Deep Bore Tunnel ongoing disaster update, or why no new transportation megaprojects should ever be designed for cars only, in one chart

I was a bit surprised to get some pushback on the notion that deep bore tunnel project in Seattle was a terrible idea in my previous post on the subject. I think I can see where it comes from psychologically–the viaduct, while ugly and extremely unsafe in an area of seismic activity, is an very efficient and useful piece of infrastructure. Seattle traffic is pretty bad, and the idea of taking something so big and important out of the picture without replacing it with something that’s also big and important is understandably alarming.

Going back far enough, the possibly plausible alternatives to the tunnel were: retrofit the viaduct, rebuild the viaduct, build a big urban waterfront freeway, build a shallow ‘cut and cover’ tunnel with downtown exits, and a ‘enhanced roads and transit’ option which contained a substantial non-freeway road in the place of the viaduct, a widening of I-5 to increase capacity on the other freeway through downtown, and a series of other improvements to help address in increase of traffic downtown. To be clear, there are no great options here: one can find serious problems and shortcomings with all of them. (The rebuild option, for example, demands an answer to the question of traffic flow during the 8-10 years it would probably take to construct it). Some of these, including retaining/rebuilding the viaduct and the urban freeway, were undoubtedly aesthetically worse options than the tunnel, but the tunnel was the worst option from a strict cost/benefit analysis of all the possibilities for what should be a profoundly obvious reason: a majority of current viaduct users use it to get in or out of downtown, not to get through it quickly, and the tunnel is too deep to include any downtown exits. Astonishingly, they decided to knowingly replace an important piece of infrastructure with the one option that had no utility for a majority of users of the viaduct.

For a variety of reasons, by the end of the process the politically viable options appeared to be the deep bore tunnel and enhanced roads and transit plan. The latter drew a lot of skepticism, including from some people who should have known better, in part because it just seemed implausible that a bunch of little things could replace one really big thing. But in order to respond to such fears, we need look no further than WSDOT’s own study of the options. Dominic Holden (in what is probably the single best summary of the comprehensive case against the tunnel) explains:

The most compelling evidence for I-5/surface/transit is buried in the state’s own SDEIS. In an appendix of a transportation discipline report, WSDOT cites the miles of travel, hours of travel, and hours of delay that would exist if we built the I-5/surface/transit option. To compare that data to the tunnel, The Stranger filed a records request with the City of Seattle to obtain the most recent data the state is using to forecast miles traveled, hours traveled, and hours of delay on the tolled deep-bore tunnel.

The data shows surface/transit performing slightly better in the downtown system overall, with fewer miles traveled, fewer hours traveled, and less delay. Specifically, hours of delay would drop from 37,500 hours with a tolled tunnel program to 36,700 hours with a surface option. Hours of overall travel drop from 106,600 to 103,600, and miles traveled drops from 2.5 million to 2.4 million.

Granted, these aren’t huge differences. But this data proves something very important: The much more expensive tunnel doesn’t perform any better than I-5/surface/transit.

That flies in the face of Governor Gregoire and other tunnel advocates who have insisted the I-5/surface/transit option won’t work. When I asked about the traffic impacts of the tunnel at a press conference on October 29, Gregoire said, “The alternative [to the tunnel] is taking 110,000 cars and putting them on I-5 and the streets of Seattle. You want to see a stalled city, you take those 110,000 cars.” But the state’s own data shows that downtown is actually more “stalled” with a tunnel.

This is not at all surprising, when we bear in mind that the majority of viaduct users will find the tunnels useless, even before we consider those who’ll avoid it for purposes of fare avoidance (more on that below). Furthermore, there are increasingly good reasons to take WSDOT’s future traffic projections with a hefty grain of salt–there’s a good chance they’re overestimating future traffic in and through downtown Seattle. Along with their national counterpart, WSDOT’s projections have been overshooting actual traffic for a good while now. A devastating chart from Sightline provides some visualization of this phenomenon:

The USDOT has been overestimating future vehicle miles traveled for a long time now. And this trend is continuing: they overestimated 2013 vehicle miles traveled by 11%. In 2012. WSDOT is not appreciably different than USDOT on this issue; their long-range forecasts for 2008 ended up overestimating VMT by well over standard deviation (see pg 9 here) but despite that error, their 2010 model predicted in the dip would return to 20th century growth rates in 2011, and continue on that trajectory until 2030, essentially replicating the now evident folly on the graph above.  But the declining trend in Washington dates back over a decade–there were fewer VMTs total (not per capita in this growing region, but total) in 2012 than 2002. And WSDOT is still using these clearly broken projection models. In some cases, this makes their projection models overestimate the utility of their projects, but in the case of of the tunnel project, where the budget relies on revenue projections from future tolls, it also means this is another way they’re underestimating the total cost.

