METHOD REPORT

An analysis of the contribution C40 cities can

make to delivering the Paris Agreement objective
of limiting global temperature rise to 1.5 degrees.
Contents
Page

1 Introduction 2
1.1 Process 2

2 Baseline Emissions and Sector Profiles 4

2.2 Sources of Emissions Inventories 5
2.3 Adjusting to baseline year 8
2.4 Mapping city emissions data 8

3 C40 Cities Carbon Budget 12

3.1 Global Carbon Budget 12
3.2 C40 Cities Carbon Budget 15
3.3 Final C40 Budgets 18

4 C40 Cities Trajectories 20

4.1 Developing BAU Trajectories 20
4.2 Climate safe emissions trajectories 21

5 2CAP 29
5.1 Functioning 29
5.2 Data 38

6 Key Data Inputs 40

6.1 Population 40
6.2 City GRP data 43

7 Bibliography 45

8 Glossary 47
1 Introduction
This document provides a summary of the method used by the C40 Cities Climate Leadership
Group (C40) and Arup research team in developing the analysis behind the Deadline 2020 report 1,
launched at the C40 Mayoral Conference in Mexico City on 1 December 2016.
Deadline 2020 was conceived as a research project under the $2 million C40-Arup partnership with
the aim of understanding C40 cities’ role in delivering a future world consistent with the
commitments and ambitions of the 2015 Paris Agreement, which entered into force in November
2016.
The Paris Agreement commits signatories to “holding the increase in the global average
temperature to well below 2°C above pre-industrial levels, and to pursue efforts to limit the
temperature increase to 1.5°C above pre-industrial levels.” While cities were not party to the state-
level agreement, hundreds signed up to the Paris Pledge for Action 2 alongside other non-state
groups such as businesses, investors, and civil society groups.
Accordingly, this research (“the Study”) investigates future greenhouse gas (GHG) emissions
burdens and the city actions associated with reducing these such that 1.5 or 2 degree-limit
temperature rise scenarios may be achieved.

1.1 Process
Figure 1 below outlines the overall process followed in the Study, which also reflects the structure
of this report.

Define C40 cities’ carbon budget

Define C40 cities’ target emissions

trajectories to 2050 and beyond

Build actions scenario model

Develop scenarios, iterate, analyse,

test

Figure 1. Overall process for Deadline 2020 research

It was first necessary to define and identify the share of a global carbon budget that could be
attributed to C40 cities under 1.5 or 2 degree scenarios. In this Study, the global carbon budget is
referred to as the estimated permissible anthropogenic GHG emissions to confidently limit global
warming within a target temperature increase from a given start date.
Following this, C40 cities were assigned one of a series of target emissions trajectories that shared
the overall C40 budget and allow for a degree of burden sharing. Alongside the budgeting and

1
www.C40.org/other/deadline_2020
2
www.parispledgeforaction.org

Page 2 of 49
trajectories, a city actions and emissions scenario model was developed to test the steps necessary
for C40 cities to achieve their targets. This model was then used to test a range of scenarios and
develop the final content in the Deadline 2020 report.
One of the main goals for this study was to develop a future-proofed methodology that uses as
much of the wealth of data reported to C40 by member cities over the last six years as possible, and
is able to benefit from additional data currently being gathered. At the time of analysis, C40
membership consisted of 84 cities (Table 1); membership has since grown and is expected to
continue to do so. Also, not all C40 cities were able to provide full actions / emissions / powers
data. As C40 membership grows, and comprehensive data gathering is progressed, the frameworks
developed in the Study will enable the analysis to be revised and refreshed for greater accuracy and
insight.
It is noted that due to the above reasons, in some instances it was necessary to make simplifications,
assumptions, or interpolations where the data was not available in the required format. Overall
conclusions, therefore, need to be read on the understanding that absolute numbers may change in
future revisions of this work, but that this is a natural part of the research process.
Table 1. List of C40 cities analysed for the Study

Accra Copenhagen Lima Rotterdam

Addis Ababa Curitiba London Salvador
Amman Dar es Salaam Los Angeles San Francisco
Amsterdam Delhi Madrid Santiago de Chile
Athens Dhaka North Melbourne São Paulo
Auckland Dhaka South Mexico City Seattle
Austin Dubai Milan Seoul
Bangalore Durban Moscow Shanghai
Bangkok Guangzhou Mumbai Shenzhen
Barcelona Hanoi Nairobi Singapore
Basel Heidelberg Nanjing Stockholm
Beijing Ho Chi Minh City New Orleans Sydney
Berlin Hong Kong New York City Tokyo
Bogotá Houston Oslo Toronto
Boston Istanbul Paris Tshwane
Buenos Aires Jaipur Philadelphia Vancouver
Cairo Jakarta Portland Venice
Cape Town Johannesburg Quezon City Warsaw
Caracas Karachi Quito Washington, DC
Changwon Kolkata Rio de Janeiro Wuhan
Chicago Lagos Rome Yokohama

Page 3 of 49
2 Baseline Emissions and Sector Profiles
The emissions “baseline”, a key data point for the analysis carried out in the Study, is the point in
time from which forward projections are made. This section details the approach taken to develop
the baseline emissions profile across all C40 cities.

2.1.1 Method
For this Study, the emissions baseline was defined as the scope I and II 3 greenhouse gas (GHG)
emissions of C40 cities in 2015, the Study’s “baseline year”. The baseline emissions data used
within the Study were:
1. Total territorial scope I and II GHG emissions reported in CO2e
2. The proportional split of those emissions across five sectors, namely Stationary Energy,
Transport, Waste, AFOLU (Agriculture, Forestry and other Land Use), and IPPU (Industrial
Process and Product Use)
Scope III emissions were not included within the baseline data because of the risk of double
counting emissions. A lack of available Scope III data was also a driver for this decision.
The Global Protocol for Community Scale Greenhouse Gas Emissions Inventories (GPC) 4 is a city-
tailored inventory standard developed and promoted by the C40 and its partners, which cities
considered in this Study have already reported against. At the time of writing, nearly 30 C40 cities
had completed GPC inventories for emissions within their city boundaries.
The total territorial emissions data was sourced from inventories for the Global Protocol for
Community-Scale Greenhouse Gas Emissions or the Carbon Disclosure Project (CDP). 5 A
prioritisation approach was developed whereby GPC data was adopted in preference to CDP data
due to higher available sectoral resolution. The reporting year of emissions data ranged from 2009
to 2016. Data was normalised to the year 2015 using city-based Gross Regional Product (GRP)
growth rate projections sourced from the Economist Intelligence Unit (2016). Those cities with
emissions data are referred to as “Primary” cities. For cities without data available from either
source, a process of “mapping” these cities (referred to as “Secondary” cities) to the most similar
Primary city was adopted. The “mapping” methodology is detailed in Section 2.3.
Detailed GPC-compliant emissions inventories are currently being compiled for all C40 cities. The
approach developed to generate emissions data for all C40 cities is a working solution to facilitate
understanding of the scale of the challenge, prior to obtaining full GPC-compliant emissions
inventories from all C40 cities.

3
Scope I: direct emissions from combustion of fuels for heating, transportation etc.; Scope II: indirect emissions from
consumption of purchased electricity, heat or steam; Scope III: indirect emissions outside territorial boundaries that are
associated with cities’ activity.
4
A comprehensive City-level carbon accounting method based on the Greenhouse Gas Protocol (GHG Protocol)
http://www.ghgprotocol.org/city-accounting/
5
A secondary source of self-reported city emissions data used when GPC data is unavailable https://www.cdp.net/en-
US/Pages/HomePage.aspx

Page 4 of 49
The method for developing a city emissions baseline is shown in Figure 2 below.

City X

Yes
GPC inventory?

No

Yes
Baseline year
CDP inventory?
Adjustment

No

“Map” emissions of
Secondary cities Database of Primary cities

Baseline inventory
Figure 2. Method for generating C40 cities baseline emissions inventories

2.2 Sources of Emissions Inventories

City-specific emissions inventories were available for 47 cities. These were identified as the
Primary cities used. The GPC and CDP inventories are briefly described below, including city
coverage and data available for the Study.

2.2.1 GPC
Established by the Compact of Mayors, the Global Protocol for Community-Scale Greenhouse Gas
Inventories (GPC) is a structured methodology for reporting city GHG emissions. At the time of
completing this study, detailed city emissions inventories had been collected for 25 C40 cities.

Page 5 of 49
The GPC inventories include the seven gases covered by the UNFCCC 6 namely CO2, CH4, N2O,
hydrofluorocarbons, perfluorocarbons, SF6 and NF3. All emissions are reported in CO2 equivalents
within the GPC and this study.
Table 2. List of cities with completed GPC inventories

City Country

Amman Jordan
Bogotá Colombia
Boston USA
Buenos Aires Argentina
Cape Town South Africa
Copenhagen Denmark
Johannesburg South Africa
London United Kingdom
Madrid Spain
Melbourne Australia
Mexico City Mexico
New York City USA
Oslo Norway
Paris France
Philadelphia USA
Portland USA
Quito Ecuador
Rio de Janeiro Brazil
Seattle USA
San Francisco USA
Stockholm Sweden
Sydney Australia
Toronto Canada
Vancouver Canada
Washington, DC USA

6
United Nations Framework Convention on Climate Change

Page 6 of 49
The GPC categorises city emissions into five main sectors, matching the GHG Protocol: 7
1. Stationary Energy
2. Transportation
3. Waste
4. Industrial process and product use (IPPU)
5. Agriculture, forestry and product use (AFOLU)
The GPC data was used to derive total scope I and II city emissions, and also the proportional split
across sectors for cities without complete emissions data.

