Electric Two- Wheelers in China: Analysis of
Environmental, Safety, and Mobility Impacts
Christopher Robin Cherry
DISSERTATION SERIES
UCB- ITS- DS- 2007- 1
Spring 2007
ISSN 0192 4109
1
Electric Two- Wheelers in China:
Analysis of Environmental, Safety, and Mobility Impacts
by
Christopher Robin Cherry
B. S. ( University of Arizona) 2000
M. S. ( University of Arizona) 2003
A dissertation submitted in partial satisfaction of the
requirements for the degree of
Doctor of Philosophy
in
Engineering- Civil and Environmental Engineering
in the
Graduate Division
of the
University of California, Berkeley
Committee in charge:
Professor Adib Kanafani, Co- Chair
Professor Robert Cervero, Co- Chair
Professor Arpad Horvath
Professor Maximilian Auffhammer
Spring 2007
1
Electric Two- Wheelers in China:
Analysis of Environmental, Safety, and Mobility Impacts
Copyright 2007
by
Christopher Robin Cherry
1
Abstract
Electric Two- Wheelers in China:
Analysis of Environmental, Safety, and Mobility Impacts
by
Christopher Robin Cherry
Doctor of Philosophy in
Engineering- Civil and Environmental Engineering
University of California, Berkeley
Professor Adib Kanafani, Co- Chair
Professor Robert Cervero, Co- Chair
Chinese cities have a long legacy of bicycle use due to relatively low incomes,
dense urban development, and short trip lengths. Because of tremendous economic
growth, increased motorization, and spatial expansion of cities, trips are becoming longer
and more difficult to make by bicycle. As a result, electric powered two- wheelers have
risen in popularity over the past five years, with sales exceeding 16 million in 2006.
Recently, policy makers have enacted bans on electric two- wheeler use, citing a poor
safety record, a large contribution to congestion, and poor environmental performance.
This study quantifies many of the safety and environmental impacts of electric two-wheelers
and balances the negative externalities by quantifying benefits to users in terms
of increased mobility and access to opportunities.
Touted by some as environmentally friendly vehicles, electric two- wheelers are
capable of traveling 40- 50 kilometers on a single charge and emit zero tailpipe emissions.
However, they do have significant environmental impacts because they use lead acid
batteries that are recharged with electricity that is predominantly generated from coal
2
power plants, but they also have significant mobility benefits that are seldom considered.
This research investigates the tremendous growth of electric two- wheelers in China and
compares their environmental and safety impacts to those of alternative modes of
transportation; such as traditional bicycles, public transportation, or personal cars. This
research also analyzes the benefits of electric two- wheelers in terms of increased mobility
and accessibility to opportunities due to their increased speed and range.
Electric two- wheelers tend to be more energy efficient and produce less air
pollution per kilometer traveled than many other modes. Also, to the extent that they
displace car trips, they improve the safety of the transportation system in Chinese cities.
Electric two- wheelers provide much higher mobility and access to opportunities than all
other low cost modes.
The impacts of electric two- wheelers on the transportation system are dependent
upon local characteristics of the transportation system. Considering alternative
transportation modes in two case studies ( Shanghai and Kunming), banning electric two-wheelers
will result in higher net energy use and greenhouse gas emissions. Moreover,
the public health impacts from traditional air pollutants and road safety would likely be
worse in a situation where electric two- wheelers are banned. The mobility and
accessibility to the city will also deteriorate significantly for users of electric two-wheelers.
However, allowing electric two- wheelers in a city results in significant
increases in lead pollution over the lifecycle, compared to alternative modes. This
research shows that while electric two- wheelers do have some problems that need to be
addressed ( namely excessive lead acid battery pollution); they provide large benefits and
can be a successful strategy toward a sustainable transportation future.
i
TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION .................................................................................... 1
1.1 Research Objective: .................................................................................................. 7
1.2 Dissertation Organization ......................................................................................... 8
CHAPTER 2: RESEARCH FRAMEWORK, METHODOLOGY, AND DATA .... 10
2.1 Research Approach ................................................................................................. 11
2.1.1 Energy Use During Production ........................................................................ 13
2.1.2 Energy Use During Vehicle Operation ............................................................ 15
2.1.3 Air Emissions from Electricity Generation ...................................................... 15
2.1.4 Converting emissions into intake ..................................................................... 18
2.1.5 Converting intake into health effects ............................................................... 19
2.1.6 Lead Pollution from Battery Use ..................................................................... 21
2.1.7 Safety ............................................................................................................... 23
2.1.8 Mobility and Accessibility Changes ................................................................ 24
2.2 Case Studies ............................................................................................................ 24
2.2.1 Kunming .......................................................................................................... 25
2.2.2 Shanghai ........................................................................................................... 27
2.3 Data ......................................................................................................................... 29
CHAPTER 3: USE CHARACTERISTICS AND MODE CHOICE BEHAVIOR ... 33
3.1 Survey Methodology ............................................................................................... 33
3.1.1 Location ........................................................................................................... 34
3.1.2 Sampling .......................................................................................................... 35
3.2 Survey Results ........................................................................................................ 35
3.2.1 Descriptive Statistics ........................................................................................ 35
3.2.2 Travel Behavior ............................................................................................... 38
3.2.3 User Attitudes .................................................................................................. 43
3.3 Factors that Influence Two- Wheel Vehicle Choice ................................................ 45
3.3.1 Choice Between Bicycle and Electric Bike ..................................................... 46
3.3.2 Choice of Alternative Mode ............................................................................. 50
3.4 Conclusion and Policy Inferences ........................................................................... 52
CHAPTER 4: ENVIRONMENTAL IMPACTS OF ELECTRIC BIKE USE ......... 55
4.1 Energy Use and Emissions of Electric Bike Life Cycle ......................................... 56
4.1.1 Production Processes ....................................................................................... 56
ii
4.1.2 End- of- Life ...................................................................................................... 59
4.1.3 Lead Acid Batteries.......................................................................................... 60
4.1.4 Use Phase ......................................................................................................... 64
4.1.5 Total Environmental Impacts of Electric Bike Lifecycle ................................ 68
4.2 Environmental Impacts of Alternative Modes ........................................................ 70
4.2.1 Energy Use and Emissions of a Bicycle .......................................................... 71
4.2.1.1 Production Phase ....................................................................................... 71
4.2.1.2 Use Phase .................................................................................................. 72
4.2.2 Energy Use and Emissions of a Bus ................................................................ 73
4.2.2.1 Production Phase ....................................................................................... 74
4.2.2.2 Lead Pollution from Bus Batteries ............................................................ 75
4.2.2.3 Use Phase .................................................................................................. 77
4.3 Modal Comparison of Environmental Impacts ....................................................... 79
4.4 Exposure of Populations to Air Pollution ............................................................... 81
4.4.1 Intake Fraction of Power Plant Emissions ....................................................... 83
4.4.1.1 Intake of Pollutants Emitted Power Plants – Kunming ............................ 85
4.4.1.2 Intake of Pollutants Emitted from Power Plants – Shanghai .................... 87
4.4.2 Intake Fraction of Vehicle Tailpipe Emissions ................................................ 88
4.4.2.1 Tailpipe Intake Fraction in Kunming and Shanghai ................................. 90
4.4.3 Normalized Emissions Considering Exposure ................................................. 92
4.5 Distribution of Environmental Impacts .................................................................. 94
4.6 Direction of Public Health Impacts ......................................................................... 95
4.6.1 Public Health Impacts of Air Pollution ............................................................ 95
4.6.2 Public Health Impacts of Lead Pollution ......................................................... 95
4.7 Policy Discussion, Conclusion and Future Work ................................................... 96
CHAPTER 5: SAFETY IMPACTS OF ELECTRIC BIKES IN CHINA ................. 99
5.1 Designing Electric Bikes for Safety ...................................................................... 100
5.2 Unsafe versus Vulnerable ..................................................................................... 104
5.3 User Perceptions ................................................................................................... 108
5.4 Policy Implications ............................................................................................... 108
CHAPTER 6: MOBILITY AND ACCESSIBILITY IMPROVEMENTS OF
ELECTRIC BIKE USERS ...................................................................................... 110
6.1 Mobility versus Accessibility ............................................................................... 110
6.2 Measuring Mobility Increases ............................................................................... 113
iii
6.3 Job Accessibility Gains: The Case of Kunming ................................................... 120
CHAPTER 7: IMPACTS OF ELECTRIC BIKE PROHIBITION ......................... 128
7.1 Kunming ............................................................................................................... 128
7.1.1 Vehicle Population and Travel Behavior ....................................................... 128
7.1.2 Environmental Impacts of Mode Shift in Kunming ...................................... 131
7.1.3 Exposure Effects of Change in Pollution Levels in Kunming ....................... 136
7.1.4 Transportation Network Safety in Kunming .................................................. 137
7.1.5 Mobility and Accessibility Advantages of Electric Bikes in Kunming ......... 140
7.1.6 To Ban or Not to Ban- Kunming? ................................................................... 142
7.2 Shanghai ................................................................................................................ 143
7.2.1 Vehicle Population and Travel Behavior ....................................................... 144
7.2.2 Environmental Impacts of Mode Shift in Shanghai ....................................... 145
7.2.3 Exposure Effects of Change in Pollution Levels in Shanghai ....................... 147
7.2.4 Transportation Network Safety in Shanghai .................................................. 148
7.2.5 Mobility and Accessibility Advantages of Electric Bikes in Shanghai ......... 149
7.2.6 To Ban or Not to Ban? ................................................................................... 151
7.3 Conclusion ............................................................................................................ 152
CHAPTER 8: CONCLUSION AND POLICY RECOMMENDATIONS .............. 155
8.1 Economics ............................................................................................................. 156
8.2 Environment .......................................................................................................... 157
8.2.1 Local Impacts ................................................................................................. 158
8.2.2 Non- Local Impacts ......................................................................................... 158
8.2.3 Global Impacts ............................................................................................... 159
8.2.4 Policy Response ............................................................................................. 160
8.3 Safety .................................................................................................................... 162
8.4 Accessibility .......................................................................................................... 163
8.5 Cost Effectiveness of Travel ................................................................................. 164
8.6 Shortcomings of Study and areas of future work .................................................. 167
8.6.1 Data Availability and Reliability ................................................................... 167
8.6.2 Other Externalities ......................................................................................... 169
8.7 Closing Remarks ................................................................................................... 172
REFERENCES .............................................................................................................. 173
APPENDIX A. 1: SURVEY INSTRUMENT ( ENGLISH) ........................................ 180
APPENDIX A. 2: SURVEY INSTRUMENT ( CHINESE) ........................................ 182
iv
APPENDIX B. 1: EAST CHINA POWER NETWORK INTAKE FRACTION
ESTIMATION PARAMETERS ............................................................................. 184
APPENDIX B. 2: INTAKE FRACTION OF POLLUTANTS FROM EAST CHINA
POWER NETWORK .............................................................................................. 185
v
LIST OF FIGURES
Figure 1.1: Bicycle Style and Scooter Style Electric Bikes ................................................ 4
Figure 1.2: Production of E- bikes and Cars For Domestic Market in China ...................... 4
Figure 2.1: Framework of Analysis of Cost Effectiveness of Electric Bikes ................... 12
Figure 2.2: Emission Rates from Chinese Power Plants................................................... 17
Figure 2.3: Map of Kunming ............................................................................................ 26
Figure 2.4: Mode splits for all trips in Kunming ( 2003) and Shanghai ( 2006) ................ 27
Figure 2.5: Map of Shanghai ............................................................................................ 28
Figure 3.1: Trip Purpose by Mode and City ..................................................................... 41
