Transportation Conformity Hot-spot Analyses

TRB Near-Road Air Quality Workshop January 16, 2020

Laura Berry

Transportation and Climate Division, Office of Transportation and Air Quality, U.S. EPA

1

Overview

• Background • Near-road health effects

• EPA’s hot-spot analysis requirements

• CO and PM monitoring information

• Lessons learned and best practices for future research

2

Background

3

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0

20

40

60

80

100

120

140

160

180

196

8

196

9

197

2

197

3

197

4

197

6

197

7

198

4

198

5

198

7

198

8

198

9

199

0

199

1

199

2

199

3

199

4

199

5

199

6

199

7

199

8

199

9

200

0

200

1

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

201

6

201

7

201

8

201

9

Cu

mu

lati

ve C

ou

nt

An

nu

al C

ou

nt

of

Pu

bli

cati

on

s

Publication Year

PubMed Articles with Keywords, "Traffic Pollution Epidemiology"

count cumulative since 1968

T r r T T T T r r T T T

r r ~ + + t + + t t t t + + I

,-

I .... + +- +- + + +- +- + + + ...... +- +- + + < I I ~ t -

I -+ +- +- + + .... +- +- + +

J -+ +- +- + + .... +- +- + +

~ ~

.... + +- +- + + +- +- + + + .. .. .. i +- •• ~ I I

~ ,-

+ t t t t t t i

i ,-

~ + t t + + t t I/ I

,-;

" + +- +- + + ....

1 ~ I I < -

I ult l t

-

Public Health Concerns • Populations living near roads have elevated rates

of health problems, including • Pediatric asthma onset and symptoms • Pediatric leukemia • Impaired lung function growth • Cardiovascular disease • Premature mortality

• Enormous body of literature has required periodic expert reviews • HEI

• In 2010, published expert panel report on literature published through mid-2008

• Now engaging new panel to review post-2008 literature, to be complete in late 2020

• CDC: 2014 meta-analysis on child leukemia • NTP: recently published review of traffic

pollution and pregnancy-associated hypertension

4

1.0 u C o 0.8 u ] 0.6

~ 0.4 0 z 0.2

1 .0 u 8 0 .8

"U ] 0.6

E o .4 0 z 0.2

0 50 100 150 200 Distance from highway (m)

. 50%

90')(,decay

~ •• /Winter •.('_

········• 0 50 100 150 200

Distance from highway (ml

70

60 •

. ' so -

40

30

20

10

1.0 u 5 0.8 u ] 0.6

~ 0.4 0 z 0.2

@10m · ... ,.. ..... Winter

···a../ .. ·.:it·•··· ...

/ ··-Summer

0.0 90')(, decay

1.0 u C 8 0.8

] 0.6 .; E 0.4 0 z 0.2

0 50 100 150 200 Distance from highway (m)

~ ... @ 10m . d) ~ \..!:<Winier

\ 1 ····1 ....... ! 50%, Summer

0.0 90')(, decay

0 50 100 150 200 Distance from highway (m)

---- --- --- -- -- ----- ------- -- --------- ------ --so 100 150 200 2SO

Distance from road {m)

300

,..._ 50 -g_ .e 0 40 z -c £ u g .0 ::,

Cl)

-c c:: ::,

30

20

O 10

E> g a:i 0

a

0.9

g 0.8

~ 0.7 C G> g 0.6 0 (.)

~ 0.5 '5 E o.4

~ 0.3 iii !§ 0.2 0 z

0.1

(a)NO

Upwind ·•

4 : •

' • •

Downwind

• • •

• • • •

~ 1.5 E

~

-c ., ~ -g 0.5

Cl)

-c C ::,

e t;JJ0.0 g xl0'

a:i

-600 -400 -200 0 200 400 600

Distance (m)

Exponential fit for distance-decay: C ( d) = a + /3 exp(-3d I k)

NO: a= 0.28 /3 = 0.71 k = 460 m r' = 0.98

0 100 200 300 400 500 600 700 800 900 1000 O

Spatial Lag to Nearest Highway (d, m)

(b) UPP

Upwind

-600 -400 -200 0

• •

200

Distance (m)

Downwind

400 600

100 200 300 400 500 600 700 800 900 1 000 Spatial Lag to Nearest Highway (d, m)

Examples of Recent Research

Baldwin et al., 2015 – From mobile Saha et al., 2018 – Field data from I-40 near Durham, NC monitoring in Detroit, MI in Winter 2012

Richmond-Bryant et al., 2017 – Field data from Las Vegas Apte et al., 2017 – Using mobile monitors around I-15 monitors in Google’s StreetView cars

5

EPA’s Hot-Spot Analysis Requirements

• CAA section 176(c) requires that federally supported transportation plans, transportation improvement programs (TIPs) and projects in nonattainment and maintenance areas cannot: • Cause or contribute to new air quality violations, • Worsen existing violations, or • Delay timely attainment of the national ambient air quality standards (NAAQS) or interim

milestones

• Transportation conformity determinations are required for non-exempt projects that receive either FHWA or FTA funding or approval