2.2.2 CDP
CDP is a not-for-profit organisation which collects and discloses emissions data from over 100
cities. 22 cities’ total territorial emissions were provided from the CDP database (listed in Table 3).
Unlike the GPC data, emissions data is reported according to a range of both international and local
reporting guidelines. As shown in Table 3, the majority of cities followed the 2006 IPCC
Guidelines, a globally recognised reporting standard.
The CDP data was not available with a sector breakdown.
The CDP city boundaries may not align exactly with C40 city members’ boundaries as considered
in this Study.
Table 3. List of cities whose emissions reported to CDP were used in study

City Country Reporting Methodology

Athens Greece 2006 IPCC Guidelines for National Greenhouse Gas Inventories
Amsterdam Netherlands No details provided
Draft Community-Scale GHG Emissions Accounting and Reporting Protocol
Austin USA
(ICLEI)
Berlin Germany Statistical Institute Berlin-Brandenburg
Caracas Venezuela 2006 IPCC Guidelines for National Greenhouse Gas Inventories
International Local Government GHG Emissions Analysis Protocol Version 1.0; and
Durban South Africa Local Government Operations Protocol for the Quantification and Reporting of
Greenhouse Gas Emissions Inventories Version 1.1
Hong Kong China 2006 IPCC Guidelines for National Greenhouse Gas Inventories
The International Basic Standard for Community-Scale GHG Emissions Inventories
Houston USA
(ICLEI/C40/WB)
Jakarta Indonesia 2006 IPCC Guidelines for National Greenhouse Gas Inventories
Los Angeles USA ICF International Inventory Guidelines
2006 IPCC Guidelines for National Greenhouse Gas Inventories (Revised 1996
Moscow Russia
Guidelines).
Rome Italy In-house methodology

7
www.ghgprotocol.org

Page 7 of 49
 City Country Reporting Methodology

The International Basic Standard for Community-Scale GHG Emissions Inventories

Rotterdam Netherlands
(ICLEI/C40/WB)
Santiago de
Chile 2006 IPCC Guidelines for National Greenhouse Gas Inventories
Chile
São Paulo Brazil 2006 IPCC Guidelines for National Greenhouse Gas Inventories
Seoul Korea 2006 IPCC Guidelines for National Greenhouse Gas Inventories
Singapore Singapore Revised 1996 IPCC Guidelines
Tokyo Japan Guidelines set by Japanese Ministry of Environment
Tshwane South Africa International Emissions Analysis Protocol (ICLEI)
Venice Italy 2006 IPCC Guidelines for National Greenhouse Gas Inventories
Warsaw Poland 2006 IPCC Guidelines for National Greenhouse Gas Inventories
Yokohama Japan Guidelines set by Japanese Ministry of Environment

2.3 Adjusting to baseline year

Incoming emissions data were not all reported for the baseline year of this Study (2015). To account
for this, an adjustment process was used to bring inventory data to 2015 levels.
An annual city GRP growth rate was used to adjust emissions to the baseline year. This approach
was adopted on the basis that:
• Emissions growth is strongly linked to economic activity, i.e. GDP / GRP growth is a driver of
emissions. 8
• Absolute GDP / GRP growth has already internalised population growth meaning changes in
population need no further consideration.
Details of the city GRP growth rates adopted are detailed in Section 6.2.

2.4 Mapping city emissions data

This section outlines the methodology for “mapping” emissions data from cities with available data,
to those without. It was necessary to map cities with no emissions inventory, and those with no
sectoral breakdown. As further data becomes available from C40 cities, this can readily replace the
mapped data used in the first iteration of the Study.

2.4.1 Mapping process

Secondary cities were mapped to Primary cities using a number of demographic, climatic and
socioeconomic indicators.

8
IPCC AR5 WGIII Chapter 6 Assessing Transformation Pathways

Page 8 of 49
Table 4 lists the indicators used to match a given Secondary city with a Primary city. The indicators
were weighted according to an assessment of relative importance (final column of Table 4). As an
example, a higher score (5) was given to a secondary city in the same country as a primary city. A
lower score (2) was given to a city with a similar rate of GRP growth.

The criteria were split into binary matches (yes or no) or bands. In the latter case, the bands of
possible values for these criteria were broken into discrete bands to enable the pairing of cities that
are similar (as indicated in Table 4). For instance, Human Development Index has been broken into
five bands (0 – 0.2; 0.21 – 0.4, etc.)
The maximum score possible with the current weights is 30. The Primary city with the highest
pairing score is used to “map” the emissions for the Secondary city.

Page 9 of 49
 Table 4. Criteria used in pairing Primary and Secondary Cities

Criteria Description Units Binary Banded Weighting

Country Nation in which city is located  5
Region Region in which city is located. The
regions considered were Sub-Saharan
Africa, North Africa, Europe, North
 4
America, Latin America, East Asia, South
Asia, Southeast Asia & Oceania and
Middle East.
CITY GRP / capita See information on sources in Section 6.2 US$ /
 4
capita
Continent Continent on which city is located  3
Climate Zone Based on the Köppen Climate
Classifications:
A – Tropical
B – Dry  3
C – Temperate
D – Continental
E – Polar
Human Development Index The Human Development Index (HDI) is a
summary measure of average achievement
in key dimensions of human development:
a long and healthy life, being  3
knowledgeable and have a decent standard
of living. HDI rankings were sourced from
UNDP 9
Population growth rate in See information on sources in Section Capita /
 2
year 2015 6.1.1 year
CITY GRP growth rate in See information on sources in Section 6.2 %
 2
year 2015
CITY GRP See information on sources in Section 6.2 Million
 1
US$
Population density See information on sources in Section 6 Capita /
 2
km2
Population in year 2015 See information on sources in Section 6  1
Maximum 30

9
http://hdr.undp.org/sites/default/files/ranking.pdf

Page 10 of 49
2.4.2 Generating emissions data
As mentioned previously, missing data included both total emissions and the sectoral split. The
Primary-Secondary city pairing was therefore made in two steps:
1. Once a Primary-Secondary city pairing was made, the total per-capita emissions of the
Primary city were multiplied by the population of the Secondary city to produce the total
baseline territorial emissions of the Secondary city.
2. The sector proportions (in percentages across the five sectors discussed in Section 2.2.1) of
the Primary city were replicated for the Secondary city.
Steps 1 and 2 were required for all cities with no emissions inventories from either source. Those
cities with CDP data only underwent the mapping described in step 2 to obtain sector profiles.
While in reality it is unlikely that two cities will have identical emissions per capita or emissions
profiles across the sectors of interest, the following method was an appropriate interim solution to
generate city specific emissions data. These mapped inventories will be superseded once full GPC-
compliant inventories are produced by all cities.

Page 11 of 49
3 C40 Cities Carbon Budget
This section details the development of the carbon budget – the estimated permissible
anthropogenic GHG emissions to confidently limit global warming within a target temperature
increase from a given start date – for the C40 cities. The following key steps were taken:
Step 1: Determine the global carbon budget for safe levels of warming of below 1.5 and 2 degrees.
Step 2: Identify an approach to allocate a fair portion of this budget to the C40 cities.
Step 3: Calculate the resulting total C40 carbon budget using the chosen approach in step 2 and the
relevant carbon budgets in step 1.

3.1 Global Carbon Budget

In line with the Paris Agreement’s aims and aspirations, two target temperature increase scenarios
were considered to reflect a range of low-carbon trajectories: 1.5 and 2 degrees Celsius warming
beyond average levels in 1870 (referred to as “pre-industrial” times).
Following a review of published global carbon budgets, those from Intergovernmental Panel on
Climate Change (IPCC) were selected to be used for this study. The IPCC is the recognised
scientific body on climate change science and its assessment is based on independent study of peer-
reviewed research. Critically, the cumulative emissions figures reported in the IPCC fifth
assessment report (AR5) satisfied the following criteria 10 for this study:
• Must include all GHGs. Some published carbon budgets report on cumulative carbon dioxide
emissions only, excluding other gases listed under the Kyoto Protocol. Note that the IPCC
provided a carbon budget including all GHGs for warming consistent with below 2 degrees, but
not 1.5 degrees, therefore the the 1.5 degree carbon budget was derived as shown in Table 5.
• Must include all anthropogenic emissions from all possible sources. A distinction is made in
the literature to emissions derived from fossil fuels and industrial activity from land use change.
The latter is significantly more complex to predict and as a result often excluded. It was
considered important that any carbon budget should fully account for all potential sources of
emissions.
• Good chance of meeting the target temperature. The selected carbon budgets have a 66%
probability of not exceeding the target level of warming. Under the IPCC terminology this is
referred to as “likely”. The figures selected have been widely published as the IPCC’s
recommended level of confidence.
• Extended out to 2100. This was established with the C40 as the desired analysis period.

10
Based on criteria developed by the Carbon Tracker Initiative (Carbon Tracker Initiative & LSE GRI, 2013)

Page 12 of 49
Table 7 outlines all carbon budgets reviewed and the reasons in favour of the IPCC figures.
The global carbon budgets as well as those used to derive them (a combination of CO2 and non-CO2
budgets) in this Study are presented in Table 5.
Table 5. Global emissions budgets consistent with below 1.5 and 2 degrees temperature rise between 1870
and 2100. Values in italics derived.

Description of Value Source Details Value Source Details

emissions budget consistent consistent
with below with below
2 degrees 1.5 degrees
CO2-only budget 2900 IPCC AR5 Synthesis 2250 IPCC AR5 Synthesis
between 1870 and 2100 Report, Table 2.2 page Report, Table 2.2 page
(CO2) 64 64
Non-CO2 budget 770 Derived from subtracting 770 Equal to non-CO2 budget
between 1870 and 2100 total carbon budget consistent with below 2
(CO2e) (below) from CO2 only degrees
budget
Carbon budget between 3670 IPCC AR5 WGI, page 27 3020 Derived by adding CO2
1870 and 2100 (CO2e) only and non-CO2 budget

The year of the baseline emissions for C40 Cities has been set as 2015. Accordingly the carbon
budget (CO2e) figures presented in Table 5 were converted to carbon budgets from 2016 to 2100 by
subtracting historic GHG emissions.
According to Le Queré (2015), the cumulative GHG emissions between 1870 and 2015 were 2,037
GtCO2e. Historic estimates of the non-CO2 GHGs were given as 596 GtCO2e based on available
data from IPCC and Emissions Data for Global Atmospheric Research (EDGAR) 11 on historic non-
CO2 GHGs. 12 These emissions were subtracted from the calculated 1870-2100 budgets in Table 5 to
produce the remaining budgets for each scenario, shown in Table 6.
Table 6. Resulting global carbon budgets between 2016 and 2100

Value consistent Value consistent

with below 2 with below 1.5
degrees degrees

Global carbon budget between 1,037 387

2016 and 2100 (GtCO2e)