Speaking of tolls! Back when the tunnel project was politically vulnerable, WSDOT did its projections for toll revenue based on $3.50/$4.00 tolls. As a matter of basic fairness, tolls should if anything probably be higher than this. But as a matter of avoiding massive toll avoidance, especially in a region with little history of and tolerance for tolling, this seemed obviously and totally unviable. Now they’re  admitting as much: the draft proposal from a meeting of the Advisory committee on tolling and traffic management two weeks ago is now online, and it shows that they’re recommending a toll of $1 ($1.25 during peak hours) to avoid massive toll avoidance. While that figure might be low enough (although some people will still avoid any toll on general principle), their projected revenue figure of 1.1 billion over 30 years requires an average of over 91,000 daily trips in the tunnel. Relevant facts: there are presently about 105,000 daily trips on the viaduct. Over half of them end or originate downtown. Overall traffic has been flat to slightly declining for over a decade. In order for the numbers to support this revenue forecast, we have to believe the same models for projecting VMTs–the very same ones that have been systematically and badly wrong for a while now–will somehow start being correct again.

WSDOT’s other current megaproject that is going quite badly in the 520 bridge replacement. For fun, here’s a projection vs. reality history of traffic on the bridge in chart form, courtesy of streetsblog:

How should we think about WSDOT and USDOT’s systematic failure to adjust to what is clearly a major change in the trendline on VMT? Clearly, simple bureaucratic interia is part of the story. But I think it’s instructive to compare WSDOT to another local bureaucratic agency currently engaging in a tunneling megaproject: Sound Transit. One commenter on the previous post criticized me for complaining about cost overruns thusly:

I challenge you to name a single $3Bln project that has come in exactly on-time and on-budget. I don’t think I’ve been on a $3Mln project that has met those two criteria (for keyboard commandos – talk to someone in tech and see what their opinion is). If you require that, you end up with people completely sand-bagging their projections, and then the project will look so bad, it won’t get started.

He’s certainly not wrong that most megaprojects have cost overruns. But commenters immediately and correctly noted that U-link, Sound Transit’s light rail tunnel from downtown to the University district, which is over 80% complete, is on track to open six months early and over 100 million under budget. (It doesn’t quite meet the commenters’ criteria for a $3 billion project, as it was only a $1.95 billion project). A bit of history: Sound Transit was created in the mid 90’s to deal with the Puget Sound region’s staggering lack of regional mass transit. They currently run a very useful and well used network of regional express buses, and heavy commuter rail from Tacoma and Everett to Seattle, but their most important charge was building a new light rail system. ST started out more like WSDOT–massively underestimating costs and underestimating time needed on the initial light rail line. Once that became apparent, in 2002 there was a major leadership shakeup. Since then, ST estimates have been sober and conservative–the bids for future tunneling expansions of light rail are already coming in well under budget. In 2008, despite the initial overpromise and despite the first light rail line not even being open yet, voters in the Puget Sound approved by wide margins round II of ST funding, which raises $18 billion over 30 years to fund, primarily, major extensions of the current light rail system. For well-crafted public transportation projects, lying about actual cost and time needed to completion is not necessary to engender sufficient political support. For roads-only projects, this is not the case: the first version of ST II failed the year prior, when it was tied to major new roads spending. And Washington politicians understand this: the Republicans in the state senate (and two “democrats” who switched side to give Republicans control of the chamber) refused to pass a bill that would give King County permission to raise taxes to close a funding gap for Metro bus service unless it’s tied to their unpopular, pork-larded, low on maintenance and high on new projects roads bill. Why? Because they know they can’t pass it unless the transit-starved voters of King County are essentially extorted into supporting it (Happily, that horrible bill was not passed, and King county found a legal way to potentially raise revenue without state authorization). WSDOT’s future as an agency of its current size and scope depends on misleading the public about the true costs of their projects (I’ve ignored here, of course, the environmental impact of building the future infrastructures for cars).

It’s heartening to see the current troubles leading not just transit advocates but conduits for the center-right version of conventional wisdom like Dori Monson jumping on the kill the tunnel bandwagon. (And he is not the only well-connected person hinting that rumors from insiders suggest the situation with Bertha is considerably more dire than has been publicly acknowledged). We need to build an infrastructure for the actual future–the one where VMT decreases are a long term trend, not a recession driven anomaly, and where global warming is real–not the future WSDOT won’t stop foolishly projecting.  WSDOT in in desperate need of a ST-2002-style shakeup. The 10% or so of the tunnel completed could make a pretty cool skate park.

….A commenter asks how confident we should be that the current VMT trend is real and not noise and recession-driven. It’s a reasonably question and I give my answer in the comments, but in poking around I found another chart, this one VMT per capita coming out of recessions. The recent recession was, of course, the worst of the bunch, with the sorriest recovery, so this isn’t necessarily unimpeachable evidence that something else is going on. That the levelling off occurred around 2004 nationally (and earlier in some places, including the Puget Sound) is more strong circumstantial evidence. Sightline daily’s “Dude, Where are my Cars?” series is good on this.