Figure 3.2: What Mode Would You Take Otherwise? ..................................................... 42
Figure 3.3: What Mode Did You Previously Use? ........................................................... 43
Figure 3.4: Why Did You Choose This Mode? ................................................................ 44
Figure 3.5: Modeling Hierarchy for Discrete Choice Models .......................................... 45
Figure 4.1: Emission Rates from Chinese Power Plants................................................... 66
Figure 4.2: Pollution of BSEB Over Lifecycle ................................................................. 69
Figure 4.3: Pollution of SSEB over Lifecycle .................................................................. 70
Figure 4.4: Pollution of Traditional Bicycle over Lifecycle ............................................. 73
Figure 4.5: Pollution of Bus over Lifecycle...................................................................... 79
Figure 5.1: Histogram of Moving Speeds ( No Stops) - Kunming .................................. 102
Figure 5.2: Histogram of Moving Speeds ( No Stops) - Shanghai .................................. 102
Figure 5.3: Histogram of Moving Speeds ( No Stops) - Kunming
20 km/ hr Limit on Electric Bikes .............................................................................. 103
Figure 5.4: Histogram of Moving Speeds ( No Stops) - Shanghai
20 km/ hr Limit on Electric Bikes .............................................................................. 103
Figure 6.1: Example Speed Data Collected in Southeast Kunming ............................... 115
Figure 6.2: Histogram of Measured Speed Data in Shanghai ......................................... 117
Figure 6.3: Histogram of Measured Speed Data in Kunming ........................................ 117
Figure 6.4: Speed Advantage of Various Alternative Modes ......................................... 119
Figure 6.5: Residential and Job Distribution in Kunming .............................................. 121
Figure 6.6: Mode Specific Jobs Access Within 20 minutes of Kunming City Center ... 123
vi
Figure 7.1: Electric Bike Ownership in Kunming .......................................................... 129
Figure 7.2: Best Stated Alternative Mode and Displaced PKT in Kunming .................. 131
Figure 7.3: Best Stated Alternative Mode and Displaced PKT in Shanghai .................. 145
vii
LIST OF TABLES
Table 2.1: Data, Units, and Sources .................................................................................. 30
Table 3.1: Demographics of Two- Wheel Vehicles Users in Kunming and Shanghai ...... 36
Table 3.2: Household Vehicle Ownership Levels ............................................................ 38
Table 3.3: Travel Characteristics, Surveyed weekday ( April- May 2006) ........................ 39
Table 3.4: Logit Model for Predicting Probability of Electric Bike Mode Choice .......... 48
Table 3.5: Logit Model for Predicting Probability of Current Electric Bike Users
Switching to Bus, Bicycle, or Walk if Electric Bikes Became Unavailable ............... 51
Table 4.1: Material Inventory, Emissions and Energy Use- Electric Bike ........................ 58
Table 4.2: Electric Bike Lead Emissions .......................................................................... 63
Table 4.3: Scooter Style Electric Bike Emissions ( g/ km) ................................................. 67
Table 4.4: Material Inventory, Emissions and Energy Use- Bicycle ................................. 72
Table 4.5: Material Inventory, Emissions and Energy Use- Bus ....................................... 74
Table 4.6: Electric Bike Lead Emissions .......................................................................... 76
Table 4.7: Emission Factors of Urban Buses ( g/ km) ........................................................ 78
Table 4.8: Lifecycle Environmental Impact Per Passenger Kilometer Traveleda ............ 80
Table 4.9: Intake fraction average and range in China ( Zhou, Levy et al. 2006) ............. 83
Table 4.10: Regression Coefficients of various pollutants ( Zhou, Levy et al. 2006) ...... 84
Table 4.11: Intake Fraction Calculations of Emissions from Power Plants in Yunnan
Provincial Power Grid ................................................................................................. 86
Table 4.12: Intake Fraction Calculations of Emissions from Power Plants in East China
Power Network ........................................................................................................... 88
Table 5.1: Safety Data from Zhejiang and Jiangsu Provinces ( 2004) ............................ 107
Table 6.1: Hardware and Software Configuration Used For Speed Collection .............. 114
Table 6.2: Job Accessibility Between Electric Bike and Alternatives ............................ 124
Table 7.1: Environmental Impacts in Kunming ( g/ pax/ km unless otherwise noted) i ..... 133
Table 7.2: Total Emission Changes Resulting From Mode Shift- Kunming i ................. 134
Table 7.3: Total iF Normalized Net Emission Changes Resulting From Mode Shift-
Kunming ................................................................................................................... 136
Table 7.4: Net Safety Impacts of Electric Bike Ban in Kunming ................................... 139
viii
Table 7.5: Time Savings From Using Electric Bike in Kunming ................................... 141
Table 7.6: Total Emission Changes Resulting From Mode Shift- Shanghai .................. 146
Table 7.7: Total iF Normalized Emission Changes Resulting From Mode Shift in
Shanghai .................................................................................................................... 147
Table 7.8: Net Safety Impacts of Electric Bike Ban in Shanghai ................................... 148
Table 7.9: Time Savings From Using Electric Bike in Shanghai ................................... 150
Table 8.1: Direction and Magnitude of Electric Bike Advantage or Disadvantage ....... 156
Table 8.2: Cost Effectiveness of Travel by Competing Modes in China ....................... 166
ix
ACKNOWLEDGEMENTS
This dissertation could not have been written without the support and assistance
of a host of family, friends and professional colleagues. First and foremost, I have to
thank my wife and children. This work is the culmination of several years of sacrifice on
their parts and I cannot thank them enough for enduring small apartments with no
backyards and cheap restaurants. Also thanks for coming to China and getting out of your
comfort zone, especially while pregnant! Julie is a true hero and Avah and Kylie are
super children and you all inspire me. I also want to thank my parents, sister,
grandparents and in- laws for the supporting our family and the decisions we’ve made. I
love you all.
I would like to thank my acting committee – Adib Kanafani, Robert Cervero,
Arpad Horvath, and Max Auffhammer for supporting and advising this work. I would
also like to thank Betty Deakin, Marty Wachs, Samer Madanat, Mike Cassidy and Carlos
Daganzo for providing valuable advice, direction, and of course funding to be successful
during my studies at Berkeley. You are all an inspiration on how to teach, advise students,
and conduct research. I could not have done it without you all. Thanks also to the support
staff in ITS, DCRP, UCTC, and Civil Engineering.
I have many students to thank and I am sure I would forget to mention some, so
thank you all for supporting my work and sanity while I was here. Particularly, I would
like to thank Jennifer Day, Wendy Tao, Allie Thomas, Juju Wang, Mike Duncan and
David Weinzimmer for helping with ideas and other support. I also would like to thank
Julian Marshall for making sure my work was good and providing a lot of input and
guidance on my analysis. Thanks also to Jonathan Weinert for interacting and
x
collaborating on this work. Thanks also to the White Stripes for providing some rhythm
to the dissertation writing.
I am especially grateful for everyone who made my work in China successful.
Professor Pan Haixiao and Yao Shengyong from Tongji University were integral to my
success. Thanks also to Jeffrey Zhen from Shanghai University of Finance and Economic
and all of the students who supported my work there. Professor Xiong Jian, Sun Jingyi,
Liao Ying, and Guo Fengxiang were essential to my success in Kunming - thank you. My
work in Beijing was supported by Professor Lu Huapu and Professor Yang Xinmiao from
Tsinghua University, as well as students Qiu Ying and Ma Chaktan. Thank you to all of
the students who supported me and my work during my short stays at these institutions.
I’ve enjoyed the collaboration. Thanks also to Sarath Guttikunda, Lee Schipper, Peter
Danielsson, and Renting Xu for helping me understand the energy, environment, and
transport sectors in China.
There are a number of individuals from the electric bike industry that educated me
quickly on the state of the industry, markets, and regulations. Several electric bike makers
granted interviews and I appreciate that. I am particularly grateful in Ni Jie’s intellectual
investment in this research. I also appreciate early support provided by Ed Benjamin.
This work was funded mostly by the Volvo Foundation through UC Berkeley’s
Center for Future Urban Transport. Some supplemental funding was provided by the
National Science Foundation and the University of California Pacific Rim Research
Program.
1
CHAPTER 1: INTRODUCTION
Chinese cities have been developing economically at a phenomenal rate for the
past decade. With this has come an increase in urbanization and motorization, which has
increased congestion and reduced urban air quality. China’s transition to a more market
based economy has effectively unbundled housing and employment, causing increased
trip lengths. Additionally, growing employment and labor markets are prompting more
multiple worker households and trip destinations throughout the urban area. Increases in
income have led to increased consumption and thus increased demand of local and
regional shopping destinations. As a result, residents in Chinese cities are spending more
time and a higher portion of their income on transportation than ever before ( Cherry
2005).
Chinese cities are investing heavily in advanced public transportation systems in
order to improve the efficiency in their transportation system ( Chang 2005). Many cities
have coupled investment in public transportation with restrictions on bicycle and
motorcycle use, presumably to improve the safety and efficiency of the transportation
system and reduce conflicts between modes. While public transportation systems are the
most efficient mode of transportation by many metrics, they do not provide door- to- door
flexibility or the short travel times of personal transportation modes ( such as bicycles)
that Chinese residents are accustomed to. Because of these inherent limitations of public
transit systems, bicycles are still widely used, despite annexed infrastructure and
increased regulation.
As a result of these trends, industry has been developing modes that can provide
low cost personal transportation that is fast, flexible and energy efficient. Particularly,
2
electric bicycles and electric scooters have gained popularity and their use has become
widespread in many Chinese cities. Electric bikes come in a range of styles and
performance specifications, but the primary technology is the same. The vast majority of
them utilize lead acid batteries to provide energy to a hub motor that is usually on the rear
wheel. Most electric bikes fall into two categories: scooter style electric bikes ( SSEBs) or
bicycle style electric bikes ( BSEBs) ( Figure 1.1). SSEBs appear much like gas scooters
complete with headlights, turn signals and horns; with large battery packs under the
footboard. BSEBs resemble bicycles, with functioning pedals and usually smaller
batteries and a lower power motor. Electric bikes are capable of speeds exceeding 20- 30
km/ hour and weigh between 40 and 60 kilograms.
Electric bikes are recharged by plugging into standard wall outlets. This is a great
advantage because there is no need for dedicated refueling/ recharging infrastructure.
Most electric bikes have removable batteries and chargers so that they can be transported
indoors and recharged during the day or night. With their increased popularity, many
apartments or workplaces are retrofitting bicycle parking areas to accommodate electric
bikes by providing electrical outlets. Batteries require 6- 8 hours to charge. Charging
electric bikes at night can increase the efficiency of the electric power generation network.
By recharging batteries overnight, excess electricity production capacity can be used to
charge batteries that will be used during the day, when electricity demand is at its peak.
This has the effect of smoothing the demand peak and could potentially require little or
no electricity generation capacity improvements.
Electric bikes are very cheap and efficient to operate. The purchase price is 1600-
2400 RMB or US$ 200- 300. Considering an average SSEB with a 350W motor and a
3
48V 14Ah battery, the energy requirement is 1.3kWh/ 100km. Electricity rates in most of
China are around 0.6 RMB/ kWh, so the cost of operating an electric bike is
0.78RMB/ 100km or about $ 0.10/ 100km. The total average cost is about 0.10- 0.12
RMB/ km ( Jamerson and Benjamin 2004), far cheaper than any other motorized mode; for
instance user costs to ride the bus is around 0.5 RMB/ km. Moreover, this cost is rarely
realized by electric bike users. They often do not pay for the recharging because they
recharge at a centralized parking lot. If they recharge the battery in their apartment, the
cost is bundled into their electric utility bill and they do not see how much is from battery
recharging. This results in difficulty regulating electric bike use through the cost of fuel
( electricity). The main expense is the purchase of batteries, which is over half of the in-use
cost ( Jamerson and Benjamin 2004).
In 2005, over 10 million electric bikes were sold in China, which is about 3 times
the amount of cars sold ( Figure 1.2) ( Jamerson and Benjamin 2004; National Bureau of
Statistics 2005). Guo ( 2000) chronicles the emergence, development, and regulation of
the electric motorcycle over the past 30 years, indicating that China is currently
experiencing its third peak in electric motorcycle use. The author cites reasons for the
current success such as better batteries, more government support, and more reliability.
Recent laws passed by China’s central government classify electric bikes as bicycles
from an operational and regulatory perspective. Driver licenses and helmets are not
required and they are allowed to operate in the bicycle lane ( China Central Government
2004). Manufacturers are required to adhere to technical standards developed by the
central government that stipulate a maximum weight of 45 kg and a maximum speed of
4
20 km/ hour ( China Central Government 1999). This standard precludes most SSEB’s
from development, but the standard is poorly enforced ( Weinert, Ma et al. 2006).
0
2,000,000
4,000,000
6,000,000
8,000,000
10,000,000
12,000,000
Production ( unit)
E‐ bikes
All Autos
Personal Cars
Figure 1.1: Bicycle Style and Scooter Style Electric Bikes
( image source: www. forever- bikes. com)
Figure 1.2: Production of E- bikes and Cars For Domestic Market in China
5
This mode of transportation has certain advantages over others, but also presents
challenges to transportation planners and policy makers. As cities expand, many origins
and destinations grow farther apart and become less accessible by bicycle. Public
transportation in many cities is underdeveloped and inefficient. Buses often operate in
mixed flow lanes and the average operating speed has decreased with increases in
congestion. The result is that electric bicycles, operated in the bicycle lane, increase
personal mobility in terms of reduced travel time and thus accessibility to goods, services,
jobs, etc. Bicycle lanes are seldom congested and offer high levels of capacity to bicycle
and electric motorcycle users. Using an electric motorcycle could be seen as a superior
mode to a car in terms of travel time and cost savings, potentially resulting in lower car
ownership. Additionally, electric motorcycles have zero local emissions and low noise
levels.
