DAPPLEDecember 2003

Dispersion of

Air

Pollutants & their

Penetration into the

Local

Environment• Background

• Activities to date

• Results - tracer & wind tunnel

• Future plans

EPSRC Infrastructure & Environment Programme

DAPPLEDecember 2003

Consortium University of Surrey Alan Robins

Imperial College Roy Colvile

University of Bristol Dudley Shallcross

University of Cambridge Rex Britter

University of Leeds Margaret Bell

University of Reading Stephen Belcher

A 4 year EPSRC funded project, supported by:London -> ALG/APRIL/EA/Local Government/TfL

DEFRA, AEQ Division

DSTL (Porton Down), HSE

Meteorological Office

DAPPLEDecember 2003

DAPPLE aims: to enhance understanding of pollutant dispersion and individual exposure to pollutants in urban areas to improve predictive ability to enable better planning and management of urban air quality and ‘incidents’

Background

“… We do not yet have the understanding needed to answer fundamental questions about pollution behaviour over short distances in the urban environment. …”

DAPPLEDecember 2003

The basic understanding gained will be used:

in the evaluation and development of appropriate decision support tools and risk assessment methodologies in the evaluation and development of best practice guidelines for their application in assessing the inherent uncertainty in their use in assessing their contribution to the sustainable development of safer, more pleasant cities.

As far as possible, DAPPLE deliverables will be of generic value and applicable both within UK urban areas and others overseas.

Output

DAPPLEDecember 2003

Street intersections - the urban environment

100

100

49 56Relative flow rates at the intersection; off-set 0.6H

DAPPLEDecember 2003

Intersection modelling

LDA mean velocity measurements within canyons

Highly unsteadyexchanges

Block size:

4H x 6H x H

Orientation 10degs

1:1

C measured

C predicted

Regulatory models (ADMS) don’t model intersections -CFD does a reasonable job for a simple 4 block geometry.

DAPPLEDecember 2003

Classic street canyons

Measured Street Canyon Pollution Levels in Jagtvej

0 90 180 270 360 Wind Direction, degrees

800

600

400

200

0

Jagtvej

NOxppb

N 300 degs

receptor

DAPPLEDecember 2003

Classic street canyons?

DAPPLEDecember 2003

How do we predict the movement of pollutants in & over a street network?

What are best practices in the urban context?

How much detail is necessary?

What area needs to be modelled and to what resolution?

How do we best go about managing & improving local urban air quality?

What tools are needed to deal with these matters?

How should these be developed for incident management?

In an urban setting:

DAPPLEDecember 2003

Activities

Vehicle movement analysis

Emissions modelling

Tracer experiments

Individual dose

measurement

Air quality monitoring

Wind tunnel studies

Model evaluation & development Marylebone Rd - Gloucester Place intersection

DAPPLEDecember 2003

Key times

2002 April: project start

2003 Spring: initial field trials

Monitoring studies: 28 April to 23 May

Tracer study: 15th May

Summer: results analysis, interim output & workshop

2004 Spring: main field trials

19th April to 28th May

2005 Continued analysis; results, workshop, dissemination

2006 March - project end; final reports and workshops

DAPPLEDecember 2003

Site features

DAPPLEDecember 2003

First field trials 28/4/03 - 23/5/03

Instrumentation installed courtesy of TfL

DAPPLEDecember 2003

York Street

LibraryBickenhall St.

Gloucester Place

Thornton Place

Bickenhall Mansions

Baker Street

Marylebone Road Marylebone Road

WU roofnot accessible due to pair of nesting falcons!

SITE 1SITE 2

SITE 3

SITE 4

SITE 5

SITE 6

SITE 7

SITE 8 Super-siteWCC

WCC INDOORS

•VOC - basement, reception & 4th floor rooms•particles - basement & first floor room

sonics street boxes

AWS

lamp post direction of bracket

EXTERNAL EQUIP

Schematic site plan

DAPPLEDecember 2003

Weather stations

AWS at 15m AWS at 64m

DAPPLEDecember 2003

York Street

LibraryBickenhall St.