11
http://edgar.jrc.ec.europa.eu/overview.php

Page 13 of 49
Table 7. Summary of review of potential carbon budgets
Source Date of Period Gas emissions Likelihood Unit Target Target Target Target Target Target Explanatory Notes Reasons for rejection
Source basket Temp Temp Temp Temp Temp Temp
1˚C <1.5˚C <2˚C 2˚C <2.5˚C <3˚C
Based on number of different scenarios including
Carbon budget period inconsistent with
with and without overshoot (the latter require
2011-2050 GHG emissions 44 - 68% GtCO2 870-1254 study period and includes scenarios with
negative emissions) and different levels of non-CO2
negative emissions.
emissions.
Based on number of different scenarios including
Chapter 6, WGIII, IPCC 2011-2100 GHG emissions 14% GtCO2 630-1180 with and without overshoot (the latter require Likelihood below 66%
2014
(Clarke et al., 2014) negative emissions) and different levels of non-CO2
2011-2100 GHG emissions 63% GtCO2 630-1180 emissions, Probability in paper was stated as a range Likelihood below 66%
of likelihood of exceeding target temp. The
probability of achieving below target temp was
Target temperature higher than stated
2011-2100 GHG emissions 81% GtCO2 540-640 therefore calculated as 1-minus the highest
target.
probability of exceeding target temp.
Limited to CO2 emissions from fossil fuels
Fossil fuel CO2 Target for fossil fuels assuming 100 GtC are restored
(Hansen et al., 2013) 2013 2011-2100 GtCO2 500 and target temperature lower than stated
emissions to biosphere through reforestation
target.
Fossil fuel CO2
2013-2100 50% GtCO2 525 Limited to CO2 emissions from fossil fuels.
emissions
Fossil fuel CO2
2013-2100 50% GtCO2 1075 Limited to CO2 emissions from fossil fuels.
emissions
There are based on assumption of lower non-CO2
(Carbon Tracker Initiative & Fossil fuel CO2
2013 2013-2100 50% GtCO2 1275 emissions and higher aerosol levels than IPCC Limited to CO2 emissions from fossil fuels.
LSE GRI, 2013) emissions
estimates that result in higher carbon budget.
Fossil fuel CO2
2013-2100 80% GtCO2 900 Limited to CO2 emissions from fossil fuels.
emissions
Fossil fuel CO2
2013-2100 80% GtCO2 1125 Limited to CO2 emissions from fossil fuels.
emissions
Inconsistent period with study period and
2000-2049 GHG emissions 63% GtCO2 1356 probabilities below satisfactory confidence
level.
Inconsistent period with study period and
Probability in paper was stated as a range of
2000-2049 GHG emissions 57% GtCO2 1500 probabilities below satisfactory confidence
likelihood of exceeding 2 degrees. The probability of
level.
(Meinshaussen et al., 2009) 2009 achieving below 2 degrees was therefore calculated as
Inconsistent period with study period and
1-minus the highest probability of exceeding 2
2000-2049 GHG emissions 49% GtCO2 1678 probabilities below satisfactory confidence
degrees.
level.
Inconsistent period with study period and
2000-2049 GHG emissions 30% GtCO2 2000 probabilities below satisfactory confidence
level.
Probability expressed as best estimate within
from 2009 confidence interval of95%. Based on concept of Limited to CO2 emissions i.e. not complete
(Matthews et al., 2009) 2009 CO2 emissions 95% GtCO2 1466.8
onwards carbon-climate response. Ignores effect of non-CO2 budget for GHG emissions.
emissions. Stated range of 0.4TtC to 1.5TtC.
Probability expressed as most likely. Ignores effect of
(Allen et al., 2009) 2009 2000-2050 CO2 emissions most likely GtCO2 400 Inconsistent period with study period.
non-CO2 emissions.

Page 14 of 49
3.2 C40 Cities Carbon Budget
This section relates the methodology for allocating a “fair” carbon budget to C40 cities from 2016
to 2100. The sub-sections follow the chronological order adopted, starting with framing the context
for allocation of budgets to sub-global entities and a snapshot of existing approaches developed by
the scientific community and government and non-governmental organisations. This information
was evaluated and a decision matrix developed to aid in selecting a suitable approach for this Study.
The results for a C40 carbon budget consistent with both 1.5 and 2 degrees are presented along with
the key implications of these.
While individual city budgets are implicit in many of the methodologies described, a key concept to
note is that the budgets discussed refer to a single, overall budget for the bloc of C40 cities;
individual city budgets are described in Section 4.

3.2.1 Context
The international community has been grappling with the challenge of reaching a global consensus
to tackle climate change since the United Nations Framework Convention on Climate Change
(UNFCCC) was created in 1992. The two main policy approaches for agreeing on global action can
be described as a top-down approach or bottom up approach. The former is based on the concept of
a “fair” distribution of the emissions reduction challenge according to agreed justice-based criteria.
The Kyoto Protocol is an example of this approach through establishment of a carbon reduction
commitment on Annex I countries only, i.e. those considered to be “principally responsible for the
current high levels of GHG emissions in the atmosphere as a result of more than 150 years of
industrial activity.” (UNFCCC, Kyoto Protocol, 2015)

In contrast, the bottom-up approach is devised from a “self-assessment” of a nations’ capacity and
willingness to contribute to the necessary emissions reduction. The recent Paris Agreement
achieved consensus based on the latter approach. Each nation pledges its contribution through the
Intended Nationally Determined Contributions (INDCs). The first draft pledges submitted to the
UNFCCC do not fall within a 2˚C target scenario (UNFCCC, 2015). These INDC’s were
established in view of the respective nations’ interpretation of equitable sharing of the carbon
problem. According to the UNFCCC, nations justified the ambition of their pledges on the basis of
“responsibility and capability, mitigation potential, cost of mitigation actions, the degree of
progression/stretching beyond the current level of effort, and the link to objectives and global
goals”. (UNFCCC, 2015).

3.2.2 Overall approach to selection of methodology

The allocation of a carbon budget to C40 Cities is a top-down approach. As indicated above, this
approach to assigning climate action on nation states has so far not been accepted by the global
community. Nonetheless, it serves as a useful means of evaluating the contribution and
effectiveness of regional climate action until data is available for a bottom-up approach.
This section presents proposed approaches for assigning carbon budgets for the C40 cities based on
a literature review. The literature review has not been exhaustive but rather focused on the most
widely promoted approaches. These were identified from three sources considered to provide a
comprehensive assessment of the current state of debate, namely:

Page 15 of 49
• A joint paper by The Centre for Climate Change Economics and Policy (CCCEP) and The
Grantham Research Institute on Climate Change and the Environment – (Averchenkova et al,
2014)
• WWF Australia commissioned report by Ecofys on Australia’s carbon budget based on global
effort sharing. (Ecofys, 2013) The proposed methodologies were used to advise South Africa as
well. (WWF, 2014)
• Concordia University (Gignac and Matthews, 2015)
Following this summary of available approaches, the initial proposed variations are reiterated in the
context of the published literature. Finally a scoring matrix is presented to aid the choice of a final
approach or set of approaches for allocating the global carbon budget to C40 Cities out to 2100.

3.2.3 Methodology for allocation of C40 Cities budget

A number of approaches have been discussed to inform climate change negotiations to align with a
conception of a “fair” distribution of the right to emit. According to Averchenkova, (Averchenkova
et al, 2014) the three principles that dominate debate on allocation of carbon budgets are:
1. “Equality, based on an understanding that human beings should have equal rights;
2. Responsibility for contributing to climate change, linked to the ‘polluter pays’ principle;
and
3. Capacity to contribute to solving the problem” (also described as capacity to pay).
Given the prominence of these principles within climate change negotiations, it is proposed that
they be adopted to frame the assessment of proposed approaches. It should be noted that the 84 C40
member cities included in this Study are split quite evenly between cities located in Annex I (38)
and non-Annex I (45) parties to the UNFCCC convention. As a result, the concerns around historic
emissions and a capacity to contribute are very relevant. Accordingly, in the following sub-section,
comment on the extent to which these principles feature in the approaches is provided.

3.2.4 Treatment of historic emissions

It is worth noting that it is considered preferable not to explicitly account for past historic emissions
in the allocation of future carbon budgets. The second principle listed above refers to accounting for
the uneven contribution both historically and currently to GHG emissions. Less developed nations
have raised the concern that their economic development will be compromised by climate
mitigation costs and some reparation for this should be made by industrialised nations that benefited
from uninhibited development. The polluter pays’ principle accords with this by assigning greater
responsibility for past atmospheric pollution to those polluters. It has been argued by Neumayer
(Neumayer, 2000) and others that the means to account for historic emissions in carbon budgets
would be to deduct a so-called carbon debt from future emissions.
On the other hand, it appears unlikely that accounting for historic emissions in this way will be
acceptable to the current major polluters. Instead the development of climate change negotiations
indicates that reparation will be enforced via greater financial assistance to less developed nations
with historically low emissions profiles. Indeed, article 9 of the Paris Agreement stipulates that
developed countries provide financial support to developing countries with respect to both
mitigation and adaptation (UNFCCC, Adoption of the Paris Agreement, 2015). Based on this, it is
recommended that past emissions are not accounted for through a deduction from the future carbon
allowance of any C40 cities in industrialised nations. Instead, as discussed in the next section, a

Page 16 of 49
compromise is available by adopting an approach for allocating future emissions which internalises
the polluter pays’ principle to a more moderate extent.

