Study Finds Limited New East-West Washington Passenger Rail Line Could Generate Substantial Ridership


On Tuesday, the Washington State Joint Transportation Committee (WSJTC) and its consultants presented findings to state legislators on the east-west intercity passenger rail study. The study evaluates several corridor and service level options as well as ridership, costs, targeted corridor improvements, and operational models. Findings by consultants show that ridership between Seattle and Spokane could top 205,000 rides per year under the most robust scenario. That would include two daily roundtrips running through Stampede Pass, Yakima Valley, and the Tri-Cities area.

How we got here

The study is the result of a 2019 request by the Washington State Legislature. The body approved a proviso in the state transportation budget mandating that evaluation of a new east-west passenger rail corridor between Auburn and Spokane by this July, essentially a more robust update of the 2001 study. Five intermediary stops were included in the study proviso, but several extra stops were added by consultants to improve overall viability and provide rational system connections to existing passenger rail service. That is why Seattle and Tukwila found their way prominently into the study options.

  • Gold line is the proposed east-west passenger rail corridor; blue lines are existing Amtrak passenger rail corridors. (WSJTC / Steer)
  • The different service options by stops and number of daily roundtrips.  (WSJTC / Steer)
  • The surveyed level of support for a new east-west passenger rail service. (WSJTC / Steer)

Overall support for additional east-west passenger rail service is very high. Steer, the study consultant, registered over 76% support from respondents; less than 5% of respondents indicated opposition to the idea. Underpinning the study are eight different service scenarios, which differ by the number of cities served by the line and number of daily roundtrips. The study limited the number of daily roundtrips to one and two scenarios. The shortest route scenario serves six cities running between Seattle and Yakima, and all scenarios would directly serve at least those six cities.

The consultants emphasized that the highest ridership projections are lower than Amtrak Cascades service, at about a quarter. But the comparison is a little misleading given that Cascades operates along a longer, denser corridor with three variations and four or eight daily roundtrips, depending upon the counting method. The ridership and costs of operating the east-west service is in line with other state-sponsored intercity passenger rail services like the Amtrak Piedmont service in North Carolina.

Key findings of the eight service options

Initial startup costs for the service generally range from $263.5 million to $388 million, which would go toward purchasing rolling stock, infrastructure improvements, and station construction. The study considers the four- and eight-coach trainsets and would likely require two locomotives per trainset since the Stampede Pass incline is so steep and challenging. Up to four trainsets could be required for the corridor, including spares in case equipment needs to be taken offline for maintenance and repair.

The consultants provided a high-level breakdown of the key startup costs, including:

  • $64 million to $75 million for new passing and siding tracks;
  • $50 million for four-coach stations and $120 million for eight-coach stations (though the costs could be higher for better quality stations); and
  • $144 million to $252 million for rolling stock equipment.

On the low end, Options 7 and 9 would run service between Seattle and Yakima, taking about four and a half hours. These options would generate very low ridership in the range of 31,000 to 43,000 riders per year. Farebox recovery, which is inclusive of food sales, would be fairly low at about 15% to 14%, respectively among the options.

On the high end, Options 1 and 2 would run service between Seattle and Spokane. The travel time for the service would be approximately eight and a half hours end to end. Annual ridership would range from 192,000 to 205,000 with a modest farebox recovery of 33% to 21%, respectively.

  • Details on Options 1, 3, 5, and 7, such as ridership, revenue, costs, speed, and asset requirements. (WSJTC / Steer)
  • Details on Options 2, 4, 6, and 8 such as ridership, revenue, costs, speed, and asset requirements. (WSJTC / Steer)

Option 3 would net the highest farebox recovery rate at 47%, but would be only one daily roundtrip between Seattle and Pasco. Ridership is still projected to be healthy though with 189,000 annual rides. Travel time for the service is assumed to be about six hours.

