The Key Bridge carried commuters, freight, and port traffic around the southern edge of Baltimore. After the collapse, most traffic was redirected to the I-95 and I-895 tunnels, while hazardous-materials trucks and other restricted vehicles were forced onto other sections of the I-695 beltway. Travelers and freight carriers began experimenting with new routes and departure times almost immediately.
Rather than just asking “Is traffic worse?”, this study looks at three more specific questions:
- Where did congestion intensify, and where did it actually ease?
- How long are delays lasting on real, door-to-door trips?
- Who is bearing the brunt of added delay?
How It Was Measured
The team used large-scale, anonymous data and compared three time periods: before the collapse, roughly one month after (short-term), and about one year after (mid-term). For each period, morning peak (6–10 a.m.) and evening peak (4–8 p.m.) were analyzed separately.
The analysis drew on several data sources and methods:
- Regional probe data (from in-vehicle GPS and cell-phone apps) to measure speeds. This was used to compute a Travel Time Index (TTI)—a ratio of actual travel time to free-flow travel time. A TTI of 2.0 means a trip takes twice as long as it would in free-flowing conditions.
- Roughly 160,000 real trips, grouped into shorter trips (about 15 miles or less) and longer trips (regional and intercity travel).
- Before/after statistical comparisons using paired tests to determine whether trip times changed in a meaningful way for the same origin–destination pairs.
The result is a picture of congestion that is both network-wide and people-centered.
Short-Term Findings
Where congestion got much worse:
The most severe changes were on I-95 and I-895 approaching the Fort McHenry and Baltimore Harbor tunnels:
- Speeds on these tunnel-approach segments dropped by roughly two-thirds during both morning and evening peaks
- TTI values more than tripled in some cases. A 10 minute trip could now take 30 at the height of congestion
- Queues stretched 8–9 miles upstream during peak periods, making these stretches the region’s dominant bottlenecks
Where conditions actually improved:
At the same time, some parts of the network got better:
- Southern I-695 between I-70 and the Key Bridge saw noticeably lighter traffic, simply because bridge-bound trips vanished
- Specific locations on I-695 — such as the inner loop near I-97 and the outer loop near I-95 — showed speeds rising back toward free-flow, with TTI dropping by nearly half
Why downtown didn’t change as much:
Many downtown arterials were already so congested before the collapse (often below 20 mph) that they didn’t attract much additional detour traffic. Most of the “new pain” stayed concentrated on the freeway approaches to the tunnels, not in the local street grid.
One Year Later
Partial Relief During the Morning Rush
On the southbound approaches to the tunnels, speeds have rebounded from the extreme lows of the first month, but conditions are still significantly worse than before the collapse. Travel times are still well above pre-event levels. For longer-distance travelers, the extra delay has shrunk, suggesting that queues may not stretch as far across the network as they initially did.
Evening Rush Hour Remains
Northbound tunnel approaches on I-95 and I-895 remain just as congested as in the first month, with speeds stuck around 20 mph or below. Some northern beltway segments on I-695 have actually become more congested in the mid-term than they were in the short term.
Why the difference?
- Morning demand has eased as some commuters shifted departure times or adopted telework, trimming peak-period pressure.
- Evening demand is harder to reduce. It mixes commuting with dining, shopping, social trips, and essential errands that even teleworkers make
- Ongoing construction on I-695, including shoulder-lane conversion and associated work-zone lane closures, pushes some truck traffic from midday into the late afternoon, adding to the PM peak load.
The result is a network that is adapting but not fully recovering. The worst freeway bottlenecks remain severe, and some improvements elsewhere are offset by new sources of delay.
Implications
Targeted capacity improvements: Projects like shoulder-lane conversions on I-695 and future upgrades near I-895 may help, but they cannot directly expand capacity in underwater tunnels where physical constraints are tight. Careful cost–benefit analysis is needed to decide how far to go with lane additions or ramp reconfigurations on detour routes.
Smarter use of arterials: Diverting some traffic onto signalized arterials (US-1, US-40) can take pressure off the tunnels, but only if those corridors are equipped with modern traffic signal management so they don’t gridlock themselves.
Demand management: Encouraging flexible work hours, telework, and transit use can reduce peak loads, especially where capacity cannot easily be increased. The travel-behavior studies linked from this project are critical to understanding what people are actually willing and able to change.