3.2.5 Literature Review

Within the literature reviewed, seven relevant distinct methodologies for allocating a carbon budget
out to the year 2100 for the C40 Cities have been identified, as follows:
1. Grandfathering: This approach is based on allocating future emissions budgets according to a
region’s current share of global emissions. The C40 cities current share of emissions is 6% and
therefore their carbon budget would be 6% of the global budget. Clearly, this is a very simplistic
means of assigning a carbon budget that does not respond to any of the ethical principles
described above. Indeed, developed cities with high current emissions benefit disproportionately
by maintaining high emissions irrespective of global population growth trends and development
needs.
2. Carbon Space: This approach calls for equitable sharing of the remaining carbon budget per
capita. In other words, it is strongly based on the equitable principle that everyone has a right to
emit the same emissions.
This approach does not explicitly take into account any responsibility for past emissions or
capacity to reduce emissions. However, by applying this equality principle, cities with
historically high emissions are forced to reduce these aggressively to match the global average.
In contrast, many developing nations with lower than average emissions are given the
opportunity to increase up to the global average.
3. Blended Sharing: This approach is a variation on the grandfathering and carbon spaced based
approaches whereby a factor w is introduced to moderate the impact of the two. The equation is
as follows:
𝐸𝐸𝑖𝑖 𝑃𝑃𝑖𝑖
𝐶𝐶𝑖𝑖 = (1 − 𝑤𝑤) + 𝑤𝑤
𝐸𝐸𝑤𝑤 𝑃𝑃𝑤𝑤
where Ci is the carbon budget for nation i, Ei and Pi are the emissions and population at a given
time of respectively of nation i and Ew and Pw are the emissions and population respectively of
the world. Raupach (Raupach et al., 2014) proposes a value of 0.5 for the blending factor w and
sets the emissions share for the present, and population share based on a future population.

This approach is effectively a compromise between an abrupt (and arguably unrealistic)

transition to equal emissions per capita and an abatement trajectory that recognises the reduction
challenge posed by current major emitters.
4. Contraction and Convergence (C&C): This approach is similar to the Carbon Space approach,
but in this a future year of convergence to equal global per capita emissions is specified. This
two-phase process developed by the Global Commons Institute (GCI, 2005) therefore results in
an “adaptation” period up to the year of convergence during which regional per capita emissions
can linearly increase up or decrease down to the global average. In contrast, the Carbon Space
approach implicitly makes the convergence year today (or agreed baseline year).
This method has been recognised as a viable means of sharing responsibility by nations and
regions of the world (GCI, 2005) and even informed the United Kingdom in setting its own
target (Gignac and Matthews, 2015). It is arguably less effective than the carbon spaced
approach in responding to principle 2 and 3 as a result of this delayed convergence. Within the

Page 17 of 49
 literature, the proposed convergence years are 2020, 2035, 2050 or between 2020 and 2050
(Gignac and Matthews, 2015) (Ecofys, 2013).
5. Common but Differentiated Convergence: This approach proposed by Hohne (Hohne et al,
2009) is a variation on the C&C approach which aims to differentiate the convergence year
according to a nation’s relation to the global average emissions per capita. Linear convergence
starts from the year a threshold percentage of global emissions per capita is reached.

This method is arguably more effective in aligning with the three principles discussed as
although it ultimately aims for “fair” sharing of the carbon reduction burden, action is delayed
for those that are less responsible for past emissions. Additionally, the industrialised nations
caught by the threshold are likely to be better able to pay for mitigation costs. There is an
element of “conditionality”, in that the entry of new countries in the convergence race depends
on the aggressiveness of the mitigation action by previous groups within the threshold
boundary. In the original proposal by Hohne, Annex I countries are made to start converging
from day one and the threshold criteria is applied on Non-Annex I countries. Crucially, this
method depends on defining an acceptable threshold rate. The literature includes thresholds in
the range of -10% to 300% above the global average emissions per capita. (Hohne et al, 2009)

6. Cost Proportional to City GRP per Capita: Under this approach, all cities are made to
contribute an equal percentage of their GRP towards the total mitigation cost. This satisfies the
principle of capability in that wealthier cities pay more towards reducing emissions. It is also
partially effective in responding to the polluter pays as in general (but not exclusively) wealth is
correlated to historically high emissions. The equality principle does not feature in this method.
Deducing the carbon budget based on this approach is more complex than other approaches as it
requires estimating the total cost of mitigation based on a realistic BAU trajectory. The latter is
inherently difficult to predict because of uncertainty on the development of exogenous drivers
such as technology, global population and city GRP and carbon intensity of the global economy.

7. Equal Marginal Cost: By this approach, responsibility for mitigation of climate change is
driven by cost effectiveness. This approach is not influenced by any ethical principles, instead
relying on regional marginal abatement costs for assigning where mitigation should occur and
by implication the nation or region responsible for implementing the recommended actions. In
other words, “in the absence of a global carbon market, or other mechanisms to transfer finance,
these costs would be borne by states and regions.” (Averchenkova et al, 2014)

3.2.6 Scoring Matrix

According to Averchenkoza (Avencherkova et al., 2014): "there is no single agreed method that
can be used to define what a country’s contribution to mitigating climate change should be." In the
absence of global consensus, a scoring matrix was devised that enabled a robust and appropriate,
though subjective, selection of approach for the purposes of this Study. The selected approach was
discussed and agreed between Arup and C40.

3.3 Final C40 Budgets

Based on the scoring described in Section 3.2.6, the Contraction and Convergence approach was
adopted to establish the overall carbon budget for C40 cities, and framing the development of
individual carbon trajectories.

Page 18 of 49
In order to apply the contraction & convergence year, a convergence year of 2030 was specified.
Previous studies have indicated that a convergence year of 2050 would not benefit many developing
countries because they are not actually given additional allowances to grow emissions per capita.
For this reason, a more ambitious convergence year of 2030 was selected, as suggested by (Hohne
et al, 2009).
The convergence value taken for 2030 was 3.2 tCO2e per capita, equal to half the current global
emissions per capita (6.4 tCO2e 13) and consistent with 2030 global emissions per capita under an
ambitious 2 degree pathway (as per IPCC AR5, 430-480ppm range 14).
The C40 budgets determined for scenarios consistent with below 1.5 degrees and 2 degrees were
determined as 22 GtCO2e and 67 GtCO2e between 2016 and 2100 respectively.
One important difference regarding the 1.5 degree scenario is that it was assumed that the only
possible means to achieve this target required negative emissions. In principle, there is no given
date for when global GHG emissions must turn negative, but later years will require far greater
negative emissions subsequently to keep total emissions within budgets. It was assumed, consistent
with research presented in analysis published in Nature Climate Change 15 that emissions should hit
zero by ~2050. Negative emissions technologies (such as bio-energy carbon capture and storage,
BECCS) are likely to be needed such that emissions of over 22 GtCO2e to 2050 in the 1.5 degree
scenario are reduced to the necessary 22 GtCO2e budget by 2100.
Table 8. Key results of C40 Carbon Budgets

Target Temperature Rise Below 1.5 degrees Below 2 degrees

with negative
emissions

C40 Cities Carbon Budget 2016

22.0 67.0
to 2100 (GtCO2e)
C40 Cities Emissions per Capita
3.2 3.2
in 2030 (tCO2e/capita)
C40 Cities Emissions per Capita
0.0 0.9
in 2050 (tCO2e/capita)

Side note: While the city-level budgets have been developed based on global budgets and trajectories that have been
statistically assessed to have minimum degrees of certainty (e.g. 66%), the methodology used for this work does not
allow us to ascribe a confidence estimate to city-level budgets. To do so would require further detailed modelling that is
outside the scope of this work. We therefore caution readers on the particular language used to discuss these budgets
and trajectories.

13
Based on Arup calculations from IPCC (2014) and UN (2015) data
14
IPCC, 2013, Summary for Policymakers WGI AR5
15
Rojelg et al, 2015, Energy system transformation for limiting end of century warming to below 1.5 degrees, Nature
Climate Change

Page 19 of 49
4 C40 Cities Trajectories
This section details the methodology for modelling each city’s business as usual (BAU) and climate
safe emissions trajectories.
There are a number of potential approaches for defining a BAU emissions scenario, such as taking
forward existing policies and / or commitments, or simply maintaining current emissions per capita
levels alongside population growth estimates.
This Study defines the BAU scenario as the case where C40 cities’ underlying trends (presented in
the next subsections) continue as expected, but, importantly, the carbon intensity of consumed
energy is not assumed to improve beyond existing levels. As such, the BAU scenario can be thought
of as a “no further climate action” scenario; that is, a worst case view. This definition allows for
consistent assessment across all cities in the C40 in this Study. It does not preclude an additional
definition of BAU being used in future.
The BAU trajectories were used an as input to the actions scenario model (see Section 5) in order to
calculate necessary reductions from each cities BAU trajectory.

4.1 Developing BAU Trajectories

The Kaya identity (Kaya, 1997) was used in order to develop BAU trajectories. This is the same
methodology used by IPCC to develop the BAU pathways presented in IPCC AR5. The Kaya
identity states that a geographical entity’s emissions are defined by its population, economic output,
energy efficiency of economy and carbon intensity of energy. This is best illustrated as an
equation:
Equation 1: Kaya Identity
𝐺𝐺𝐺𝐺𝐺𝐺 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝐶𝐶𝐶𝐶2
𝐶𝐶𝐶𝐶2 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 = 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 × � �×� �×( )
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝐺𝐺𝐺𝐺𝐺𝐺 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸
These four variables and their change over time are illustrated in Figure 3.

Page 20 of 49
Figure 3. Kaya identity variables (source: IPCC, 2013, AR5 WGIII Chapter 6 Figure 6.1)

For the Study, the first three variables of the equation: population, city GRP per capita and energy
per unit city GRP, were projected based on available forecasts from sources including the UN
(2015), Economist Intelligence Unit (2016) and IPCC (2014). These were based on moderate
forecasts such as medium fertility rate. For energy per unit city GRP, the median projection from
the IPCC Figure 6.1 (dark grey line shown in bottom left quadrant, Figure 3) was used. This
projection implies that even under a BAU scenario, technological progress will enable energy
efficiency improvements over time irrespective of the climate agenda.
The key variable in terms of framing a “no climate action” BAU scenario was the assumption that
the carbon intensity of energy will not improve significantly (as per the median case in the bottom
right quadrant of Figure 3). As such, within the equation, the carbon intensity of energy was treated
as constant between 2016 and 2100.

4.2 Climate safe emissions trajectories

The “climate safe” trajectories are city-specific emissions trajectories that in aggregate enable C40
cities to meet the target C40 carbon budget consistent with a given target temperature i.e. 1.5 or 2
degrees (see Section 3).

Page 21 of 49
The methodology for generating trajectories was based on categorising C40 cities into four
typologies using a number of criteria, including baseline emissions level and City GRP, as shown in
Table 9.
Each of the four trajectories was mapped to a logistic growth function, or “S-curve”, that governed
the overall shape of the trajectory.
Section 4.2.1 details the method for categorising cities by typology, the mathematical function used
to generate plausible emissions trajectories and an illustration of the results.