During the presentation, the consultants noted that the real ridership maker for the corridor is the central portion, specifically in the Yakima Valley. Including Yakima and Toppenish in the model showed that ridership would be significantly boosted with trips running between those two cities and cities in Kittitas County, representing nearly half of all ridership. The consultants noted that this was strongly indicative of pent up demand in the corridor for a high quality transit service.

Competitiveness, connections, and constraints

On the competitiveness front, the train would generally be slower than many other options–at least under perfect conditions on paper–from Seattle to Pasco–and other destinations in between. But the reliability of driving, flying, or taking the bus suffers during certain times of the day and year.

Destination from SeattleTrainBusPlaneCar
Cle Elum3.00 hours1.25 hours
Ellensburg3.50 hours2.50 hours1.75 hours
Yakima4.25 hours3.00 hours2.50 hours2.25 hours
Toppenish5.00 hours5.25 hours3.00 hours2.50 hours
Pasco6.00 hours5.00 hours3.00 hours3.50 hours

It is not uncommon for Snoqualmie Pass–the main all-year thoroughfare between the two sides of the Cascade Mountains–to close for hours during the winter or be clogged with general recreational congestion on summer weekends. Congestion in the Seattle area is also quite awful on a nearly daily basis. And regional airline service is notoriously unreliable, expensive, and requires substantial time traveling to the airport to check in and collect baggage. So the train may win out on reliability and cost for many travelers, but the ridership analysis likely underweights this consideration significantly in the western corridor.

The east-west rail line could be well positioned to connect with other major rail services in five cities along the corridor. Virtually all cities have other connecting transit options, such as local, regional, and intercity bus services. It bears mentioning, however, that Empire Builder service can wind up arriving in Spokane and Pasco in the middle of the night, making it a less than convenient option. Relatedly, state legislators asked if Steer looked into additional Seattle-Spokane service via Stevens Pass since existing Empire Builder service is not terribly ideal, to which consultants essentially said “no” because of the limited mandated study parameters.

Connecting Rail ServiceSeattleTukwilaAuburnPascoSpokane
Amtrak Cascades*XX
Amtrak Empire Builder**XXX
Amtrak Coast Starlight**X
Sound Transit Sounder North*** X
Sound Transit Sounder South***XX
Sound Transit Link****X

Table Notes: * intercity regional passenger rail, ** long-distance passenger rail, *** commuter passenger rail, and **** regional light rail.

The stretch between Auburn and Cle Elum is particularly sluggish as tracks wind their way through the suburbs, foothills, and Cascades. Travel time between the two stops is estimated to be just under three hours (approximately two hours and 56 minutes) according to the sample schedules. Many of the curves and sections have very low speed restrictions, which would reduce the average speed to 26 mph even with higher speeds allowed for passenger trains.

The journey time timetable and average speed between cities. (WSJTC / Steer)
The journey time timetable and average speed between cities. (WSJTC / Steer)

The study does identify sections of the corridor between the two stops that could benefit from straightening tight curves, siding tracks (increasing tracks to two), and snow protection improvements. The costs of these improvements would not come cheaply, but they could help shave off quite a bit of time.

Speeds above the typical limit of 79 mph were not analyzed in the study, but if special approval is granted on higher rated sections of the corridor east of Cle Elum, then travel times could be further reduced. Passive tilting technology, high-level platforms for boarding, automatic doors, and no checked baggage for intermediate stops could also help reduce general travel times.

The study suggested if the full suite of track, equipment, and operational improvements are used, travel times could be reduced by an hour between Seattle and Spokane. However, the increased ridership was marginal, rising slightly from 205,000 to 215,000 on the high end. Perhaps the time savings were not drastic enough, but generally ridership modeling is sensitive to travel times and reliability–though Greyhound advertising the bus trip from Seattle to Spokane at five and a half hours (using the more direct I-90 route) is likely a factor, too.

