The Self-Driving Dilemma
A Google self-driving car is seen on display at Google's I/O conference in Mountain View, Calif. Eric Risberg/AP
Autonomous vehicles could spark a cleaner, cheaper urban mobility revolution—or they could make it tougher to combat sprawl, congestion, and climate change.
The promise of autonomous cars has struck an especially jubilant chord with a chorus of futurist urban thinkers. The big transformative hope: We can break the death grip of car-centric urban design and planning, which has been something of a disaster for most American cities in the 20th century. In the near future, self-driving cars will simply circulate through cities, freeing road space and liberating millions of acres of parking lots for more useful purposes. Combine that with the ongoing electrification of the vehicle fleet, and it might look as if we are nearing an urban transportation utopia.
But the dream of cheap, clean mobility in cities might run up against some harsh realities—soaring energy consumption, supercharged sprawl, and intensified traffic congestion—if AVs are simply deployed to encourage more driving.
That’s one message from a new report prepared by the University of California Davis’s Institute of Transportation Studies and the Institute for Transportation and Development Policy (ITDP), a nonprofit organization that develops bus rapid transit systems and promotes environmentally friendly urban planning. They’ve been crunching the numbers on how to avert warming the planet with carbon emissions while also reducing gridlock and increasing mobility.
The report looks at three possible scenarios for vehicle use by 2050 and compares their energy demands. Option one: We continue with privately owned internal-combustion cars the way they are. Or, there’s a “two revolutions” model, where both electric and automated vehicles come into common use by 2030 and 2040. Then there’s the triple-revolution scenario, which introduces widespread ride-sharing by 2030, as explained by this handy infographic.
First, the energy and emissions angle: If we continue using gas engines in vehicles, we’ll dump 4,600 megatons of CO2 in the atmosphere by 2050. The second scenario combines electric and autonomous vehicles and results in 63 percent fewer emissions with 1,700 megatons of CO2 total. That math figures in the fact that the combination of two technologies will not reduce the 2.1 billion cars expected to be on the road in 2050. By the ITDP’s estimates, introducing AVs might increase vehicle travel by 15 to 20 percent.
The press release for the study outlines how many megatons each big global emitter’s urban vehicles stand to contribute, emissions-wise, by 2050.
In implementing the Paris Agreement, countries need to cut their CO2emissions in half in order to achieve the goal of preventing a 2-degree Celsius increase in global temperature. Consider those changes in the context of the United States in 2015: The EPA estimates that the United States was responsible for about 659 gigatons of carbon (a gigaton is 1,000 megatons) with transportation overall producing about 27 percent of those emissions compared to other economic sectors.
gas emissions by economic sector
CityLab asked lead author Lewis Fulton, who co-directs the Institute of Transportation Studies’s Sustainable Transportation Energy Pathways (STEPS) program at UC Davis, how much urban transportation factors in this climate change equation. “Cars in cities are like one-tenth of [the transportation] number,” Fulton says. “Cars in general, is probably one-fifth of transportation’s carbon footprint.”
One barrier to getting that number down has been the slow adoption of electric vehicles. The average age of a U.S. car is over 11 years, and prices on new electric vehicles remain generally higher than traditional ones. Right now, the United States is on target for about one million all-electric cars on the road, a small fraction of the overall fleet. “That's still only one million out of close to 300 million vehicles,” he says. “It's going to take more than 15 to 20 years to get the full stock to turn over. If we somehow miraculously get very high sales shares of electric cars by 2030, we still might not get rid of internal combustion engine vehicles until 2050.”
To reduce transportation’s carbon impact by 80 percent, down to 700 megatons by 2050, we’ll need to do big-time ride-sharing, according to Jacob Mason, a transportation researcher at ITDP who co-authored the report. And we’ll need to continue to rely on mass transit modes—think, yes, buses and subways. “A four-person shared UberPool or Lyft Line is not going to replace 40 people on a bus,” Mason says. “It certainly won’t ever replace a whole fleet. Sharing a car was never going to replace 1,000 people on a train or a bus rapid transit system.”
One school of thought on autonomous vehicles holds that they’ll increase traffic and sprawl: Without expensive fuel and the soul-gnawing investment in time wasted commuting behind the wheel, people will feel the pain of driving less, live even further away from urban centers, and basically use robot-cars for all manner of frivolous pursuits. In fact, AVs might produce a rebound effect, increasing both energy consumption and traffic congestion. Once we hit the coveted Stage 5 automation, we could even see “zero occupancy” driving—empty vehicles roaming the city doing errands for their masters, without people inside. (Forgot your lunch? No problem—send your car home to pick it up!)
“With a much lower cost of traveling,” Mason says. “People [will be able to live] farther away with longer commutes, and there will also be more low-occupancy vehicle trips.” Once driving these vehicles becomes cheaper, the challenge will become encouraging people to share instead of driving alone.
The model in the study also assumes that a transition to cleaner energy will be slower if electrification comes through individual vehicles. “It will be much more expensive to bring decarbonization to the grid,” Mason says. “It will be harder to make sure that all of that energy production is through carbon-neutral sources, so we assume there is a much lower likelihood that we’re going to decarbonize all of the energy production.”
Even if each individual’s driving costs drop, the external costs that society pays may continue to grow. ”One of our big policy recommendations is setting pricing for travel that corresponds to the benefits and costs to society,” Mason says. “We [now] subsidize driving in a number of ways, from fuel subsidies to free parking and free streets. We haven’t priced any of those costs into our current transportation system or urban development.”
And even in the electric AV future, the researchers say, cities will need to move people via mass transit—don’t fill in those subways. But there could be ways to add flexibility to encourage more sharing. “What’s often talked about is microtransit or a van pooling,” Mason says. “With vehicle automation and better technology, those could really supplement the existing public transport network in lower-demand areas. It is about getting where you need to go in a reasonable amount of time and having cities that are functional and sustainable in the big sense—socially, environmentally, and economically.”
Fulton sounds a big note of caution on the timing of these multiple revolutions, whatever shape they take. “It could take 30 or 40 years to phase in electric cars and reach this driverless car world where everything is clean and automated in a perfect symphony,” he says. “But we have to figure out how to live for decades where you have many kinds of vehicles out there, unless we made some kind of decision as a society to accelerate the change.”