Air Quality :Modeling GroupAir Quality :Modeling Group

R t d i kRecent and ongoing workGroup members, 31/8/2007

Ari Karppinen Mikhail Sofiev Juha NikmoJari Härkönen Kari RiikonenMarke Hongisto Leena KangasMi P hj l Ilkk V lkMia Pohjola Ilkka ValkamaJukka-Pekka Jalkanen Joana SoaresMervi Haakana Mari Kauhaniemi (Kuopio)J i Jä i Minna Rantamäki (25%)Jouni Jäppinen Minna Rantamäki (25%)

Noora Eresmaa (mat. leave) Leena Partanen (mat leave)Leena Partanen (mat. leave)

Personell: 14.25 + 2

FMI-Budget: 4.75 py+ projects: 9.5 py (300 k€ + 500 k€)Total: 0.8 M€/year (14.25 py + ~100k€ other costs)

http://www.fmi.fi/organisaatio/yhteys_34.htmlhttp://www.fmi.fi/research_air/air_2.html

Dispersion modelling

Development and evaluation ofair quality modelsair quality modelsCombination of meteorological

d l d di i d lmodels and dispersion modelsApplication of models and di i ti f i f tidissemination of information

Focus areas in modelling

1 Integrated modelling systems (from1. Integrated modelling systems (from emissions to impacts, from street canyon to global scale)global scale)

2. Combined utilisation of meteorological models and dispersion modelsmodels and dispersion models

3. Health effects of air pollution, especially d lli f th t ti f dmodelling of the concentrations of and

exposure to particulate matter

Modelling system - FMIWeather prediction models

Dispersion models -long-range, regional

Dispersion and effects models – urban, local

g y

ECMWFHIRLAMRCR

SILAM LRT, meso, radioactivity

UDM FMI b

CAR-FMI, roadsidey

OSPM (NERI), street canyon

UDM-FMI, urban

HILATAR

HIRLAMMBE

AROME,LAPS

LRT, meso ,chem.

ESCAPE, chemical id

BUOYANT, fires

FLUENT, CFD code Aerosol process models: UHMA (U Helsinki, FMI)

MONO32 (U Helsinki, Stadia)

accidents

EXPAND (FMI YTV)( , ) EXPAND (FMI, YTV) population exposure

MPP-FMI, Meteorological pre-processing model

A centre of expertise at the Kumpula campus: wild-land firesFMI Air Quality Earth Observation Climate and Global Change Kuopio Unit and

Ilmanlaadun seuranta-asemat 2003

FMI Air Quality, Earth Observation, Climate and Global Change, Kuopio Unit, and Univ. Helsinki

Joensuu

Sevettijärvi

Oulanka

Hailuoto

Kevo

Raja-JooseppiPallas

Hietajärvi

SodankyläEMEPWMO/GAWEGAPKansallinenYhdennetty seurantaErikoisasemaRadioaktiivisuusOtsoni

Ground-based observations of fire plumes at selected stations and in the TROICA

Ähtäri

Kotinen

Virolahti

Utö

Hietajärvi

Tikkakoski

Nurmijärvi

Helsinki

Evo

Ilomantsi

Jokioinen

Combineduse of data

Earth observation of wildfires: Fire

se ected stat o s a d t e O Ccampaign

Data assimilationand analysis

use o data

ECMWFHIRLAMAtlantic

HIRLAMNorth-Europe

HIRLAMLINUX

ECMWFHIRLAMAtlantic

HIRLAMNorth-Europe

HIRLAMLINUX

Top of the boundary layerPassive plume

Buoyantplume

wind

Gaussianmodel

K-theorymodel

Top of the boundary layerPassive plume

Buoyantplume

wind

Gaussianmodel

K-theorymodel

Earth observation of wildfires: Firespread and plume dispersion

analysis

.

. ..

p

Air flow

Burning region

Deposition

...

.. ...

.. . . .

Particles

.

. ..

p

Air flow

Burning region

Deposition

...

.. ...

.. . . .

Particles

Modelevaluation

Modelling of fire spread, plume rise, atmosphericdispersion and source apportionment

Controlled forest burningexperiment

NOx emissions from marine trafficJukka-Pekka

R l ti it i t• Real-time monitoring system for ship emissions• Position reports from p

transponder messages• NOx estimate based on

available technical dataavailable technical data• Current speed vs. design

speed • Possibility to track

emissions ship by ship• Combine emission data with• Combine emission data with

3D atmospheric models to model long range transport

Practical example

Mari, Mervi

Modelling NOx and NO2 concentrations in different urban environments

Vehicle Type Distribution

Models:• CAR-FMI (open area line source model)• OSPM (street canyon model)

Vehicle Type DistributionVehicle Type2

Vehicle Type3

Hourly Distribution of Traffic

0.060.070.080.090.10

ffici

ent

OSPM (street canyon model)