Although electric vehicles produce no local emissions, they do require electrical
energy, which in the case of China is almost exclusively generated by coal- fired power
plants ( National Bureau of Statistics 2005). Electric motorcycles require 0.9- 1.3 kWh of
electric energy per 100 km. Electric bikes will have different emission rates based on
regional location and energy mix. Emissions from one point source ( powerplant) are
easier to manage and regulate than emissions from multiple sources ( tail- pipes) and likely
have lower public health effects because of their rural location.
Currently, most electric motorcycles are powered by lead- acid batteries and each
battery has a lifespan of approximately 300 charges or 10,000 km. Generally, a battery
lasts one to two years. Battery disposal and recycling is a serious environmental
consideration, as improper disposal can lead to contamination of soil or groundwater and
6
inefficient production and recycling processes can lead to high emissions of airborne lead
pollution. The recycling process and the negative effects of lead- acid batteries in the
developing world are well documented ( Lave, Hendrickson et al. 1995; Yeh, Chiou et al.
1996; Suplido 2000; Cortes- Maramba, Panganiban et al. 2003; Mao, Lu et al. 2006).
China currently does not have a well regulated and institutionalized disposal and
recycling program for lead- acid batteries. This is a serious consideration when
considering an appropriate policy for Chinese cities.
The growth of this mode has prompted local and national policy makers to
question the impact of electric bikes on the transportation system and pursue policies to
regulate them. Taiwan promoted and even subsidized electric bike use in the 1990’ s to
provide a clean alternative to gas powered scooters ( Taiwan EPA 1998; Chiu and Tzeng
1999). Despite this subsidy, electric bikes competed directly with gas scooters and the
performance characteristics were not competitive enough to induce a large market shift.
Although they were promoted in Taiwan, several cities in mainland China, notably
Beijing and Fuzhou, have attempted to ban electric bikes altogether, citing lead pollution
and safety issues ( Beijing Traffic Development Research Center 2002; Weinert, Ma et al.
2006). These policies are being implemented with little information about who is using
this mode and what impact it has on the transportation system. Taiwan attempted to shift
from gas scooters to electric bikes, but little is known about who is riding electric bikes in
China and from which modes they are shifting.
7
1.1 Research Objective:
Policy makers are making decisions based on perceived environmental and social
costs, but little research has been done that carefully quantifies these costs and also looks
at benefits that electric bikes provide to the urban transportation system. Little is known
about the life cycle energy use and environmental impacts, safety impacts or accessibility
effects experienced by electric bike users. Policy makers may cite environmental
concerns regarding electric bikes, but they often do not consider the environmental
impacts of alternative modes, if electric bikes became unavailable.
The research question addressed in this dissertation is: Compared to the predominant
alternative modes, bus and traditional bicycle-- under what conditions do electric bikes
provide a greater relative benefit in terms of mobility and accessibility improvements
than relative costs in terms of energy use, environmental impacts and safety?
Since many of these impacts are local in nature, two case studies are carried out in
Kunming and Shanghai, two cities with very distinct differences, but similar electric bike
use. Several research activities are carried out that address the primary research question.
1) Investigate electric bike user demographics, vehicle use characteristics, and
factors that influence mode choice through a user survey. Calibrate a choice
model that identifies factors that influence current mode choice.
2) Conduct a life cycle assessment ( LCA) of electric bikes and compare energy use
and emissions outcomes to those of alternative modes, namely bicycles and buses.
8
3) Identify safety impacts of electric bikes and develop mode shift scenarios that
influence the overall safety of the transportation system.
4) Quantify mobility and accessibility changes in terms of origin to destination travel
time differences and jobs access, compared to bus and bicycle use.
These activities represent the primary costs ( emissions, energy use, and safety)
and benefits ( accessibility) of electric bikes in China. The metrics of these analyses
reflect the difficulty in developing environmental, economic and equitable sustainable
transportation policy. These metrics are not comparable in the sense that one could make
direct comparisons which would result in an objective, deterministic policy solution.
There will likely be trade- offs that will differ, depending on the goals of the policy maker.
For instance, accessibility will be measured in terms of jobs access increase, as a
proportion of increase compared to alternative modes. Emissions will be measured in
terms of total pollutants or public health effects. While it is difficult to compare these
metrics, they must both be considered in the decision making process, and the policy
recommendation will differ, depending on the goals of the policy maker. This research
quantifies these costs and benefits so that the decision making process is more informed
and transparent.
1.2 Dissertation Organization
This dissertation is composed of nine chapters, including the introduction. The
second chapter of this dissertation will build a research framework and discuss the
methodology and data collection techniques for each of the research activities. It will
9
review relevant literature on each of the topics and give introductions to the case study
cities. The third chapter will discuss the user demographics and use characteristics of
electric bike use in Kunming and Shanghai and discuss the development of a discrete
choice model that predicts mode choice based on individual and mode specific variables.
The fourth chapter identifies major contributors to the environmental impact of electric
bikes and alternative modes ( buses and bicycles). An LCA is conducted that accounts for
the environmental impact of production, use and end- of- life phases of the life cycle. The
life cycle impacts are compared across all three modes. Public health impacts are
calculated from electric bike and bus emissions from the use phase of the life cycle.
Chapter five discusses the safety data of electric bikes. Crash and fatality rates are
compared across modes in different cities and regions and scenarios are developed in the
event of modal shift. Chapter six synthesizes collaborative research conducted by
Chinese partners on the effect of electric bikes on congestion. The seventh chapter
discusses the results of mobility studies in Kunming and Shanghai and extends the results
of those studies to accessibility gains. The eighth chapter summarizes the results of the
analysis in the context of the Kunming and Shanghai case studies. The ninth chapter
extends this analysis to a national context and develops a framework from which to
analyze electric bike impacts in any city. It discusses some shortcomings of this study
and future research directions are presented that improve on this methodology and extend
this work to other modes and cities.
10
CHAPTER 2: RESEARCH FRAMEWORK, METHODOLOGY, AND DATA
Different cities or regions have different electricity use patterns, travel mode
patterns, demographics and transportation regulations that influence the use of electric
bikes. If a majority of electric bike users would otherwise be using bicycles, then the net
environmental impact is negative. If electric bikes replace motorized vehicle use, then
there is possibly a positive environmental benefit. One must weigh the environmental
impacts against the economic benefits, which are realized through increased mobility.
This research will develop a framework within which to analyze relative impacts of
electric bikes in any Chinese city. Different Chinese cities have different data reporting
practices and thus different ways to approach this analysis. Generally, this framework
involves identifying the following:
Environmental Impacts:
• Number of electric bikes in a Chinese city and the approximate daily vehicle
kilometers traveled ( vkt)
• Safety impacts of a shift from alternative modes to electric bikes.
• Emissions generated by the production of an average electric bike, a bus, and a
bicycle.
• Energy mix and subsequent emission factors for fossil fuel power plants serving a
city where electric bikes are operated
• Human exposure of airborne emissions
• The amount of lead emitted into the environment during the production, recycling,
and disposal processes of batteries
11
• Proportion of electric bike users that would otherwise use bicycles or transit if
electric bikes were prohibited
Mobility and Accessibility Impacts:
• Difference in operating speed and thus travel time from origins and destinations
between competing modes
• Change in accessibility to jobs, goods or services
2.1 Research Approach
The framework for analysis is outlined in Figure 2.1. One of the difficulties
associated with conducting a full analysis of the costs and benefits of a new mode is to
bound the research to include the most significant costs and benefits. This research will
not include a full cost and benefit accounting, but will consider what are seen as the
greatest impacts; those associated with vehicle life cycle emissions and energy use, safety
impacts, and mobility and accessibility changes. These environmental and safety impacts
are commonly cited by electric bike opponents, but opponents rarely acknowledge
mobility and accessibility gains.
The research approach first involves identifying case cities, Shanghai and
Kunming. These cities have given demand characteristics of electric bikes and alternative
modes. They have city specific operating speeds for electric bikes, buses and bicycles.
Each city also has somewhat static electricity mix. Given these inputs, primary costs and
benefits can be calculated. Environmental production costs will be incurred in the
provinces where electric bikes and their components are manufactured and will be
constant across all cities where electric bikes are used. The costs imposed by operating
12
electric vehicles will be distributed among the population affected by power plant
emissions serving that particular city. Air emissions can be converted to population
exposure and thus mortality and morbidity changes associated with electric bike use. The
safety and mobility impacts will be experienced by users of electric bikes. Mobility
changes can be expressed as changes in accessibility, given transportation network and
land use data for a given city. In short, environmental externalities will be external to the
electric bike user, and are social costs, while safety and mobility changes are internal to
the user. The following sections will discuss the research approach of each component of
the costs and benefits to be evaluated. A more thorough methodology section will be
given in each of the respective chapters.
Environmental
Emissions
• Production, Use
Lead Emissions
Safety Impacts
Public
Health
Impacts
Mobility Changes
Benefits
Quantify
Benefits
in Terms of
Accessibility
Changes
City Level Data
E- Bike Use Characteristics
Electricity Mix
Mode Displace
Average Speed by Mode
Energy Use
• Production, Use
Externalities
Figure 2.1: Framework of Analysis of Cost Effectiveness of Electric Bikes
13
2.1.1 Energy Use During Production
To identify the effects of the development of the electric bike market and industry
in China and the effect of regulation in different cities, the entire life- cycle of the electric
bike must be investigated. This includes identifying the production processes for each
unit and identifying resource, energy use, and environmental impacts during production.
The production function will likely vary between factories, but since most factories are
located in Zhejiang or Jiangsu Province ( near Shanghai) their access to resources should
be similar.
When conducting a environmental life cycle analysis of a vehicle, five
components of the vehicle’s life should be considered ( Sullivan, Williams et al. 1998).
1) Raw materials acquisition and processing
2) Part and Subassembly Manufacturing
3) Vehicle Assembly
4) Vehicle Use and Operation
5) Disposal
Sullivan and Williams et al. ( 1998) found that the vast majority of personal car’s
energy use ( 84%) is from operation. Raw material production and manufacturing account
for 14% of energy use. In terms of air emissions, vehicle operation accounts for 87% of
CO2, 94% of CO, and 90% of NOx. The material production and manufacturing
components account for 65% of particulate emissions and 34% SO2 emissions. This is
primarily because the production and manufacturing components use the most electricity
and thus coal emissions of the life cycle phases. Vehicle disposal uses very little energy,
but is the greatest contributor to solid waste of all other stages of the vehicle’s life. The
authors do not consider infrastructure, building construction, transportation costs of
14
distribution or secondary inputs into production processes. It is generally accepted that
these inputs are very small in relation to the overall costs. This approach is used to
determine the environmental impact of electric bike use in Chinese cities.
Electric bikes, buses and bicycles all have different fuel technologies. Buses are
most closely related to the personal car example, with most of the environmental impacts
occurring during the use phase. Alternatively, the production of a traditional bicycle
accounts for nearly all of its environmental costs, so when comparing these two modes,
the environmental costs of a bicycle should be very carefully measured.
Since electric bikes use electricity from a power plant, which more efficiently
generates and transfers primary energy into movement than burning gasoline or diesel
internal combustion engines, electric bikes have lower use phase environmental impacts.
A greater proportion of an electric bike’s environmental impact is imposed during the
manufacturing phase.
An electric bike, like most vehicles, is made from hundreds of parts and
components. Comprehensive component lists that include the weight and material of
various components are supplied by industrial partners. The major parts/ component
manufacturers and processes that likely use energy and produce emissions are: batteries,
motors, tires, steel frame welding and forging, and plastic manufacturing. While this is
not an exhaustive list of the components, these produce the most pollution. These
components are manufactured and shipped to a final assembly plant where the electric
bike is finally produced. Aggregate environmental data on these processes are readily
available in statistical yearbooks. These costs can be divided over the life of the vehicle
to identify energy use per kilometer. Once primary, first- order production costs are
15
calculated, sensitivity analysis can be conducted to evaluate the potential effects of the
second- order costs that were omitted or estimated, such as distribution or infrastructure.
2.1.2 Energy Use During Vehicle Operation
During electric vehicles’ operation, they emit zero local air pollution, but they do
use electricity ( about 0.9- 1.3 kWh per 100 km). For example, consider an average SSEB
with a 350W motor, a 48V/ 14Ah battery and 50km range.
Current= Power/ Voltage= 350W/ 48V≈ 7.3 A
Drain Time= 14Ah/ 7.3A= 1.9 hours
Energy= Power* Time= 350W* 1.9 h= 672Wh
Energy/ Range= 672Wh/ 50km= 13Wh/ km
This energy use varies by different motor/ battery combinations. The weighted average
electricity use per kilometer can be calculated based on fleet composition in a city or
nationwide.
