Institute of Transportation Studies at UC Berkeley

Evacuations are a primary transportation strategy to protect populations from natural and humanmade disasters. Recent evacuations, particularly from hurricanes and wildfires, have exposed three critical evacuation challenges: 1) persistent evacuation non-compliance to mandatory evacuation orders; 2) poor transportation response, leading to heavy congestion, slow evacuation clearance times, and high evacuee risk; and 3) minimal attention in ensuring all populations, especially those most vulnerable, have transportation and shelter. With ongoing climate change and increasing land development and population growth in high-risk areas, these evacuation challenges will only grow in size, frequency, and complexity, further straining transportation response in disaster situations.

Energy use and emission factors for passenger transportation modes typically ignore the total environmental inventory which includes vehicle non-operational components (e.g., vehicle manufacturing and maintenance), infrastructure components, and fuel production components from design through end-of-life processes. A life-cycle inventory for each mode is necessary to appropriately address and attribute the transportation sector’s energy and emissions impacts to reduction goals instead of allowing tailpipe emissions to act as indicators of total system performance.

The contributions of U.S. passenger transportation modes to national energy and emissions inventories account for roughly 20% of U.S. totals, mostly attributed to gasoline consumption. Furthermore, world consumption of primary energy amounted to 490 EJ in 2005 with the U.S. responsible for 110 EJ, or 21% of the total. This means that passenger transportation in the U.S. accounts for roughly 5% of global primary energy consumption annually. With a predominant fossil fuel energy base, the impacts of U.S. passenger transportation have strong implications for global energy consumption, U.S. energy security, and climate change. Furthermore, criteria air pollutant emissions from transportation (passenger and freight) are also significant, accounting for 78% of national CO, 58% of NOX, 36% of VOCs, 9% of PM2.5, 2.6% of PM10, and 4.5% of SO2 emissions. These emissions often occur near population centers and can cause adverse direct human health effects as well as other impacts such as ground-level ozone formation and acid deposition.

To appropriately mitigate environmental impacts from transportation, it is necessary for decision makers to consider the life-cycle energy consumption and emissions associated with each mode. A life-cycle energy, greenhouse gas, and criteria air pollutant emissions inventory is created for the passenger transportation modes of automobiles, urban buses, heavy rail transit, light rail transit, and aircraft in the U.S. Each mode’s inventory includes an assessment of vehicles, infrastructure, and fuel components. For each component, analysis is performed for material extraction through use and maintenance in both direct and indirect (supply chain) processes.

For each mode’s life-cycle components, energy inputs and emission outputs are determined. Energy inputs include electricity and petroleum-based fuels. Emission outputs include greenhouse gases (CO2, CH4, and N2O) and criteria pollutants (CO, SO2, NOX, VOCs, and PM). The inputs and outputs are normalized by vehicle lifetime, vehicle mile traveled, and passenger mile traveled. A consistent system boundary is applied to all modal inventories which captures the entire life-cycle, except for end-of-life. For each modal life-cycle component, both direct and indirect processes are included if possible. A hybrid life-cycle assessment approach is used to estimate the components in the inventories. We find that life-cycle energy inputs and emission outputs increase significantly compared to the vehicle operational phase. Life-cycle energy consumption is 39-56% larger than vehicle operation for autos, 38% for buses, 93-160% for rail, and 19-24% for air systems per passenger mile traveled. Life-cycle greenhouse gas emissions are 47-65% larger than vehicle operation for autos, 43% for buses, 39-150% for rail, and 24-31% for air systems per passenger mile traveled. The energy and greenhouse gas increases are primarily due to vehicle manufacturing and maintenance, infrastructure construction, and fuel production. For criteria air pollutants, life-cycle components often dominate total emissions and can be a magnitude larger than operational counterparts. Per passenger mile traveled, total SO2 emissions (between 350 and 460 mg) are 19-27 times larger than operational emissions as a result of electricity generation in vehicle manufacturing, infrastructure construction, and fuel production. NOX emissions increase 50-73% for automobiles, 24% for buses, 13-1300% for rail, and 19-24% for aircraft. Non-tailpipe VOCs are 27-40% of total automobile, 71-95% of rail, and 51-81% of air total emissions. Infrastructure and parking construction are major components of total PM10 emissions resulting in total emissions over three times larger than operational emissions for autos and even larger for many rail systems and aircraft (the major contributor being emissions from hot-mix asphalt plants and concrete production). Infrastructure construction and operation as well as vehicle manufacturing increase total CO emissions by 5-17 times from tailpipe performance for rail and 3-9 times for air.

