Holy cow, how did this not get noticed before? The Washington Department of Transportation’s 2010 deep-bore tolling study claims that tolling the deep-bore tunnel will actually increase driving through downtown Seattle.
Take a look at the gem of a chart to the right, from page 31 of that report. The chart compares forecasted 2030 traffic volumes from a toll-free tunnel (the column to the left) with a tolling scenario in which peak-hour travelers would pay $3.50 in the morning and $4.00 in the afternoon to drive through the tunnel, and drivers at other times would pay at least a buck a trip.
Ignore the details for a moment. What caught my eye, and made me scratch my head, was the difference in total traffic volumes (the numbers at the top of the chart) between the toll-free and the tolled scenario.
Yes, that’s right, the state is arguing that tolling will INCREASE TOTAL TRAFFIC in the downtown corridor. It’s a modest increase—from 504,000 cars without tolling to 509,000 cars with tolls. But it still doesn’t make much sense. Why on earth would tolls—requiring tunnel commuters to fork out up to $7.50 for a round trip, and increasing congestion on city streets and I-5—lure more drivers through downtown Seattle?
If you’re mystified by this, join the club. Maybe there’s some sort of semi-rational explanation—say, their models predict that a tolled tunnel will entice drivers with the massive 1-2 minute time savings (the p. 33 of the report)? But mostly this just seems like nonsense: the sort of result that should have raised red flags early on in the modeling process.
To be clear: I’m not saying that tolling will increase traffic. But I am saying that an anomaly like this makes the state’s traffic models seem mighty fishy. (If you’re a transportation geek craving details, read on…)
The chart compares forecasted 2030 traffic volumes from a toll-free tunnel (the column to the left) with a tolling scenario in which peak-hour travelers would pay $3.50 in the morning and $4.00 in the afternoon to drive through the tunnel, and drivers at other times would pay at least a buck a trip. I don’t include the legend (it’s too wide), but the lightest blue segments on top represents tunnel traffic, the second and largest bar is I-5 traffic; the third is city streets; and the dark blue segments at the bottom represent traffic on the Alaskan Way surface street.
Unsurprisingly, the WSDOT study finds that about a third of potential tunnel users will avoid tolls by shifting to free alternatives. Traffic on Alaskan Way would rise from 26,000 cars per day in the untolled scenario to 35,000 with tolls; traffic on city streets would rise from 114,000 cars per day to 129,000; and traffic on I-5 would jump 14,000. That all sounds reasonable enough—certainly speculative, but not outright crazy.
No, the crazy only rears its head when you look at the overall traffic volumes.
Viaduct junkies will no doubt have noticed the bignewbrouhahaover Seattle’s proposed deep bore tunnel, with a brand new report from Nelson/Nygaard consulting arguing that tolls on the tunnel will divert huge amounts of traffic onto city streets.
I suspect that the public perception of this debate is going to devolve to a sort of he-said-she-said argument, with dueling “experts” arguing about the arcana of their own traffic models, and ordinary folks relegated mostly to the sidelines. But maybe—just maybe—obvious anomalies like this in the state tolling report could convince some folks that they should view traffic projections with a healthy dollop of skepticism.
Buster G.
Have the soothsayers at WSDOT factored into their projections the likelihood that by 2030 the price of gasoline will be $!0/gal. (in today’s dollars) or more and that that would lead to marked reduction in auto use?
John Niles
In the world of computer-based traffic modeling, 504K and 509K forecasts for twenty years in the future are equivalent given the limits on precision inherent in the methodology.The differences in the divisions between tunnel and surface flows are what’s big enough to be significant, as described in the posting above.
Clark Williams-Derry
John – Agreed that the two are essentially identical—but the direction of change does remind me of just how squishy and, yes, unreliable transportation models can be. They may be the best tools that we’ve got, but they can be pretty crummy tools.A friend suggested that the increase might have been due to a small number of cars crossing a “screen line” more than once; apparently, the Seattle transportation model is detailed enough that it might catch someone traveling, say, south on a one-way street and circling back north on another one-way street—which, in certain cases, can result in one car being counted twice. I don’t know if that’s happening here, but it’s at least one plausible account…