4.2.1 Categorisation of Cities into groups

The methodology used in the Study and model is based on the premise that the C40 cities’
emissions per capita projections can be broadly categorised into four typologies. This general
categorisation was developed using the following logic:
• It is unrealistic to assign all cities the same trajectory, there must be some variation based on
Capacity (this was identified as one of the three principles underpinning allocation of
contribution to climate action, as described in Section 3.2.3).
• For developmental purposes, developing cities should be allowed to continue to emit
emissions for longer than developed cities where living standards are generally higher. This
relates to the Responsibility principle (see Section 3.2.3) whereby nations with low historic
contributions to climate change are given the opportunity to benefit from unrestricted 16
economic development.
• Within developing cities, there should be some variation in the year their emissions are
expected to peak. This rule differentiates between developing cities with high baseline
emissions (early peaking required) and developing cities with very low per capita baseline
emissions (late peaking).
• There should be a differentiation between developed cities with high and low baseline
emissions. Initial model iterations established that it was necessary for developed cities with
high baseline emissions per capita to reduce their emissions at faster rate than other
developed cities in order for cumulative emissions between all cities to stay within the target
C40 carbon budget. It was therefore decided that developed cities would be categorised as
either having a ‘steep’ or ‘steady’ emissions reduction trajectory.
Side note: It is acknowledged that the most accurate approach would be to model a required city trajectory for every
city based on city-specific data and context. Given data and time constraints, it a generalised approach was instead
taken, with the functionality built in to the modelling environment for the input of more detailed city-specific
information as the C40 research and data gathering agenda progresses.

4.2.2 Methodology for categorising cities

Cities were categorised according to their emissions per capita and city GRP. Firstly, city GRP per
capita was considered, classifying cities into two categories: those with the capability to reduce per
capita emissions immediately and those where per capita emissions will most likely peak in the
future, following continued increased economic development.

16
In this context this refers to conventional economic development that is not constrained by the low carbon agenda.

Page 22 of 49
A threshold city GRP value of $15,000 was used to categorise into ‘peaking’ or ‘declining’ groups.
This value was used because it aligns with the UN development classification for countries moving
from low income to middle income. The UN officially uses gross national income (GNI) as a
measure to classify development status but this data is not available at a city scale (United Nations,
2012).
The city’s current level of emissions was then used to determine the rate of decrease required. Cities
with emissions higher than average C40 emissions (5.1 tCO2e/capita) and with a city GRP per
capita over $15,000 were assigned a steep decline. Those also with high city GRP but emissions
lower than the C40 average were assigned a steady decline trajectory. Cities with city GRP per
capita lower than $15,000 were either assigned a late peak or early peak trajectory, early peak for
cities with emissions higher than the C40 average and late peak for those with lower emissions.
Table 9. Examples of process for assigning city typology. Cities marked with * reported via CDP.

GHG/Capita City Assigned typology Example cities

GRP/capita

Toronto
High
Steep Decline Melbourne
High (>$15,000/capita)
New York City
(>5.1 tCO2e/capita)
Low Cape Town
Early Peak
(<$15,000/capita) Durban*
Stockholm
High
Steady Decline Seoul*
(>$15,000/capita)
Low London
(<5.1 tCO2e/capita) Quito
Low
Late Peak Caracas*
(<$15,000/capita)
Amman

4.2.3 Trajectory Curve Modelling

Each of the four trajectories shown in Table 9 relate to a mathematical function that governs the
overall shape of the trajectory. The process to develop the final trajectories was an iterative process
that considered the following factors:
1. Growth rate until emissions are expected to drop: it was found that allowing peaking
cities to increase their emissions on a per capita basis either meant they had to peak very
soon or declining cities had to reduce emissions at a very fast rate. This resulted partly
because developing countries still have a very fast population growth, meaning that a low
gradient per capita emissions growth trajectory still results in very large overall emissions
growth.
2. Peak year: similar to the growth rate, deciding on a peak year for each category was a
balance between allowing developing cities sufficient time before reductions are required
and not assigning unrealistic reduction rates to developed cities.
3. The maximum rate of emissions decrease: this was an important consideration because
the final trajectories needed to have a plausible year-on-year reduction rate. Although it is

Page 23 of 49
 very hard to predict what this value might be, the maximum limit used to develop these
trajectories was a 20% annual reduction.
Key differentiating factors between the typologies are shown in the table below.
Table 10: Peak years assigned for city typologies. *i.e. these cities must already have peaked.

Trajectory Peak Year Trend up to peak year Rate of emissions

decrease
Steady decline 2016* n/a Steady

Steep decline 2016* n/a Steep

Early Peak 2020 Linear increase Steady post peak year

Late Peak 2025 Linear increase Steady post peak year

Logistic growth functions

All city emissions trajectories were modelled with logistic growth functions (S-Curves). According
to a recent study on the use of logistic growth functions, historic emissions and projections are
‘adequately’ profiled using these functions (Suarez & Menendez, 2015). S-curves are typically used
to describe the uptake and growth of a technology before maturity slows the rate of growth.
S-curves are typically used to describe the uptake and growth of a technology before the growth rate
slows when the technology is mature. The mathematical function combines two kinds of
exponential growth: the first reflects exponential decay increasing at an exponential rate, and the
second reflects the decay decreasing at an exponential rate.

Figure 4 Example logistic (S-curve function) (UKERC, 2015)

In practice, cities that were assigned a decreasing emissions per capita typology were modelled
using only a logistic growth function. Cities assigned a typology with growing emissions before
decline were assigned a period of linear growth up until the peak year, after which logistic negative
growth was applied to meet the target emissions level (TEL).

Page 24 of 49
Target emissions level
The TEL was set at 0 tCO2 / capita, ensuring that all C40 cities reach the same, equitable emissions
per capita by 2050 for a 1.5 degree scenario.
The TELs for the 1.5 and 2 degree temperature targets were informed by modelled emissions
pathways. For instance, according to IPCC AR5, pathways consistent with below 2 degrees
warming show annual net CO2 emissions at or below zero between 2070 and 2100 and GHG
emissions near 0 GtCO2e per annum by 2100 17. Under 1.5 degree consistent scenarios, net-zero
CO2 emissions are suggested to occur as early as, or even earlier than 2050. 18 Example pathways
from these sources are shown in Figure 5.

Figure 5. Global emissions pathways across a range of scenarios (left) and their corresponding mean
temperature rises (right), presented by Rogelj et al (Rogelj, 2015). Note: Pink: medium 2 degree scenarios—
limiting warming during the twenty-first century to below 2 °C with 50–66% chance; orange: likely 2 degree
scenarios—limiting warming during the twenty-first century to below 2 degrees with below 66% chance;
blue: 1.5 degree scenarios—limiting warming in 2100 to below 1.5 °C with greater than 50% chance. Thin
black lines are scenarios included in the IPCC AR5 scenario database.

17
Non-CO2 emissions are often assumed to stay positive because of the difficulty in removing these altogether therefore
requiring negative CO2 emissions to offset them.
18
Rogelj et al, 2015, Energy system transformations for limiting end-of-century warming to below 1.5 degrees, Nature
climate Change

Page 25 of 49
Example city trajectories
Some example trajectories by typology are shown in Table 11.
Table 11. City emissions trajectories

City Example trajectory

typology
description
Steady
Per capita emissions

decline
(tCO2e/capita)

2015 2035 2055 2075 2095

Year
Steep
Per capita emissions

decline
(tCO2e/capita)

2015 2035 2055 2075 2095

Year
Early Peak
Per capita emissions
(tCO2e/capita)

2015 2035 2055 2075 2095

Year

Late Peak
Per capita emissions
(tCO2e/capita)

2015 2035 2055 2075 2095

Year

Page 26 of 49
4.2.4 Absolute Emissions Trajectories
Absolute emissions trajectories were obtained by multiplying the annual emissions per capita by
population growth in the corresponding year. As a result, the aggregate emissions were calculated
as the total area under the trajectory curve for each city as follows.
Equation 2
𝑛𝑛 2100

𝐸𝐸 = � � � 𝑇𝑇𝑛𝑛 � = 𝐶𝐶40 𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶 𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵

0 2015

Where, 𝐸𝐸 = 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝐶𝐶40 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸

𝑇𝑇 = 𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇
𝑛𝑛 = 𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁 𝑜𝑜𝑜𝑜 𝐶𝐶40 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐

The following figures show an example absolute emissions trajectory graph (Figure 6), and a
cumulative emissions graph (Figure 7).

Page 27 of 49
 3
Annual emissions (GtCO2e/year)

0
2015 2035 2055 2075 2095
-1

-2

Figure 6. All cities annual emissions

60
Cumulative emissions (GtCO2e)

50

40

30

20

10

0
2015 2035 2055 2075 2095

Figure 7. Cumulative emissions

Page 28 of 49
5 2CAP
This section presents the methodology, functionality, and assumptions behind the C40-Arup
Partnership Climate Action Pathways (2CAP) model. The purpose of this model is to develop an
actions pathway for each C40 city to meet the target climate safe trajectory developed out of the
trajectory analysis (see Section 4).

5.1 Functioning
The spreadsheet based model develops an action pathway for each C40 city to meet an assigned
target carbon emissions reduction trajectory. City by city, the model takes baseline city and action
data, as well as other key inputs such as grid carbon intensity, and dispatches action to try and meet
the climate safe trajectory on an annual basis. The high-level functioning is described in Figure 8,
with further detail in the subsequent sections.

Scenario Cond it ions Act ion Carb on

& Cont rols Im p act s

Baseline Dat aset

84 Cit ies
New d at a
A ct ions t aken
ent ry
Pow ers d at a Act ions Result ant
BA U t raject ory Disp at ch Em issions Collat e Cit ies
Map p ed Nat ional
energ y carb on Mod ule Traject ory
charact eris
t ics charact erist ics Cit y A
… Cit y B
Cit y C

Loop
Ext ernal t hroug h
Drivers all cit ies

Grap hical
Dashb oard s
and Out p ut s
A nalysis

Figure 8. 2CAP outline structure and functionality

5.1.1 Data inputs

The following list shows all the data required in order to develop action pathways.
• Modelled climate safe per capita emissions trajectories per city
• BAU per capita emissions trajectories per city to 2100
• City baseline action data (year and scale of action already being taken, with actions
classified as per C40’s Climate Action in Megacities action framework). Adaptation actions
are not included; these are assumed not to deliver additional carbon benefits.
• Individual Action criteria
o Impact (% of carbon reduction achievable against specific emissions sub-sectors)
o Roll-out time (years at Pilot & Planning and Significant scales before action is City-
wide and delivering maximum Impact)

Page 29 of 49
 o Programme allocation and Vital / Non-vital status
o Any other criteria including Replicability, cost, co-benefits that could be used to rank
actions. Note, cost and co-benefits information was not used to dispatch actions in
this iteration of the 2CAP model.