A case study comparing the environmental performance of metropolitan regions is presented as an application of the inventory results. The San Francisco Bay Area, Chicago, and New York City are evaluated capturing passenger transportation life-cycle energy inputs and greenhouse gas and criteria air pollutant emissions. The regions are compared between off-peak and peak travel as well as personal and public transit. Additionally, healthcare externalities are computed from vehicle emissions. It is estimated that life-cycle energy varies from 6.3 MJ/PMT in the Bay Area to 5.7 MJ/PMT in Chicago and 5.3 MJ/PMT in New York for an average trip. Life-cycle GHG emissions range from 480 g C02e/PMT in the Bay Area to 440 g C02e/PMT for Chicago and 410 g C02e/PMT in New York. CAP emissions vary depending on the pollutant with differences as large as 25% between regions. Life-cycle CAP emissions are between 11% and 380% larger than their operational counterparts. Peak travel, with typical higher riderships, does not necessarily environmentally outperform off-peak travel due to the large share of auto PMT and less than ideal operating conditions during congestion. The social costs of travel range from 51 cent (in 2007 cents) per auto passenger per trip during peak in New York to 6 cents per public transit passenger per trip during peak hours in the Bay Area and New York. Average personal transit costs are around 30 cents while public transit ranges from 28 cents to 41 cents.

This dissertation was completed with Professor Arpad Horvath serving as the advisor. This document supercedes the University of California, Berkeley, Center for Future Urban Transport papers, vwp-2007-7 and vwp-2008-2. Additional project information can be found at http://www.sustainable-transportation.com.

Various theories have been proposed to describe vehicular traffic movement in cities on an aggregate level. They fall short to create a macroscopic model with variable inputs and outputs that could describe a rush hour dynamically. This dissertation work shows that a macroscopic fundamental diagram (MFD) relating production (the product of average flow and network length) and accumulation (the product of average density and network length) exists for neighborhoods of cities in the order of 5-10km2. It also demonstrates that conditional on accumulation large networks behave predictably and independently of their origin-destination tables. These results are based on analysis using simulation of large scale city networks and real data from urban metropolitan areas. The real experiment uses a combination of fixed detectors and floating vehicle probes as sensors. The analysis also reveals a fixed relation between the space-mean flows on the whole network and the trip completion rates, which dynamically measure accessibility. This work also demonstrates that the dynamics of the rush hour can be predicted quite accurately without the knowledge of disaggregated data. This MFD is applied to develop perimeter control strategies based on neighborhood accumulation and speeds and improve accessibility without the uncertainty inherent in today’s forecast-based approaches. The looking-for-parking phenomenon that extends the average trip length is also integrated in the dynamics of the rush hour.

Electric powered two-wheel bicycles, while extremely popular in China, have been recently banned by policy makers due to safety, congestion, and environmental concerns. This study investigates the tremendous growth of electric two wheel bicycles in China and compares and quantifies their environmental and safety impacts with the impacts of alternative modes of transportation, such as traditional bicycles, public transportation, or personal cars. The research also analyzes the benefits of electric two wheel bicycles, such as increased mobility and access to opportunities. Additionally, the author looks at the impacts of prohibiting the use of electric bicycles. Two case studies are carried out in Kunming and Shanghai, cities that have similar electric bicycle use but with very distinct differences.

