TRAKER: A Method for Fast Assembly and Update of Paved and … · 2015. 9. 10. · TRAKER Results to Entire Network Winter - Ada - Rural 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 5 10 15 20 25

TRAKER: A Method for Fast Assembly and Update of Paved and Unpaved

Road Dust Emission InventoriesV. Etyemezian, H. Kuhns, G. Nikolich, A. Gertler

Desert Research Institute , Las Vegas, NV

D. Fitz, K. BumillerCE-CERT, Riverside, CA

R. Merle, and R. LangstonClark County DAQEM, Las Vegas, NV

Outline

• Road Dust Introduction• Mobile Platform for Road Dust Measurement

– The TRAKER Concept– TRAKER Characteristics and Calibration

• Examples of completed TRAKER Applications– Las Vegas, NV– Lake Tahoe, NV/CA– Treasure Valley (Near Boise), ID

• Conclusions and Plans for Future Work

Road Dust Emissions

• PM emissions of fugitive dust from paved and unpaved roads due to vehicle travel

• Road Dust Emission Measurement– Towers upwind and downwind of road

segment– Silt loading/content

• Calibrated against upwind/downwind measurements

– Vehicle-based technologies (DRI, CE-CERT)

Road Dust Measurement

• Planning agencies largely use silt loading (paved roads) and silt content (unpaved roads) to determine emission factors (AP-42 guidance document)

• Silt loading (paved roads)– Lane closure– Vacuum/sweep material– Use sieves to determine silt– Per Sample/Location ~ 10 labor hours + planning +

hazards

Testing Re-Entrained Aerosol Kinetic Emissions from Roads (TRAKER)

Fron

tTop ViewFr

ont Top View

Fron

t Side View

Fron

t Side View

Background Monitor

Background Monitor

Influence Monitor

Influence MonitorFron

tTop ViewFr

ont Top View

Fron

t Side View

Fron

t Side View

Background Monitor

Background Monitor

Influence Monitor

Influence Monitor

Background Monitor

Background Monitor

Influence Monitor

Influence Monitor

TRAKER

• Particle Sensors– TSI DustTrak

5820– Grimm Particle

Size Analyzer 1.108

• GPS

Data Acquisition and Processing• Lab View program

displays and logs data from – 6 DustTraks– 3 Grimms– 1 GPS

• Uniform time stamp applied to all data for synchronization

• Data tables are loaded into MS Access for processing and analysis

TRAKER Signal vs Vehicle Speed

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

0 5 10 15 20 25 30 35

Speed (m/s)

TRA

KER

Sig

nal (

mg/

m3 )

Treasure Valley RegressionT=0.00017*speed2.96

R2 = 0.972

0

0.5

1

1.5

2

2.5

0 5 10 15 20 25 30 35 40

Speed (m/s)

TRA

KER

Sig

nal (

mg/

m3 )

Ft. Bliss RegressionT=0.00012*speed2.75

R2 = 0.923

• T = Ctire – Cbkgrnd• T = a S3

• On the same paved road the TRAKER signal increases with the speed cubed

• Factoring out speed leaves a signal proportional to the emission potential of the road.

TRAKER Calibration with Upwind/Downwind Flux Towers

T=TT-TB

Unpaved Road EMF (g/veh-km) = 8.36 T1/3

Paved Road EMF (g/veh-km) = 0.33 T1/3

Example TRAKER Use: Las Vegas

• 150 km loop in Las Vegas Valley• Covering Road Types, Locations,

Construction Influence• TRAKER operated:

– 6/30/04, 7/1/04 (Loop 1)– 2/14/05 – 2/17/05 (Loop 2)

Emission Potentials

(g/veh-km) on 2/14/05

Average of PM10Emission Factor

(g/vkt) on a road link basis for all 4 days of sampling (2/14/05 –

2/17/05)

0

50

100

150

200

-100 -80

-60

-40

-20 0 20 40 60 80 100

% Difference from Average

Freq

uenc

y

0%

20%

40%

60%

80%

100%

120%02/14/05 2.0%

0

50

100

150

200

250

-100 -80

-60

-40

-20 0 20 40 60 80 100

% Difference from Average

Freq

uenc

y

0%

20%

40%

60%

80%

100%

120%02/15/05 4.3%

0

50

100

150

200

-100 -80

-60

-40

-20 0 20 40 60 80 100

% Difference from Average

Freq

uenc

y

0%

20%

40%

60%

80%

100%

120%02/16/05 0.1%

0

50

100

150

200

250

-100 -80

-60

-40

-20 0 20 40 60 80 100

% Difference from Average

Freq

uenc

y0%

20%

40%

60%

80%

100%

120%02/17/05 -6.9%

0

50

100

150

200

-100 -80

-60

-40

-20 0 20 40 60 80 100

% Difference from Average

Freq

uenc

y

0%

20%

40%

60%

80%

100%

120%02/14/05 2.0%

0

50

100

150

200

250

-100 -80

-60

-40

-20 0 20 40 60 80 100

% Difference from Average

Freq

uenc

y

0%

20%

40%

60%

80%

100%

120%02/15/05 4.3%

0

50

100

150

200

-100 -80

-60

-40

-20 0 20 40 60 80 100

% Difference from Average

Freq

uenc

y

0%

20%

40%

60%

80%

100%

120%02/16/05 0.1%

0

50

100

150

200

250

-100 -80

-60

-40

-20 0 20 40 60 80 100

% Difference from Average

Freq

uenc

y0%

20%

40%

60%

80%

100%

120%02/17/05 -6.9%

Comparison of Link-averaged emissions potentials from 4 consecutive days of TRAKER measurement in Las Vegas

0

2

4

6

8

10

12

2/14

/05

0:00

2/14

/05

6:00

2/14

/05

12:0

02/

14/0

5 18

:00

2/15

/05

0:00

2/15

/05

6:00

2/15

/05

12:0

02/

15/0

5 18

:00

2/16

/05

0:00

2/16

/05

6:00

2/16

/05

12:0

02/

16/0

5 18

:00

2/17

/05

0:00

2/17

/05

6:00

2/17

/05

12:0

02/

17/0

5 18

:00

2/18

/05

0:00

Time

Win

d Sp

eed

(mph

)

.