Operational models

The consultants highlighted three operational models in the study for the state to consider, which include options for a public operator concession, private operator concession, or a statutory state operator. Each of these models have their pros and cons, but the public operator concession is the option most familiar to the state. This would essentially be identical to the Cascades service sponsored by the state and operated and maintained by Amtrak. Assets such as rolling stock, stations, tracks, and facilities would be directly owned by the state, except where leased from Amtrak or private railroads.

The private operator concession would be open to any other third-party railroad company. Notable examples in the American market include Keolis, Herzog, Deutsche Bahn, and RATP. Virgin Trains USA, which operates the Brightline service in Florida and is pursuing high-speed rail between Los Angeles and Las Vegas, might also be interested in something like this. This model could deliver substantial operational cost savings and even enhanced rider experience over Amtrak. There are also possible equipment procurement advantages with the model.

The main downside the private operator concession is the negotiation disadvantage it presents in acquiring slots on Burlington Northern Santa Fe’s tracks, which could essentially wipeout cost savings. Going to this model would best be served if the Cascades service was also folded under the same operator. Not mentioned in the presentation by Steer’s project manager is the seamless ticketing and reservation system that sticking with Amtrak presents, which may be useful to many riders.

The third model with a statutory state operator would be fairly unique for such a long regional intercity passenger rail line, but it could be done. Initial startup cost would be significant to establish the entity and build institutional know-how, but there are upsides in streamlining the process of designing, building, and operating the line. Doing this would most be prudent if also integrating the Cascades service in order to benefit from economies of scale and shared human capital.

Missed opportunities and future

The study does miss out on some opportunities to possibly improve ridership along the corridor and may be worth a closer look down the road. In the Yakima Valley, there is Prosser which represents a notable population center rivaling Toppenish and could provide a good catchment for local ridership. Between Pasco and Spokane, intermediate stops at Connell, Ritzville, and Cheney also offer additional ridership opportunities and destinations. And then west of the Cascade Mountains, stops in Covington or Maple Valley could prove useful to capturing more of the urban Puget Sound population.

Of course, each stop that is added does introduce more dwell time and reliability risk to a schedule. Those tradeoffs might be warranted though if there are consequential ridership and cost recovery increases. On top that, these stops could justify higher levels of investments and alternative schedules with more than two daily roundtrips.

Ultimately, this study puts the state in a good position to chart the next steps in planning and executing a new east-west passenger rail line. But will the governor and state legislators move to act? Or will this study just end up on the shelf among the heaps of other state transit studies? Time will tell.

We hope you loved this article. If so, please consider subscribing or donating. The Urbanist is a 501(c)(4) nonprofit that depends on donations from readers like you.

Stephen is an urban planner with a passion for sustainable, livable, and diverse cities. He is especially interested in how policies, regulations, and programs can promote positive outcomes for communities. Stephen lives in Kenmore and primarily covers land use and transportation issues for The Urbanist.

Inline Feedbacks
View all comments
AJ McGauley

The problem with this study is every option included Seattle. This is falling under the same “elite projection” error that requires every urban rail line to serve the airport. If you want strong ridership, serve daily and weekly trips, not “oh yeah I could see myself using that a few times a year” trips.

Thankfully, the consultants noted “opportunity to improve intercity transit services within the Kittitas and Yakima valleys.” This can probably be done just fine with a bus given the general absence of freeway congestion, but if we want to try out a rail service, I would pilot between Yakima and Pasco, to avoid both weak links Zach identifies, Stampede Pass and Yakima Canyon. And by avoiding Stampede pass, I think you can also eliminate half the locomotives needed to start up?

If we realize the point of this line is to serve those along the way, not to give people quick access to Seattle (if they want that, they should fly or drive), we can add stops like Wapto and Prosser before extending beyond the Yakima-Pasco core.

If ridership was good enough to keep going, then we can look at extending to Kittitas and then to Spokane and/or Seattle. Seattle to Spokane trips will never be competitive with flying, so the success of a Seattle to Spokane route will be built on trip demand to stops along the way.