• For project-level conformity determinations, sometimes a hot-spot analysis is required: • In PM2.5 and PM10 areas, only for those projects with a significant number or a significant

increase in diesel vehicles • All projects in CO areas need some type of hot-spot analysis

6

C 0

+,,I

ro !,,_

+,,I

C QJ u C 0 u

+,,I

C ro

+,,I :::,

0 Cl. tl.() C V')

ro QJ !,,_

u C

Wind Direction )

.... -,, ;

; ;

; _______________________ ...,.,, Highway Contribution

Background + Background Contribution

Highway

Distance Normal to Highway

------------

What is a hot-spot analysis? The transportation conformity regulation (40 CFR 93.101) defines hot-spot analysis as an estimation of likely future localized pollutant concentrations and a comparison of those concentrations to the relevant NAAQS

• Assesses impacts on a scale smaller than the entire nonattainment or maintenance area -the area substantially affected by the project

determine the effects of emissions on air

(40 CFR 93.123(c)) • Uses an air quality dispersion model to

quality

7

PM Hot-spot Analyses to Date

• Requirement for quantitative hot-spot analyses in effect since 2012

• Since then, there have been about a dozen PM hot-spot analyses done for transportation conformity purposes

• Examples include • I-70 expansion in Denver;

• Gordie Howe International Bridge in Detroit;

• South Mountain Freeway in Phoenix;

• I-69 Section 5 in Indianapolis

8

For More Information

• EPA web site for project-level conformity and hot-spot analyses: • https://www.epa.gov/state-and-local-transportation/project-level-

conformity-and-hot-spot-analyses

• Includes links to: • PM Hot-spot Guidance

• Guidance on Using MOVES for Project-level CO Analyses

• FHWA’s Categorical Hot-spot Finding (for CO)

• Guidance on New R-LINE Additions to AERMOD

• Hot-spot training information

• FAQs

9

90% of sites have concentrations below this line

'[I NaicriaJ Standard

8 Average among all sites e

2 10% of sites have concentrations below this line 0 ........,_ ________________ ____,,---_____________________________ ,.............

CO and PM Monitoring Information Key for next slides

Co

nce

ntr

atio

n

Time (in years)

https://www.epa.gov/air-trends/air-quality-trends-how-interpret-graphs 10

co Air Quality, 1980 - 2018 co Air Quality, 201 o - 2018 (Annual 2nd Maximum 8-hour: /J.~w"' .. .,. ·. e) (Annual 2nd Maximum 8-hour: /J.!da · ·. e)

Nati onal Trend based o · ~ Sites Nati onal Trend based or1 213 Sites

16 16

14 E

14 -0... 0...

12 E 12-c' 10

0... 0... a 10-:i,...., c'

rn 8 a :.... :p +-' 8-C

6 rn

(LI :.... +-' u C

C 4 (LI 6-a u

National Standard

u C

2 a 4-u

o ............... ~~~~~~~~~~~~~~~~~~~~---.---.' ~ 2-1 ~~~~~~~~~~~~~~~~~~~~§§§§§§§§§§§§§§§§§§§ ....,,. I _____________________ ~,

888888888899999999990000000000 111111111 ° -l---lr------.-l--~l----.-l--~l---..--l----,-I----I

0 1 2 3 4 S 6 7 8 9 0 1 2 3 4 S 6 7 8 9 0 1 2 3 4 S 6 7 8 9 0 1 2 3 4 S 6 7 8 201 0 2011 201 2 201 3 201 4 201 5 201 6 201 7 201 8

1980 to 2018 : 83% decrease in Nati onal Average 2010 to 2018 : 15% decrease in Nati onal Average

https://www.epa.gov/air-trends/carbon-monoxide-trends

11

PM10 Air Quality, 1990 - 2018 (Annual 2nd Maximum 24-H OJ 1r: _A 110 age)

Nati onal Trend based on 121 Sites

200~-------------------~

~ National Standard ....._ 1 50 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -CTl ::::,

1a 100 p C (]J u C 0 u

soL ~--~......._.,.--•- --...... "------.. ------- ,__..._.........,...,...._,.__. -

PM10 Air Quality, 2010 - 2018 (Annual 2nd Maximum 24-H OJ 1r- . ~.• .... r:;3ge)

Nati onal Trend based or 556 Sites

200~-------------------~

M National Standard E 1 50 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -...... CTl ::::,

c· 0 P 100 ro. :..... +-' C

C 0 50 u ~ I

o........._~~~~~~~~~~~~~~~~~~~-----,, l'---~ - ----.............,_ _ _ ____________ __JI

~ ~ n 1 ~ ~ ~ ~ ~ ~ n n n n n n n n n .. o-------r----r---~-----.----.--------r---.------9 9 9 9 9 9 9 9 9 9 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8

1990 to 2018 : 26°/o decrease in Nati onal Average

2010 2011 2012 2013 2014 2015 2016 2017 2018

2010 to 2018 : 2% decrease in Nati onal Average

https://www.epa.gov/air-trends/particulate-matter-pm10-trends

12

PM2 .5 Air Quality, 2000 - 2018 (Seasonally -Weighted Annu~I ··· ~10 r-:age)