Conditions:• 4 open area line source cases

Vehicle Type1

0.000.010.020.030.040.05

OSP

M c

oefp

• 6 street canyons cases• 2 years• 3 heights in street canyons

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24timeWeekdays Saturdays Sundays

3 heights in street canyons• NOx and NO2 emission factors

Daily and Monthly Distribution of Traffic

1.2

1.3

s0.7

0.8

0.9

1.0

1.1

1.2

OSP

M c

oeff

icie

nts

Example of the street canyon in the OSPM

0.5

0.6

1 2 3 4 5 6 7 8 9 10 11 12time

O

Mon Tue Wed-Fri Sat Sun

Mari

Publication (in progress for Atm. Env.)Evaluation and application of a modelling system forEvaluation and application of a modelling system for predicting the concentrations of PM2.5 in an urban areaKauhaniemi M1, Karppinen A.2, Härkönen J.2, Kousa A.3, Alaviippola B.3, Koskentalo T.3, Aarnio P.3, Elolähde, T.3 and Kukkonen J.2 (1 = FMI Kuopio, 2 = FMI Helsinki, 3 = YTV)

VALLILAweekdays (in 2002)

combust12.7 %

KALLIOweekdays (in 2002)

non-combust

8.5 %

cold start

combust4.7 %

non-combust22.8 %

cold start3.5 %

long-range transport61.0 %

cold start1.3 %

long-range transport85.5 %

VALLILA

40

45

50

55

2:1 1:1

KALLIO

40

45

50

55

2:1 1:1

15

20

25

30

35

Obs

erve

d (µ

g/m

3 )

1:2

15

20

25

30

35

Obs

erve

d (µ

g/m

3 )

1:2

y = 0.94x - 0.630

5

10

0 5 10 15 20 25 30 35 40 45 50 55

Predicted (µg/m3)

R2 = 0.54N = 360

y = 0.95x + 1.150

5

10

0 5 10 15 20 25 30 35 40 45 50 55

Predicted (µg/m3)

R2 = 0.60N = 358

Joana

Traffic iF for BZ for different urban environmentsstreet canyon (OSPM) Helsinki downtown area (PC-CAR) and the whole Helsinkistreet canyon (OSPM), Helsinki downtown area (PC CAR) and the whole Helsinki Metropolitan Area (URB-CAR).

• PM2.5 iF spatial distribution from a single source: Salmisaari cogeneration PP

Ilkka Valkama/FMI

• Benzene regional dispersion modelling• Aerosol dynamics implementation

SILAM

Kari

Salmisaari power plant: iF of PM2.5

Intake fraction =

Concentration x Breathing rate x ActivityEmission

Leena

Intake fraction (iF)fraction (iF) studyfor PMfor PM2.5emissions from thefrom the cogenerating power plantpower plant Salmisaari in HelsinkiHelsinki

PM2.5 iF for 2005 calculated by EXPAND

Leena

Intake fraction (iF) studyfor traffic emitted benzene in Helsinki Metropolitanfor traffic emitted benzene in Helsinki Metropolitan area

1 6

1

1.2

1.4

1.6

0.4

0.6

0.8

µg m

-3 ModelledMeasured

0

0.2

Kallio Tikkurila Lintuvaara

Comparison of modelled and measured annual average benzene concentrations Benzene iF for 2005 calculated by EXPAND

Leena

Local scale dispersion calculations for fine particle emissions from domestic combustion and road trafficfine particle emissions from domestic combustion and road traffic

Leena, Mia

Dispersion studies over a noise wall

southern wind, ~2 m/s

Dispersion 30.3.2007

300

S

150

200

250

9 (u

g/m

3)

N

0

50

100PM9

stationary measurements

-2 10 20 30 40 50 60 70

Distance from the wall

stationary measurements on the sidewalk

Measurements will be compared with model resultsMeasurements will be compared with model results

Kari,Ilkka,Juha

NBC Modelling & Simulation

OVERALL OBJECTIVE• Develop improved model chains for simulation of NBC-scenarios for

planning and decision support

• SPECIFIC GOALS FOR THIS PROJECT• SPECIFIC GOALS FOR THIS PROJECT• Clarify uncertainties in NBC M&S

- Compare results from national model chains for reference scenarios- Emergency response (very fast) ---- studies (“unlimited” time for calc.)- Identify key weaknesses

M d lli G id li• Modelling Guidelines• Best Practice• Possible start of model developmentPossible start of model development

Ilkka

Particle-cloud formation

- in SILAM (above)- compared to ”traditional” cylinder

Ilkka Valkama/FMI

ycloud (left)

Kari,Ilkka,Juha

EXPLOSION OF AN IMPROVISED NUCLEAR DEVICE (FI)