Daily ridership within the Yakima valley is necessary to sustain this service to then allow us in Seattle a lovely train ride over the mountains.


Interesting study. But, the unfortunate conclusion is obvious – spending hundreds of millions of dollars to run a train that will be slower than a Greyhound bus just doesn’t make sense.

And, the travel-time difference may be even larger than it appears. The 45-minute difference between bus vs. car between Seattle and Ellensburg is particularly suspect. There is no reason for a bus driving nonstop down the same freeway as a car to take 45 minutes longer. Greyhound buses drive the same 70 mph down I-90 as cars do. (I suspect what’s really going on, is the study author is confusing traffic padding artificially added to the official schedule as how long the trip actually takes).

The issue about traffic over the mountain passes is also overblown. Once in a blue moon (e.g. on the last day of a long holiday weekend), there can be backups as everybody returns home at once. But, it’s usually smooth sailing and, even on the worse days, the travel times aren’t remotely as long as what’s proposed for the new train.

If we really want to improve transportation options across the mountains, the simplest option is to just run more buses down the existing highways. It’s not sexy, but it is something we can do today, which will make a real difference in people’s lives. Maybe if a bus between Seattle and Ellensberg ran every couple hours, rather than once per day, people would feel more inclined to try riding it.


MATH: 205,000 total annual riders works out to an average of about 570 total daily riders. Divide that by 4 train trips (2 EB, 2 WB) works out to an average of about 145 riders per train. This appears to be a vanity project for rail buffs, certainly not worth the investment of >$350m taxpayer dollars.

Michael G

While there were many reasons why the Chicago, Milwaukee, St. Paul & Pacific Railroad’s successful efforts a century ago to build an electrified rail corridor over the Cascades via the Snoqualmie Tunnel were not viable for the long term (especially since the Milwaukee Road’s route between Ellensburg and Missoula missed nearly every population center except Othello, if that counts), jeez, wouldn’t it be nice today if an earlier generation of decisionmakers had somehow intervened to keep the electrified rail corridor connecting Seattle to the Snoqualmie Pass ski areas and points east? But the Milwaukee Road’s electrified Pacific Extension rail corridor between Puget Sound and Montana was mostly ripped up in the 1980s and now all we have in the I-90/I-82 corridor is the slower, diesel BNSF line via Stampede Pass. But that train left the station long ago, so to speak. (Yes, the repurposed Milwaukee Road corridor for trail use is nice, too. Hopefully the Milwaukee’s old Columbia River bridge at Beverly can be rehabilitated and reopened for Palouse-to-Cascades trail recreational use someday.) I for one would happily take a train from Seattle to Yakima to pick up pupusas, to Toppenish for taco trucks and to Prosser for wine tastings.

AJ McGauley

Well, they did keep the corridor intact, didn’t they? They just looked at it and said, “there’s no reason to run a train when a car or bus works better, so let’s make it a trail.” Because really this study is about “trains are fun,” and catering to the bike lobby is much cheaper than catering to the train lobby.


15 million vehicles pass over Snoqualmie Pass every year, and it is only rarely congested. The freeways east of the pass area, I-90 to Spokane and I-82 to Yakima and Pasco, are never congested. Adding 205,000 rides per year on a train that takes twice as long to get anywhere seems like a complete waste of money. Don’t get me wrong, I’m a big fan of trains. I suggest a pilot project, Seattle to Cle Elum on a significantly upgraded Stampede Pass line. Get the travel time down to 1.5 hours or so and it will be competitive with driving. Cle Elum could then be more integrated into the Seattle metro area. There is a lot of empty space in downtown Cle Elum next to the tracks that could be used for transit infrastructure, like parking, a nice station, etc.

AJ McGauley

Given half the projected ridership is within the Yakima valley, wouldn’t a Cle Elum to Pasco service be the more logical pilot line?

