Presentation on "Challenges for Environmental Product Declarations (EPDs) Implementation" delivered by Dr. John Harvey of the University of California Pavement Research Center (UCPRC) at the CalAPA EPD workshop on Oct. 25, 2023 in Sacramento.

1. Challenges for EPD Implementation
John Harvey, Ali Butt
University of California Pavement Research Center
City and County Pavement Improvement Center
National Center for Sustainable Transportation, ITS-Davis
CalAPA
EPD Workshop
Sacramento
25 October 2023

2. • https://escholarship.org/uc
/item/3fn4n3q6

3. Procurement System
• Should be practical in terms of ease of use and complexity
• Should be cost-efficient to maximize environmental improvement per taxpayer
dollar spent; cutoffs for material quantities, which materials are included in EPD
program; other considerations
• Should be sufficiently robust that the risk of unfair comparisons or unwanted
outcomes is sufficiently low to avoid disputes
• Should be documented and communicated well so readily accessible and easily
understood;
• Should provide quantitative feedback to both material producers and the
agency about whether or not EPD program is resulting in improvement of
environmental impacts

4. Benefits & Caveats of Using Cradle-to-Gate EPDs
in Materials Procurement Decision-Making

5. Benefits of State & Local Governments Requiring
EPDs for Construction Products
• Providing information on potential midpoint indicators
• Allowing meaningful quantitative comparisons of the environmental
impacts of materials
• Encouraging industry to become more efficient and less impactful to the
environment
• Providing a means to building out open LCA data
• Informed Decisions - providing a mechanism for measuring
improvement in the environmental impacts of materials through
procurement

6. Caveats of Requiring EPDs for Construction
Products
In context of the typical current approach for specifying materials or selecting
materials using EPDs
1. Thresholds based on national averages
2. Go/no-go specifications
3. Comparability in terms of functionality: sufficient differentiation of performance-
related properties in product categories
4. Variability due to less prescriptiveness in the PCR especially background data
5. Incomplete information in the EPDs (e.g. additives production impacts)
6. Agency knowledge

7. Recommendations for Improving the Benefits
of Using Environmental Product Declarations
in Procurement Decision-Making

8. 1. Recommended Approach to Set Appropriate
Thresholds/Benchmarks
• Transparent documentation/publication of the method to calculate and update
thresholds
• Thresholds should be based on regional data where there are significant regional
differences
• Either using initial collection of published EPDs from the pool of the agency’s likely suppliers
• Or strong agency-specific data based on prior use of materials
• Improving EPDs should be used to periodically set new thresholds
• Thresholds may initially get worse rather than better as data and systems capture in PCRs
become more complete (better data shows things are worse, more things getting counted)
• Set thresholds considering the total quantity of emissions determined from the
project-weighted sum of emissions across all EPD-required products in the project
• Set the benchmark based on all covered materials and their quantities in project
• Let contractor find most cost-effective means to maximize impact reductions for whole project

9. 2. Recommended Approach for Use of
Incentive/Disincentive Specification
• Apply a financial penalty for emitting more than the threshold or a bonus for
emitting less, instead of go/no go
• Where prime is not the material supplier, consider language to require that
incentive/disincentive for a specific material be given to or taken from subcontracted
materials suppliers, or shared between prime and materials suppliers
• Consider sufficient bonuses, up to 5%, and penalties, up to 25% to 50% of the unit
cost of the material to incentivize improvements in industry practices without
causing problems to project budgets
• Thresholds for incentive/disincentive should initially be set near the 50th percentile
of the range of emission values
• Incentive/disincentive should be based on the net sum of the reduction in impacts
for the project, rather than material by material
• Can start by setting net sum for all materials in a project under the same PCR
• Likely to be delivered by same subcontractor/supplier
• Examples: all asphalt materials, all concrete materials, all steel materials

10. 3. And 4. Comparability and Variabililty of EPDs
• Material having lower cradle-to-gate impact may not always be the right
choice
• Not last longer
• Require more frequent pavement maintenance cycles
• Not be recyclable
• A4 product transportation to the consumer not included; example: impact of
transportation of RAP may be greater than impact benefit of using RAP
• Two EPDs for the same material may not have same impacts
• If PCR allows use of different systems boundaries, cutoffs, and especially
background data, variability of up to 20% has been identified (MIT study)

