As promised, I’m returning to my ideas on high speed rail policy and plans. This time, we are going to consider the options for my adopted home of Oregon, with the start of our high speed network beginning in Portland and stretching south into the Willamette Valley. Unlike my series on Wisconsin, there will be substantial segments that need to be built from scratch – largely because of the specifics of railway history and geography in Oregon.

What Even Is High Speed Rail?

I’ve talked about this at length in other posts, but as a reminder when we talk about high speed services we are really talking about trains with top speeds of at least 200 kph (125 mph), preferably up to 250 kph (155 mph). Typically, a high speed line will be mostly (though not necessarily entirely) grade-separated for safety purposes, and will be electrified with overhead catenary wires. Most high speed lines are passenger exclusive, though in some situations there are times and places where local freight trains run as well (such as the Northeast Corridor in the eastern US).

While the general impression for high speed rail involves Shinkansen style services (literally new main line in Japanese), the large and often redundant railway network in the US offers the opportunity for a less flashy approach focusing on upgrades with some limited new lines being built too. With that overview out of the way, let’s dive into the specifics of Oregon.

Railways of the Willamette Valley

Historically, there were two major players in passenger rail service in the Willamette Valley, the Southern Pacific (SP) and the Oregon Electric (OE). I may use Oregon Electric and Northern Pacific (NP) slightly interchangeably, since the OE was a subsidiary of the NP for most of its history – but the corporate history of each line in the valley is not really the purpose of this post and is only tangentially relevant.

Rather, the relevant thing here is that there were three (or four) roughly parallel main lines, two (or three) controlled by SP and one controlled by OE/NP. SP (below, left) had two different routes from Portland to Eugene, the central line, one obviously less direct further east and a secondary west side main line which terminated in Monroe. Meanwhile, the OE (below, right) had just the singular main line to Eugene – it was electrified and largely exhibit qualities of a super-interurban. SP also ran electrified services, though mostly on the west side line – with the main route from Portland to Eugene (central route, passing through Salem and Albany) never being electrified – it was a key “steam line” for SP between the Pacific Northwest and California.

In the years since these maps, large portions of them have been abandoned, but the two principal main lines remain – mostly at least. Union Pacific, which merged with SP in the 1990s still owns and operates the main line between Eugene and Portland and this is the route which Amtrak uses for Cascades and Coast Starlight services to Seattle and California. This line has remained intact, which is unfortunately more than we can say for the OE.

Since the OE was a relative latecomer to Portland area railroads, it’s primary access to the major rail terminal north of Hoyt Street (now the Pearl) was on streets through downtown. SP being a primary competitor did not allow access via its former route on Front (Naito), and OE was also locked out of using Union Station as its terminal (though the former North Bank Road Depot is now posh apartments in the Pearl).

This meant that the OE route to downtown essentially operated as a passenger focused spur, and when passenger traffic declined in the 1930s and 1940s it was not practical to run freight services along the alignment. Ultimately, because of this (and other factors), the entire main line from the junction in Tigard to downtown Portland was abandoned, and I5 was built on top of it. However, parts of the former OE alignment do still see passenger traffic – TriMet’s WES runs from Tigard to Wilsonville on the old alignment (north of Tigard, the route is on SP’s west side line). Additionally, the state of Oregon owns large parts of the right of way – which is leased to Portland & Western for freight (and WES!) service.

Which Route to Take?

Between the two routes, there are advantages and disadvantages to both. The UP (Union Pacific) main line sees substantial traffic – though it is not as important in terms of freight as UP’s transcontinental line terminating in Portland (along the south bank of the Columbia) - but it has a direct connection to Portland’s lovely Union Station. The OE (Portland & Western) main line sees far less traffic, but is in a worse state of repair and does not connect directly to Union Station. I favor the OE line – mostly because of traffic, but also because the state already owns a significant portion of it. Plus, service could reach the most important centers in the Willamette Valley without much in the way of new tracks laid – with the notable exception of a connection to Portland and between Salem and Albany.

If we examine ODOT’s diagram of 2030 rail traffic projections, we see that the OE line is projected to get less than 5 trains per day in 2030. I don’t see a compelling reason to keep passenger service on a busy line – projected for 25 to 50 trains a day and predominately single tracked and owned by Union Pacific – when the old OE is right there. Of course, there would need to be substantial upgrades made for higher speed service, more than would be needed on the UP line, but given the amount of upgrades that would be needed on both lines (double tracking, electrification) the difference is relatively minimal.

What’s more, is that ODOT also projects slightly higher ridership (on conventional speed rail) on the OE than it does on the UP line. There are frustratingly little details around these projections contained in the state rail plan, with the only commentary on the OE study met with substantial opposition from the state’s passenger railway advocacy group, AORTA.