Two freeway bottlenecks, each with a distinct geometry, have been investigated in an effort to understand traffic conditions leading to capacity losses (i.e., breakdown). One bottleneck is formed by a horizontal curve and the other by a reduction in travel lanes. These bottlenecks are shown to exhibit breakdowns after queues form immediately upstream. The vehicle accumulations that arise near these bottlenecks are shown to be good proxies for the mechanisms that trigger breakdowns. Evidence is provided to show that these losses can be recovered, postponed or even avoided entirely by controlling the accumulations. An algorithm for estimating vehicle accumulations has been developed in this dissertation. This algorithm? estimates are obtained from the counts made by ordinary detectors (e.g. inductive loops) placed in series. The accumulations estimated are those that arise on the intervening (freeway) segments between the detectors. These estimates can be obtained in real-time at short intervals of a second or so. The systematic errors (i.e., bias) that invariably arise in detector counts are automatically corrected when traffic is freely flowing. The algorithm is thus well suited for monitoring accumulations near a bottleneck prior to capacity drops and the estimates it furnishes can, in turn, dictate control actions (e.g. metering rates) that prolong higher outflows from the bottleneck. The estimates that the algorithm furnishes can also be used for incident detection and delay estimation.

This research describes field studies of how on-ramp metering can increase the capacity of freeway merges. Some effects of on-ramp metering have been known for a long time. We have known that on-ramp metering can 1) increase freeway flow and speed upstream of a merge; and 2) reduce system-wide delay by alleviating gridlock-causing queues that have blocked off-ramps. However, past studies have not conclusively shown that on-ramp metering can increase the maximum outflow (capacity) of freeway merges. The experiments conducted in the present study verify that on-ramp metering can increase freeway merge capacities. Detailed traffic data collected from videos for more than 30 rush periods at two merge bottlenecks unveil six major research findings: 1) merge capacity diminishes after merges became active bottlenecks; 2) the mechanism of "capacity drop" has been identified and was found to be reproducible across all days and it both sites. By metering the on-ramp in certain strategic ways, the capacity drop mechanism can be 3) reversed; and 4) even averted; 5) such metering strategies can be fully automated using loop detector measurements; and 6) control strategies other than ramp metering also hold promise for increasing merge capacities. These findings provide much-needed information concerning how to control freeway traffic. They also offer basis for more realistic theories of merging traffic flow.

Transit properties in the USA have historically experienced loss of market share and low levels of farebox recovery. They resorted to service expansion to maximize subsidies. Experience suggests that: (a) fare increases have not had the desired effect; (b) fare reductions can boost ridership but can also reduce revenue and increase subsidies. The challenge lies with the adoption of such strategies as deep discount group pass programs that can produce more marginal revenue than cost. Deep discount transit pass programs provide groups of people with unlimited-ride transit passes in exchange for a contractual payment for or on behalf of pass users by an employer or other organizing body. Although successes of deep discount group pass programs are documented, there is substantial skepticism toward their wide-scale deployment because transit management perceives them as "special treatments" or "favors" to participants. Management fears such perception could raise questions about equity because they fail to see the fundamental difference in the fare structure of the "group pass" from individual ticket purchases. Group passes operate in a manner analogous to insurance programs. The deep discount program cases studied consistently revealed either higher revenues per boarding than the system-wide average or higher total revenues from target markets with the program than without it. Employment-based programs yielded the highest net revenues to operators. Although agencies recognize the factors for price determination, research reveals that no systematic methodology exists and pass prices are largely determined by watching what others have done. This dissertation has developed a methodology to aid operators in determining deeply discounted but favorable pass prices. The methodology considers: revenue lost from existing riders at prevailing fares; level of patronage in the primary location of transit use; any additional costs necessitated by the program; attractiveness of program terms to participants; and a policy goal of increasing operating revenue. The methodology permits the investigation of alternative objective functions and thus can serve as a common tool for transit agencies, employers and other constituents who may choose to maximize or minimize either the price of the pass or the number of participants subject to sets of constraints.