5.1.2 Programme allocation

In this Study, C40 developed a series of “Programmes”. 410 possible climate actions were mapped
into 62 Programmes covering five Sectors: Buildings, Energy, Transit, Waste, and Planning. Within
each Programme, Actions were assigned either Vital (crucial to the success of the Programme) or
Non-vital (non-essential but supporting) status. The same Action may feature in multiple
Programmes, but is only taken once, depending on the Action dispatch order described in the next
section.

5.1.3 Action dispatch

The first stage of the model creates a dispatch order for all Actions. This is done in three main
stages.
1. Weighting of Actions: Each action is given a weighting based on a number of different
criteria that the user can add to. The model currently uses three criteria; Power, Impact and
Replicability scores to give actions a weighting.
Carbon impact was allocated the highest weighting of any of these criteria. The distribution
of carbon impacts was plotted and the modal value was 3.5% reduction. To factor Power and
Replicability scores to ensure that carbon would generally be the highest component in the
weighting formula, the normalised values (i.e. between 0 and 1) for Replicability and Power
scores were multiplied by 3.5%.
𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑡𝑡𝑖𝑖𝑖𝑖𝑖𝑖 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅 𝑜𝑜𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡
𝑀𝑀𝑀𝑀𝑀𝑀 𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐𝑐 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑜𝑜𝑜𝑜 + × 3.5% + × 3.5% = 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤ℎ𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
𝑀𝑀𝑀𝑀𝑀𝑀 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑀𝑀𝑀𝑀𝑀𝑀 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑜𝑜𝑜𝑜𝑜𝑜 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡
𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎

2. Determine Programme Rank: Programmes are ranked by summing the weighted actions
within the programme. The programme with the largest total sum is the highest rank. Note:
Weighting the overall programme score by number of actions was also considered but it
was felt that if a programme had many more actions (and carbon reduction potential) it
should be ranked higher than one with fewer actions.

� 𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴𝐴 𝑊𝑊𝑊𝑊𝑊𝑊𝑊𝑊ℎ𝑡𝑡𝑡𝑡 = 𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂𝑂 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

3. Rank actions within the programme: Actions within each programme are then ranked by
two criteria.
• Vital actions are always ranked higher than Non-vital actions.
• Actions are then ranked by their weighting score.
The process described above is used to determine the overall dispatch order of actions that are taken
in order to meet required per capita trajectories.

Page 30 of 49
5.1.4 Actions Dispatch Methodology
Figure 9 shows the logic that the model follows in order to meet the required emissions trajectories.

Figure 9. Model dispatch logic

Starting at the top left of the figure, in a given year N the model runs through all actions in the
dataset. The first test checks whether the Action has already been taken (either pre-2016 or as a
result of dispatch by the model itself). If the Action has been taken and is ready to be scaled up
(based on its roll-out assumptions) it is scaled.
The model next focusses on the highest ranked Programme P, running through the weighted rank
order of Vital and Non-vital actions, checking whether if the next-highest rank Action will
contribute towards necessary target savings T, or exceed these.
There are two key aspects of the model that are described in more detail below.

Page 31 of 49
 1. Skipping actions: All actions in each programme have either been defined as Vital or
Secondary actions by C40 Initiative leads. 19 The ‘Bus Rapid Transit’ programme can be
used as an example of how actions are distinguished. The action ‘Increase routes, frequency
and night services’ would be a Vital action because it should be implemented as part of
every BRT programme. ‘Reduce fares’ and ‘Smart ticketing’ would be considered as
secondary because these are not crucial to the delivery of the programme.
After all the Vital Actions in a programme have been taken, the model decides whether to
move onto the next Programme or continue taking Non-vital Actions.

𝐼𝐼𝐼𝐼(�(𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉𝑉 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤ℎ𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑖𝑖𝑖𝑖 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝

> 𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤ℎ𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑜𝑜𝑜𝑜 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎) 𝑻𝑻𝑻𝑻𝑻𝑻𝑻𝑻 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑡𝑡𝑡𝑡 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝
𝑰𝑰𝑰𝑰 𝑵𝑵𝑵𝑵𝑵𝑵 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑎𝑎𝑎𝑎𝑎𝑎 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝
2. Checking future emissions reduction: The model decides whether to take an Action based
on the future emissions reduction potential of Actions already taken and the next ranked
Action already being considered. This is an important functionality because Pilot Actions
are assigned a lower Impact score than City-wide Actions. The model sets a threshold of +/-
10% the target emissions in the full roll-out year of the action being considered.
Figure 10 and Figure 11 exemplify how the model would decide whether or not to take an
action. The figures show that Action 1 and Action 2 have already been taken and the model
needs to decide whether action 3 should be taken or not. Figure 10 shows an example of
when the active action (Action 3) would be taken. The full roll out year of Action 3 is 2033;
in order to meet the required trajectory this action should be taken in 2020. Figure 11 shows
if Action 3 was taken in 2020 the actual trajectory would overshoot the trajectory in 2033. In
this case the Action would not be taken and the model would move to the next ranked
Action to investigate whether it was more appropriate when trying to meet the required
trajectory.

19
C40 currently has 17 networks organised under six initiative areas covering mitigation, adaptation and sustainability
topics. The initiatives engaged in this work included Urban Planning & Development, Solid Waste Management,
Energy, Transportation, and Finance and Economic Development

Page 32 of 49
Figure 10. Example of when additional action would be taken

Figure 11. Example of when active action would not be taken

During model development, it was decided that estimation of future action benefit would not take
into account the level of grid decarbonisation in the actions’ full rollout years. This was because
doing so would result in over-reliance on future grid decarbonisation and has the result of cities not

Page 33 of 49
“choosing” to take Actions, relying instead on energy decarbonisation to deliver the work
externally. This is representative of a dilemma faced at the city and national level around the world
today.

5.1.5 Modelling carbon impacts

The C40 Action database has a total of 410 Actions (not including Adaptation) grouped in 74
Action Areas and 12 sectors. The GPC emissions reporting methodology has five sectors and 22
sub-sectors. It should be noted that there are many instances where one Action will reduce
emissions across multiple sectors. The implications of this are discussed in Section 5.1.6.

After Actions were mapped to GPC categories, Actions were categorised based on their relationship
with other Actions and Sectors. Two broad categories of actions were defined:
Product Actions: These are Actions where the absolute emissions reduction potential is affected by
the introduction of another Action. Emissions reductions from these actions are multiplied by each
other to determine the overall emissions saving. Two examples of product actions are ‘Energy
efficiency retrofit measures: Installation of energy efficient lighting (Direct)’ and ‘On-site
renewable energy generation: Solar electricity’.
P action 1 = 10% reduction, P action 2 = 50% reduction

𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 = 1 − �(1 − 0.1) × (1 − 0.5)� = 55% 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

Where saving refers to the savings in subsector or subsectors where these Actions interact.
Within Product Actions, we have identified two types of Action that categorise how the emissions
reduction is achieved. Direct emissions reduction actions (e.g. energy efficiency) describe those
where the exact emissions reduction can be quantified from the action being taken. For example,
household energy efficiency directly reduces energy demand which reduces fuel consumption and
leads to lower emissions. Enabling Actions are those where the exact emissions reduction is harder
to quantify but it is very likely that through the introduction of these Actions that emissions are
reduced, albeit indirectly. Two examples of enabling Actions are ‘introduction of carbon reporting’
and ‘establishment of green enterprise zones’. Both of these Actions can plausibly lead to lead to
emissions reduction through behavioural change or encouraging other lower carbon activities but
evidence for the links is weak and the impacts are much more difficult to measure.
Sum Actions: The emissions reduction potential of a Sum Action is completely independent of
other Actions. The total emissions reduction potential of a city taking several Sum Actions is the
sum total of them. Two examples of Sum Actions are ‘Rooftop Farming’ and ‘Tree planting /
afforestation’.
S action 1 = 10% reduction, S action 2 = 5% reduction

𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 = 10% + 5% = 15% 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

Emissions saving of P1, P2, S1 & S2

𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 = 1 − �(1 − 0.55) × (1 − 0.85) � = 61.75% 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠

Page 34 of 49
5.1.6 Sector / sub-sector boundaries of influence
An important consideration for Product Actions is level of influence that a particular Action has on
others. For example, changing electricity carbon intensity will act as a multiplier for all actions that
are related to electricity use across all sectors. Improving the efficiency of heat and cooling systems
in commercial buildings will only act as a multiplier for commercial buildings. This detail
complicates the emissions reduction calculation because baseline emissions data is only available
for at the GPC category level (stationary energy, transportation) and some data at GPC sub-category
(residential, commercial). Although it may be possible to estimate the emissions split within a
subcategory (stationary energy split by residential / commercial etc.) it becomes more difficult
when considering actions within residential buildings that only impact the heat or electricity related
proportion of emissions. Figure 12 shows a proposal for how a particular action’s influence may be
treated. Figure 12 also shows a diagram of how the total emissions reduction of several actions
would be calculated. Key points to raise with this example include:
• Although ‘switching to biomass’ and ‘solar PV’ would not logically be multiplied together,
the granularity of baseline emissions data (for instance domestic direct emissions are only
available in aggregate) means that these cannot be separated and individually taken off a
baseline emissions per capita value. The same logic applies to ‘dedicated cycle routes’ and
‘switch buses to electric engines’.
• Carbon intensity of the grid impacts all the three sub-categories so it acts as a multiplier for
all of them. However as each sub-sector will have a different amount of electricity-related
emissions an additional factor is applied for this. See section 5.1.7 for detail on external
drivers.

