I-35W bridge collapse had complex effect on metropolitan traffic flows, researchers find

Our I-35W Bridge study is now out, the short article from CTS is below: I-35W bridge collapse had complex effect on metropolitan traffic flows, researchers find

The collapse of the Interstate 35W bridge over the Mississippi River in Minneapolis on August 1, 2007, instantly transformed the Twin Cities’ transportation network. Thousands of commuters were forced to revise their daily travel routes literally overnight, resulting in dramatic changes in traffic patterns around the busy downtown area. Recognizing that the tragedy afforded researchers a unique opportunity to study real-world responses to sudden network disruption, University of Minnesota researchers including associate professor and Richard P. Braun/CTS Chair in Transportation Engineering David Levinson and civil engineering assistant professor Henry Liu initiated a suite of research projects designed to capture and analyze data on travel behavior in the immediate aftermath of an unexpected large-scale disruption. Findings from these studies may help the Minnesota Department of Transportation (which sponsored the research) and other transportation agencies prepare for and respond to catastrophic network disruptions. Levinson and graduate student Shanjiang Zhu used a variety of data sources to understand the changes in traffic flows resulting from the collapse, including traveler surveys, GPS tracking of study participants’ travel, and aggregate data on traffic volumes, traffic controls, and transit ridership. Data collection incorporated both the post-collapse period and, insofar as possible, the pre-collapse period.
The researchers found that an unexpected disruption produces an avoidance response among travelers whose routes are affected. Drivers initially avoid the area around the disruption site until the perceived risk of traveling through it is reduced with time. This response produces an oscillation in travel patterns, as traffic levels on links near the disruption drop precipitously and then rebound as travelers adjust to the altered topology of the travel network.
Comparing this phenomenon to the effects of preplanned disruptions such as the closure of bridges or highway segments for reconstruction, the researchers found that the impacts of such expected closures were much smaller. The researchers speculate that the psychological shock of a sudden collapse or other catastrophic event is much more powerful than that produced by a “normal” network disruption, and suggest that rapid implementation of an effective system of detours may be key to minimizing this effect.
Network redundancy–the availability of alternate routes, including other bridges across the Mississippi–was a critical factor in accommodating the excess traffic produced by the bridge collapse. Mn/DOT was able to detour traffic along alternate freeway routes including I-94/Minnesota Highway 280 soon after the collapse, mitigating some of the negative effects of the event. However, Levinson and Zhu note in their research report, if the I-94 bridge had collapsed instead, the asymmetrical nature of the road network in the area would have made the I-35W bridge route much less able to absorb excess traffic. This finding appears to have important implications for analyses of network robustness. The addition of a temporary fourth lane on the I-94 bridge also proved to be very important to maintaining effective traffic flow in the area.
Based on their analysis of travel demand data, Levinson and Zhu conclude that the new I-35W bridge (which opened one year after the collapse with greater capacity and faster average travel speeds than its predecessor) helped reduce travel costs most of the time, but that this benefit was fairly small–on the order of 0.2 to 0.3%. This finding is consistent with a preliminary study by Levinson and graduate student Feng Xie using planning models developed at the University of Minnesota. This agreement between the models and observed travel demand data, the researchers say, suggests that forecasting models incorporating elastic demand (varying in response to travel cost) can provide good first-order estimates of the impacts caused by network disruptions. “Quick-response” travel demand models could also be useful in developing mitigation plans for planned network disruptions.
Traffic Flow and Road User Impacts of the Collapse of the I-35W Bridge over the Mississippi River (Mn/DOT 2010-21) is available from the CTS Web site. More information on University of Minnesota research on the bridge collapse is also available online.