The Yale Climate Initiative: GHG emissions from transportationMarco ButtazzoniKathryn A. Zyla
2
Presentation Overview
• Study background– Yale Climate Initiative (YCI)– Yale’s Inventory
• GHGs from transportation– Boundaries and methodology– Emissions calculations approaches
• Conclusions
3
YCI Overview: Purpose
Student-initiated study to:– Understand Yale’s greenhouse gas (GHG) emissions
drivers– Develop a GHG emissions inventory for Yale– Analyze approaches to make the University more
climate friendly
An opportunity to reflect on existing inventory methodologies and tools
4
Greenhouse Gases at Yale
CO2CH4N2OHFCs
CO2e offsets?
PFCsHFCsSF6CO2N2O
CO2CH4N2O
CO2CH4N2O
CO2. CH4, N2O, CFCs
5
Yale’s GHG Emissions: Year 2002Yale University GHG Emissions (2002)
Total 291,696 Tons CO2 equivalent
38,653
13,254
246,080
-6,291
-50,000
0
50,000
100,000
150,000
200,000
250,000
300,000
Power Plants/Buildings Transportation Other Sources Sinks
tons
of C
O2
equi
vale
nt
Other sources 4%
Transport. 13%
Power Plants/ buildings 83%
6
Transportation: System boundaries
Yale owns 366 vehicles, over half of which are
trucks or vans
• Institutional (WRI Scope 1)– Yale-owned vehicles
• Work-related (WRI Scope 3)– Conferences, meetings, research trips(flights, train, ground transportation)
• Commuters (WRI Scope 3)– Daily (faculty, students)– Travel home (students)
Yale spends about $20 million in travel each year
Yale commuters travel over 57 million miles per year
System boundaries chosen to:– Understand Yale’s total footprint– Explore methodologies for quantifying more than the usual direct
fleet emissions
7
Uncertainty
Importance of uncertainty– General need to understand
range of potential values – Very large for transportation
emissions– Several assumptions needed
to calculate emissions
The ‘inventory community’ has an opportunity to define more standardized factors and procedures
Our approach– Develop calculation tool– Estimate the variability
deriving from single variables– Evaluate total uncertainty– Identify most significant
variables and assumptions
8
Institutional Emissions: Data
• Vehicles owned and operated by Yale University– Vehicle purchasing decisions made by Yale– Fuel purchases made by Yale drivers
Data sources– Kept centrally
– Up-to-date list of vehicles– Accurate fuel purchase information from credit card transactions– Inconsistent identification of vehicles being refueled– Inaccurate vehicle mileage data
9
Institutional Emissions: Analysis
Diesel Fuel Consumption
Unleaded Fuel Consumption
Activity Data+/- 5%
Emissions Factors
Calculated Emissions
+/- 3%
Total CO2Emissions
1,622 t
Diesel Emissions
Factor
Unleaded Emissions
Factor
+/- 8%
aaa Activity Data Assumptionsbbb ccc Calculations +/- x% Uncertainty
10
Institutional Emissions: Conclusions
• Low uncertainty for total institutional emissions, but • High uncertainty on the emissions and efficiency of
individual vehicles and users→ Difficult to translate to mitigation opportunities
Case studies and “Best management practices” to help companies gather and analyze more granular data?
11
Work-related emissions: Data
• Travel directly linked with university activities – Conferences, meetings, research trips
• Includes many modes of transport– Flights, train, ground transportation
Data sources– Financial records (from expenditure reports etc.)– Yale’s travel agency records (with good granular information
about flights)– Some purchases not made through centralized process– Many types of trips are aggregated monetarily, not separated
by mode
12
Administration expenditure
data
% spent per
mode
Air and train
Car rental
Ground transp.
Lodging & food
Miscellan.
Travel agency data
% spent per mode
Emissions per $
Emissions (air, train, rented car,
ground transpor.)
Not considered
aaaData
Assumptions
Calculations
bbb
ccc
Work-related emissions: Analysis
Train
Air
13
Air travel analysis v. 1
# InternationalSegments
# DomesticSegments
CityDistances
GallonsConsumed
Miles/ gallon
Total Energy
Consumed
Energy/ gallon
Travel Agent CO2
Emissions
Tons CO2/GJ
Total CO2Emissions
8,792 t
TA as % of
total $Travel
Agency data
+/- 3%+/- 5% +/- 38% +/- 2% +/- 11%
Miles per segment
(web)
+/-59%
bbbaaa cccData Assumptions Calculations +/- x% Uncertainty
14
Air travel analysis v. 2
Total CO2Emissions
8,384 t
TA as % of total $
+/- 11%# International
Segments
# DomesticSegments
Travel Agent CO2Emissions
CityDistances
Miles per segment (web)
Emissions/ mile long
trips
Emissions/ mile medium
trips
Emissions/ mile short
trips
Travel Agency data
+/- 5%
+/- 10%
+/- 10%
+/- 10%
+/- 26%
bbbaaa cccData Assumptions Calculations +/- x% Uncertainty
15
Work-related emissions• Using financial data to calculate GHG emissions
required a variety of assumptions. E.g.ParameterMiles/$ train
GHG/passenger-mile train
$/day car rentalMiles/day car rental
GHG/pass. mile car rental
Miles/$ ground transportation
GHG/pass. mile ground t.