Page 35 of 49
Figure 12. Example of action influence (see Section 5.1.7 for detail on electricity multiplier) Note: The
values used in this are for example – see section 5.2.2 for detail on the methodology to calculate these
numbers.

5.1.7 External drivers

There are two key external drivers that impact the carbon reduction in a particular sub-sector; grid
decarbonisation, and the proportion of electricity dependant emissions in a given sub-sector.

Electricity dependent emissions (electricity multiplier)

This value indicates the proportion of a sector’s emissions that are directly dependent on the carbon
factor of electricity supplied. For example, any buildings emissions coming from lighting, electric
heating and cooling and appliance use are dependent on the carbon intensity of the electricity
supplied. Likewise, electric vehicles and electrically powered trains would be included in the
proportion of transport emissions dependent on electricity.
In order to determine the current level of electrification for a particular sector, IEA energy balance
data at a country level was applied to each city. From this starting point, an electrification rate
modelled on an S-Curve was applied to each sector.

Page 36 of 49
 100
90
80
Starting Electrification %

70
60
50
40
30
20
10
0
2015 2025 2035 2045 2055 2065 2075 2085 2095

0 0.1 0.2 0.3 0.4

Figure 13. Example electrification rates from different starting points

Carbon factor of energy

The second external factor that impacts the overall emissions savings is the carbon intensity of
energy being supplied. The formula applied to calculate overall savings from changes to grid
intensity is:
[Assuming 10% carbon intensity reduction from baseline year and 20% electricity dependant
emissions]
𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠 = 0.2 × 0.1 = 2% 𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠𝑠
As with electrification rates, carbon intensity of energy projections are also modelled on an S-
Curve, as shown in Figure 14.

Page 37 of 49
 0.9
National Grid Carbon Intensity

0.8
0.7
0.6
kgCO2/kWh

0.5
0.4
0.3
0.2
0.1
0
2006 2011 2016 2021 2026 2031 2036 2041 2046

World US China

Figure 14. Example projected grid decarbonisation rates converted to absolute intensity

5.1.8 Total emissions saving

The total emissions saving formula including actions and external factors is:
Emissions in year X = BAU Emissions in year X × Carbon factor of energy × proportion of
electricity dependant emissions × Emissions reduction from actions

5.2 Data
Section 5.1.1 listed the key data requirements for the model. This section provides more detail on
each of the inputs and where the data was sourced from.

5.2.1 City Baseline Action data

City action data has been sourced from the Climate Action in Megacities database. This database
contains key city action information reported by C40 cities in either 2011, 2013, 2015 and 2016.
This database contains information on the following information that is used in the model:

Powers
The C40 Powers database contains information on city powers over 70 city ‘Assets’ and
‘Functions’. An example of an asset is ‘Traffic lights and signals’. An example Function is ‘Public
health’. There is a maximum power score of 12 and this is broken down into four main categories,
each with a score from 0-3, where 3 is the highest level of power:
1. Own / operate
2. Set / Enforce Policies and Regulation
3. Control Budget
4. Set Vision
In the 2CAP model, only the total power score is referred to.

Page 38 of 49
Replicability
In this study an Action’s Replicability is estimated as a measure of how regularly that particular
Action is reported in the CAM database within a particular region. High incidence of an Action
within a region suggest it may be more appropriate than Actions of low incidence in that particular
region. This is a normalised value that is used within the Action weighting criteria.

Scale of action
In addition to reporting whether an action is being taken, cities also provide detail on the scale of
the action reported. This is used in the model as a measure of how much additional savings can be
achieved from a particular action. If for example, a city is ‘piloting’ a bike sharing programme then
the model assumes that the city can roll out the programme over a wider area, thereby increasing the
emissions savings. If a city has already rolled out a BRT programme at a city wide scale then no
additional savings can be achieved from taking this action. This is a particular assumption that will
be tested in further research.

Cost
Cities also report the cost to deliver particular actions. Cost data was not used in the dispatch
methodology at this stage of research. The model’s functionality allows this to be readily integrated
in due course, once further data becomes available.

5.2.2 Actions Emissions Savings

In order to calculate emissions savings from actions as described in the previous section, each
action had to be assigned an emissions saving potential. This saving is either applied to one sub-
sector only or multiple sub-sectors. The emissions saving for a particular action was calculated
using the World Bank’s CURB modelling tool, provided by C40. The values are currently based on
single-city CURB model, but in future can be updated with similar assessments for other regionally-
representative cities or city-specific estimates from CURB. As emissions savings are recorded as
percentage reductions, this method is still relevant to cities with different absolute emissions
breakdowns and magnitudes.

Page 39 of 49
6 Key Data Inputs
This section presents a brief review of the principle data sources other than baseline emissions
(details in Section 2) and their treatment for this Study. The analysis for this Study was highly data-
reliant, and therefore faced challenges of data quality, incompleteness or availability. Weaknesses
in the dataset are reviewed under each data type.

6.1 Population
A multi-source approach was adopted to obtain both current and future city populations because at
present no single source has been identified which covers all cities within the C40 to the necessary
resolution. Population data was collected to align with the emissions reporting boundaries of cities.
In general, this coincided with administrative boundaries.

6.1.1 Baseline population data

The baseline city populations for the year 2015 were derived from the following four sources in
order of preference:
• GPC population data was available for 25 cities. These were provided for different years and
therefore adjustments using city specific growth rates sourced from UN data (see Section 6.1.2)
were used to establish these for the model baseline year of 2015.
• CDP population data was available for 50 cities.
• UN city proper data – population data was used for six cities. These were adjusted to the year
2015 using city-specific growth rates sourced from UN data (see Section 6.1.2). (United Nations
Statistics Division, 2016)
• Regional government statistics were used for the populations of Dubai and Dhaka.
It was also necessary to obtain data for the global population. This was sourced from the UN
Department of Economic and Social Affairs (2015).
Table 12. C40 Cities population data for the administrative boundary and metropolitan area boundary

City Population Source

according to
administrative
boundary
Accra 1,838,022 CDP
Addis Ababa 3,272,000 CDP
Amman 3,485,000 GPC
Amsterdam 822,272 CDP
Athens 656,144 CDP
Auckland 1,569,900 CDP
Austin 888,204 CDP
Bangalore 12,877,554 CDP
Bangkok 5,696,409 CDP

Page 40 of 49
City Population Source
according to
administrative
boundary
Barcelona 1,604,555 CDP
Basel 197,005 CDP
Beijing 20,855,748 UN
Berlin 3,456,459 CDP
Bogotá 8,215,714 GPC
Boston 648,584 GPC
Buenos Aires 3,086,897 GPC
Cairo 8,462,959 UN
Cape Town 4,048,387 GPC
Caracas 3,518,590 CDP
Changwon 1,086,852 CDP
Chicago 2,735,187 CDP
Copenhagen 587,320 GPC
Curitiba 1,949,425 CDP
Dar es Salaam 5,137,095 CDP
Delhi 12,732,961 UN
Dhaka North City 3,957,302 DNCC
Corporation Statistics
Dhaka South City 7,000,000 DSCC
Corporation
Dubai 2,446,675 Dubai
Statistics
Centre
Durban 3,555,868 CDP
Guangzhou 13,500,000 CDP
Hanoi 7,500,000 CDP
Heidelberg 143,533 CDP
Ho Chi Minh City 8,244,400 CDP
Hong Kong 7,305,700 CDP
Houston 2,306,973 CDP
Istanbul 14,657,434 CDP
Jaipur 3,405,343 CDP
Jakarta 10,075,300 CDP
Johannesburg 4,919,726 GPC
Karachi 25,828,287 CDP
Kolkata 14,502,447 CDP
Lagos 20,205,638 CDP

Page 41 of 49
City Population Source
according to
administrative
boundary
Lima 8,930,398 CDP
London 8,624,819 GPC
Los Angeles 3,938,458 CDP
Madrid 3,295,728 GPC
Melbourne 119,336 GPC
Mexico City 9,138,130 GPC
Milan 1,354,443 CDP
Moscow 12,252,703 CDP
Mumbai 13,645,883 UN
Nairobi 3,853,385 CDP
Nanjing 5,780,671 UN
New Orleans 389,617 CDP
New York City 8,542,025 GPC
Oslo 647,521 GPC
Paris 2,369,846 GPC
Philadelphia 1,570,468 GPC
Portland 789,136 GPC
Quezon City 3,015,007 CDP
Quito 1,713,125 GPC
Rio de Janeiro 6,552,682 GPC
Rome 2,868,347 CDP
Rotterdam 617,685 CDP
Salvador 2,944,966 CDP
San Francisco 830,153 GPC
Santiago de Chile 7,314,176 CDP
São Paulo 12,060,284 CDP
Seattle 656,484 GPC
Seoul 10,297,138 UN
Shanghai 22,470,340 CDP
Shenzhen 10,755,868 CDP
Singapore 5,500,000 CDP
Stockholm 928,862 GPC
Sydney 199,852 GPC
Tokyo 13,513,734 CDP
Toronto 2,836,391 GPC

Page 42 of 49
 City Population Source
according to
administrative
boundary
Tshwane 3,157,664 CDP
Vancouver 629,264 GPC
Venice 262,290 CDP
Warsaw 1,622,795 CDP
Washington, DC 668,335 GPC
Wuhan 10,607,700 CDP
Yokohama 3,719,589 CDP

6.1.2 Projection of Future Population

The population estimates for cities were projected from the reported year out to 2100 using both UN
(UN Department of Economic and Social Affairs, World Urbanization Prospects: The 2014
Revision, 2014) city figures and national growth rates (UN Department of Economic and Social
Affairs, 2015). The city growth rates extended out to 2035, hence the need to use national-level
growth rate projections out to 2100. The global population was projected using UN annual global
growth rates. Note that the UN national-level projections have several growth scenarios including
low, medium and high fertility variants. For the purposes of this study, the medium fertility variant
was used.

6.1.3 Identified uncertainty in data

One of the key risks to the multi-source approach undertaken is that these values may not be
consistent with the GPC data once completed. We have attempted to mitigate this issue by selecting
baseline population data that covers the administrative boundary as it was found that on the whole
the GPC population data coincided with the administrative boundary. Indeed, for those cities for
which CDP and GPC data were available, the values for population were very similar.
Another consideration is that the national growth rates applied to cities for post-2035 may differ
considerably from cities’ actual population growth rate. Urban populations are expected to grow
faster than their national averages because of increasing urbanization rates and cities’ attractive
force for international migrants.