11. 3. and 4. Comparability and Variability of EPDs
• Recommendations:
• Move to performance related tests and
specifications; especially as innovative
concrete and asphalt materials are
coming on line
• Such as in BMD
• Require that PCRs adhere to ACLCA PCR
Guidance which will result in greater
prescriptiveness in PCRs; particularly
regarding background data
• Push for standardization of reporting of
EPD impacts, etc
• Future inclusion of transportation (A4,
A5)

12. 5. Incomplete Information
• There is currently incomplete information in many EPDs,
particularly for additives and admixtures, also use of national
averages for some important materials (asphalt binder)
• Recommendation:
• Voice support to program operators and additive/admixture producers
to supply complete information (EPDs)
• Consider letting them know will not accept incomplete EPDs from
future PCRs that don’t require high level of completeness (NAPA PCR
requires a label showing incompleteness on EPD first page)
• Note that more complete data may increase the impacts shown in
EPDs; this needs to be explained to EPD program sponsors and public

13. 6. Agency Knowledge
• If EPDs are not being reviewed for data quality, completeness, etc
they are not providing best value
• Like collecting QA data but not using it to support decisions
• Recommendation:
• Support continuous improvement of agency knowledge and ability to
interpret information and act on it
• And to write better technical specifications for EPDs
• And to push for improvements in PCRs

14. Deriving Benchmarks for Construction
Materials Based on EPDs
John Harvey, PhD, PE
University of California Pavement Research Center
UC Davis
FHWA EPD Community of Knowledge Webinars
and other EPD information
https://www.fhwa.dot.gov/pavement/sustainability/epds/resources/

15. Outline
• Objectives:
• Learn about methodologies and key considerations when developing benchmarks
• Understand the necessity and importance of benchmarks in the context of Buy Clean Policies
• Participate in an engaging, open discussion on how to collaborate between public agencies and
industry to develop benchmarks
• Questions of the Quarter:
• What is background data and how does it affect benchmarking?
• What is the current state of practice on developing benchmarks in the US?
• What are some of the recommendations/gaps/needs that public agencies and
industry associations have identified to develop benchmarks?
• References:
• Previous webinar (7 Feb 2023, Chait Bhat) on background data:
https://usdot.zoomgov.com/rec/play/OeQ4OkV-8HLYKbxGxnlctDHLRXbcii5oG_qLh4HmCYPpKV53M-
obQB4F2N9ij-88Yv1E3yJthZBAqK9f.hM8I53bpy_xck75X?continueMode=true
• White Paper from National Center for Sustainable Transportation:
https://ncst.ucdavis.edu/research-product/recommended-approach-use-cradle-gate-
environmental-product-declarations-epds

16. What are benchmarks used for?
• What is a benchmark?
• A benchmark is a threshold
value for an environmental
impact that an agency sets to
communicate with producers
• Typically, global warming
potential (GWP) calculated by
TRACI (US EPA) method for civil
infrastructure materials in USA
• Agencies use the benchmark
when procuring materials in Buy
Clean type legislation based on
producer’s environmental
product declaration (EPD)
impact value(s)
• What about cradle-to-gate EPDs
should be considered when using
them in procurement?
• Performance categories
• Variability
• Completeness
• Regionality, timeliness of data
• Is there only one type of benchmark
and one way of using them?
• There are different ways of setting
benchmarks
• There are different ways of using
benchmarks
• Are there other considerations for
agencies as they move ahead?
• Yes

17. Environmental Impacts over the Pavement Life Cycle
What do EPDs cover?
GWP
Years
Initial Preservation Major Rehab
Analysis Period
Use Stage: Building smoother
pavements during construction,
preservation, reduce rolling
resistance
Design and
Construction Stage:
Material
performance
properties,
pavement
structural design,
better construction
quality control,
preservation reduce
number of
treatments
EPDs address
materials impact
only in each design
and construction
event
Need for maintenance
and rehabilitation and
time to reconstruction
can be more important
than initial impacts

18. Performance related properties of materials
must be in same category for comparison
• A material with low GWP on their cradle-to-
gate EPD may produce more GWP over the
life cycle of the infrastructure
• Example: A material may have 15% less GWP in
its EPD than benchmark
• But if it is 25% less durable, it will be replaced
more frequently, and emit more GWP over the
life cycle
• Greater use of EPDs in procurement will require
greater use of performance related specifications
and tests to categorize materials to avoid this
potential unintended negative consequence
Analysis period = 60 yrs
Material A Material B
0.85 GWP 1.0 GWP
15 year life 20 year life
15 20
30 40
45
Total GWP Total GWP
3.4 3.0

19. Some example pavement materials
performance related properties
• Asphalt concrete:
• Stiffness
• Rutting
• Aging
• Fracture
• Fatigue
• Moisture damage
• Portland (or other hydraulic) cement concrete
• Flexural strength (can be related to compressive strength)
• Drying shrinkage
• Coefficient of thermal expansion
• Chemical reaction (ASR, sulfate, chloride) and freezing durability
• Aggregate base
• Shear strength
• Moisture sensitivity
Which of these do
current specifications
consider?