SourcesTA, Market prices
WRI
TA, Market pricesTA benchmark
WRI
Expenses mile rebate Market prices
WRI, DOT
NotesSignificant variability in train pricesOnly one factor for the US at the time of the analysis (now improved)
Based on limited random sampleBased on industry data
Based on limited sampleFew data available
None of these parameters influenced the work related emissions by more than 3.3% of the total
16
Work related emissions: Conclusions• Several parameters can affect uncertainty when
calculating work related emissions• The inventory community could improve consistency, if
not accuracy, if it agrees on parameters such as:– Miles per $ spent for various transportation methods (ideally at
state or local level)– Emissions per mile for different transportation methods (requires
aggregating different options)• With transportation companies looking at offering GHG
emissions offset services we should also think about information protocols to transfer relevant data between transportation companies and their customers
17
Personnel commuting: Data• Overall Yale employees
about 12.500 people– Faculty– Researchers– Management – Support
Where employees live Zip codes data
0%
5%
10%
15%
20%
25%
30%
less than1
between1 and 5
between5 and 10
between10 and 25
between25 and 50
more than50
Miles from Yale
% e
mpl
oyee
sData sources
– Zip code records from Human Resource Department
18
Zip code data
Correction for record mistakes
CT-DOT transport
parameters
Edited zip
code data
Mode of transport
ation
Total miles
traveled
aaa Data
Assumptions
Calculations
bbb
ccc
# of trips per
week
Miles per trip
Miles traveled
per mode of
transport
Emissions parameters
Emissions per mile
per mode
Emissions per mode
of transport.
Car occup. rate
FES & parking dep. data
Personnel commuting: Analysis
19
Zip code data
Correction +/- 2%
CT-DOT transport
parameters
Edited zip
code data
Cars only vs. PT
and bike
Total miles
traveled
aaa Data
Assumptions with change in the parameter
Calculations
bbb
ccc
# trips +/- 1%
Miles+/-5%
Miles traveled
per mode of
transport
Emissions parameters
Cars +22 -22%PT +120 –50%
Total CO2Emissions16,665 t
+ 19%- 12%
xx% Change in emissions
+ 6%- 9%
+ 4.8%- 4.8%
+ 1.5%- 3.7%
FES & parking
dep. data
Occ. Rate
+/- 17%
+ 20%- 14%
car +21.6%- 21.6%
PT +3.24%- 1.35%
+ 80%- 50%
Personnel commuting: Analysis
20
Personnel commuting: Conclusions• Work related emissions estimate could have benefited
from– A mobility study for the University– More precise and current data on personnel and students
residence– Vehicle-specific parking data (dept name doesn’t matter, but
type of data desired seems important)– State level and local data on transportation behavior– State level and local emission parameters (e.g. for car, train or
bus emissions)– Average (non-technology-specific) automobile emissions
parameters
21
Summary: Emissions from Transportation
4151,700 1,688
16,700
8,3945,4002,734
1,622
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
Inst
itutio
nal
trave
l
Wor
k re
late
dai
r tra
vel
thro
ugh
trave
lag
ent
Oth
er w
ork
rela
ted
trave
l
Em
ploy
ees
com
mut
esan
d vi
sits
Stu
dent
sco
mm
utes
and
visi
ts
Stu
dent
sre
turn
ing
hom
e (d
om.)
Stu
dent
sre
turn
ing
hom
e (in
t’l.)
Tran
spor
t'o
ther
activ
ities
'
tons
CO
2 eq
uiva
lent
Other sources 4%
Transport. 13%
Power Plants/ buildings 83%
22
Conclusions: Data Quality
• Lack of Data– Define management systems and templates for data
gathering and emissions calculation– Prepare standard templates for mobility studies for
Universities– (In the future?) Define communication protocols
between the travel industry and end users• E.g. For airline or train companies to communicate CO2
emission data to their users
23
Conclusions: Emissions factors
• Create a repository with agreed/standard parameters for: – Miles per $ spent per mode of transportation– Emissions per mile traveled per mode of transportation– Regional and local vehicle use parameters
• Occupation rate• Distance traveled for commutes• Mode of transportation chosen
– Average automobile emission factors
Parameters of increasing level of granularity can be provided and associated with decreasing level of uncertainty
24
Conclusions: Methodology
• Current calculations tools for transportation emissions can be a basis for improving scope and precision– New calculation tools and protocol can broaden calculations– More granular analysis is possible if suitable parameters are
available (and provided by reputable sources)– More explicit uncertainty analysis
• Case studies, ‘best practices’ and training to help companies improve:– Internal data management systems and – Processes and incentive structures
Thank you!Kathryn Zyla, [email protected] Buttazzoni, [email protected]