AORTA’s brief on the proposal has many good points, but most of the issues arise in areas where the OE has legacy street running sections, in the connection to Union Station in Portland, and relating to track upgrades. In the context of existing Amtrak Cascades service (a few trains a day), this makes sense. But in the context of high speed service, it makes less sense given the sheer amount of investment needed to get the line up to snuff.

Realistically, the OE is the only choice for high speed rail service. Union Pacific will absolutely not cooperate with the minimum levels of service required to make an investment in electrified high speed trains worth our time and money (at least hourly). So let’s use that 2010 AORTA brief as a way to frame the particular issues with the OE right of way, and how we can work around it.

And while it’s unlikely that Union Pacific would allow for electrification and high speed upgrades on their own rails, I am operating on the assumption that they would be willing to share rights of way in and around Salem. If this is prohibitively expensive, Salem will have to have a freeway area station. Not a deal breaker, but we should at least float a better idea for Salem.

Below is an overview of the general plan for the route though, feel free to skip ahead to “Ridership Guesstimates” if the ins and outs of the alignment are beyond what you really care about. Long story short, our grand total for cost is about $4.5 billion, and travel times between Portland and Eugene are under and hour with stops in Wilsonville, Salem, and Albany. Red lines show tunnels, orange lines show freeway right of ways, orange lines show railroad right of ways, and yellow is major bridges.

Portland to the West Side

Currently, the only real connection between Union Station and the OE is via Milwaukie and Lake Oswego – and requires sharing tracks with potentially the busiest part of UP’s Oregon network, the area between Brooklyn Yard, Albina Yard, and the Steel Bridge. This is a major source of delays for Amtrak currently, and given that the diversion to OE doesn’t avoid this any passenger service proposed to run on those tracks requires a work around. Luckily, there is a 200 foot wide state-owned corridor that can be utilized. And it used to be a railroad right of way too!

I’m talking, of course, about Interstate 5. Even without the peculiar history of the OE having originated the right of way, I5 makes a lot of sense. It’s very wide, state owned, and a two track railway could fit into it without substantial property acquisitions or even roadway narrowing. A line roughly paralleling I5 between Harbor Drive and Wilsonville should be reasonably affordable by megaproject standards, though it would also have to be paired with a cut and cover tunnel through downtown. A 1.8 mile (3 km) shallow tunnel in downtown under Naito shouldn’t break the bank, and could even be a surface level alignment north of the Hawthorne Bridge to allow for regional service to serve downtown more directly.

In total, this alignment is about ten miles long and should be paired with a SW regional service as well (more on that later). How much would this cost? It’s hard to say for sure – but using these Australian estimates (adjusted for inflation and currency), we arrive at a figure around $850 million, including about $250 million for upgrades to Union Station. This works out to about $30 million/km, in line with typical construction costs in France. This is about 11% of the cost of the Interstate Bridge Replacement project – we will call it “0.11 IBR$”, and will use that unit as a helpful heuristic.

West Side to Salem

From the West Side suburbs to Salem, our route is straightforward. Passing along the still existing right of way is easy – with only one real issue to navigate. The OE line street runs in downtown Salem – and while a waterfront station is intriguing, it probably makes more sense to try to serve the existing Amtrak location if we can manage it.

Things get more interesting north of Salem, since the OE right of way is very narrow – just wide enough for one track. Demolishing a row of homes to widen the right of way when there isn’t exactly a station we can just roll up to doesn’t seem wise. Instead, we can utilize I5 to crossover to the UP right of way for a few miles, before constructing a cut and cover tunnel to get us to the existing Salem Amtrak station. This would require the cooperation of Union Pacific – something I was hoping to avoid – but there isn’t a substantially better option other than bypassing downtown Salem. I think it’s worth our time to serve the seat of the state government directly.

All told, this route from Wilsonville to Salem would come in around $600 million, or 0.08 IBR$ if you prefer. These figures are derived from the handy table from earlier, with conversions and other notes in this spreadsheet for the curious

Salem to Albany

From Salem to the next city of note, Albany, the OE is not appealing. Even if we were to follow the OE all the way to the Salem waterfront, the routing around the Salem Hills is circuitous, curvy, and very close to the river. It would be pretty, but I think we ought to find a more direct route. Luckily, the UP right of way we are currently in intersects with I5 not too far down the line. This route will require a tunnel through the relatively steep Salem Hills, but it has the benefit of far higher potential speeds than either the OE or UP lines.