2.1.3 Air Emissions from Electricity Generation
In China, About 75% of electricity is generated by coal- fired power plants. Much
of China’s power is generated locally by small, inefficient power plants, with a limited
regional or national power grid and distribution network ( Zhu, Zheng et al. 2005). There
are currently 15 power grids that serve different parts of China. Cities throughout China
are served by different proportions of power sources ( coal, natural gas, hydro, wind and
16
nuclear). For instance, the construction of the Three Gorges Dam provides a large
amount of clean hydro power, although its capacity still comprises a small proportion of
China’s overall capacity. In general the power generation capacity in northern China is
almost exclusively coal powered because of abundant coal supply. The power generation
capacity in southern China has much higher hydro- electric capacity. The wind, solar and
nuclear power generation capacity in China is negligible.
The following graphs illustrate the emissions of primary pollutants from average
existing coal- fired power plants, new coal- fired power plants, and gas turbine power
plants.
emiss
produ
lifesp
bike,
kilom
Once emi
sions rates p
uction, the e
pan of the v
the amoun
meter.
a= Avera
F
issions rates
per kilomete
electricity us
vehicle. To c
t of electric
age Chinese Co
Figure 2.2: E
( E
s have been
er. In order
sed per elec
calculate the
city used pe
oal Boiler b= Ne
Emission Ra
Energy Fou
17
determined
to calculate
ctric bike pro
e emissions
er kilometer
ew Coal Boiler
ates from C
ndation Ch
per MWh, t
e the emissio
oduced mus
rate for the
can be con
c= Gas Combin
hinese Pow
hina 2005)
they can be
ons per kilo
st be divided
operation o
nverted to e
ned‐ Cycle Turb
er Plants
converted t
ometer due t
d by the tota
of the electri
emissions pe
bine
to
to
al
ic
er
18
Different emissions rates can be calculated using various energy mix
combinations of hydro generation, coal generation, or gas generation. Additionally, there
is a spectrum of technologies that must be considered on a case by case basis to
accurately estimate the emissions per kilometer of electric bike use ( Larson, Wu et al.
2003; Wang, Mauzerall et al. 2005).
The production emissions can be calculated using the average power plant
emissions and energy mix in the East China power network sector, where most of the
production facilities are located ( Anhui, Zhejiang and Jiangsu Provinces and Shanghai
Municipality). The emissions for operating the electric bike would be calculated using the
average power plant emissions and energy mix of the sector in which the city is located.
2.1.4 Converting emissions into intake
Electric bike policy is highly dependent on the energy profile of a city or region
and different scenarios of future electricity generation. In addition, the exposure of people
to pollutants depends on proximity of power plants to population centers and
meteorological conditions. Cities with urban power plants are more likely to expose
higher populations to airborne toxics, while rural power plants will not have the same
negative health impacts.
One of the techniques that has recently been developed to measure the exposure
of people to pollutants is the intake fraction ( Bennett, McKone et al. 2002; Marshall and
Nazaroff 2004). The intake fraction is defined as the proportion of the pollutants that are
emitted that are actually inhaled and can be calculated as follows:
19
Q
P C BR
iF
N
i
i i ( )
1 Σ=
× ×
=
Where P is population of zone i, C is pollutant concentration of zone i, BR is the average
breathing rate or the volume of air inhaled per unit time of the population and Q is the
total mass of pollutant emitted into the environment. The intake fraction is unit- less and
can be a powerful tool to identify health impacts due to incremental changes in emissions
such as pollution controls on power plants or added emissions due to electric bicycle use.
It is also helpful to compare public health impacts of various alternative technologies
without calculating public health end- points. That is, a technology that results in twice
the intake fraction of an alternative will have twice the public health impacts.
2.1.5 Converting intake into health effects
Intake can be extended to public health impacts. Epidemiologists ( Xu, Gao et al.
1994; Xu, Li et al. 1995; Wong, Ma et al. 2001; Pope III 2002; Brajer and Mead 2003;
Chen, Hong et al. 2004) have developed dose response functions for different pollutants;
primarily particulates, sulfur dioxide, and nitrogen dioxide, which are the most hazardous
to human health. These researchers report relative risk factors, which are defined as a
percent increase in mortality or morbidity per unit increase in pollutant. For instance,
there is a 0.7% increase in mortality per μg/ m3 of PM2.5 concentration increase, and 0.084%
increase per μg/ m3 of PM10. Similar numbers have been reported for morbidity, which
result in increased hospital and doctor visits. Ultimately one would like to know the
number of mortalities or sicknesses that are incurred as a result of increased pollution.
20
This is calculated using a concentration response function that was developed by the US
EPA ( 1997).
ΔC = C( ebΔP − 1)
Where ΔC is the change in mortality or morbidity, C is the baseline mortality or
morbidity rate, b is the response coefficient and ΔP is the change in pollution
concentration level. Baseline mortality and morbidity rates are known for various cities
or China in general. The change in pollutant concentration is modeled using pollutant
transport models or back- calculated using the intake fraction methodology. The response
coefficient, b, is related to the relative risk factor as follows ( Brajer and Mead 2004).
b= ln( relative risk)/( change in pollutant)
Using this methodology, the total health effects of an increase in emissions and
thus an increase in concentration of a pollutant or set of pollutants can be quantified in
terms of additional lives lost as a direct result of increased power plant emissions.
Alternatively, if the net change of air emissions for different pollutants are
determined and the relative public health impacts between each of those pollutants can be
identified, then the direction of the public health impact can be estimated. For instance, if
policy is enacted that doubles the amount of SO2 and halves the amount of NOX emitted
from the transportation sector ( controlling for exposure), then the public health impact of
such a policy would be positive. Since NOX has more severe public health impacts than
21
SO2, halving its emissions would produce more public health benefit than the negative
impact associated with doubling SO2 emissions ( Health Effects Institute 2004).
2.1.6 Lead Pollution from Battery Use
Perhaps the most significant environmental disadvantage electric bikes have is the
use of lead acid batteries. According the Electric Bikes Worldwide Report ( Jamerson and
Benjamin 2004), 95% of all electric bicycles and scooters in China are powered by lead
acid batteries. Chinese electric bikes use 24 or 36V, 7- 12Ah batteries. The batteries
weigh between 9 and 15 kilograms. Batteries typically have a lifespan of 300 charges, or
about 10,000 kilometers. Electric bicycle manufacturers typically cite the lifespan of a
battery is about 2 years, depending on use, maintenance, and recharging protocol. Recent
developments have made Nickel Hydride and Lithium batteries more feasible for future
uses, but the prospects for use of these batteries is uncertain and these types of batteries
also have negative environmental implications.
Recent research has shown that equivalent of 70- 100% of lead content of a battery
is emitted into the environment in China through the mining, manufacturing, recycling
and disposal processes ( Mao, Lu et al. 2006). It is unclear what portion of this is emitted
into the air, ground, or water. However, lead is classified as a hazardous material that
decays slowly, so all emissions could eventually have public health effects.
The Center for Disease Control ( CDC 1991) and the World Health Organization
( WHO 1995) have identified the lead poisoning blood concentration threshold for
children ( 10μg/ dL), men ( 40 μg/ dL) and women ( 30μg/ dL). If a person’s lead
concentration exceeds this value, they are in danger of experiencing symptoms of lead
22
poisoning. Lead poisoning manifests in many ways that are difficult to quantify. Children
experience long term developmental disorders, low IQ, and physical growth impairments
( Shen 2001). There have been a couple of studies in the context of battery recycling and
manufacturing plants in Asia and their health effects on workers and people nearby.
Suplido and Ong ( 2000) found that workers at battery recycling shops and children of
workers in the Philippines had much higher lead levels than control groups ( 330% higher
for adults and 400% higher for children). The blood lead concentration is five times the
WHO guidelines for children. Cortes- Maramba et al. ( 2003) found that populations living
within five kilometers of a large battery recycling plant (> 14,000 batteries per year)
experienced significantly higher blood lead concentrations than control groups living
outside of the five kilometer radius ( 20% higher for adults and 30% higher for children).
In terms of quantifying the health impacts, they identified that adults living within five
kilometers of the plant had a 23.1% history of hospitalization, compared to 4.2% for the
control. Likewise, 37.5% of the affected children have a history of hospitalization,
compared to 11.8% of the control group.
The US EPA ( 1997) identified the public health impact of removing lead from
fuel. The report identifies several quantifiable public health impacts of lead pollution,
including mortality, lower IQ, hypertension and stroke. These effects are a function of the
blood lead levels, not air concentration as in the previous section. Given absence of blood
lead levels, approximations can be made based on studies made by Cortes- Maramba et al.
( 2003) or Suplido and Ong ( 2000).
For the near term, lead acid batteries will be the primary source of power for
electric bikes and policy must be developed that encourages more environmentally
23
benign batteries and establishes disposal and regulation policy. The negative
environmental impacts can be quantified in terms of lead emissions during the production
and recycling processes. Public health effects can be calculated using hospitalization
rates near lead recycling plants or estimates of blood lead concentration increases and
thus public health effects. These are imperfect measures without more advanced medical
screening for specific cases, but could give an estimate of the effects of lead pollution.
2.1.7 Safety
Safety is a primary concern of Chinese government officials. In each of the last
three years, China has exceeded 100,000 road fatalities, where most of the victims are
vulnerable road users such as pedestrians or bicyclists ( National Bureau of Statistics
2005). One of the motivations cited for regulating the use of gasoline powered
motorcycles is safety. Beijing officials cited safety as one of the main reasons to ban
electric bikes as well. The China Bicycle Association ( electric bike advocates) countered,
citing the crash rate ( percent of vehicles involved in a crash per year) for electric bicycles
is 0.17% and 1.6% for cars ( Ribet 2005). The primary question is whether electric
bicycles result in a decrease of safety of the entire transportation network, in terms of
fatalities and injuries per person kilometer traveled, or if the incidence of fatalities is
higher for electric bike users because they are vulnerable road users. For safety
considerations, electric bikes’ operating speed is limited so that they can safely operate in
bicycle lanes. Moreover, if we assume that the traveler will take the trip regardless of
mode, what are the safety implications of switching to an alternative mode, bicycle or
transit?
24
2.1.8 Mobility and Accessibility Changes
The reason we tolerate environmental externalities as a society is because the
benefits that activities provide outweigh their externalities. In the case of transportation,
mobility is the primary benefit. Mobility can be defined as average operating speed or
travel time between two points. Mobility by itself does not provide economic benefits,
but it provides access to jobs, goods and services. Mobility differences between modes
can serve as a proxy for accessibility differences between modes in a static, uniformly
distributed built environment. That is, given an origin and a set of destinations, a mode
with higher operating speed than an alternative can access proportionately more
destinations. If origins and destinations are clustered, accessibility increases could be
higher than simply the increase in speed.
Floating vehicle studies using a global positioning system ( GPS) interfaced with a
geographic information system ( GIS) are conducted for bicycles and electric bikes in the
city. These data give an accurate distribution of speed for each mode. They also indicate
a spatial distribution of speeds throughout the urban area. This speed is used in
conjunction with spatial distribution of jobs and housing using an accessibility index
( Cervero 2005) to identify the difference in accessibility between modes.
2.2 Case Studies
China has 660 cities and three quarters of its urban population lives in small and
medium sized cities by Chinese standards ( 0.5- 4 million people) ( Cherry 2005). However,
many of the cities facing the greatest transportation challenges and which are looked to
25
for best practices are China’s megacities, notably Beijing, Shanghai and Guangzhou. To
represent a large portion of the population and investigate differences between two sizes
of cities, the authors decided to investigate Kunming, a medium sized city with an urban
population of about 3 million and Shanghai, a megacity of 15 million.
2.2.1 Kunming
Kunming is the capital of Yunnan province in southwest China ( Figure 2.3). It is a
gateway for trade with Southeast Asia and also a major tourism destination. It has an
urban population of 2.5 million, but the population of the metropolitan area exceeds 5
million. The per capita gross domestic product of urban residents was 31,700 RMB1/ year
in 2004 ( China Data Online 2006). This is significantly lower than the national average
of 37,000 RMB/ year, which is indicative of western China’s lagging economy, compared
to coastal areas.
1 8 RMB= 1 USD
26
Although it has no urban rail transit system, Kunming was the first city in China
to build a bus rapid transit system ( Joos 2000; Kunming Urban Traffic Research Institute
2004). Its road network features three east- west arterials, four north- south arterials, and
two ring roads. A third ring road is currently under construction. Motorcycles are
prohibited within the first ring road and trucks and rural vehicles are prohibited within the
second ring road ( with some exceptions).
The municipal area of Kunming contains about 45 passenger vehicles/ 1000
people ( National Bureau of Statistics 2005). The mode splits for all trips in Kunming are
Figure 2.3: Map of Kunming
27
shown in figure 2.4 ( Kunming University of Science and Technology 2003; Li 2006).
Non- motorized modes, bicycle and walk trips, clearly dominate. The data presented in
Figure 2.4 classifies electric bikes as a non- motorized mode, or a bicycle.