This dissertation presents dynamic stochastic optimization models for Air Traffic Flow Management (ATFM) that enables decisions to adapt to new information on evolving capacities of National Airspace System (NAS) resources. Uncertainty is represented by a set of capacity scenarios, each depicting a particular time-varying capacity profile of NAS resources. We use the concept of a scenario tree in which multiple scenarios are possible initially. Scenarios are eliminated as possibilities in a succession of branching points, until the specific scenario that will be realized on a particular day is known. Thus the scenario tree branching provides updated information on evolving scenarios, and allows ATFM decisions to be re-addressed and revised. First, we propose a dynamic stochastic model for a single airport ground holding problem (SAGHP) that can be used for planning Ground Delay Programs (GDPs) when there is uncertainty about future airport arrival capacities. Ground delays of non-departed flights can be revised based on updated information from scenario tree branching. The problem is formulated so that a wide range of objective functions, including non-linear delay cost functions and functions that reflect equity concerns can be optimized. Furthermore, the model improves on existing practice by ensuring efficient use of available capacity without necessarily exempting long-haul flights. Following this, we present a methodology and optimization models that can be used for decentralized decision making by individual airlines in the GDP planning process, using the solutions from the stochastic dynamic SAGHP. Airlines are allowed to perform cancellations, and re-allocate slots to remaining flights by substitutions. We also present an optimization model that can be used by the FAA, after the airlines perform cancellation and substitutions, to re-utilize vacant arrival slots that are created due to cancellations. Finally, we present three stochastic integer programming models for managing inbound air traffic flow of an airport, when there is adverse weather impacting the arrival capacity of the airport along with its arrival fixes. These are the first models, for optimizing ATFM decisions, which address uncertainty of future capacities of multiple NAS resources.

Accurate facility deterioration models are important inputs for the selection of Infrastructure Maintenance, Repair, and Reconstruction (MR & R) policies. Deterioration models are developed based on expert judgment or empirical observations. These resources, however, might not be sufficient to accurately represent the performance of infrastructure facilities. Incorrect deterioration models may lead to wrong predictions of infrastructure performance and selection of inappropriate MR & R policies. This results in higher lifecycle costs. Existing infrastructure MR & R decisionmaking models assume that deterioration models represent the real deterioration process of infrastructure facilities. This assumption ignores the uncertainty in empiricallyderived facility deterioration models. This dissertation presents a methodology for selecting MR & R policies for systems of infrastructure facilities under uncertainty in the deterioration model parameters. It is assumed that inspections reveal the true conditions of facilities. Based on the inspection results, the deterioration model parameters can be updated to express the deterioration process more accurately. It is expected that more appropriate maintenance policies will be selected as a result. In the first part of this dissertation, it is assumed that facility inspections are performed at the beginning of every year. The model parameters are updated and MR & R policies are selected every year using the updated deterioration models. In the second part, the assumption is relaxed and alternate inspection frequencies are considered. In this case, the updates of the model parameters and the selection of optimal MR & R policies are executed only after an inspection. The results of the parametric analyses demonstrate that updating the deterioration models reduces the expected system costs. The results also show that relaxing the facility inspection frequency can reduce the total costs further.

Thirty years ago scholars first presented convincing evidence that local officials use biased travel demand forecasts to justify decisions based on unstated considerations. Since then, a number of researchers have demonstrated convincingly that such forecasts are systematically optimistic-often wildly so-for reasons that cannot be explained solely by the inherent difficulty of predicting the future. Why do modelers-professional engineers and planners who use quantitative techniques to predict future demand for travel and estimate its potential impact on built and proposed transportation facilities-generate biased forecasts and otherwise tolerate the misuse of their work? On initial consideration, it is tempting to surmise that corrupt modelers are responsible for biased forecasting. Indeed, corruption is the most common explanation of forecasting bias and tales of mercenary behavior are all too common in the field. Data from in-depth interviews with twenty-nine travel demand forecasters throughout the United States and Canada, how-ever, suggest new and different ways to understand the suspect behavior of transportation planning professionals. Those most likely to introduce bias and invite misuse of travel forecasts assume that their technical analyses have little, if any, impact on policy making. For many, this leads to disillusionment and requires responses to cope with feelings of marginalization. Others, untroubled by their apparent lack of influence, are complacent and need ways to avoid the ethical questions of practice. Both types of practitioners circumscribe professional roles and rely on the self-deceptive strategies of evasion and excuse making to mute their own disquieting realities that undermine positive concepts of self. The disillusioned wish not to see that they do not matter and the complacent that they do. Bias and misuse seem to be the unintentional byproducts of these attitudes. Beyond enhancing the understanding of the systemic failures of travel demand modeling, this research suggests practicable steps to reform and outlines an agenda for future work. Attention to these matters is important, not just to avoid expenditures on projects and programs that cannot be justified on the basis of sound utilitarian calculations, but also to restore and preserve the credibility of a profession.