Gloucester Place

Thornton Place

Bickenhall Mansions

Baker Street

Marylebone Road

X tracer release tracer samplers (10)

Marylebone Road

x

Super-site

WCC

sonics street boxes

AWS

EXTERNAL EQUIP

Schematic site plan

DAPPLEDecember 2003

Trial achievements

• Meteorology - 7 Sonic anemometers

- at 4 and 8m on lamp posts

- 4, 12 and 15m on WCC Hall)

- logged at 20Hz or 5Hz

- 3 Automatic Weather Stations

- 30 secs wind, temperature and humidity

- synoptic - forecasts & back trajectories (Met Office)

- meso-scale model output, data for tracer release

• Pollution - 10 Street boxes (CO & NO2)

- 4 & 7m giving 5 minute averages; along

Marylebone Rd between Super-site & WCC Hall

- ITS instrumented vehicle

DAPPLEDecember 2003

Ultrasonic anemometers

Lamp post and building mounted

DAPPLEDecember 2003

Trial achievements

• AURN Super-site - CO, PM10, PM2.5, NO2, O3, SO2, Wind, etc.

• WCC Hall Site - NO, NO2, PM10 + Opsis

• Traffic - SCOOT data (TfL)

- Leeds instrumented vehicle and traffic counts

• Tracer - SW winds from York Street (15TH May)

- 10 sampling boxes deployed within 2-3 blocks

• Exposure - 4 people, 3 times a day (particulates & CO)

- different routes & modes of transport, 40 minute exposure

- indoor/outdoor VOC & particles (MSc students)

DAPPLEDecember 2003

Example day - 15 May

AWS mean 30s data for wind speed, direction, temp & humidity

WCC Rooftop AWS

Tracer exp. started at 5pm

DAPPLEDecember 2003

SCOOT

DAPPLE

5 mins, hourly & daily average traffic flows

Verification by manual counts

Emission estimates

Traffic flow modelling

2 w

ay d

ata

netw

ork

I/O

pro

gs

Onlinetraffic model

Optimiser

Signal timingssignal

controllers

trafficdetectors

Operator

DAPPLEDecember 2003

CO box data

0.00

0.50

1.00

1.50

2.00

2.50

3.00

Week of 12th May 03

1 (C 3m)

1 (C 7m)

Mon. Tues. Wed. Thurs. Fri. Sat. Sun.

Morning & eveningrush hour peaks

Night timelows

CO at 7m

= 0.77

CO at 3m.

DAPPLEDecember 2003

CO box data - 15 May

0

0.5

1

1.5

2

2.5

3

Time on 15th May 03

4 (S)

3 (N)

2 (C)

Light winds Stronger SSW winds

NS

CO, 4 (S)

CO, 2 (C)

CO, 3 (N)

Circulation:

high CO to S

----

low CO to N

DAPPLEDecember 2003

Wind field - SSW winds

Site 1

Site 2

Site 3

Typical 10Hz wind direction data for SSW winds

SSW winds

DAPPLEDecember 2003

Wind field

Marylebone - Rd Gloucester Place intersection

Street wind conditions

‘roof’ level wind from SSW

Generally:

Wind channelling and direction switching:

e.g.

Gloucester Place wind direction changes by 180 degrees depending on above roof wind having north or south component.

DAPPLEDecember 2003

Wind channelling dominates horizontal motion

Weak vortex-like motion driven by cross-flow only apparent in time averaged results.

Vertical mixing - other processes

Analysis continues

DAPPLEDecember 2003

Comparison between field data and wind tunnel simulation

(reference velocity is roof level d=51.35 o

)

-1

0

1

2

3

4

5

6

0 1 2 3 4 5 6 7 8 9 10 11

Site number

simulation mean (83 puffs)

PDF band of simulation

maximum of one field measurement

Cumulative PDF from 10% to 90%

DAPPLEDecember 2003

OBJECTIVES

a) CAMPAIGN1 – Collect co-ordinated data-set of traffic flow (SCOOT data and manual counts for verification purposes) and air quality data (Learian streetboxes);

b) Verify the SCOOT data using manual count data and automatic count site data;

c) Process and average the SCOOT data into 5, 15 and 60 minute time intervals;

d) Calculate Emissions using NAEI factors;

e) Implement ETEM (Enhanced Traffic Emission Model – uses SCOOT data, congestion algorithm and ADMS-Urban to predict roadside concentrations);

f) Implement microscopic traffic simulation model in the study area using as demand input, the raw SCOOT detector data. The simulations will predict emissions using the transient CMEM emission factor database; working with TfL staff

Conduct an integrated analysis of traffic activity, emission prediction, meteorology and ambient concentrations.