Day 1: +2.0%

Day 3: +0.1%

Day 2: +4.3%

Day 4: -6.9%

What about wind

effects?

Correction for Cross-winds

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

0 5 10 15 20 25 30 35

TRAKER Speed (m/s)C

oeffi

cien

t of V

aria

tion

Ft BlissTreasure Valley

Avoid cross-wind effects by considering data where TRAKER speed > 10 mph

Example TRAKER Use: Lake Tahoe

• 13 runs over a mountain pass

• 9 circuits around the lake

• Comparison with flux tower measurements

Emission Factors Using Stationary Flux Tower

Sampling Period

Condition PM2.5 EMF (mg/km)

PM10 EMF (mg/km)

1 Baseline 76 229

2 After salting 99 310

3 1st dry day after storm

112 612

4 2nd dry day 133 660

5 After sweeping

211 735

Emission Factor Using TRAKER

0

0.1

0.2

0.3

0.4

0.5

0.63/

31/2

003

4/7/

2003

4/14

/200

3

4/21

/200

3

4/28

/200

3

5/5/

2003

5/12

/200

3

5/19

/200

3

5/26

/200

3

6/2/

2003

6/9/

2003

6/16

/200

3

6/23

/200

3

6/30

/200

3

7/7/

2003

7/14

/200

3

Date of Measurement

Emis

sion

Fac

tor (

g/vk

t)

TotalBest Linear Fit to Data

Paved Road CalibrationUnpaved EF = 8.36 T1/3

0.1

1

10

100

1000

0.1 1 10 100 1000 10000

Traker Signal (mg/m3)

Em

issi

on F

acto

r (g/

vkt)

Lake Tahoe PavedEF = 0.33 T1/3

Example TRAKER Use: Treasure Valley, ID

• Winter/Summer TRAKER measurements over a closed loop

• Develop emission inventories based on road characteristics and traffic speeds/volumes

• Assess effect of road sanding, street sweeping, seasons

Time Series of Emissions from RoadsPrincipal Arterial Emissions Factors (g/VKT)

0123456789

102/

27/2

001

3/1/

2001

3/3/

2001

3/6/

2001

3/10

/200

1

3/15

/200

1

3/17

/200

1

7/12

/200

1

7/14

/200

1

7/19

/200

1

7/20

/200

1

7/22

/200

1

Date

EF (g

/VK

T)

Winter Summer

End of Road Sanding for Winter

Use Roadway Properties To Extend TRAKER Results to Entire Network

Winter - Ada - Rural

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20 25 30 35Speed (m/s)

Emis

sion

s Po

tent

ial [

g/vk

t/(m

/s)]

TRAKER SpeedsPosted SpeedsExponential Fit

Summer - Canyon - Rural

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 5 10 15 20 25 30 35Speed (m/s)

Emis

sion

s Po

tent

ial [

g/vk

t/(m

/s)]

TRAKER SpeedsPosted SpeedsExponential Fit

Statistically significant differences found based on Season, County, Setting (urban/Rural), and Road Speed.

Use Traffic Demand Model to Obtain Link-level Emission Inventory

Another Interesting Finding About Street Sweepers

0

0.000005

0.00001

0.000015

0.00002

0.000025

0.00003

QUARTERBefore

QUARTER After STILLWELLBefore

STILLWELLAfter

Spe

ed C

orre

cted

TR

AK

ER

Sig

nal

Left PM10Right PM10

SWEPT NOT SWEPT

Conclusions• Vehicle-based methods for road dust emissions

measurement are efficient and comparatively fast– Cover 100’s of mile of road– No need to worry about “picking” representative sections of road

• For TRAKER– Las Vegas Study shows day to day differences (precision) is on

the order of 5% overall and generally better than 20% on a link by link basis

– Important to recognize the effect of cross-winds– Can use to

• Efficiently assemble emission inventories for road dust• Test effects of parameters : sweeping, sanding, season, location,

construction, traffic volume, etc• Technology has been around for 6 years and multiple

studies

Planned Future Work• Increase # calibration data points for paved roads using

upwind/downwind technique (Las Vegas, summer 2005)• Quantitatively compare with CE-CERT SCAMPER

(summer 2005)• Redesign towards achieving turnkey application (Winter,

2006). Features for TRAKER III will include:– New Platform: Dodge Sprinter– On-the-fly span and zero for DustTraks onboard– On-the-fly switching between paved and unpaved road sampling– Real-time mapping of dust emission potentials and factors

Acknowledgements

• Clark County Department of Air Quality and Environmental Management

• Idaho Department of Environmental Quality

• DoD – Strategic Environmental Research and Development Program

• California Air Resources Board