20. Types of EPDs
Want plant and product specific
• Types of benchmarks
• National average
• Are specifications the
same across country?
• What are built-in inputs
to local production
compared to national
average inputs?
• Who calculates this and
how often updated?
• Regional average
• Can be based off first 1-
3 years of collecting
EPDs from agency’s
suppliers for
information only
• Update periodically
using recently collected
EPDs

21. What could go wrong with national average
benchmarks for regionally sourced materials?
• Regionally sourced materials may have regional supply chain constraints
• Asphalt and concrete can only be sourced within a small finite radius of the
construction location (time to set for concrete, time to cool for asphalt)
• Different regions have different:
• Electrical energy supply sources
• Different transportation distances and available modes (truck, rail, barge) to obtain
recycled and other lower impact materials
• Different specifications for materials to meet local climate conditions
• Bad outcomes:
• All your suppliers are better than the national average (no improvement)
• None of your suppliers can become better than the national average (no suppliers)

22. Variability: What is the variability of an EPD value for a
given product and how can it affect procurement?
Environmental Impact
Deterministic EPD
Results for Alternative A
Deterministic EPD
Results for Alternative B
Range of LCA Results due to uncertainties
in input data for Alternative B
Range of EPD Results due to uncertainties
in input data for Alternative A
“B is Better A”
Bhat, C. G., & Mukherjee, A. (2019), “Sensitivity of Life-Cycle Assessment Outcomes to Parameter Uncertainty: Implications for
Material Procurement Decision-Making”, Transportation Research Record, 2673(3), 106–114.
19
Equivalence Interval

23. GATE/ PLACEMENT
Asphalt Mixture
Asphalt Binder
Aggregate
Diesel
Natural Gas
Electricity
Stakeholder: Plant
Manager
Foreground Data
Electricity from
Coal
Electricity from
Wind
Electricity from
Gasoline
Background Data
Crude Oil
Rock
Coal Deposits
Complete Supply-Chain for Asphalt Mixture
At plant
At Grid
At Source
4
Source of
Variability
Producers know
their own primary
data
What background
data is in the EPD?

24. GATE/ PLACEMENT
Asphalt Mixture
Asphalt Binder
Aggregate
Diesel
Natural Gas
Electricity
Stakeholder: Plant
Manager
Foreground Data
Electricity from
Coal
Electricity from
Wind
Electricity from
Gasoline
Background Data
Crude Oil
Rock
Coal Deposits
Complete Supply-Chain for Asphalt Mixture
At plant
At Grid
At Source
4
Source of Bias
and Variability
Is the EPD
“complete”
i.e. was
information
available for all
ingredients?
Additive?

25. Uses of benchmarks in procurement
go/no go vs incentive/disincentive
• Goals:
• Improving environmental outcomes
• Keep a healthy pool of suppliers, with all
competing to improve
• Go/no go procurement specifications
• Set benchmark and only accept materials with
GWP < benchmark
• Does not differentiate just under the benchmark
from substantially under the benchmark
• Incentive is to do the minimum improvement
• Risk when setting benchmark of too few suppliers
can meet it, or nearly all meet it
X material C
Benchmark X material D
X material A
X material B
C, D cannot supply
A, B get paid the same

26. Uses of benchmarks in procurement
go/no go vs incentive/disincentive
• Incentive/disincentive specifications
• Incremental reward or penalty for how much
below or above benchmark
• Set second benchmark above
incentive/disincentive benchmark where
define unacceptable material
• Everyone incentivized to get better to compete
against benchmark
• Similar system used in QC/QA
• For both go/no go and
incentive/disincentive
• There must be enough confidence in the EPD
system that claims are not filed when supplier
cannot supply or receives disincentive
• See previous notes about variability and bias
2nd Benchmark X material C
Benchmark X material D
X material A
X material B
C cannot supply
D gets paid less
A gets paid more
B gets paid even more