We will still serve the existing Albany Station, with future Corvallis service done via a conventional or higher speed branch line. More on branches in a later post, but I think serving Corvallis via a branch transfer is preferable to building a new North Albany station – especially since we need to connect back to the OE in a tunnel to avoid the street running section anyways..

All told, we can expect this choice of alignment to come in around $750 million dollars, though the specific cost may be slightly lower depending on the specific tunneling needs in the Salem hills (I went for the longest tunnel I felt was reasonable). That’s 0.1 IBR$, bringing our project subtotal to 30% of one single highway expansion program.

Albany to Eugene

Our final segment, from Albany to Eugene is the longest and our old friends at AORTA do not recommend using it owing primarily to issues surrounding the street running sections in Harrisburg and Junction City. I am proposing shallow cut/cover tunnels here, which raises costs over the parallel UP line but has the benefit of fully grade separating the alignment – something worth pursuing in populated areas. Other than these, there are no real challenges to overcome – it would mostly just be a matter of rebuilding tracks and slightly widening some of the right of way.

Once we are in Eugene, there is a short segment of shared right of way with Union Pacific, but this is wide enough for two tracks at least, and three or four in the vicinity of the terminal station. All told, this is the least expensive stretch of the project in per-mile costs, since most of it is in existing right of way and there are no long tunnels required. At just over $750 million, our final section works out to be 0.1 IBR$, bringing our subtotal for a high speed line (without trains or property acquisition) to 0.4 IBR$.

Other Required Costs

Of course, we would have to pay Portland & Western a fair amount of money to acquire the rest of the right of way and likely to buy out portions of the UP right of way as well. Plus, we are going to need high speed trains to run on the line. I’m ball-parking $1 billion for the P&W buyout and assorted other property acquisitions, while high speed and regional trains should come in at around $50 million/trainset.

Leg	Express Time	Regional Time
Portland to Wilsonville	9m 14s	16m 30s
Wilsonville to Salem	14m 5s	21m 23s
Salem to Albany	11m 40s	15m 34s
Albany to Eugene	18m 0s	28m 41s
Total	53m 0s	1hr 22m 12s
Portland to Eugene Nonstop	43m 53s	n/a
Portland to Salem Nonstop	19m 0s	n/a

With travel times like shown above, we would only need about six to eight trainsets to provide half-hourly service between Eugene and Portland, though it’s likely there would be some fairly complex service patterns to serve intermediate regional stations (like at Barbur TC, or Keizer) since the service is so much faster than any alternative. We need ten total trainsets to provide the service pattern identified (in the sample schedule section), but we should have 3 or 4 more in reserve – bringing our “other” costs to $1.7 billion, or 0.2 IBR$. In total, it’s $4.7 billion – or 0.63 IBR$.

Ridership Guesstimates

No matter how you slice it, $4.7 billion is a lot of money. But luckily, our trains will be providing tons of travel options for Oregonians who will be paying money to use the service. Just how many riders would use this service is a bit of guesswork, but let’s return to a graph from earlier referenced in the 2050 rail plan to take a stab. In that graph, we see projected yearly ridership of 400,000 on six trips a day – and increase of 100% over the base case of two trips a day (all of this for the current 3 hour trip time between Portland and Eugene).

A 200% increase in train trips only yielding a 100% increase in ridership is a bit troubling to me (it actually seems bogus on its face, but I’m not an expert on that front), but our given that our proposed 32 trains per day represent a 1,500% increase in trains we would still expect a 750% increase in ridership over the base using this simplistic model. This works out to 1.7 million rides a year if we assume exactly the same number of people will find a less than one hour trip as good as a three hour trip.

It’s an outrageously conservative estimate, especially considering that more than 100,000 people use I5 in the Portland area, and north of 60,000 in the Eugene area. Let’s consider a more aggressive ridership projection – one that every seat on every train will be sold exactly once. This sounds like the most aggressive ridership projection possible – but keep in mind that trains routinely sell seats more than once, since a passenger could board in Portland and deboard in Salem on a Portland – Eugene trip, thus leaving their seat open for a Salem – Eugene traveler. For a high speed trainset with 300 seats (the current Acela trains have 304), this works out to 3.5 million rides a year on 32 trips a day. Interestingly, this 3.5 million also implies similar per-train ridership as the base case in the ODOT study (it’s about a 1,500% increase in ridership).

I think there’s plenty of reason to believe that ridership would be far, far higher – especially between Portland and Wilsonville.

This Is The SW Corridor We Deserve

In 2020, Portland area votes struck down a light rail (and other things) funding measure that would have constructed a light rail line to the SW suburbs. Primarily along Barbur Boulevard and I5, we can think of the first segment of our high speed line as an upgrade to this shelved plan. And it’s a big upgrade.