Brian Nelson

“Regional airline service is also notoriously unreliable, expensive, and requires substantial time traveling to the airport to check in and collect baggage.“

Gonna call you out here, at least to/from Spokane. Used to make this flight several times a month and its incredibly reliable and cheap – $200 covers your roundtrip ticket and there’s flights leaving nearly every hour. Alaska is also great about swapping out flight times for free, especially out of Spokane, if you need to leave a couple hours earlier or later.

Found that flying from Seattle to Spokane total trip time w/ a carryon takes about 2.5 hrs (Lyft, security, flight) and costs about $260 after Lyft rides to and from the airport.

Driving takes about 4.5 hrs and costs about $100 in roundtrip gas. Bonus is you have your car with you.

AJ McGauley

I think that comment is much more directed at Yakima and Pasco. Spokane is a large enough market that air travel is frequent and affordable.


Right. Nobody is going to take the proposed route to go from Seattle to Spokane, or vice versa.

Brian Nelson

@AJ, @Bill, yeah Stephen rewrote the “Competitiveness, connections, and constraints” section of the article after I posted my comment.

FWIW, my experience with traveling to the Tri-Cities is similar to Spokane. Several flights a day to choose from, takes about 2.5 hrs from stepping out of my front door to stepping out of the airport in Pasco, and costs around $200 roundtrip (flights and Lyfts). I’ve had two 40ish min flight delays over eight trips, bumping my total travel time on those segments up to about 3 hrs – still half of what a train ride would be.

At 3-3.5 hrs, driving can be a wash compared to flying. It’s a pretty scenic drive and like to Spokane, takes about two tanks of gas roundtrip. Again, bonuses are you can leave whenever you want and you already have a car once you arrive.

Not really seeing the benefits of spending $200+ million to get people from once place to another in 12 hours round trip when it’s already being done in 5-7 hours…


One could argue that since Stampede Pass and the Yakima Canyon are the weakest links, not really fixable travel time wise without realigned or new track…that you could give up on Seattle-Yakima for now and look at daylight Yakima-Spokane service. At 4 hours it’d be competitive with driving and could hit all the ag or college towns that share economic and family ties etc, Toppenish, Prosser, Pasco, Connell, Ritzville, and Cheney. Long-term I don’t think Seattle-Spokane via Pasco is viable for through-trips, though maybe it could still work if scheduled that way for all the intermediate trip pairs it would enable.

Even if you could optimize Stampede and the river canyon, I think Seattle-Spokane service could only really be competitive if you build the Ellensburg-Lind connector via Milwaukee Road ROW, for Seattle-Tukwila-Auburn-Cle Elum-Ellensburg- Othello-Ritzville-Cheney-Spokane service. But that’d be hundreds of millions to rebuild 113 miles of track to Class IV track or above.


Only bother if its going to be high-speed rail. Doesn’t seem helpful to build a 6hr train ride to Pasco. If we are going to put capital investment into cross-state transportation, lets have it be better than the existing options – that will really get cars off the road.


Maybe you missed the part about it being better than many existing options. Capital costs for high-speed rail would be orders of magnitude higher. And maybe that would be worth it, but there’s no reason to not move ahead with a manageable program like this now. We spend all sorts of crazy money on highways that have bigger subsidies than this could ever hope for with Eastern Washington being the poster child for that. Why do we have to think small and set stupid goal posts?


I mean even this is think small and stupid goalposts. It’s a decent chunk of money that gets you a train twice a day; good luck trying to get a train that’s convenient for you. The improvements in the study improve travel by an hour, which makes it barely competitive with the bus.

Given that the bus is faster, it would be way more convenient to subsidize affordable, hourly or half-hourly buses that are appropriately sized for the capacity rather than one big train twice a day. Frequency is freedom, which is something the airlines discovered last decade and why 747s and A380s are sight unseen at US airlines anymore.