Nati on al Trend based or 41 2 Sites

30~-------------------~

M 25 E -.. 0)

:::::, 20

:µ 15 ro L

~ 10 t---------------------- ·. National Standard · . -----------------

C ,,___ 0 ----- ----... __ ....______ _ ___J1 L) 5

0

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

2000 to 2018 : 39% decrease in Nati onal Average

PM2 .5 Air Quality, 201 0 - 2018 (Seasonally -Weighted Annu;:3 8~ 10 r-· ae)

Nati on al Trend based o 649 Sites

30~-------------------~

25 M E -.. ~ 20

c · 0 ~ 15 f ir---~ Nati onal Standard ~ f ·---~- -- -------------------------------------~ 10 I -- ----- I B I

5 .. 0 1 I I I I I I I I

2010 2011 2012 2013 2014 2015 2016 2017 2018

2010 to 2018 : 16% decrease in Nati onal Average

https://www.epa.gov/air-trends/particulate-matter-pm25-trends

13

Lessons Learned and Best Practices for Future Research

14

Lessons Learned to Date

• Model-to-monitor studies based on emissions from traffic are difficult to do well: • Since traffic data underlies the entire analysis, study should focus on obtaining

detailed and accurate data • Analysis of data must be done appropriately, e.g., averaging data such as vehicle

speeds, temperatures, or wind speeds not appropriate

• These studies are not conducted in same way or for same purpose as a hot-spot analysis

• For advancing the science of modeling, the most useful research would focus on • traffic data and vehicle operating modes, and • tracer gas studies

15

Model-to-Monitor Studies

• These studies seek to compare model results with measured data

• Two main types, based on either • emissions from traffic, or

• tracer gas

• Each of these types of studies has advantages and disadvantages • Important to consider before embarking on research

16

Model-to-Monitor Studies

Based on emissions from traffic:

• May be able to use data sources established for other purposes, e.g., near road monitoring data or traffic monitoring data

• Uncertainty about emissions: even with good data for speed and number of vehicles, usually need to make assumptions, e.g., vehicle types, ages, fuel used, drive cycles

• Uncertainty about background: even with a monitor representing background, there may be other sources influencing concentrations

• May need to match averaging periods when using traditional PM monitors

Based on tracer gas:

• Source emissions rate and other characteristics are known: reduces uncertainty in traffic, emissions, and background concentrations

• Usually more monitors deployed, so greater spatial coverage

• Limited by length of study, number of met conditions evaluated, and logistics of making sure wind is the “right” direction

• Expertise needed, e.g., outfitting vehicles to release tracer gas correctly

17

What can model-to-monitor studies inform?

• Studies based on traffic emissions: • Because of inherent uncertainty, not as well-suited for assessing model

accuracy

• May be more useful for evaluating gradients predicted, i.e., rate of decrease in concentration the model predicts over distance

• May be more useful for evaluating what contributes to error: are errors larger in certain hours, under certain meteorological or traffic conditions?

• May be useful for evaluating sensitivity to assumptions

• Studies based on tracer gas: • Can generate data either for model algorithm development or evaluation

18

Best Practices for Studies Based on Emissions from Traffic • Robust traffic data collection is needed:

• If the plan is to model each lane as a source, data by lane is necessary • Need to know not only counts, but vehicle types, speeds • Even when known, speed data does not reveal operating mode

• Ideally, use video and analyze it to obtain information about both vehicle type and activity • Activity should not be averaged: at any moment, some vehicles accelerating, some

decelerating, some cruising • Hour by hour congestion will differ, which will affect vehicle numbers, speeds, and activity

• License plate studies, connected to VINs, would be helpful to characterize the fleet as accurately as possible • Could identify actual vehicle types and fuel type used (e.g., are some passenger cars diesel?

Are some electric? Which trucks are gasoline vs. diesel? Etc.) • Could indicate whether high-emitters are present (one or two could skew results) • Would provide accurate age distribution • If not available, need to think carefully about whether county average is appropriate

19

Best Practices for Studies Based on Emissions from Traffic, continued • High-resolution meteorological data is needed

• On-site meteorological data is important: met data, such as wind speed and direction, can differ across small distances

• Even hourly data may be too coarse: some hours may not be clearly upwind or downwind

• Wind vectors should not be averaged across a day

• If upwind monitors are measuring higher concentrations than downwind monitors, these data should not be used in the comparison • “Downwind” monitors can be higher due to other sources around them • Dispersion models cannot produce negative numbers due to mass

conservation 20

What type of research would be valuable?

• More research and data collection about traffic: composition, age, activity • Currently difficult to QA/QC traffic data • For hot-spot analysis, would be useful to have operating mode distributions for

various types of traffic conditions

• More research about travel modeling: how well do these models predict future traffic volumes and speeds? • How can these models and their inputs be improved? • What are the best ways to communicate model choices transparently? • How can the features of the most accurate models be available to more agencies?

• Additional tracer gas studies • Producing independent data sets for use in developing model algorithms, or for

evaluation of AERMOD algorithms still ALPHA or BETA

21