2.2.2 Shanghai
Shanghai is one of China’s megacities, and the municipality is one of the four
municipalities that is classified on the prefecture level ( Figure 2.5). With an official urban
population of a 13 million in 2004, some estimate the entire municipal region to contain
20 million inhabitants. Shanghai’s economy was boosted in the mid 1980’ s when the
central government invested in and developed it as a major economic hub. Since then,
Shanghai has become the industrial and economic center of China. The per capita GDP
exceeded 57,000 RMB/ year in 2004, making it one of the most productive regions in
China.
Kunming
Bus
10%
Car
14%
Motorcycle
4%
Taxi
5%
Walk, Bike, or
E- bike
67%
Shanghai
Bus
16%
Car
9%
Taxi
5%
Subway
3%
Motorcycle
5%
Walk, Bike, or
E- bike
62%
Figure 2.4: Mode splits for all trips in Kunming ( 2003) and Shanghai ( 2006)
28
Shanghai’s transportation system consists of two major grade separated ring roads
and a north- south and east- west elevated highway crossing the center city. The city center
is composed of a highly dense historic road network. Pudong, on the east side of the
Huangpu River is being developed as the new financial center of Shanghai, with a
superblock arterial grid pattern in addition to new subway service. Pudong is connected
to the west bank by tunnels, bridges, subways and ferries. Shanghai currently has four
metro lines, primarily serving the historic city center, Pudong, and the northern and
southern suburbs. Shanghai is undergoing a massive infrastructure development plan for
the 2010 World Fair. This plan will expand the existing rail network to a total of 311 km,
where 30% of the city and 50% of the population will be within a 600 meters of a station.
Pudong New
District
Historic City
Center
Figure 2.5: Map of Shanghai
29
The recent mode split is displayed in figure 2.2. Motorcycles are also heavily restricted in
Shanghai’s city center. Shanghai’s private car ownership rate is 47 passenger
vehicles/ 1000 people, which is considerably lower than some Chinese cities because of
rationed vehicle registrations and license distribution and high registration fees ( National
Bureau of Statistics 2005). When Shanghai’s taxi fleet converted to LPG, the
infrastructure became available for the growth of LPG scooters. As a result, Shanghai is
the only city in China where LPG scooters have gained a significant share of the market.
They are not restricted from the city center and are required to operate in the bicycle lane.
2.3 Data
Through partnerships with Tsinghua University, Tongji University, Kunming
University of Science and Technology and electric bike industrial partners, primary and
secondary data were collected to conduct the research outlined above. Secondary data
sources, particularly for environmental impacts, and bus operations come from statistical
yearbooks, electronic databases, and transit agencies. Primary data were collected,
including interviews with electric bike manufacturers, public security bureaus, surveys of
bicycle and electric bike users, and floating vehicle speed studies. Table 2.1 shows the
main data collected and sources.
30
Vehicle production processes and energy use are obtained from partnerships in
the electric bike industry. Yearly power and resource usage can be divided by yearly
output and a production function can be developed for each electric bike produced.
Detailed material inventories are collected and the energy and emissions of those
materials for each vehicle are calculated using statistical yearbook data. From these data,
energy use and emissions during the production process can be estimated.
Table 2.1: Data, Units, and Sources
Data Units Source
Local City Level Data for Case Study
Energy Mix ( local power
network) % coal, % gas, % hydro ( National Bureau of Statistics 2005)
Power Plant Emission
Factors
μg pollutant/ kWh by
pollutant ( Energy Foundation China 2005)
Power Plant Locations latitude and longitude for
GIS
( International Institute for Applied
Systems Analysis, World Bank et
al. 1999)
Population distribution GIS ( population/ county) ( All China Marketing Research Co.
Ltd. 2003)
Job distribution GIS ( population/ county) ( All China Marketing Research Co.
Ltd. 2003)
Battery Recycling rates % of batteries from virgin or
recycled lead ( Mao, Lu et al. 2006)
Crash Rates fatality and injury per
million veh km Local Public Security Bureaus
Average Speed by Mode km/ h by mode ( bicycle, e-bike,
bus)
GPS/ GIS floating vehicle travel time
study on major corridors, transit
agencies
Mode Shift % of e- bike users who
otherwise use bicycle/ transit travel survey
Average e- bike and
bicycle use per day vkt per day travel survey
Production Data
Electricity Use per e-bike
and bicycle
kWh per year and vehicle
production per year
Interview managers of major
components of bicycles and electric
bike
Energy Mix ( East China
Power Network) % coal, % gas, % hydro ( National Bureau of Statistics 2005)
Energy Intensities of
production processes
Tonne Coal Equivalent
( tce)/ ton product
( National Bureau of Statistics
2005),( Lawrence Berkeley National
Laboratory 2004)
Emission Factors ( East
China Power Network)
μg pollutant/ kWh by
pollutant ( Energy Foundation China 2005)
Power Plant Locations
( East China Power
Network)
latitude and longitude for
GIS
( International Institute for Applied
Systems Analysis, World Bank et al.
1999)
31
Emissions of Chinese power plants have been documented along with scenarios
for future fuels and technologies. Greenhouse gas emissions and conventional pollutants
such as CO, SO2, NOx, and particulates are considered because their public health effects
and treatments vary. Provinces generate electricity from different power sources. The
National Bureau of Statistics ( 2005) keeps yearbook data on the proportion of power
generated by various means for all provinces and major cities in China. Using the
combination of power generation mix and emissions from each type of power generation
by province or city ( or power network at a more aggregate level) can aid in the decision
making process to determine how much electricity is used and the conventional and
greenhouse gas emissions generated per kWh, which can be translated to emissions per
vehicle kilometer traveled by region and growth scenario.
Energy and emissions data should be considered for each of the alternative modes
available to the user. This includes both production and operating pollution. Since this
research will consider the primary shift from bicycle and bus to electric bicycle, I
explicitly investigate the production cost of traditional bicycles and buses, using the same
methodology as that used to calculate electric bike impacts.
Lead loss rates are quantified in Mao et al. ( 2006) Formulations proposed by
other researchers to quantify the effects on public health, such as the increase in
hospitalization as a function of distance to a recycling plant ( Cortes- Maramba,
Panganiban et al. 2003), can be generalized in the Chinese case; considering various
changes in recycling, disposal and battery technology.
Safety records are collected, but data is often reported in aggregate number of
fatalities. Estimates of exposure are extrapolated by converting these totals into a rate
32
( fatalities/ million vkt), using survey data for annual vehicle kilometers traveled by mode.
From these data, estimates of safety impacts can be determined by considering shifts
between modes.
Many of the factors that are required for the above analysis require information on
electric bike use characteristics, particularly average trip length and number of trips, and
thus daily VKT. Additionally, information on alternative mode choice is required to
evaluate the impact of electric bike regulation. These metrics are identified through a
travel survey conducted in Shanghai and Kunming ( see Appendix A). This survey
includes questions related to:
1) Demographic information
2) Origins and Destinations of all daily trips
3) Trip purposes
4) Average travel time and costs of trips
5) Other modes available
6) Alternative mode if current mode were unavailable
Spatial distribution of jobs and housing is provided in GIS format from academic
partners in China. These data are average residential and job density in a census tract in
Shanghai and residential and job points in Kunming. Both maps represent the same
information. Bus routes and headways are attained from bus agencies. Bicycle and
electric bike travel times are collected using a GIS/ GPS based floating vehicle speed
study. These data feed into the accessibility analysis. Specific descriptions of data
collected are included in the subsequent chapters.
33
CHAPTER 3: USE CHARACTERISTICS AND MODE CHOICE BEHAVIOR
In order to understand characteristics of users of electric bikes and other modes in
the choice set, a survey of two- wheeled vehicle users was conducted in a Chinese
megacity- Shanghai and in a medium sized city- Kunming. This chapter discusses the
results of two surveys of electric bike, traditional bicycle, and liquefied petroleum gas
( LPG) scooter users carried out in these two cities. The first section presents
transportation and demographic information on both cities. This is followed by a
discussion on the survey methodology and sampling approach. The results of the survey
and descriptive statistics of electric bike users in these cities are then discussed. Next
structural models that predict mode choice based on user and mode characteristics and
stated preference responses are presented. The final section of this chapter discusses
conclusions and policy inferences.
3.1 Survey Methodology
Two surveys were conducted in Kunming and Shanghai in early April 2006 and
late May 2006, respectively. See Appendix A for a sample survey form. The surveys
targeted electric bike and bicycle users. In the case of Shanghai, LPG scooter users were
also surveyed. The survey contained two parts, a travel diary for the previous day’s travel,
which asks information about trip origins and destinations, travel times and alternative
modes. The second part asks household and individual demographic and attitudinal
questions. The surveys for all modes and both cities are identical, except for a few
location and mode specific differences. Conducting a random household survey in China
is logistically and institutionally difficult. As a result, targeted intercept surveys were
34
conducted at locations that contain a representative sample of urban two- wheel vehicle
users, specifically centralized parking facilities of major activity centers and trip
generators throughout the urban area. These activity centers contain employment, social
activities, and shopping that serve all demographic groups. In both cities, university
students were hired from local universities to conduct the survey.
3.1.1 Location
In Kunming, surveyors were stationed at five major trip generators in the city
center and around the 1st ring road. These locations included major shopping centers that
cater to all demographics of users as well as centralized bike parking facilities
surrounding a large pedestrian mall in the center of the city that contains shopping,
entertainment, and employment. Importantly, most of the survey sites were within the gas
motorcycle restricted zone.
A similar approach was taken in Shanghai. Surveyors were positioned at six
major trip generators throughout the city, including locations in city center, Pudong, and
residential districts. Additionally, several of the survey sites were also near subway
stations, so some respondents utilized two- wheeled vehicles to access the subway. Again,
locations were chosen that served all demographics. Shopping centers often have a major
“ anchor” store and dozens of other smaller stores surrounding the anchor, all served by a
centralized bike parking lot. Often the bike parking lot has capacity to store thousands of
bikes.
35
3.1.2 Sampling
Since bicycle parking is rarely free, most bike parking lots have a single entrance
or exit point, where parkers can pay the attendant. Surveyors were instructed to position
themselves at the entrance of the parking lot and ask every adult entering, regardless of
age or gender, if they would participate in the survey. If people arrived while completing
a survey, they would skip those individuals and ask the first person arriving after he or
she returned to the gate. This sampling method minimized bias. Surveyors conducted the
survey during the middle of the week, from Tuesday to Friday, so that the previous day
travel diary would represent a “ typical” weekday ( Monday to Thursday) and during the
periods of heaviest activity, from mid- morning to evening. After the survey was
completed, survey respondents were offered a small gift ( parking fee payment) as a token
of appreciation. In Shanghai, 696 responses were collected and in Kunming, 502
responses were collected.
3.2 Survey Results
3.2.1 Descriptive Statistics
Overall, people who use bicycles, electric bikes, and LPG scooters come from
similar populations. There are some differences between household characteristics,
particularly wage, household income, and education. Table 3.1 shows the household
demographics of bicycle, electric bike and LPG scooter users in Shanghai and Kunming.
36
Shanghai
Mean value of:
Gender
(% F) Age ** Education
( index) 1 ***
HH Income
( RMB) 2* Wage ( RMB)* HH Size
Bicycle 41% 35.3 ( 14.7) 2.424 ( 1.235) 52626 ( 29756) 2080 ( 1722) 3.49 ( 1.13)
Electric
Bike 41% 36.4 ( 12.8) 2.352 ( 1.111) 59209 ( 29418) 2563 ( 1862) 3.70 ( 1.27)
LPG
Scooter 29% 38.2 ( 11.1) 2.623 ( 1.131) 66000 ( 29572) 3270 ( 1779) 3.56 ( 1.23)
Kunming
Mean value of:
Gender
(% F) Age Education
( index)*
HH Income
( RMB)* Wage ( RMB)* HH Size
Bicycle 50% 34.2 ( 12.0) 2.293 ( 1.010) 29761 ( 16774) 1652 ( 1022) 3.47 ( 1.41)
Electric
Bike 51% 33.1 ( 9.6) 2.551 ( 1.003) 37734 ( 19411) 1905 ( 1101) 3.47 ( 1.22)
Note: t- statistics were calculated to identify differences between samples
* P< 0.05 all modes different ** P< 0.05 bike- lpg different *** P< 0.05 ebike- lpg different
Note: Standard deviation in parenthesis
1 In calculating the index, the following ordinal values were used: less than high school ( 1), high school ( 2),
some college ( 3), college degree ( 4), and graduate study ( 5)
2 Stated yearly income of all workers in the household
3 Monthly wage of individual survey respondent
The Shanghai survey included LPG scooter users, which were significantly
different than bicycle and electric bike users on most metrics. However, bicycle and
electric bike users are significantly different only in wages and household income. The
majority of bike, electric bike and LPG scooter users are male, in the mid 30s. There is no
statistical difference between the education of bicycle and electric bike users although
LPG scooter users have significantly higher education than electric bike users. Household
income and wage are significantly different across all modes, with LPG scooter users
having higher incomes than electric bike users and bike users as expected.
