Prediction/Causality Tradeoffs and Data Size Issues in Transportation Modeling Webinar

Fred Mannering

Prediction/Causality Tradeoffs and Data Size Issues in Transportation Modeling: The Example of Highway-Safety Analysis Paradigm shift towards smart and healthy cities —systems innovation at the nexus of transportation, environment, and public health

September 24, 2021 at 12:00 p.m. ET

Webinar link: Join Microsoft Teams Meeting (Conference ID: 495 510 442#)

H. Oliver Gao
Professor, Director,
Center for Transportation, Environment and Community Health (CTECH)
Cornell University

Abstract: Transportation-related air pollution, GHG emissions and energy problems are a significant issue in the U.S. and across the world. The World Health Organization estimates that urban air pollution causes 200,000 deaths per year worldwide. How do we meet the mobility needs in urban transitions without sacrificing environment sustainability and global health? In this talk Dr. Gao takes a systems approach to study the nexus of transportation and environment/health systems. We’ll examine the broad spectrum and necessary depth of models, tools, and insights for trans-disciplinary systems research in support of integrative transportation, environment, and health systems planning and finance/policy innovation such as public private partnership.

Bio: Dr. Gao, Director of the Center for Transportation, Environment, and Community Health (CTECH), is a Professor with the School of Civil and Environmental Engineering at Cornell University. Gao is an international leading expert in urban infrastructure systems research and policy innovation for healthy living in sustainable communities. His research focuses on modeling and development of systems solutions for sustainable and intelligent infrastructure systems, low carbon and low emission transportation systems, and human-centered design for environment and public health. Before joining Cornell, Gao was a QUANT in the mathematical and econometrical modeling division at the Rohatyn Group, LLG, a Wall Street hedge fund specializing in emerging markets including the BRIC countries.

Long commutes, home crowding tied to COVID transmission

traffic cars driving in city

By Blaine Friedlander

Long commute times and household crowding may be good predictors for a higher number of transmissible coronavirus cases in metropolitan settings, according to Cornell urban planning, architectural and public health researchers, in a July study published in the journal Buildings and Cities.

Neighborhoods that had populations with predominantly longer commute times to work – from about 40 minutes to an hour – were more likely to become infectious disease hotspots, the research said.

“We are trying to determine how the built environment influences coronavirus propagation,” said senior author Timur Dogan, assistant professor of architecture in the College of Architecture, Art and Planning.

“We found that high residential density and high percentage of people commuting by public transit do not relate to a higher COVID-19 case rate,” Dogan said. “Household overcrowding and longer commute times appears to impair the pandemic resilience of individual families, medically vulnerable communities and cities, as a whole.”

The study, “Urban Design Attributes and Resilience: COVID-19 Evidence from New York City,” was published July 6, to offer guidance for short-term responses in the safe recovery from the COVID-19 crisis, as well as long-term urban design and planning decisions for a resilient, inclusive and sustainable urban environment in future public health emergencies.

In-home crowding and urban density are two related, but different concepts, when considering pandemic-resilient design and planning. Crowding indicators, which include the number of units per building and the number of occupants per room, correlate with coronavirus transmission, according to the paper.

Neighborhood residential density, on the other hand, was not related to the daily COVID-19 case rate in New York City, as high density often entails other beneficial urban features that are advantageous for the pandemic resilience.

For example, dense and well-mixed neighborhoods can mean that job sites are closer to home, giving rise to a shorter commute time and improved mobility conditions – such as accessible travel modes like walking or riding a bike to work, Dogan said.

“High-density neighborhoods aren’t necessarily bad from a disease transmission perspective,” said Dogan. “A well-mixed neighborhood in a city could be beneficial.”

The group used ZIP code tabulation area data, and then combined it with other available urban information, to determine how population density and crowding affected infection rates, and how the spatial distribution of points of interest – such as grocery stores, shopping centers and parks, for example – impacted infection rates.

The points of interest-related mobility data in this study was derived by a computer-aided design software utility called Urbano, developed by lead author Yang Yang, a doctoral student in systems engineering and a design researcher at Dogan’s Environmental Systems Lab.