Helsinki RN-ScenarioApproximate evolution of the cloud( j ti th d l l f ll h i ht )(projection on the ground level from all heights)

Ilkka

SILAM Nuclear Fallout Modulecloud parametres- cloud parametres

- particle formation- activity distribution

CLOUD dimensions

140

80

100

120

ht (k

m)

Cloud radius(obs)SILAM

0

20

40

60

heig

h SILAM

42.097.0 WRadius =

00.1 10 1000 100000

yield (kT)Ilkka Valkama/FMI

Ilkka

Particle-cloud formation

- in SILAM (above)- compared to ”traditional” cylinder cloud (left)

Ilkka Valkama/FMI

Jari

Further development of MPP-FMI utilizing p gsatellite data (ESA)

Available data:

• MODIS: T-profile product every 6th hour (in cloud-f diti )

Available data:

free conditions)• RS-data twice a day• Ground surface measurements at least once a hour

Jari

Polar orbiting satellites:Aqua Terra

HELSINKITESTBED:MODIS

Test PlanAqua, TerraRS - RASS

- SODAR

Test Plan

700700 km

SO- CEILOMETER- Meteorological

0

k

700 km and air qualitymonitors

m0-7 km

100 km100 km

Jokioinen Helsinki

Mia

VIEME-project:• Computational fluid• Computational fluid

dynamics study of dispersion of particulate matter pfrom a highway over a noise wall

• PM will be simulated as an inert substance

• 2D, standard k-ε

90 m behind the noise wallnoise wall

Mia

COST-732

• “QUALITY ASSURANCE AND IMPROVEMENT OF MICRO-SCALEMICRO SCALE METEOROLOGICAL MODELS”

• 23 members• 23 members• Test case of 120

containers on a field• To be developed:

Guideline on how to model dispersion from thedispersion from the experiences obtained

Mia

SRIMPART: Source-receptor and inverse modelling to quantify urban particulate emissionsquantify urban particulate emissions

• Aims:• Aims:• To improve emission estimates with emphasis on

emissions from domestic wood burningemissions from domestic wood burning.• To improve emission estimates for wood burning

applied in the EMEP model and in other internationalapplied in the EMEP model and in other international modelling systems

• To intercompare and assess uncertainty in p yindependent methodologies for assessing emission rates of particulates from urban sources

Mia

• AIM: provide an overview of the latest research findings on air quality and health to support European sustainable development action plans and strategies

• WP 4: “Report on current state-of-the-art research in urban air quality and health”

CAIR4HEALTH• The Air Quality Key Questions have been decided on, and now

answers are being drafted for them:H U b Ai P ll t t LEVELS CHANCING i th l t• How are Urban Air Pollutant LEVELS CHANCING in the long term - past 10 years and future trends, including expected effects of climate change? (Editor: UH, writers UH, FMI, AUTH)A th t ll t t li it l tt i bl th ff ti• Are the current pollutant limit values attainable – the effectiveness of REDUCTION MEASURES? (Editor: UH, writers UH, FMI, AUTH)Wh t th t hi t f INTEGRATED• What are the current achievements of INTEGRATED ASSESSMENT TOOLS for estimating the health impacts of air pollution? (Editor: AUTH, writers AUTH, TNO, FMI)Wh t i k b t th CHEMICAL COMPOSITION AND SIZE• What is known about the CHEMICAL COMPOSITION AND SIZE DISTRIBUTION OF URBAN PM and their HEALTH EFFECTS? (Editor: FMI, writers: UH, FMI)

Backup slidesBackup slides

CLOUD dimensions

70 2248.03104 WH =

30

40

50

60

ght (

km)

Cloud-top(obs)

310.4 WHtop

305.0917.3)20( WktWH =≤

0

10

20

30

0 1 10 1000 100000

hei

SILAM2-param

1774.0362.6)20( WktWH =>

0.1 10 1000 100000

yield (kT)CLOUD dimensions

40

2296.04932 WH = 2025303540

ht (k

m)

493.2 WHbottom =

3433.0166.2)20( WktWH =≤ 05

101520

heig

h

Cloud-bottom(obs)SILAM21594.0439.4)20( WktWH =>

00.1 10 1000 100000

yield (kT)

2-param

Ilkka Valkama/FMI

Activity-particle size distribution

22)(l (1)(l (1 ⎤⎡ ⎞⎛⎤⎡ ⎞⎛

Radioactivity-particle size distribution (bi-modal log-normal distribution)

23 )(ln(21)(ln(

21

)2()1(

)2()( ⎥

⎥⎦

⎤

⎢⎢⎣

⎡⎟⎟⎠

⎞⎜⎜⎝

⎛ −−

⎥⎥⎦

⎤

⎢⎢⎣

⎡⎟⎟⎠

⎞⎜⎜⎝

⎛ −−

−+= βα

βα

βπβπ

r

tv

r

tv e

rA

fer

AfrA

Log-normal distributions for volatile (left) and volatile (right) radionuclides.Ilkka Valkama/FMI