In the SW Corridor plan, a 30 minute travel time is quoted between Bridgeport Village and Portland State. Currently, this is a 12 to 20 minute drive, though worst case traffic tends to be worse than Google will tell you. Bridgeport Village is near my hypothetical Tualatin station – which has an estimated high speed regional service travel time to Union Station of just 11 minutes. That’s about twice as fast as driving in medium traffic, and 5 minutes faster than driving with no traffic at all.

As a reminder, the SW Corridor plan had a budget in the neighborhood of $2.5 billion. I like transit expansion projects, but at some point the Portland region needs to consider higher order modes. High speed regional service competes favorably with driving in the off peak right now! Not just at some future hyper-congested peak hour roadway condition.

But in any case, TriMet’s daily ridership projections for this line are in the neighborhood of 12,000/day – which works out to 4.4 million a year or so. If our high speed line can manage blended SW suburban regional service (which it absolutely can!) with all-day frequencies in the 15 minute range I think it’s fair to assume our combined ridership will be closer to 8 million/day. While our longer distance high speed trainsets will have capacity in the 300 person range, regional style service can and should have higher capacity trains. A six car train on the Long Island Railroad has a capacity closer to 650, and something analogous would be wise on a corridor like this to avoid crush loads.

Sample Schedule

Between Portland and Wilsonville, we would be running 15 minute regional service while Portland to Eugene gets half-hourly trains with limited stops (in Wilsonville, Salem, and Albany). Throughout the day, various other services could be run as well – particularly regional service to Salem (stopping in Donald and Keizer), and regional service between Albany and Eugene (to serve Harrisburg, Junction City, and North Eugene).

This schedule for intercity express services conveniently only requires 6 trainsets, with assumptions of at least 20 minutes for end of line turning and cleaning procedures (no actual turns would be done since the trains could be driven from either side) and 90 second dwell times. This is a fairly aggressive schedule, but there is ample padding at the end of runs to account for the happenings of the world.

On the Portland to Wilsonville regional schedule, we need to work around this base case intercity service. Ideally, we should schedule our trains in such a way to minimize potential conflicts between regional and intercity services. This also gives us the benefit of having a situation where Portland to Wilsonville intercity trains can operate as express regionals, offering the super-fast 9 minute trip as opposed to the just fast 16 minute trip. With a fair amount of padding, we can manage this schedule with just four train sets. Additional stations at OHSU and the Morrison Bridge would push the time out a little more, likely into the 20 to 25 minute range but that should still be workable with four trains.

Final Thoughts

Ultimately, this post can be viewed as a criticism of both light rail and intercity rail policy at the regional and state level. Our state DOT is more rail-friendly than many others, but the bulk of physical improvements and service upgrades to the Cascades service over the years has been either monumentally incremental or just between Portland and Seattle. And our regional bodies like Metro and TriMet seem hellbent on making every major transit project a MAX extension. I like the idea of an expanded light rail network, but I think it’s short-sighted to not seriously consider other alternatives.

The historical context of the SW suburbs and their relevance owes to the existence of Interstate 5, which was constructed where it was because it was the shortest route between Portland and Salem. This historical circumstance should not be taken for granted or thrown out the window. Any transit project that serves the area will always be improved by extending to Salem or Portland, but a MAX extension to Salem would be foolish. Though come to think of it, it would really put the MAX firmly into the “modern interurban” category of rail transit service (something it arguably already is).

In any case, by leveraging the advantages of the SW suburbs have in the broader context of the state and metropolitan area, we can deliver a project that fits much more harmoniously with the history of urban development in the Portland area. And this matters for reasons far outstripping esoteric historical trivia – it means the things that people like about the SW suburbs in terms of location will be accentuated which should broaden the base of popular support. Looking forward, by building our high speed spine primarily in an existing rail corridor we maximize potential higher speed branch line connections to Corvallis, Newberg/McMinnville, Lebanon, Hillsboro, and Springfield (more on that in a following post).

And as a parting note, the speculative $4.7 billion dollar price tag is enough to give a $5,000 subsidy to 940,000 Oregonians purchasing qualifying electric cars. Great, I love it here.

The high speed service described in this post represents a new age in passenger rail service in Oregon, and one that is by the standards of high speed rail construction total costs fairly affordable. Even if we can’t match Spanish or even French costs per km, Oregon can build a world-class system for less than our share of one highway expansion project. Let’s find a way to get it done.

Bonus: here’s a diagram of future high speed rail services in the Willamette Valley. Light purple lines are connecting branch lines that terminate at a station on the main line.