“The Urbano software assists with collecting, simulating, and analyzing urban mobility data,” Yang said. “It allows mobility-aware decision-making for designers and planners in building a sustainable and resilient city.”

Said Dogan: “This is where we start the idea of the 20-minute city, a concept where a person can fulfill all the daily errands, work and daily needs within a 20-minute walk or bike ride.”

“This kind of urban design paradigm promises benefits that make our cities more livable, sustainable and resilient,” he said. “Professional urban planners say that active mobility is a healthy thing to do. If we can reduce vehicle traffic, we can reduce pollution and reduce energy demand, we can get a healthier population.”

Dr. Nathaniel Hupert, associate professor of population health sciences, at Weill Cornell Medicine and co-founder of the Cornell-Oxford COVID-19 International Modeling (CoMo) Consortium, said, “This research helps us to see what features of the built environment might be beneficial or detrimental to health during this and future pandemics.”

In addition to Dogan, Yang and Hupert, co-authors are Yihong Li, a faculty member in the Cornell Master of Public Health Program; and Katharina Kral, a lecturer on architecture (AAP). Dogan, Kral, Li and Hupert are fellows at the Cornell Atkinson Center for Sustainability.

Group members Samitha Samaranayake, assistant professor in the College of Engineering, and Nikhil Saraf ‘22, helped to develop the Urbano software.

The research was funded by a COVID-19 rapid response grant from Cornell Atkinson.


How to reach 50 percent zero-emission vehicles


The Biden administration has set a goal for 50 percent of cars and light trucks sold in 2030 to be zero-emission vehicles. This is a critical step in the fight against climate change. The transportation sector accounts for nearly 30 percent of U.S. greenhouse gas emissions, and more than half of transport emissions are from light-duty vehicles. With electric vehicle prices falling and more models being introduced, the transition to electric vehicles is within grasp — but only if it is supported by the right mix of policies.

Two main types of policy support are on the table. The bipartisan infrastructure deal moving through Congress includes $7.5 billion for charging stations. The $3.5 trillion budget reconciliation package includes a to-be-determined sales incentive aimed at consumers, like (but different than) the current electric vehicle (EV) tax credit. The sales incentive is likely to have a high price tag: light duty vehicle sales average around 16 million annually, so if one-fourth of those sales are EVs and the tax credit is the current $7500, the annual fiscal cost would be $30 billion — a sum that increases with the EV sales share.

The question for Congress, then, is how to allocate funds across EV rebates and cost-shares for charging stations.

In principle, one can argue for both programs. Subsidizing EV sales will tempt consumers to try an EV and will provide a public benefit through expediting the decarbonization of light duty vehicles.

Subsidizing charging stations addresses a different challenge: the chicken-and-egg problem that few EVs beget few charging stations and vice versa. There are more than 150,000 gasoline stations in the US, but fewer than 5,000 level-three (“Fast DC”) chargers. While there are roughly 40,000 public level-two stations, which work with all EVs, they can take 8 hours or more for a full charge.

To examine the tradeoff between public spending on charging stations and on rebates, we undertook an economic modeling exercise that simulates the battery EV and charging station markets under different policy scenarios. We varied the size of the subsidies and total program budgets for both vehicles and charging stations. From this exercise, we obtain the share of battery EVs, the reduction in greenhouse gases, total governmental outlays, and the program costs measured in the standard units of dollars per ton of CO2 emissions abated.

Our main finding is that charging stations are key to the rapid electrification of this sector — especially if working under fiscal constraints. For example, if the charging station budget is fixed at $7.5 billion, EV rebates of $11,000 (roughly as proposed in the Clean Energy for America Act) would achieve approximately a 45 percent EV share in 2050, at a cost of $400 billion. In contrast, increasing charging station spending to $30 billion while halving the per-car rebate would achieve a 2030 EV share of 50 percent at a fiscal cost of $170 billion. 

