriving conditions in Copenhagen based on street and roof-level observations. Atmospheric Environment 37,2735-2749.

Ketzel, M., Wåhlin, P., Kristensson, A., Swietlicki, E., Berkowicz, R., Nielsen, O. J., and Palmgren, F. (2003b):Particle size distribution and particle mass measuremnts at urban, near-city and rural levl in the Copenhagenarea and Southern Sweden. Atmos.Chem.Phys.Discuss.(submitted for publication).

Palmgren, F., Wåhlin, P., Berkowicz, R., Ketzel, M., Illerup, J. B., Nielsen, M., Winther, M., Glasius, M., andJensen, B. (2003): Aerosols in Danish Air (AIDA), Mid-term report 2001-2002. National EnvironmentalResearch Institute, Roskilde, Denmark. NERI Technical Report No. 450.

Pohjola, M., Pirjola, L., Kukkonen, J., and Kulmala, M. (2003): Modelling of the influence of aerosol processes forthe dispersion of vehicular exhaust plumes in street environment. Atmospheric Environment 37, 339-351.

Vignati, E. (1999): Modelling Interactions between Aerosols and Gaseous Compounds in the Polluted MarineAtmosphere. PhD Thesis. Risø National Laboratory, Denmark, Report: Risø-R-1163(EN).

Vignati, E., Berkowicz, R., Palmgren, F., Lyck, E., and Hummelshoj, P. (1999): Transformation of size distributionsof emitted particles in streets. The Science of The Total Environment 235, 37-49.

Experimental and Modelling Studies ofUltrafine Particle Concentrations in

Urban Street and Background Environment7th ETH Conference 2003

Matthias Ketzel, Peter Wåhlin, Ruwim Berkowiczand Finn Palmgren

National Environmental Research Institute, Roskilde, Denmark

Funding: Danish EPA

Outline of the presentation■ Objective

■ Particle measurements in Copenhagen(follow up of last years presentation)

■ Dispersion modelling– time scale analysis– kerbside modelling– urban scale modelling

Objective

■ ambient measurements of nanoparticles– ‘real world’ particle emission factors +

size distribution– separated for vehicle classes (gasoline/diesel; LD/HD)– analyse dependence on meteorological parameters

■ include particles in our dispersion models– forecast of pollution levels– scenario calculations– ...

Measuring station at Jagtvej, Copenhagen

HCØ (background)particles trace gas meteorologyDMPS 10-700nm CO wind direction(UCPC > 3nm) NO wind speedPM10 - Teom NOx temperature

O3 RHglobal radiation

Jagtvej (street canyon)particles trace gas divDMPS 10-700nm CO traffic countsPM10 - Teom NO rainPM10 - filter packs NOx

26 000 veh./day; 6-8% HDVDanish diesel: 50 ppmS; gasoline 60-70 ppm SPair of stations: kerbside -- urban background

0 6 12 18 24 30 36 42 48 54 60 66

-200

0

100

400

1000

3000

4000

8000

10000

20000

40000

100000

200000

dN/dlog Dp (cm-3)

Dp

in

nm

working days Saturdays Sundays

0

5

10

15

20

25

30

0 6 12 18 24 30 36 42 48 54 60 66Rat

io C

O/N

Ox

(pp

b/p

pb

)

RatioCO/NOx

■ constant modediameter during day– diesel ≈ gasoline

■ night time shift tosmaller sizes– diesel taxi + oxicat.– nucleation ?!

0 6 12 18 0 6 12 18

time (‘hour of the week’)

Average weekly profile of size distribution, Difference Street - backgroundperiode 15-05 to 23-11-2001 ca. 12 weeks of data (1h averages)

Ketzel et al. 2003 Atmos. Env. 37, 2735

??

Fitting of the particle size distribution with 3 log-normal modes

0

10000

20000

30000

40000

50000

60000

1 10 100 1000

MeasuredSum FitMode1: 63.6 nm, GSD= 1.97Mode2: 23.9 nm, GSD= 1.57Mode3: 11.7 nm, GSD= 1.42

Day: 6 ;Hour: 4

0

10000

20000

30000

40000

50000

60000

1 10 100 1000

MeasuredSum FitMode1: 63.6 nm, GSD= 1.97Mode2: 23.9 nm, GSD= 1.57Mode3: 11.7 nm, GSD= 1.42

Day: 5 ;Hour: 7

0

10000

20000

30000

40000

50000

60000

1 10 100

MeasuredSum FitMode1: 63.6 nmMode2: 23.9 nmMode3: 11.7 nm

Day: 5 ;Hour: 22

Source Profiles

0

100

200

300

400

500

600

700

800

1 10 100

Sum FitMode1: 63.6 nm, GSD= 1.97Mode2: 23.9 nm, GSD= 1.57Mode3: 11.7 nm, GSD= 1.42

Factor 2 (modern diesel; Taxi)

0

100

200

300

400

500

600

700

800

1 10 100 1

Sum FitMode1: 63.6 nm, GSD= 1.97Mode2: 23.9 nm, GSD= 1.57Mode3: 11.7 nm, GSD= 1.42

Factor 1 (normal diesel)

Dp in nm

(dN

/dlo

gd)/

[NO

x] (

cm-3

/ppb

)Factor Analysis

■ Average weekly variation of– NOx– CO

– Strength of Mode 1-3

(all corrected for background !)