Kunming does not have LPG scooters and there was a much more notable and
significant difference between the demographics of bike and electric bike users,
particularly education and income. There was about a 50% gender split for both modes
Table 3.1: Demographics of Two- Wheel Vehicles Users in Kunming and Shanghai
37
and users were in their mid 30’ s on average. The education and income metrics were all
significantly higher for electric bike users than bicycle users
Household vehicle ownership rates of survey respondents are shown in Table 3.2.
As expected, the household ownership of vehicles who were responding to the survey
were significantly higher than those who were not ( i. e. bicycle ownership of bicycle
respondents is much higher than bicycle ownership of non- bicycle respondents).
Surprisingly, in Shanghai there is no statistically significant difference in car and
motorcycle ownership between modes, despite progressively higher incomes of electric
bike and LPG scooter users. This is most likely due to Shanghai’s restrictions on
automobile registration and ownership. Owners of LPG scooters have more electric bikes
in their household than bicycle users.
In Kunming, electric bike users have more than twice the amount of cars available
to the household than bicycle users, which is likely the effect of higher incomes. The car
ownership of electric bike households is 75 vehicles per 1000 people, which is about the
same as the city average.
38
Shanghai
Surveyed
User:
Average number of vehicles in the household:
Car Motorcycle Bicycle** Electric Bike* LPG Scooter***
Bicycle 0.140
( 0.378)
0.234
( 0.487) 1.504 ( 0.886) 0.187 ( 0.409) 0.259 ( 0.493)
Electric Bike 0.155
( 0.363)
0.163
( 0.402) 0.737 ( 0.807) 1.060 ( 0.573) 0.223 ( 0.463)
LPG Scooter 0.156
( 0.380)
0.228
( 0.425) 0.731 ( 0.749) 0.269 ( 0.458) 0.946 ( 0.562)
Kunming
Surveyed
User:
Average number of vehicles in the household:
Car* Motorcycle Bicycle* Electric Bike*
Bicycle 0.111
( 0.359)
0.151
( 0.386) 1.452 ( 0.988) 0.432 ( 0.039)
Electric Bike 0.257
( 0.544)
0.178
( 0.462) 0.782 ( 0.913) 1.234 ( 0.028)
Note: t- statistics were calculated to identify differences between samples
* P< 0.05 all modes different, ** P< 0.05 bike and others different, *** P< 0.05 LPG and others different
Note: Standard deviation in parenthesis
3.2.2 Travel Behavior
Differences in mode share have significant impact on travel demand, road
capacity, environmental impacts and in the long term, urban form. As travelers choose
faster modes, trip length and frequency will likely increase, creating more demand on the
transportation infrastructure. Faster speeds also promote the spatial separation of land
uses. Alternatively, people may choose modes like electric bikes to provide “ easier”
mobility, not necessarily to travel faster or more or access more destinations.
The surveys asked travelers to list characteristics of their previous day’s travel by
bicycle, electric bike, or LPG scooter. Questions were asked related to trip purpose,
modal choice set, primary alternative mode, previously used modes, trip length, and
travel time. Table 3.3 shows the characteristics of travel by each mode.
Table 3.2: Household Vehicle Ownership Levels
39
Shanghai
Number
of Trips1
Average Trip Lengths ( km): Average trip:
Total
Trips2*
Work
Trips***
Other
Trips
Travel Time
( min) 3
Speed
( kph) 4*
Weekday
VKT5
Bicycle 2.06 4.29
( 4.39)
4.94
( 4.86)
4.07
( 4.21)
26.31
( 22.35)
11.38
( 7.07) 8.84
Electric
Bike 2.00 4.83
( 4.25)
5.66
( 4.37)
4.50
( 4.16)
25.56
( 18.75)
13.04
( 7.25) 9.66
LPG Scooter 2.06 6.64
( 5.96)
7.78
( 6.77)
6.16
( 5.53)
28.75
( 19.81)
14.57
( 7.94) 13.68
Kunming
Number
of Trips
Average Trip Lengths ( km): Average trip:
Total
Trips*
Work
Trips
Other
Trips
Travel Time
( min)
Speed
( kph)*
Weekday
VKT
Bicycle 2.23 3.38
( 1.91)
3.54
( 1.79)
3.28
( 1.97)
22.95
( 12.29)
10.45
( 5.74) 7.54
Electric
Bike 2.54 3.63
( 2.08)
3.75
( 2.06)
3.55
( 2.09)
20.28
( 11.29)
11.85
( 5.90) 9.22
Note: t- statistics were calculated to identify differences between samples
* P< 0.05 all modes different ** P< 0.05 bike and others different *** P< 0.05 LPG and others different
Note: Standard deviation in parenthesis
Note: All distances in kilometers
1 Trip number is defined as a one way trip, so a trip to work and back would constitute two trips. The
number of trips should be at least two for any travel diary that had any trips. A few of the respondents
reported no trips on the previous day.
2 Estimated network distance from stated origin and destination
3 Stated total travel time of trip estimated by respondent
4 Average Speed is calculated as the measured trip length divided by the stated travel time of trip
5 Total VKT ( vehicle kilometers traveled) is total trip length times the number of trips.
The trip length is calculated as the network distance between stated origins and
destinations. The trip lengths increased, corresponding to increases in speed, with LPG
scooters taking the longest trips and bicyclists taking shorter trips. In Shanghai, the work
trip length is about 20% longer than the length of other trips. In Kunming, the work trip
length is not statistically longer than other trips. This could be because of Kunming’s
compact development and relatively short commute distance, compared to Shanghai.
When considering economic productivity, the total number of vehicle hours spent
traveling ( VHT) is an important metric to understand how much productive time people
lose while commuting. The travel time from origin to destination is stated for each trip
Table 3.3: Travel Characteristics, Surveyed weekday ( April- May 2006)
40
and interestingly, there is no significant difference in perceived travel time between
modes ( implying increased speed). This is consistent with time budget theory stating
people are willing to accept thresholds of travel time and people will choose origins and
destinations based on the maximum travel time they are willing to accept, not necessarily
based on distance. This question is problematic because people often know and report
door- to- door travel time. This includes access and egress time, which would have the
effect of underestimating on- vehicle speed of faster modes. Also, people often round to
the nearest 5- minutes and given the short trip distances, estimates of speed from stated
travel time could be biased. Even with these considerations, the stated speeds of electric
bikes are higher than bicycles by 15% and 10% in Shanghai and Kunming, respectively.
LPG scooters in Shanghai are 12% faster than electric bikes. A floating vehicle travel
time study conducted in Shanghai and Kunming compared bicycle and electric bike
speeds and showed a 30- 35% increase in average speed of electric bikes over bicycles.
Perhaps the most important metrics related to externalities generated by two-wheeled
vehicles is the daily vehicle kilometers traveled ( VKT) and vehicle hours
traveled ( VHT). Daily VKT is usually associated with roadway capacity needs, pollution,
energy use and safety. As expected, the VKT of electric bikes is 9% and 22% higher than
bicycles in Shanghai and Kunming, respectively. The daily VKT of LPG scooters is 41%
higher than electric bikes in Shanghai. This increase in VKT could be an indication that
travelers of higher speed modes choose to travel farther or more to access more
destinations. It could also be a result of self selection, that is, people who were already
traveling far on a previous mode switched to electric bikes or LPG scooters because of
their distant travel, i. e. they are not traveling any farther than before, just faster.
41
Interestingly, the average lengths of all trips are significantly different among all
modes, but the average trip length of work trips between electric bikes and bicycles in
both cities is not significantly different. This indicates that most of the additional VKT is
due to traveling farther for non- work trips, or discretionary trips. Work trip length of LPG
scooters is significantly higher than bicycles and electric bikes in Shanghai. Trip purpose
by mode and city is shown in Figure 3.1, with work trips constituting the overwhelming
majority on all modes in both cities.
In order to identify relative impacts of different mode choices, alternative modes
must be estimated. Respondents were asked what mode they would take in the absence
( or regulation) of their current mode for each trip. Overwhelming, people responded that
they would take a bus as the alternative mode, followed by bicycle and walking ( Figure
3.2 and 3.3). Of electric bike users, bus is the best alternative for about 55% of trips in
Shanghai and 58% of trips in Kunming and bicycle is the best alternative for about 12%
0%
10%
20%
30%
40%
50%
60%
70%
sh bike
sh e‐ bike
sh lpg
km bike
km e‐ bike
Figure 3.1: Trip Purpose by Mode and City
42
of trips in Shanghai and 21% of trips in Kunming. LPG scooter users are the least likely
to choose a bus and most likely to choose a taxi, which is representative of their higher
incomes.
When asked what mode they used before they used their current mode, the most
frequent response again was bus. Interestingly, a large portion of electric bike users used
to use bicycles for their current trip, but would use bus now if they could not use electric
bikes. This implies that a large group of travelers shifted from bicycle to electric bike in
place of shifting from bicycle to bus. In most cases, over 50% of the travelers rode the
bus before using an electric bike.
0%
10%
20%
30%
40%
50%
60%
70%
80%
sh bike
sh e‐ bike
sh lpg
km bike
km e‐ bike
Figure 3.2: What Mode Would You Take Otherwise?
43
Knowing the alternative mode is essential when developing policy regarding the
regulation of electric bikes or LPG scooters. If banning electric bikes causes a significant
increase in bus ridership during peak hours, service expansion may be required resulting
in significant public investment. Alternatively, if most people used to and would
otherwise use non- motorized modes, little public investment would be required and
energy and emission impacts would be significantly reduced.
3.2.3 User Attitudes
Several attitudinal questions were asked in this survey; particularly to find out the
reasons people use different two- wheeled modes and what how people perceive electric
bikes. When electric bike and LPG scooter users were asked why they chose the mode,
most people responded that high speed was a primary reason. Also respondents cited that
these motorized modes require less effort than alternative modes, such as bus or bicycle.
0%
10%
20%
30%
40%
50%
60%
70%
80%
sh bike
sh e‐ bike
sh lpg
km bike
km e‐ bike
Figure 3.3: What Mode Did You Previously Use?
44
Identifying factors that influence attitudes can help explain mode choice. The distribution
of responses is shown in Figure 3.4.
In order to find out how other users of the bicycle lane perceive electric bikes,
respondents were asked if electric bikes should be allowed and developed as a viable
mode in the city. Surprisingly, 70% of Kunming bicycle riders and 77% of Shanghai
bicycle riders think that electric bikes should be developed more. Over 85% of electric
bike and LPG scooter riders think that electric bikes should be developed more. This
shows that electric bikes are popular in the bike lane and even bicyclists do not have a
poor opinion of them.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
fast
Less effort
safer that
motorcycle
PT too crowded
cheaper than
auto
PT too expensive
ride in bike lane
access to moto‐restricted
areas
moved so longer
commute
new job so longer
commute
shanghai e‐ bike
shanghai lpg
kunming e‐ bike
Figure 3.4: Why Did You Choose This Mode?
45
3.3 Factors that Influence Two- Wheel Vehicle Choice
In order to gain a better understanding of the factors that influence electric bike
use discrete choice models were specified on the survey responses to predict electric bike
use based on demographic factors ( such as income, age, and gender) and alternative
specific characteristics ( such as travel time and cost of alternative modes). Two research
questions are presented:
1) What factors influence the trip mode choice between electric bikes and bicycles?
2) Given that a user has chosen electric bikes, what factors influence their best stated
alternative?
These questions can be represented by the mode choice hierarchy represented in Figure
3.5.
In order to answer these questions, discrete choice models were specified on the
survey responses. A logit modeling framework was used. In general, the logit model
Auto Bus Walk Bike
Electric
Bike
Traditional
Bike
Mode Choice
Distribution
Question 1
( Section 3.3.1)
Question 2
( Section 3.3.2)
Figure 3.5: Modeling Hierarchy for Discrete Choice Models
46
predicts a discrete, unordered outcome ( y) by a series of explanatory variables ( X). The
general functional form of the logit model is:
Σ =
j
x
x
ni nj
ni
e
P e β
β
Where Pni is the probability of individual n choosing alternative i, and xnj is the vector of
observed demographic and alternative based explanatory variables for all alternatives j.
One of the assumptions of the logit model is independence from irrelevant alternatives
( IIA). This assumption allows analysts to model subsets of the choice set. For a thorough
discussion of discrete choice modeling techniques and assumptions used in this section
see ( McFadden 1974; Ben- Akiva and Lerman 1985; Train 2002).