6.2 City GRP data

As no single existing source was identified that covered all cities, a multi-source approach was
adopted to establish both baseline City GRP, and to project these into the future. This approach is
set out below.

6.2.1 City GRP baseline data

The 2015 baseline city City GRPs were derived from the following two sources in order of
preference:

Page 43 of 49
• The Brookings Institution data - this data provided for year 2014 was used for 70 cities. These
were adjusted to the year 2015 using city specific growth rates. (The Brookings Institution,
2015)
• The Economist Intelligence Unit (EIU) national City GRP – these were converted into national
City GRP per capita figures using the UN national population estimates. (Economist
Intelligence Unit, 2016)
Note that the Brookings Institution data was used in preference over GPC data to minimise the
number of data sources. This was important because the baseline CITY GRP values were used for
matching baseline emissions between primary and secondary cities and so consistency of
methodology was critical.

6.2.2 City GRP projection

The city GRPs were projected using annual city GRP per capita growth rates for the corresponding
countries. A multi-source approach was used as the EIU projections extended only as far as 2050.
For certain developing countries, projections were limited to 2020. In these cases, future GDP
growth was based on OECD (OECD, 2012) data and used as follows:
• A non-OECD annual growth rate of 4.7% for the period 2020 to 2030 used for nations without
EIU data from 2020 to 2030.
• A global city GRP annual growth rate for the period 2030 to 2060 of 1.9% was adopted for the
country city GRP growth rate from 2050 to 2100. It was assumed that the 1.9% growth rate
would be maintained from 2060 onwards.
The global city GRP was projected to 2100 for all cities using the growth rates stated above.

6.2.3 Identified uncertainty in data

The sources of uncertainty in the city GRP data are similar to those for population projections,
namely:
• Future trends are highly uncertain, in particular when extending as far out as 2100
• City GRP growth rates are likely to deviate from national growth rates
The future city GRP data is used to create values for carbon intensity (CO2e per unit city GRP) used
in modelling BAU trajectories (see Section 4.1). As such, the accuracy of the city GRP forecast is
not as important as the general trend it illustrates.

Page 44 of 49
7 Bibliography

Allen et al. (2009). Warming caused by cumulative carbon emissions towards the trillionth tonne.
Averchenkova et al, a. S. (2014). Taming the beasts of ‘burden-sharing’: an analysis of equitable
mitigation actions and approaches to 2030 mitigation pledges. London: CCCEP and The
Granthem Research Institute on Climate Change adn the Environment.
Baden-Wurttemberg Statistisches Landesamt. (2014). http://www.statistik.baden-
wuerttemberg.de/BevoelkGebiet/Bevoelkerung/99025010.tab?R=GS221000.
C. Le Quere. (2015). Global Carbon Budget 2015. Earth System Science Data, 439-396.
Carbon Tracker Initiative & LSE GRI. (2013). Unburnable carbon 2013: Wasted capital and
stranded assets. London.
Clarke et al. (2014). Assessing Transformation Pathways. In: Climate Change 2014: Mitigation of
Climate Change. Contribution of Working Group III to the fifth Assessment Report for the
Intergovernmental Panel on Climate Change.
Dubai Statistics Centre. (2015). Retrieved from https://www.dsc.gov.ae/en-us
Ecofys. (2013). Australia's carbon budeget based on global effort sharing. Cologne: ECOFYS
Germany GmbH.
Economist Intelligence Unit. (2016). Economist Intelligence Unit Country Data. Economist
Intelligence Unit.
EIA. (2016, June 01). Energy Intensity. Retrieved from Energy Information Administration:
https://www.eia.gov/cfapps/ipdbproject/iedindex3.cfm?tid=92&pid=46&aid=2&cid=regions
&syid=1997&eyid=2011&unit=BTUPUSDM
Enerdata. (2016, June 01). Energy Intensity of GDP. Retrieved from Global Energy Statistical
Yearbook 2015: https://yearbook.enerdata.net/energy-intensity-GDP-by-region.html
GCI. (2005). CGI Briefing: "Contraction & Convergence". CGI.
Gignac and Matthews, R. a. (2015). Allocating a 2°C cumulative carbon budget to countries. IOP
Environmental Research Letters.
Hansen et al. (2013). Assessing 'dangerous climate change': Required reduction of carbon emissions
to protect young people, future generations and nature. PLOS ONE.
Heinonen, J. (2012). The Impacts of Urban Structure and the Related Consumption Patterns on the
Carbon Emissions of na Averaeg Consumer.
Hohne et al, N. (2009). Common but differentiated convergence (CDC): a new conceptual approach
to long-term climate policy. Earthscan Climate Policy, 181-199.
IPCC. (2013). Assessing transformation pathways.
IPCC. (2014). Climate Change Synthesis Report 2014.
Kaya, Y. &. (1997). Environment, energy, and economy : strategies for sustainability. Tokyo:
United Natrions Univ. Press.
Matthews et al. (2009). The proportionality of global warming to cumulative carbon emissions .
Nature.
Meinshaussen et al. (2009). Greenhouse-gas emission targets for limiting global warming to 2C .
Nature.
Neumayer, E. (2000). In defence of historical accountability for greenhouse gas emissions. Elsevier
Ecological Economics, 185-192.
OECD. (2012). Looking to 2060: Long-term global growth prospects.
Raupach et al. (2014). Sharing a quota on cumulative carbon emissions. Nature Climate Change.
Research Initiatives Bangladesh. (n.d.). Report of the Study of the Proactive Disclosure of
Information. Dhaka.

Page 45 of 49
Rogelj, J. e. (2015). Energy system transformations for limiting end-of-century warming below
1.5C. Nature Climate Change, http://dx.doi.org/10.1038/nclimate2572.
Sattherthwaite, D. (2009). The Implications of Population Growth and Urbanization for Climate
Change. In D. Sattherwaite, Population Dynamics and Climate Change (pp. 45-63).
London: International Institute for Environment and Development.
Suarez, R., & Menendez, A. (2015). Growing green? Forecasting CO2 emissions with
Environmental Kuznets Curves and Logistic Growth Functions. Environmental Science and
Policy.
The Brookings Institution. (2015). Global Metro Monitor 2014: An Uncertain Recovery. The
Brookings Institution.
UKERC. (2015). Innovation timelines from invention to maturity.
UN Department of Economic and Social Affairs. (2014). World Urbanization Prospects: The 2014
Revision.
UN Department of Economic and Social Affairs. (2015). World Population Prospects: The 2015
Revision.
UNFCCC. (2015). Adoption of the Paris Agreement.
UNFCCC. (2015). Kyoto Protocol. Retrieved from http://unfccc.int/kyoto_protocol/items/2830.php
UNFCCC. (2015). Synthesis report on the aggregate effect of the intended nationally determined
contributions. New York: UNFCCC.
United Nations. (2012). Statistical Annex: Country Classification. United Nations.
United Nations Statistics Division. (2016). The Demographic Yearbook. United Nations.
WWF. (2014). Understanding Carbon Budgets. WWF Briefing.

Page 46 of 49
8 Glossary

Action Climate Actions are defined as the measures and initiatives cities
take to reduce the severity of climate change (mitigation), or their
exposure to the effects of climate change (adaptation).

Carbon budget For this Study, the maximum permissible cumulative anthropogenic
GHG emissions between 1870 (or 2016) and 2100.

City-wide (action) An action that is in place at a transformative scale across the whole
city.

CO2 equivalent (or CO2e) As stated in the GPC, CO2e is the universal unit of measurement to
indicate global warming potential (GWP) of each GHG, expressed
in terms of the GWP of one unit of carbon dioxide (CO2). It is used
to evaluate the climate impact of releasing (or avoiding releasing)
different greenhouse gases on a common basis.

Baseline emissions Defined for this Study as the emissions of cities considered in this
study in year 2015.

Baseline year The starting year for future projections; defined for this Study as
2015.

Direct (action) Actions that have direct, quantifiable emissions reductions.

Enabling (action) Actions that may logically lead to emissions reductions in some
sector(s), but where this is more indirect and difficult to accurately
quantify.

GHG Greenhouse Gas.

GHG Protocol A global protocol on the measurement, management and reporting

of GHG emissions.

Global warming As per the GPC, GWP is a factor describing the radiative forcing
potential (GWP) impact (degree of harm to the atmosphere) of one unit of a given
GHG relative to one unit of CO2.

GRP Gross Regional Product: similar to Gross Domestic Product but

referring to a region within a nation, in the case of this Study a city.

Impact (carbon) Carbon emissions reduction potential of a given action in a given

emissions subsector or subsectors.

Power (of a city) The Power of a city mayor over one of 70 Assets (such as traffic
lights and signals) and Functions (such as public health). Rated

Page 47 of 49
 across dimensions of ownership & operation, setting and enforcing
policies, control of budget, or the ability to set the vision for that
Asset or Function. This is data collected by C40. More information
can be found in Powering Climate Action, a report by C40 and
Arup.

Planning & Pilot (action) An action that is still at the stage where its full implementation is
being planned, and / or it is being piloted or under construction.

Product Action Actions where the absolute emissions reduction potential in a sector
is affected by the introduction or presence of another Action
impacting that sector; Impacts are multiplied not summed.

Replicability An inferred measure of an Action’s Replicability, calculated as the

frequency of a particular Action in a region, which is assumed to
imply its appropriateness in other cities within the region.

Scope I emissions GHG emissions from sources located within the city boundary

Scope II emissions GHG emissions occurring as a consequence of the use of grid-

supplied electricity, heat, steam and/or cooling within the city
boundary.

Scope III emissions All other GHG emissions that occur outside the city boundary as a
result of activities taking place within the city boundary.

Significant (action) An Action that is in place at a significant scale across the city.

Sum Action Actions where the absolute emissions reduction potential in a sector
is not affected by the introduction or presence of another Action
impacting that sector; Impacts are summed not multiplied.

Page 48 of 49
C40
North West Entrance, City-Gate House
39-45 Finsbury Square, Level 7
London, EC2A 1PX
United Kingdom.

[email protected]

ARUP
13 Fitzroy Street,
London, W1T 4BQ
United Kingdom.

[email protected]