These results strongly suggest that the $7.5 billion for charging stations in the bipartisan infrastructure deal is not enough. The results also make sense: A 2020 survey showed that driving range and the availability of public charging are key determinants of EV owners’ satisfaction. If you live in an apartment building or cannot otherwise install your own level-two charger, owning an EV currently simply isn’t an option.

Getting this right can help the U.S. catch up to other major economies in electric vehicle adoption. In 2020, the share of EVs including both battery EVs and plug-in hybrids among new vehicle sales was only about 2 percent in the US, compared to over 5 percent in China and over 10 percent in many European countries (a whopping 70 percent in Norway with most of them being battery EVs).

Our analysis underscores the importance of focusing on charging infrastructure. The current EV tax credits have largely subsidized the well-to-do, but jump-starting charging stations helps all consumers. While the costs of EVs are still higher than traditional gasoline-based vehicles today, as battery costs fall and EVs get closer to cost-parity, the main hurdle from large-scale adoption is going to be a lack of recharging infrastructure. Removing this hurdle requires a change in policy focus if we are to achieve deep EV penetration by 2030.

Christopher R. Knittel, Ph.D., is the George P. Shultz professor of applied economics at the Massachusetts Institute of Technology (MIT).

James H. Stock, Ph.D., is the Harold Hitchings Burbank professor of political economy, Faculty of Arts and Sciences and member of the faculty at the Harvard Kennedy School.

Shanjun Li, Ph.D., is a professor of applied economics and policy and holds the Kenneth L. Robinson Chair in the Dyson School of Applied Economics and Management at Cornell University.



Transportation innovations could boost public health

By David Nutt

Researchers say the future of transportation will be shaped by three “revolutions” – vehicle electrification, driverless cars and ride-sharing – that could result in fewer automobiles on the road, less fossil fuels extracted from the Earth and less pollution in the air. While the environmental gains may seem self-evident, the health benefits are difficult to quantify.

Now for the first time, a Cornell-led team has used transdisciplinary systems modeling to calculate those health benefits in the United States. By 2050, these innovations could potentially slash petroleum consumption by 50% and carbon dioxide emissions by 75% while simultaneously preventing 5,500 premature deaths, with an annual savings of $58 billion.

“There are all these important emerging trends in the development of transportation, and they are becoming a reality in the near future,” said Oliver Gao, the Howard Simpson Professor of civil and environmental engineering in the College of Engineering, who led the project.

“Have you ever thought about what all these revolutions mean for your health, for our climate, and for our environment, and for our energy systems?” Gao said. “These externalities don’t necessarily come directly in the mind of the general public, the travelers, or even the decision-makers.”

The group’s paper, “Shared Use of Electric Autonomous Vehicles: Air Quality and Health Impacts of Future Mobility in the United States,” published June 26 in Renewable and Sustainable Energy Reviews. The paper’s lead author is former postdoctoral researcher Shuai Pan.

“It is worthwhile to understand the effectiveness of these mitigation strategies, as deep de-carbonization is needed in the transportation sector,” Pan said.

Co-authors include Lewis M. Fulton from the University of California, Davis, and Yunsoo Choi and Jia Jung from the University of Houston. The research was supported by the U.S. Department of Transportation’s Center for Transportation, Environment and Community Health, and by Nanjing University of Information Science and Technology.

While previous studies have looked at certain facets of transportation innovation, such as the impact of electric vehicles on fuel usage and emissions, this is the first time anyone has employed a transdisciplinary systems approach that factored in human health and the associated economic benefits, according to Gao.

Gao’s systematics research group – which uses modeling to understand complex global challenges in engineering, business, societal well-being and sustainability – is uniquely positioned for such a task.

“A transportation engineer cannot address these questions,” Gao said. “Environmental science cannot address these questions. A health researcher cannot address these questions. However, this transdisciplinary group can do this.”

Pan, Fulton and Gao built an integrated assessment system that included a technical-economic mobility model, a chemical transport model and a health impact assessment tool. Then they projected the vehicle stocks, distance traveled, energy usage and carbon dioxide emissions in the continental U.S. through 2050, and quantified the impacts of changing emissions on concentrations of fine particulate matter in the atmosphere, as well as the ensuing health and economic benefits of populations in 10 major metropolitan areas.