Available (/developed in Ubicasting): dense obs. N t k ( t & AQ) &

Now : offline met-data

Network (met & AQ) & high resolution met-modeling =>

Dispersion forecasts for toxic industrial chemicals and other hazardous releases (terrorism etc) GOAL :

Realtime dispersion & forecasts (1- 6 h)p ( )

New met-model-> OPERATIVEdispersion modellingdispersion modelling

Tailored AQ dispersion tool

• real time tool for assessing

ValitsekaasuLeviämis-

for assessing accidental releases

+1h

+2h

+1h

+2h

kaasu

kaasu 1

kaasu 2

ennuste

•Based on real-time met-

+3h+3h kaasu 3

kaasu 4

measurements

Säätilanne

Anturi1

Anturi2

Common goal: IL, YTV , Vaisala

C bi IL & YTV V i l t & k h• Combine IL & YTV, Vaisala measuremetns & knowhow :- Not only information on stations: also real-time AQ-

situation for the region & 1- 2 day forecastsg y- Prequisite: also the dispersion models need to be modified

– to be able to utilize all available (high-resolution) data

current AQ

Warnings

AQ & exposure -forecast

Warnings

Practical examples

Meteorological data:•FMI 3 meas.stations•MetPP-FMI/HIRLAM NO spatial variation in the met-pfields

bi iUbicastingMeteorological data•Helsinki Testbed (>100 st.)•Hirlam/Arome (3-9 km resolu.)•LAPS (1 km resol.)( )Met-resolution up to 1 km

Web & wapReal time information only at the YTV-stations

Ubicastingg

Web, wap, sms, mmsmmsReal time information for user defineduser defined (/GPS) )locations & periods

OSCAR:Need for an Integrated approach

Graphical User Interface/Control Module

PC OSCAR ASSESSMENT SYSTEM

Graphical User Interface/Control Module

PC OSCAR ASSESSMENT SYSTEM

Visual Basic/Visual.NET

Emissions Database Met Database AQ DatabaseEmissions pre-processor Met pre-processor AQ Data pre-processor

Input pre-

Visual Basic/Visual.NET

Emissions Database Met Database AQ DatabaseEmissions pre-processor Met pre-processor AQ Data pre-processor

Input pre-

Level I ModelsCorrelations of C U Traffic

Level II ModelsEg DMRB, CAR Int PEARL

Level III ModelsEg CAR-FMI, OSPM GAMMA

Numerical Models ADREA-HF MIMO

External ModelsEg Noise

Input preprocessor

Level I ModelsCorrelations of C U Traffic

Level II ModelsEg DMRB, CAR Int PEARL

Level III ModelsEg CAR-FMI, OSPM GAMMA

Numerical Models ADREA-HF MIMO

External ModelsEg Noise

Input preprocessor

C, U, Traffic Int, PEARL OSPM, GAMMA HF, MIMOEg Noise

Spatial/temporal distributions – CO, NO2, PM10, PM2.5, Assessment parameters

C, U, Traffic Int, PEARL OSPM, GAMMA HF, MIMOEg Noise

Spatial/temporal distributions – CO, NO2, PM10, PM2.5, Assessment parameters

Raw data:- Off line- Excel- Access

Visualisation:- Embedded - GIS-VIS5D, PAVE

Scenario Analysis:- Transport- Emissions

AQ Assessment:- Limit values- Annual means

Raw data:- Off line- Excel- Access

Visualisation:- Embedded - GIS-VIS5D, PAVE

Scenario Analysis:- Transport- Emissions

AQ Assessment:- Limit values- Annual means

WEB OSCAR ASSESSMENT SYSTEMWEB OSCAR ASSESSMENT SYSTEM

CAR-FMI : further development• EU/OSCAR (->2005)EU/OSCAR ( >2005)

• An ”international” version of CAR-FMI was integrated in to the OSCAR system

• Main issue: • users can easily use their own emission data

• ONGOING work:• Stand-alone international version

• Co-ordinate systemsy• Several options for (emission) input format• Routines to complement for missing meteorological p g g

parameters• Links to ArcInfo(GIS) / now only MapInfo

SummaryDevelopment and testing of disperion models ongoingDevelopment and testing of disperion models ongoing

in all scales : form microscale (CFD) to regional scale

Lot of effort is put on ”generalizing” the existingLot of effort is put on generalizing the existing models for wider user community -> from research models to ”easy-to-use” tools & services (strong co-

ti ith th /i tit t / i )operation with other groups/institutes/companies)

Linking dispersion models with new meteorological models and health risk assessment models increasingly important part of the work : final aimincreasingly important part of the work : final aim always a complete chain from political decisions to health effects