DAPPLEDecember 2003

We have developed, and attempted to justify a “simple correlation” for dispersion in urban areas for use on the neighbourhood and city scales ( see DAPPLE web site). This was used in the design of the first field tracer experiment.

We have compared the “simple correlation” with the experimental data from the first field tracer experiment (only PMCH tracer). The correlation overpredicts by about a factor of two.

We have compared the baseline dispersion model for urban areas developed in Hanna, Britter and Franzese (2004) with the experimental data from the first field tracer experiment (only PMCH tracer). The correlation overpredicts by about a factor of two.

DAPPLEDecember 2003

We have worked with Bristol to review the tracer experimental data collected in the first DAPPLE study to quality assure them and then to archive them on the DAPPLE web site.

This has been completed for the PMCH tracer release and the data should appear in an archived for on the DAPPLE web site imminently.

A problem was found with the SF6 data, ascribable to an unsuspected non-linear response at high concentrations. The results are currently being recalculated.

DAPPLEDecember 2003

DAPPLEDecember 2003

The PM2.5 and ultra-fine particle count personal exposure databases are now complete and are undergoing final post-processing checks. They have been constructed by Surbjit Kaur in MS Excel and will be available to the consortium via the ftp site at the beginning of February 2004. Accompanying transcripts for the real-time data will also be made available. The construction of the CO database has started; estimated date for availability of the preliminary database is end of February 2004. Statistical analysis in progress.

Currently, pollutant concentration fluctuations at a fixed receptor have been modelled for these steady emissions from a single point source. This indicates that much of the short-term exposure variability (on a timescale of seconds to minutes) is attributable to the unsteadiness of the flow field under constant meteorological conditions, provided real traffic sources are not so homogeneous as to remove much of the variability by spatial averaging.

Regulatory and exposure use of ADMS etc

Exposure modelling - empirical

DAPPLEDecember 2003

good data capture levels throughout

SCOOT calibration needs further attention to allow for loops that cross two lanes which can give ambiguous data when heavily congested - statistical methods being developed

DAPPLEDecember 2003

Personal exposure

5 modes of transport:

bus, taxi, car, cycle, foot

3 x 40 minute ‘trips’ per day

measuring:

CO, Fine particles & PM2.5

plus

video & voice records

DAPPLEDecember 2003

EquipmentCO T15 Langan

Measures carbon monoxide every 10 seconds

TSI P-Trak

Measures ultra-fine particles every second

Casella Ultraflow Pump + Filter

Measures average PM2.5 concentration for the route

DAPPLEDecember 2003

Personal exposure

Walkers

Cyclists

DAPPLEDecember 2003

Tracer study

Sampleintake

Sourceoutlet

DAPPLEDecember 2003

Tracer study

Tracers: PMCH, SF6

15 minute release

SF6 1.5 minute offset

10 sampling units

10x5L tedlar bags per unit

3 min samples

NICI mass spectrometry

negative ionchemical ionisation

perfluorocarbon

DAPPLEDecember 2003

Tracer study - I

Three minute averaged concentration values

DAPPLEDecember 2003

Tracer study

York Street

Library

Bickenhall Street

WestminsterCouncil House

Gloucester Place

Thornton Place

Bickenhall Mansions

Baker Street

Marylebone Road

x 1

3 2

6 7

8

5

4

9

10

Distance ~ 125m

Time of flight ~ 4.5min

Travel speed ~ 0.5m/s

Reference wind, 3m/s

Source

0.14

0.160.24

0.06

0.11 failed

0.07

0.02 ~ 0

~ 0

CUHH2/Q

114 mgrelease

DAPPLEDecember 2003

Tracer study - II

Travel speed= 0.10 to 0.35U(H)