3.3.1 Choice Between Bicycle and Electric Bike
The initial hypothesis was that electric bikes are an intermediate mode on China’s
motorization pathway. That is, bicycle riders will evolve into electric bikes and then into
other personal motorized modes, particularly cars. The survey discussed above was used
to develop a binomial logistic regression of the probability of choosing an electric bike
instead of a bicycle. The data were adjusted to represent linked trips into a single home-based
trip tour. A tour is defined as a series of trips that begins and ends at home. For
example, a trip from home, to work, to the grocery store then back home is defined as
three trips linked into a single tour. Each observation in the model is a tour. This removed
potential bias from the model in two ways: 1) the level to which individuals were
sampled more than once was minimized. For example people make more than 2 trips per
47
day, but most people only make one trip chain, to work and back. The individual specific
parameters are therefore independent between choice situations ( trips). This reduced the
need to correct for this dependence with a mixed logit approach ( Train 1998). 2) The
dependence between trip links is included within the trip. For example, if a person chose
to ride an electric bike to work, the probability of choosing an electric bike to travel home
is very high, and not independent of his/ her choice to choose an electric bike for the
previous trip. Combining all linked trips into a trip tour assumes that the individual
makes choice decisions based on the entire trip tour, not just the first link.
The results of the logistic regression are shown in Table 3.4. The bicycle is the
base unit of comparison, so the coefficients ( β) measure the change in electric bike use
relative to choosing a bicycle. Variables related to vehicle performance, user
demographics and attitudes entered the model.
48
Number of obs = 669
Log likelihood = - 170.329 Pseudo R2 = 0.566
Variable β Std
Error Z P> z Odds
Ratio
Std
Error
Difference in Travel time for trip tour between
bicycle and e- bike ( minutes) a 0.027 0.013 2.03 0.043 1.028 0.013
Number of e- bikes in household 3.736 0.311 12.00 0.000 41.919 12.550
Number of bikes in household - 0.756 0.203 - 3.73 0.000 0.470 0.080
Number of Cars in Household 0.700 0.291 2.41 0.016 2.014 0.703
Pro- ebike attitude ( 1 if pro- ebike, 0 otherwise) b 1.144 0.343 3.34 0.001 3.140 1.137
Perceive mode as low effort ( 1 if low- effort, 0
otherwise) c 1.469 0.490 3.00 0.003 4.347 2.147
Age 0.267 0.065 4.11 0.000 1.306 0.094
Age^ 2 - 0.004 0.001 - 3.97 0.000 0.996 0.001
Gender* Age ( 1 male, 0 female) - 0.077 0.030 - 2.54 0.011 0.926 0.028
Gender* Age^ 2 ( 1 male, 0 female) 0.002 0.001 2.39 0.017 1.002 0.001
CONSTANT - 3.488 1.206 - 4.95 0.000
a This is the total network distance of the trip tour divided by the empirically measured average speed of each
mode using a GPS floating vehicle study ( Cherry 2006), it does not use the travel time reported by
respondents.
b Respondents answered a question asking if they think that electric bikes should be encouraged in the city. If
they answered favorably, they were coded into the dataset as “ pro- ebike”
c Respondents stated that one of the reasons they chose a particular mode is because of the low effort required
This model shows that household ownership of various vehicles increases or
decreases the probability of choosing that mode. As expected, ownership of an electric
bike greatly increases the probability of choosing an electric bike. Bicycle ownership
decreases the probability of choosing an electric bike. Car ownership also increases the
probability of choosing an electric bike. This could be an indication that electric bikes act
as “ second cars” for families with multiple wage earners, or that household members are
accustomed to personal motorized mobility and thus more likely to use an electric bike
instead of a bicycle. It could also be a proxy for household income or value of time. As
expected, the respondents who share the attitude that electric bikes should be encouraged
and those who value low effort when making mode choices are more likely to choose
electric bikes. The older the person is, the more likely they are to choose an electric bike
Table 3.4: Logit Model for Predicting Probability of Electric Bike Mode Choice
49
up to a certain point, and then they are more likely to choose a bicycle. This is probably a
result of the oldest members of the population unwilling to adopt new technology.
Gender enters into the model when interacted with age. The sign on the two interaction
variables indicates that the concave curve of electric bike choice as a function of age is
flatter for men – that is, across all age categories, men are generally less likely to opt for
electric bikes then women. Finally, the longer the trip or the larger the travel time
difference, the greater the likelihood of choosing an electric bike.
Factors of note that did not enter the model ( due to statistical insignificance) are
gender alone, city ( dummy variable), household income, household size, level of
education, trip purpose and monetary trip cost. These are important findings, particularly
the non- appearance of a fixed- effect city variable and monetary cost variable. The failure
of the relationships of difference between cities suggests the results could be
generalizable to other similar Chinese cities, regardless of local GDP. Also, bicycle and
electric bikes users do not pay a large out- of- pocket marginal cost when making a trip or
tour. The major cost of operating a bicycle is largely a one time purchase price and the
cost of operating an electric bike is paid monthly through electricity bills and when
batteries are replaced, normally every year or two.
Electric bikes were oversampled to gain an adequate number of electric bike
responses, while not requiring an overly large sample of bicycles. Of the final sample of
669 trip tours that entered the model, 183 were bicycle trips and 486 were electric bike
trips. The true ratio of bicycles to electric bikes is about 4.5: 1 in Shanghai and Kunming.
Choice based sampling causes biased estimates of the alternative specific constants and is
corrected by the following equation ( Train 2002):
50
* ( ˆ ) ln( / )
j j j j α = E α + A S
Where α*
j is the true constant and E( αj) is the biased estimated constant. The true
population proportion for alternative j is Aj and the sampled proportion is Sj. The constant
presented in Table 3.4 represents this adjustment.
3.3.2 Choice of Alternative Mode
A very relevant question to determine environmental impacts of electric bike
policy is determining the alternative mode in the absence of electric bikes through
regulation. If electric bikes are banned, the implications of environment costs and
mobility benefits are very dependent on the alternative mode. A fixed- effects logit model
was specified to understand factors that influence a traveler’s choice of a low cost
alternative mode. Again, trips were categorized into trip- chains and the entire trip chain
was modeled as an independent observation. The problem of over- sampled individuals
was reduced using this technique. In this case, the three low- cost modes, bus ( 60%),
bicycle ( 16%), and walk ( 6%), with the highest response rate among electric bike users
for specific trip chains were included in the choice set. The model is shown in Table 3.5.
Walk trips were set as the base case, so the coefficients ( β) measure the change in bus or
bicycle use relative to choosing to walk. The cost of the trip did not enter significantly
into this model primarily because the marginal cost difference observed by users is small
for all modes.
51
As expected, travel time enters into the model with a negative sign, indicating the
greater the travel time of a particular mode, the lower the probability of choosing that
mode. Age of prospective bus riders does not significantly enter into the equation,
indicating age does not influence the choice between walking and bus riding. Age of
bicycle users is significantly positive, while age^ 2 is negative, indicating that people are
more likely to use a bicycle ( instead of walk or bus) as they age, up to a point and older
individuals become less likely to choose to bicycle. Interestingly, travelers who share the
opinion that public transit is too crowded are more likely to take the bus than walk, and
slightly more likely to take a bus than ride a bicycle ( although this difference is
statistically insignificant). Finally, electric bike users who have a pro- ebike attitude are
more likely to take the bus in the absence of electric bikes than walk or ride a bicycle.
Table 3.5: Logit Model for Predicting Probability of Current Electric Bike Users
Switching to Bus, Bicycle, or Walk if Electric Bikes Became Unavailable
Number of obs = 423
Log likelihood = - 298.29 Pseudo R2 = 0.3396
Variable β Std Error Z P> z Odds
Ratio Std Error
Alternative Specific Constant- Bus 1.628 0.352 4.62 0.000 5.094 1.794
Alternative Specific Constant- Bicycle - 3.034 1.542 - 1.97 0.049 0.048 0.074
Trip Chain Travel Time ( min) a - 0.042 0.010 - 4.07 0.000 0.959 0.010
Age of Bicycle Choosers 0.173 0.086 2.01 0.044 1.189 0.102
Age^ 2 of Bicycle Choosers - 0.003 0.001 - 2.27 0.023 0.997 0.001
Perceive Public Transit is Crowed ( 1 if PT
Crowded, 0 otherwise)- Bus Choosersb 2.172 1.028 2.11 0.035 8.774 9.016
Perceive Public Transit is Crowded ( 1 if PT
Crowded, 0 otherwise)- Bicycle Choosersb 2.306 1.055 2.19 0.029 10.033 10.581
Pro- ebike attitude ( 1 if pro- ebike, 0
otherwise)- Bus Choosersc 0.655 0.332 1.97 0.049 1.925 0.640
a For the bike option, travel time was estimated as the total network distance of the trip tour divided by the
empirically measured average speed of bicycle mode using a GPS floating vehicle study ( Cherry 2006) .
Walk times assume 6.5 km/ hr walk speed. Public transit agencies provide data on bus travel times that
include access and egress time, wait time, transfer time and in- vehicle time for the bus option.
b Respondents stated that one of the reasons they chose electric bike is because they perceive public transit
to be too crowded.
c Respondents answered a question asking if they think that electric bikes should be encouraged in the city.
If they answered favorably, they were coded into the dataset as “ pro- ebike”
52
Unfortunately, this model does not accommodate predictions based on most
demographic variables. For the most part, demographic variables, including education,
gender, wage, household income, household size, and vehicle ownership were not
significantly different from each other across the three choices, with the exception of age
affecting bicycle use. The factors that have the greatest influence on mode choice are
travel time and attitudinal variables. If policy makers want to influence choice, they
should focus on decreasing the travel time of the desired choice.
3.4 Conclusion and Policy Inferences
Electric bike use has grown at extraordinary rates over the past few years and
little is known about who uses electric bikes and how electric bike users make mode
choices. Policy makers in different cities are treating electric bikes differently. Some
cities have embraced them as a low cost form of high mobility, complementing other
transportation options. Other cities have pointed to environmental and safety problems
and heavily restricted their use or banned them.
In order to develop environmentally sustainable and equitable policy regarding
electric bikes, a policy maker has to understand what populations are using electric bikes,
how they are using electric bikes and what they would choose in the absence of electric
bikes. This research has identified characteristics of electric bike users in two different
cities in China, Kunming and Shanghai. Although there are significant socio-demographic
differences between these two cities, electric bike use characteristics are
similar between them. Electric bike users are generally more educated than bicycle users
and have higher incomes. Commuters do not use electric bikes in the same way as
53
bicycles. Electric bike users take more and longer trips in an average weekday than
bicycle users and LPG scooter users take much longer trips. The result is increased daily
VKT and thus energy use and air pollution, compared to bicycles.
User attitudes also affect the reason people choose electric bikes. Users primarily
cite speed, effort, safety, and crowded transit as reasons to choose electric bikes.
Interestingly, most bicycle riders do not have a poor opinion of sharing the lane with
electric bikes and would recommend developing electric bikes as a mode in the city.
User attitudes, demographics and vehicle performance are all significant factors
that influence mode choice in the logit models specified above. The model specified in
Table 3.4 predicts the choice between electric bike and bicycle use, based on survey
responses in Kunming and Shanghai. Demographic factors such as wage, age, gender and
household vehicle ownership all influence mode choice. One of the more significant
factors that can be controlled by policy makers through regulation is the difference in
travel- time between the two modes. As expected, the higher the travel time difference,
the higher the likelihood of choosing an electric bike. Travel time differences are linked
to speed, which is a function of congestion levels in the bike lane, network ( traffic signal)
density, and electric bike performance. Electric bikes are loosely regulated to a maximum
speed of 20 km/ hr, in which manufacturers rarely comply. As electric bikes become
faster, the travel time differential will change and more people will shift from bicycles.
Speed is likely the factor that policy makers have most control over that has the
greatest influence on mode choice, either through performance regulation or traffic
control. In the cities studied, electric bike users spent a larger portion of their travel time
stopped at signals than bicycles, as expected because of their higher free- flow speeds. A
54
way to increase electric bike use would be to consider control strategies that limit the
number of stops for both modes, through signal coordination or grade separated
intersection crossings, thus increasing the travel time advantage of electric bikes.
Travel time of a trip also significantly influences alternative mode choice. Electric
bike users would switch to a bus for most trips if electric bikes were banned from cities.
Some cities have made an effort to reduce two- wheeled vehicle traffic by providing high
quality transit. Signal priority and exclusive right of way for buses will increase ridership
by decreasing travel time.
Factors that influence mode choice are important inputs into policy analysis when
attempting to influence travel behavior. This chapter sheds light on this topic so that
policy makers can make more informed decisions regarding the regulation or promotion
of electric bike use in their cities. The findings of this analysis will help identify the
significance of mode specific impacts that will be investigated in the following chapters.