Their simulations show that, depending on how widely the three “revolutions” are adopted, reductions in emissions from passenger travel could prevent between 2,300 and 8,100 premature deaths annually in the U.S. in 2050.

The largest number of prevented deaths coincided with large metropolitan areas, such as Los Angeles and Chicago. At the state level, California, Texas, New York, Ohio and Florida would see the largest decreases in premature mortality.

The associated economic benefits could range from $24 billion to $84 billion annually.

The study hangs on a number of assumptions and uncertainties. After all, driverless cars are not yet commercially available, and sales of electric vehicles lag far behind conventional gas guzzlers.

“Another key finding is that for carbon mitigations and health benefits, vehicle electrification is by far the most important piece, followed by shared mobility (ride-sharing) and then automation, ” Pan said. “The net energy impacts of self-driving vehicles are highly uncertain and automation alone may not dramatically affect energy use, emissions or vehicle-related pollution. ”

A complicating factor is that the efficiency improvement and projected cost reduction from automation could actually lead to increased travel and offset other gains.

“If we automate the vehicles, you might make the transportation system more efficient, but probably more people will travel longer distances,” Gao said. “So there is a balance, there is a trade-off.”

The study concludes that policymakers can help encourage the transition to electric vehicles and boost ride sharing, for example, by issuing tighter fuel economy standards, creating economic incentives for shared mobility and investing in charging infrastructure and technological developments.

A future of autonomous flying taxis

Of course, actually creating such transportation innovations is not possible without first determining their viability.

Another research project from Gao’s lab – published July 6 in Transportation Research Part A: Policy and Practice – explores the feasibility of an airport shuttle service that uses autonomous flying taxis as a means to mitigate urban congestion. The paper’s lead author is Emily Lewis ’20.

“While you are stuck in traffic from JFK [International Airport] to Manhattan, have you ever thought, oh, I wish I could be a bird, just to fly there. Actually, that dream is not too far away,” said Gao, who directs Cornell’s Center for Transportation, Environment and Community Health. “But how do you even architect a whole system, from the technology to market prediction and to operation? Would such an idea make economic sense at all?”

The study focuses on the concept of urban air mobility – essentially a transportation service for low-altitude airspace in metropolitan areas that features autonomous unmanned aerial vehicles.

Gao’s team – which included co-authors Jesse Ponnock ’20, Qamar Cherqaoui ’20, Scott Holmdahl ’20,Yus Johnson ’20 and Alfred Wong ’20 – focused on the three busiest airports in the U.S.: Atlanta, Los Angeles and Dallas.

They used a holistic, system-architecture analysis to identify each area’s key stakeholders and the goals that meet their needs, such as fleet management, infrastructure, traffic control, safety, user experience, financial viability and performance. The modeling also took into account the relationships between annual profit, mean time between safety incidents, upfront costs and the number of passengers shuttled per day.

“Because of its geographic, meteorological and also demand factors, Los Angeles turns out to be the best case for a pilot city,” Gao said.

The analysis identified wealthy commuters, long-distance commuters, business executives, event attendees, emergency transportation and vacationers as potential early adopters of an air mobility system.

What would such a system actually look like from a passenger’s perspective? It might not be too different from the ride-sharing services of today. The analysis recommended the system use FIFO (first in, first out) queuing and a smartphone interface for passengers, which may sound familiar to anyone who has ever hailed an Uber on their phone.

Also recommended: a hybrid energy source that incorporates electric energy for the autonomous vehicles.

But vehicles and apps are only part of it. For an air mobility system to become a reality, it would need the infrastructure to support it.

“This is not actually as mature as electrification or even automation,” Gao said. “This is even further away down the road. We are not comparing urban air mobility to other modes or arguing this is a better mode. We’re just saying that now, given the interest, first you need to be able to architect this. And then you will have a better sense about cost.”

Cornell Chronicle: ‘Transportation innovations could boost public health”