C ~ 1/R2

R/H=3.4

R/H=12.5

DAPPLEDecember 2003

Tracer study

Conclusions from an ensemble of one

For R/H between 3.4 and 20:

Maximum concentration, CmaxUH/Q ~ 5/R2

Travel time, TUH/H ~ 30 to 41

- 3.6 to 5 minutes

Travel speed, U/UH ~ 1/10 to 1/3

DAPPLEDecember 2003

Tracer study

Enhancements for next trials:

Capability for releasing three different perfluorocarbon tracers from three locations

Development of 1 minute real time instrumentation

Ditto for SF6

Improved logistics

Greater freedom in choice of source locations (limited by traffic and other concerns) and sampler deployment

DAPPLEDecember 2003

Wind tunnel study

EnFlo wind tunnel20x3.5x1.5m working

section, 0-3.5m/s

DAPPLEDecember 2003

Wind tunnel study

18 channelair samplingsystem

FIDsystem

Flowcontrolsystem

Chilled watersupply ~ 10oC

Inlet and heater section - 15 layers400 kW capacity; ambient to ~ 100oC

20 x 3.5 x 1.5 m working section0 - 4 m/s

Mechanicalsimulation devices

Twin fans

Heat exchanger

Cooled rough wall ~ 10oC

Gassupplies

Heatercontrol

Source

Rough wall not cooled

Computerised control, data collection & data analysis

Thermo-couplesystem

Speed control

Traverse and turntable control

traverse system

turntable

DAPPLEDecember 2003

Wind tunnel study

Scale 1:200

Reference wind speed 2.5m/s, direction SW

Source in York Street

Ground level concentration data

DAPPLEDecember 2003

Simulation of field trial

DAPPLEDecember 2003

Wind tunnel study

Upper bound:CUref/Q = 50/R2

DAPPLEDecember 2003

Wind tunnel study

Upper bound:CUref/Q = 50/R2

DAPPLEDecember 2003

Salt Lake City

Upper bound:CUH/Q = 50/R2

DAPPLEDecember 2003

Wind tunnel study

Block averaging

T, s c/<C> c99/<C>0.005 0.55 3.80.5 0.38 2.35 0.18 1.4

3 0.22 1.6

raw data

3 s blockaverage

Concentration - Time

300 secs

DAPPLEDecember 2003

Wind tunnel study

Short duration emissions 0.5 to 10s

Determine travel, rise and decay time scales, dilution rate

DAPPLEDecember 2003

Wind tunnel study

Uadvection ≈ UH/4

R is straight line separation

Field: R/H = 3.4 to 12.5, Tnon = 30 to 41

DAPPLEDecember 2003

Wind tunnel study

DAPPLEDecember 2003

Wind tunnel study

DAPPLEDecember 2003

Wind tunnel study

Conclusions

Plumes: CmaxUH/Q ~ 50/R2

Significant variations across streetsPuffs: Uadvection ≈ UH/4

Rise time/Travel time ≈ 1/3; ditto decay timePeak/Mean ≈ 3.5 - 4

Future work: • detailed mapping of flow and concentration fields• flux exchange at intersection• exchanges with flow above roof level• effects of modelling detail

DAPPLEDecember 2003

Computer modelling

Themes:Simulations of field and wind tunnel exp; Urban Air Quality and Accidental releases Evaluation of model performance - sensitivity studies - best practice

Models: Simple rules of thumb Operational models (e.g UDM, ASUDM, ESUDM, ADMS-URBAN) CFD models (e.g. RANS, LES)

Purposes: Prediction/forecasting Design of response strategies

DAPPLEDecember 2003

State of play

First field trials a complete success - all data on web site for project members (not yet publicly available)

Data analysis still in progress

Important wind tunnel results on travel and decay times, dilution rates and fluctuations

Second field trials: 19th April to 28th May 2004 - open for collaborative third party work

Current news and project updates are on our public site at:

www.dapple.org.uk