Social distancing on mass transit
CSE researchers study COVID-19 risk on buses
Imagine a quarter of a million people riding public transit in the Twin Cities. That was a typical weekday pre-pandemic. Today, ridership on buses and trains is down—not just in Minneapolis and St. Paul, but around the world—because concerns of catching the novel coronavirus loom large. So, how can transit authorities keep commuters safe? That’s the question fueling Alireza Khani’s current research.
Khani, an assistant professor in the Department of Civil, Environmental, and Geo- Engineering (CEGE) at the University of Minnesota, has developed a mathematical simulation that can analyze the relative risk of traveling on buses and inform decision-making for public transportation agencies. The study is funded by the Office of the Vice President for Research COVID-19 Rapid Response Grants and the CEGE department.
“We know that COVID-19 transmits in close proximity,” he said. “If people are close to each other and cough, the virus gets transmitted more easily. So, it’s important to social distance. Some areas where social distance is violated are confined areas such as restaurants, grocery stores, offices, and public transportation.”
In 2017, Khani partnered with researchers at the University of New South Wales in Australia to create an epidemiological model that identifed transit system hot spots and target them for outbreak control. When the novel coronavirus hit closer to home, Khani decided to recalibrate the simulation component he developed for that research. He’s been working with Metro Transit since spring on how its service could be optimized to address current challenges.
“We had weekly meetings May to June, and it was an excellent collaboration,” he said. “Metro Transit was willing to share APC data that it collects for other purposes.”
APC, or automatic passenger count, devices, which are located on the doorway of a bus, only give a researcher the number of passengers who get on and off the bus. They don’t identify personal information or connect the boarding and alighting locations of individual riders. Having a rider count was enough, though, for Khani and his graduate assistant, Pramesh Kumar, to do what they needed.
Kumar, a Ph.D. candidate in civil engineering, and Khani combined that information with other data from Metro Transit—bus size, passenger capacity, and run frequency (how often the bus runs). They then built an algorithm for visualizing passenger movement, which allowed them to estimate the relative risk of potential virus exposure along a transit route. Risk, in this case, was defined as the number and duration of contacts a person has with other riders during a ride.
Khani shared their findings in a webinar on the “Impacts of the COVID-19 Pandemic on Minnesota's Traffic and Transit Networks” hosted by the University of Minnesota’s Center for Transportation Studies last month.
In his presentation—based on data from Metro Transit’s Route 5 (see heat map above)—he said, “if a passenger rides the busy section of their route, for example, getting on the bus before the downtown area and getting off after downtown, the risk is relatively higher. However, riding the bus on other sections [where it’s not busy and] where the busload is below 15, the risk is lower.”
Since mid-April, Metro Transit has limited the number of passengers on its buses to help customers maintain social distancing. No more than 10 people are allowed on a 40-foot bus, and only 15 can board the 60-foot bus.
The College of Science and Engineering researchers also calculated the risk another way—and came to the same conclusion. Their colleagues at Metro Transit wanted to know if the length of a bus ride played a role; so, Khani and Kumar tested a typical bus ride of five to 10 stops.
“No matter how many stops you ride, the number of people that you encounter is pretty much the same,” Khani said. “But the duration of contacts will increase, which can increase the relative risk.”
Moreover, when higher passenger loads are allowed on a bus, it increases the relative risk—for both passenger and the driver.
“So, it’s really important that we keep the load below the recommended threshold,” he noted.
Hop on, hop off safely
The idea of limiting the number of people on a bus, subway or light rail is similar to what airlines are doing. However, one effect of cutting capacity on public transit is longer wait times. Riders would be forced to wait for the next bus, or maybe even two or three busses.
“Most of the commuters during the pandemic are low-income, under-represented people who are at a higher health and financial risk,” Khani said. “Our main goal is to provide transit to people who rely on it for going to work and buying food. So, the concern now is how to provide the safest transportation, especially for those who don’t have other options like using a bike or car, or walking.”
Since Khani and Kumar’s study uses the most recent data and information that change daily with ridership, transit authorities could adjust their service as needed to get passengers where they need to go.
“By studying traffic patterns and looking at how many people are in a confined area, we can quantify the relative risk and suggest the recommended capacity,” Khani said.
“This is a really hard time for everyone, and it is not a preferred time for public transit,” he added. “As researchers, we are doing our best in collaboration with transit agencies to provide a safe and secure way for people to use public transportation.”
Khani continues to work with Metro Transit, which is evaluating the findings from this study.
For more on this research, watch the CTS webinar, “Impacts of the COVID-19 Pandemic on Minnesota's Traffic and Transit Networks.”
Read about Khani’s past work in “Approach could help control disease outbreak in public transit systems.”
Story by Pauline Oo
If you'd like to support this work and other COVID-19-related activities, visit the CSE Response Fund.