27. Adding flexibility to the use of benchmarks
Use of whole-project emission benchmarks
• Goals:
• Minimize GWP emissions for a given project
• Make it easier for the contractor to meet that goal
• Current typical approach: use a benchmark and acceptance for materials on a one-by-
one basis
• This may be hard to achieve for some materials and very easy to achieve for others
• Alternative approach: use material benchmarks and quantities to calculate a project
benchmark
• Calculate: sum of material benchmark (GWP emissions limit) x units of material in project
• Contractor optimizes combination of materials they deliver to maximize the reduction for the
overall project, and to maximize their incentive
• Can start by doing project material budget for all materials under the same Product Category Rule,
such as all concrete materials, all asphalt materials, all steel materials, etc
• Requires consideration and language about how to distribute the incentive/disincentive to general
contractor then to the materials supplier subcontractors

28. Can benchmarks go the wrong direction in the future?
• Some legislation requires that benchmarks must always be
improving
• Some reasons they can legitimately go the wrong direction at
times:
• More higher emission suppliers start publishing EPDs
• EPDs become more complete (gaps are filled)
• Higher quality background data are required that show higher emissions
• Plants become able to partition energy use to different products; some
will go up and some will go down
• Plant specific EPDs become available for important ingredient materials,
are used instead of national averages
• Example: asphalt binder
• Good thing if getting more realistic data even if benchmark goes
up
• Agencies should understand what is happening in their collected EPDs and
participate in PCR development
• Legislators and regulators should understand this

29. How might benchmarks change with improved
data
• The Asphalt Institute published a
national (USA/Canada) average LCA
for asphalt binders in 2019
• Most EPDs for asphalt mixes are
using those national average binder
GWP values
• If plant or regional specific EPDs
become available for asphalt binder
then those values would likely be
different for different binder
suppliers
• Binder drives mix GWP; crude
source drives binder GWP
Mukherjee for NAPA, 2021
Thinkstep for Asphalt Institute 2019

30. Different crude sources have different GWP
• GWP depends on extraction
method, flaring of gas, and
transportation
• AI LCA is heavy on Canadian oil
sands; unconventional onshore
extraction (see Bhat webinar)
• UCPRC has estimated
differences for PADD5 and
California refined binders vs
national average
• Not yet peer reviewed
• Regional benchmarks will
likely change; supply chains
depend on ability to
transport crude and binder
Thinkstep for AI 2019
Ostovar et al UCPRC for Caltrans 2023 in press

31. Some current benchmarking practices
• Buy Clean California (2017)
• Benchmarks: national averages (plate steel, reinforcing steel, plate glass, mineral
wool), those above cannot supply (go/no go)
• Procurement implemented July 2023
• Benchmarks cannot go down
• Developing EPD program for asphalt, concrete and aggregate materials
• Concerned that data may show increase as more EPDs come in
• Buy Clean Colorado (2021)
• Collecting EPDs since summer 2022
• No benchmarks set yet, looking at regional and national data
• Go/no go specification, not sure yet if average or percentile
• Concerned that data may show increase as more EPDs come in

32. Some current benchmarking practices
• Washington (2021-2022)
• The maximum acceptable GWP must be set at the 20th percentile value for each
eligible product category, determined by consulting with nationally or internationally
recognized databases of EPDs of like performance and quality materials
• Must report benchmarking method to the legislature by January 1, 2024
• Oregon (2022)
• Limits are set approximately 45% above the National Ready Mix Concrete
Associationʼs Pacific Northwest GWP Benchmarks, same as the City of Portland
• Estimated that will allow roughly 80% of the mixes with EPDs in the Oregon market
to meet the proposed limits now
• Minnesota Buy Clean and Buy Fair Minnesota Act (2023)
• Industry average benchmark by 2025 or 2027 depending on material, considering
nationally or internationally recognized databases
• Benchmarks cannot go in reverse

33. Conclusions
• EPDs are an important tool for improving environmental outcomes for
pavement, including use in procurement
• Not the only tool, consider the whole life cycle and whole project delivery
process
• EPDs can provide better data to whole life cycle LCA
• Recommend better differentiation of materials performance
• Should work towards improved EPDs
• Less variability, more complete data
• Benchmarks can be set different ways
• Benchmarks can be used different ways
• Benchmarks can move in different directions for valid reasons

34. Thank you!
Questions
• Note: author is solely responsible for information and opinions in this presentation
• Thanks to Chait Bhat and Chris Senseney for some of the information presented