■ Possible Solution:– Factor 1: Mode 1+2; NOx

– Factor 2: Mode 1+3; NOx

– Factor 3: - NOx; CO

kerbside (Jagtvej) 2001-2002

y = -4.0785x + 522.88

R2 = 0.2241

0

100

200

300

400

500

600

700

-10 0 10 20 30

Temp [oC]

To

N /

NO

x [p

t/cm

3 /pp

b]

Temperature variations - particle numberlower particle / NOx ratio at lower temperatures

0100200300400500600700800900

1 10 100 100d (nm)

(dN

/dlo

gd)/

[NO

x] (

cm-3

/ppb

)

0 C < t < 5 C5 C < t < 10 C10 C < t < 15 C15 C < t < 20 C20 C < t

Ratios between particle number concentration (versus size)and the NOx concentration at HCAB (street station).

by P.Wåhlin

low temp.night traffic

Conclusions - ambient measurements:■ fleet emission factors could be estimated (last years talk)

■ 3 - modal structure of the size distribution

■ possible factor solution:– 2 factors (NOx + particles)– 1 factor (NOx + CO, no particles)

■ temperature dependence

Modelling of ambient ultrafine particles

■ Which particle transformation processes arerelevant at the different scales?

1.0E-03

1.0E-02

1.0E-01

1.0E+00

1.0E+01

1.0E+02

1.0E+03

1.0E+04

1.0E+05

1.0E+06

1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08 1.E+09

Concentration in #/cm3

Tim

e sc

ale

in s

Coagulation monodisperse

Coagulation polydisperse

Dilution

Deposition

kerbsideroad tunnel

exhaust plume

background

Time scales for several processes

Gidhagen et al. 2003- 77% loss for Dp < 10nm- 41% loss for Dp 10-29nm

Vignati et al. 1999Pohjola et al. 2003 :-no effect of coag.

Gidhagen et al.poster here

Modelling the dilution at kebside

Roof level wind

Leewardside

Windward side

Background pollution

Recirculating air

Direct plume

■ Operational Street Pollution Model (OSPM); Berkowicz et al. 1989...2003

C = F(wind direction, wind speed, vehicle speed ...) * Q [pt/km/s]

0

20

40

60

80

100

120

140

160

180

200

18-06-01 25-06-01 02-07-01 09-07-01

NOx_Str_ImodcNOX_str_obs_2

0

50

100

150

200

250

0 50 100 150 200 250 300

NOx_Str_Imod

Linear (NOx_Str_Imod)

y = 0.68

R2

0

10000

20000

30000

40000

50000

60000

70000

80000

0 20000 40000 60000 80000 100000

ToN_mod

Linear (ToN_mod)

-10000

10000

30000

50000

70000

90000

18-06-01 25-06-01 02-07-01 09-07-01

ToN_modToNdiff_obs

R2=0.64

R2=0.72

1:1

1:1

Modelling at kebside for NOx and total particle number■ OSPM and emission factors from inverse modelling■ inert particles (no particle dynamics)

measured

modelledbackgroundsubtracted

Temperaturedependent Emissions

y = 0.6197x + 2298.1

R2 = 0.6998

y = 0.7306x + 3457.9

R2 = 0.6806

y = 0.5311x + 923.22

R2 = 0.62

-10000

0

10000

20000

30000

40000

50000

60000

70000

80000

90000

0 20000 40000 60000 80000 100000 120000

class=-7

class=-2

class=2

class=7

class=12

class=17

class=23

Linear (class=12)

Linear (class=17)

Linear (class=7)

y = 0.6615x + 2413

R2 = 0.7019

y = 0.6721x + 3184.8

R2 = 0.6852y = 0.6601x + 1111.7

R2 = 0.63

0

20000

40000

60000

80000

100000

0 20000 40000 60000 80000 100000

class=-7

class=-2

class=2

class=7

class=12

class=17

class=23

Linear (class=12)

Linear (class=17)

Linear (class=7)

R2=0.67

E(T) = E(15C)*[1 - 0.034 (T-15C)]E(0)/E(15) ~ 150%

R2=0.64

Measurements at urban, near-city and rural levelNote on ...

Background

Lille ValbyHCOE

Vavihill

X

XX

10

100

1000

10000

100000

1 10 100 1000

Diameter [nm]

dN

/dlo

gD

p [

#/cm

3 ]

JGTV_2001HCOE_2001HCOE_2002LVBY_2002VVHL_2002

Modelling of the urban background

0.01 0.1 1Dp [nm]

0E+000

2E+003

4E+003

6E+003

8E+003

1E+004

dN/d

logD

p [c

m-3] LVBY near-city

no removal processesDeposition not CoagCoag. + Deposition

20% removal

shift of max. due to size dependent removal■ ‘plume’ model for vertical dispersion■ background from measurements■ emission factor + size distrib.

from kerbside measurements■ sectional model for particle

dynamics including■ processes modelled:

– background– emission

– vertical dilution

– deposition

– coagulation

Conclusions - modelling:■ street level

– coagulation is too slow to alter the size distribution– dilution is the dominant process at street level

– OSPM (without particle dynamics) can be used also for particles

– temperature dependent e-factors increase correlation observed vs.modelled results

■ urban scale– coagulation and deposition account for ca. 20% loss of particles

– size dependent removal leads to a slight increase in ‘max’ - diameter