55
CHAPTER 4: ENVIRONMENTAL IMPACTS OF ELECTRIC BIKE USE
The growth of electric bikes has caused concern for government officials,
transportation engineers and city planners who are attempting to promote development of
sustainable and efficient transportation in their cities. The environmental impacts of
electric bikes are unclear and the benefits they provide to the transportation system are
ambiguous. It is clear that they emit zero tail pipe emissions at their point of use and that
their overall energy efficiency is higher and emissions per kilometer are lower than
gasoline scooters and cars; but most electric bike users might not otherwise use cars or
gasoline scooters. The environmental costs of this mode are largely related the alternative
mode, should the electric bike be prohibited or restricted. Taiwan promoted and
subsidized electric bikes in the 1990’ s ( Chiu and Tzeng 1999) in order to induce a shift
away from dirtier gasoline scooters. This chapter presents analysis of the environmental
costs of electric bikes and alternative modes and can help inform policy that will affect
millions of users.
This chapter begins by discussing the production processes and some of its energy
use and environmental characteristics. The following section discusses the environmental
impacts of electric bike use and attempt to quantify the largest sources of energy use and
pollution. Environmental impact analysis is conducted for dominant alternative modes as
a unit of comparison. Exposure differences of urban versus non- urban pollution sources
are identified to serve as a proxy for public health effects of air pollution.
56
4.1 Energy Use and Emissions of Electric Bike Life Cycle
4.1.1 Production Processes
There are hundreds of electric bike manufacturing companies in China, ranging
from small assembly factories to large component makers and assembly factories. In
order to understand the production processes, five electric bike factories in Shanghai,
Jiangsu, and Zhejiang provinces were visited. These factories ranged in production
output from 12,000 bikes/ year to over 150,000/ year. Production capability ranged from
simple e- bike assembly ( e- bikes are assembled from components produced by other
companies off- site), while others produced some main components in- house such as the
motor, controller, and frame.
Assembly of an e- bike typically requires one main assembly line where the frame
is passed through various stages of assembly until fully assembled. E- bike assembly
lines have the capacity to produce one e- bike every 5 minutes. Individual components
and processes of the e- bike are produced and performed off- line, such as assembling
wiring systems, brake systems and painting.
Through interviews with factory owners and publicly reported statistics on energy
use and emissions from the manufacture of raw materials, estimates are made regarding
the environmental implications of the production process of electric bikes. To avoid the
intensive work of calculating the environmental effect of each process in a factory, the
overall energy use of all processes is obtained and included in the energy use calculation.
Other estimates of energy use and emissions are made using the weight of raw materials
required to produce an electric bike and the energy and pollution intensities of producing
57
those materials in China. Some data are omitted because of lack of availability or the
expectation that their impacts are small compared to other impacts.
There are few energy intensive processes associated with the assembly of an
electric bike. Almost all energy use is in the form of electricity required to run the
machinery of the factory. Perhaps the most energy intensive processes of the assembly
process are steel frame construction and painting ( large dryers are required). One of the
larger e- bike manufacturers in China reports that in 20052, they produced 180,000 electric
bikes and used 1,278,545 kWh of electricity, or 7.1kWh per bike. The processes included
in this value are frame welding and bending, painting, assembly, assembly of controllers,
vehicle inspection and testing, packaging and general electricity use of the factory.
Another energy intensive process is the manufacture of lead acid batteries. A large scale
electric bike battery manufacturer was also interviewed regarding energy consumption.
The total energy consumption per 12V electric bike battery is approximately 2 kWh, so a
36V battery would require 6kWh and a 48V battery would require 8kWh3.
The energy required by the assembly process is very small compared to the
energy requirements of the raw material manufacturing, such as steel, plastic, and rubber.
Table 4.1 is an inventory of electric bike components, the material they are composed of,
the weight, and the energy required to produce those products. National statistics and
literature on Chinese steel and lead industries are used to calculate the amount of energy
used per unit weight of a product are then used to estimate the energy use of the
manufacture of a component ( Price, Phylipsen et al. 2001; National Bureau of Statistics
2 Interview with electric bike factory owner 3- 4- 2006
3 Phone interview with electric bike battery factory manager 3- 4- 2006
58
2003; National Bureau of Statistics 2004; National Bureau of Statistics 2005; China Data
Online 2006; Mao, Lu et al. 2006).
Weight of Electric Bike Materials ( kg/ bike)
BSEB SSEB
Total Steel 18.15 46.1% 26.18 46.5%
Total Plastic 5.67 14.4% 15.22 27.0%
Total Lead 10.28 26.1% 14.70 26.1%
Total Fluid 2.94 7.5% 4.20 7.5%
Total Copper 2.55 6.5% 3.46 6.1%
Total Rubber 1.14 2.9% 1.22 2.2%
Total Aluminum 0.52 1.3% 0.58 1.0%
Total Glass 0.00 0.0% 0.16 0.3%
Total Weight 41.25 65.73
Associated Energy and Emissions of Manufacturing Processes
Energy Use ( tonne SCE) 0.179 0.261
Energy Use ( kWh) 1456 2127
Air Pollution ( SO2) ( kg) 1.563 2.198
Air Pollution ( PM) ( kg) 5.824 8.173
Greenhouse Gas ( tonne CO2eq) 0.603 0.875
Waste Water ( kg) 1488 2092
Solid Waste ( kg) 4.463 7.139
The weight of each material was estimated using weights of typical components of each
style of electric bikes. These components were categorized into materials in which there
are readily available data on energy use and emissions.
Several assumptions and omissions were made to develop Table 4.1. This table
includes energy and environmental impacts due to the mining and production of ferrous
and non- ferrous metals, and the production of plastic and rubber. It does not include the
impacts of battery electrolyte production or fillers in rubber production ( particularly
carbon black). It also does not include transportation impacts. The values presented in
Table 4.1 should be considered lower bounds. The solid waste only includes solid waste
Table 4.1: Material Inventory, Emissions and Energy Use- Electric Bike
59
of the production process, not end- of- life waste, which will be discussed later. The
numbers above also include the manufacture of replacement parts, specifically five sets
of batteries, three sets of tires and two motors over the lifespan of the electric bike4.
4.1.2 End- of- Life
Because of the relatively recent appearance of electric bikes in the transportation
system, little is known about the fate of electric bikes that have become obsolete or non-operational.
Many of the earliest models of electric bikes were simply modified bicycles,
so if components failed, the electric bike could still operate as a standard bicycle. More
recent models would be inoperable if vital components failed. In order to calculate the
end of life solid waste, the recyclable components of the electric bike needs to be reduced
from the total weight. Additionally, replacement parts must be considered; five batteries,
three sets of tires and two motors.
Steel, which is the heaviest component of electric bikes has a high recycling rate,
79.9% in 2002 ( National Bureau of Statistics 2003). This is the recycling rate of the
entire steel industry, and might not reflect the actual recycling rate of the steel in electric
bikes. Likewise the entire copper industry has a recycling rate of 88.5% in 2002. If these
materials are recycled and the other materials, including replacement parts of the electric
bike enter the waste stream, BSEBs and SSEBs produce 17 and 30 kilograms of solid
waste, respectively. This does not include lead waste from batteries, which will be
discussed in detail in the following section.
4 Personal communication with electric bike manufacturers and their estimation of component reliability
60
4.1.3 Lead Acid Batteries
Lead acid battery pollution is the most cited reason for regulation of electric bikes
by policy makers. Approximately 95% of electric bikes in China are powered by lead
acid batteries ( Jamerson and Benjamin 2004). Based on interviews with manufacturers
and service facilities, the life span of an electric bike battery is considered to be one to
two years or up to 10,000 kilometers. BSEBs typically use 36V battery systems, on
average weighing 14 kilograms. SSEBs typically use 48V battery systems weighing 18
kilograms. The lead content of electric batteries is 70% of the total weight, so BSEB and
SSEB batteries contain 10.3 and 14.7 kilograms of lead, respectively.
This is perhaps the most problematic issue for electric bikes and is the same
problem that influenced the demise of electric car development in the United States in the
early 1990’ s ( Lave, Hendrickson et al. 1995). Because of the relatively short lifespan of
electric bike batteries, an electric bike could use five batteries in its life, emitting lead
into the environment with every battery. Lead is emitted into the environment during four
processes: 1) Mining and smelting lead ore 2) Battery manufacturing 3) Recycling used
lead and 4) Non- recycled lead entering the waste stream. Loss rates can be expressed in
terms of unit weight of lead lost per unit weight of battery produced for each process.
Lave and Hendrickson ( 1995) cite that, in the USA, 4% ( 0.04 tons lost per ton of battery
produced) of the lead produced is lost using virgin production processes, 1% is lost
during the battery manufacturing process and 2% is lost during the recycling process. So,
a battery composed of 100% recycled lead emits 3% of its lead mass into the
environment. A battery composed of 100% virgin material emits 5% of its lead content
into the environment. In most industrialized countries, lead recycling rates exceed 90%.
61
China’s lead acid battery system is very different from industrialized countries.
Mao et al. ( 2006) investigated the Chinese lead acid battery system. They found that 27.5%
of the lead content of a battery is lost during the mining, concentrating, smelting and
recycling process. This value can be broken down into two components, emissions of
concentration and primary refining of virgin ore and secondary refining of recycled scrap,
which have emission rates of 31.2% and 19.7%, respectively. In addition to these losses,
4.8% is lost during the manufacturing process. The reasons for these very high loss rates
are mostly due to poor ore quality and a high proportion of lead refined at small scale
factories using outdated technology. The official recycling rate of lead in China’s lead
acid battery industry is 31.2%. Mao et al. ( 2006) estimate that the actual number is
approximately double that, 62% because of informal, small scale recyclers. This value
feeds into the proportion of recycled lead in each battery. The authors indicate that lead in
a battery is made up of 22% recycled lead and 78% virgin lead.
Mao et al. uses data from 1999, before electric bike batteries were a significant
share of the market. Several of the values ( specifically recycling rate) are estimates and
could have changed since electric bikes entered the market. In 2004, electric bike
batteries constituted 8% of the market, with car and motorcycle batteries comprising 74%
of the total battery market ( Unknown 2006). Because electric bikes use batteries quickly,
some informal recycling and collection practices have developed. In most cases, an
electric bike customer can exchange an exhausted battery for ¼ the price of a new battery,
or around 60 RMB ( US$ 7.50), which is a significant amount of money in most Chinese
cities. The dead batteries are then collected from service centers and sent to lead
recycling factories. This institution could increase the average recycling rate of all lead
62
acid batteries. Interviews with factory owners estimate that 85- 100% of electric bike
batteries are recycled5.
The values in Table 4.2 are generated using the loss rates presented above. Lead is
lost to the environment in three processes. Lead is lost during production in process I,
during battery manufacture in process II, and by disposal ( lack of recycling) in process III.
The proportion of recycled material that contributes to the content of a battery is
dependent on previous years’ recycling rates and the growth rate of lead demand ( 15-
20%) ( China Data Online 2006). It is assumed that all new demand is met by virgin lead
production. Additionally, all lead that is lost to the environment due to recycling is also
met by virgin production. The maximum amount of recycled content in lead acid batteries,
assuming 100% recycling rates, would be about 60% ( considering loss rates from
previous time periods and increased demand). Mao et al. ( 2006) estimate 22% recycled
content of lead acid batteries, which could be considered a minimum. The manufacture
loss is constant, regardless of source material and the recycling rate is estimated based on
the official and estimated values.
5 Interview with factory owners and managers May 15- 18, 2006
63
Table 4.2: Electric Bike Lead Emissions
BSEB SSEB
Battery Weight ( lead content) kg 10.3 14.7
I
Lead Production Loss
(% Recycled Material)
0% 3.21 4.59
( Mao, Lu et al. 2006) 22% 2.95 4.21
44% 2.69 3.84
60% 2.50 3.57
II Manufacture Loss 0.49 0.71
( Mao, Lu et al. 2006) 4.8%
III
End- Of- Life Loss
( Recycling Rate)
0% 10.30 14.70
( Mao, Lu et al. 2006) official 31% 7.11 10.14
( Mao, Lu et al. 2006) estimate 62% 3.91 5.59
( E- bike manufactures) 85% 1.55 2.21
100% 0.00 0.00
Scenarios
( Production, Manufacture, EOL)
Scenario A ( 0%, 4.8%, 0%) 14.01 19.99
Scenario B ( 22%, 4.8%, 31%) 10.55 15.06
Scenario C ( 44%, 4.8%, 62%) 7.10 10.13
Scenario D ( 60%, 4.8%, 85%) 4.54 6.48
Scenario E ( 60% 4.8% 100%) 3.00 4.28
In the worse case scenario ( A), there is no recycling ( all lead is virgin material
and all batteries enter the waste stream), a 10.3 kilogram battery ( BSEB) and a 17.4
kilogram battery ( SSEB) emit 14 and 20 kilograms of lead, respectively. As expected,
these values are higher than the lead content of the battery ( emissions= battery weight +
manufacture loss + production loss). More realistic scenarios B and C assume moderate
recycling rates reported by Mao et al. ( Mao, Lu et al. 2006). Scenarios D and E assume
very high recycling rates as reported by electric bike manufacturers. The actual lead loss
is likely between scenario C and D.
A conservative estimate of battery life is up to 300 cycles or 10,000 kilomete