In 2017, transportation became the largest single source of greenhouse gas emissions from the United States. Globally, the 2014 Intergovernmental Panel on Climate Change report found that, without far more aggressive policies, “transportation emissions could increase at a faster rate than emissions from other energy end use sectors” reaching 12 Gt CO2-eq/year by 2050 (Sims et al., 2014). The overwhelming challenge of combatting these emissions is made far more difficult by the fact that so little is known about transportation behavior.
To use a cliché – if we can’t measure it, we can’t manage it. And transportation must be managed if we are to avoid the most catastrophic consequences of climate change. In this dissertation, I propose that better data collection is necessary to achieve reduction of transportation-related emissions. Happily, advances in technology make this more feasible today than at any time in the past. The costs of massive computing resources have gone down, the world is swarming with mobile devices like smartphones and connected cars collecting massive (if messy) amounts of data, and new techniques in data science and machine learning have emerged to help get clean answers out of all that data in a privacy-appropriate manner. In some cases, these new techniques will displace older ones. In other cases, the old ways have inherent advantages. In other cases yet, fusing new and old techniques will yield the most productive results.
In Chapter One, I lay out a framework to organize the types of transportation behavioral data that must be collected regularly to adequately measure and manage transportation’s impact on climate. This builds on classic climate impact frameworks, adapting them to the particular measurement challenges presented by transportation. In Chapter Two, I provide a history of US transportation data collection since World War II as well as a review of traditional, modern, and emerging transportation data collection technologies. I then map each technology onto each behavioral data collection need identified in Chapter One, matching each behavior to the best respective data collection technique.
Chapters Three and Four provides an example of analysis done using the traditional data collection techniques, notably Household and Commercial Travel Surveys, to explore changes in PMT related to shopping and retail freight since 1969, as well as freight for fuel transportation. They demonstrate and take advantage of the key benefits of traditional techniques: that they go back in history, that they collect clearly stated trip purposes, vehicle occupancies, demographics (including gender, an important demographic but particularly difficult to deduce from the new data collection sources), trip distances, chaining behavior, commodities logged, and more. As it turns out, these benefits are critical: the historical trends of the past 40 years allow behavioral insight that would not have been possible with a shorter term study, and gender dynamics are key to understanding the behaviors at hand.
However, the analysis in Chapters Three and Four also highlights some of the key limitations of survey-based analysis. The fact that data was only collected every five to ten years severely limits the analysis, such as limiting the exploration that can be done on the impacts of the Great Recession. In addition, fallibilities in human memory are especially pronounced in short trips, trip chains, and non-work related trips, all of particular importance to this study.
Chapters Five lays out theoretically, and then Chapter Six demonstrates via case study in India, how personal GPS diary devices can be used to log detailed data about individual trips. It demonstrates the key benefit of this data – highly individualized characteristics. Taking the example of vehicle electrification, this chapter demonstrates two ways such granular data is important: in one example, such data to give feedback to an individual to influence their car buying behavior. In the second, the granularity found with this new data collection techniques reveals the importance of highly localized policy making and emissions modeling based on driving patterns in different cities.
Chapter Seven uses the emerging technology of mass amounts of locational data, collected passively via smart phones, to explore how urban density at home and work interacts with total, work-related, and non-work-related miles driven. This demonstrates the great strength of this type of data – massive sample size combined with high spatial granularity and longitudinal data collection. These strengths enable the analysis at statistically meaningful scale of patterns across many geographies, individuals, and times of year. Thus, this data can shed light on questions about the relationship of density and miles travelled which previously have not been answered conclusively due to data constraints.
|Advisor:||Kammen, Daniel M|
|Commitee:||Callaway, Duncan, Deakin, Elizabeth|
|School:||University of California, Berkeley|
|Department:||Energy & Resources|
|School Location:||United States -- California|
|Source:||DAI-A 81/9(E), Dissertation Abstracts International|
|Keywords:||Big data, Climate change, Electric vehicles, Locational data, Transportation data, Urban density|
Copyright in each Dissertation and Thesis is retained by the author. All Rights Reserved
The supplemental file or files you are about to download were provided to ProQuest by the author as part of a
dissertation or thesis. The supplemental files are provided "AS IS" without warranty. ProQuest is not responsible for the
content, format or impact on the supplemental file(s) on our system. in some cases, the file type may be unknown or
may be a .exe file. We recommend caution as you open such files.
Copyright of the original materials contained in the supplemental file is retained by the author and your access to the
supplemental files is subject to the ProQuest Terms and Conditions of use.
Depending on the size of the file(s) you are downloading, the system may take some time to download them. Please be