WMATA Capacity Expansion
If we're going to do a train to Georgetown, do it properly
A couple of days ago, WMATA unveiled their Blue-Orange-Silver capacity study, which was intended to study ways in which the capacity of those three lines could be improved. The question has been on WMATA’s institutional mind since the 2014 opening of the silver line. The silver line was a branch, peeling off of the interstate 66 orange line section and heading northwest towards Tysons, a fringe city in Virginia. With the second phase open, the line now reaches all the way to Dulles International Airport (the city’s main long-distance airport) and the server centers in Ashburn.
The problem, of course, was that branches require trains. The more branches a subway system feeds into its core, the more congested its central section will be. Further, there’s a hard minimum on what that number of trains is, because each branch requires a minimum frequency when unassisted by other lines. The silver line meant adding a third line to the downtown east-west crosstown branch, and this has begun to cause problems. Trains at peak times frequently trip over each other merging at Rosslyn. Off-peak, the reverse problem happens - running basic service levels on all three lines results in a constant bombardment of too-frequent but uselessly empty trains in the center section.
The solution that WMATA has come up with (but of course) is to build another downtown through route which allows trains of one line to bypass the Rosslyn chokepoint altogether. WMATA proposes four different options (pictured below). I have several problems with this, first being that WMATA has a rail capacity crunch but not an actual passenger capacity crunch, and while capital budgets aren’t real a new line will do material damage to WMATA’s operating budget. WMATA should instead focus on pursuing fixes which are operating budget-neutral, and encourage densification around their existing stations rather than encouraging new suburban sprawl with new extensions.
The Press REACTS
Before, I’ll start by noting that almost every news report had the same lede: Possible Subway to Georgetown. While the news is generally not the best way to reveal popular opinion, I think this does get at the fact that the biggest problem lots of people have with the Metro’s coverage is that it does not go to Georgetown. Everyone I have talked to who lives in DC roughly agrees with this sentiment, so it makes sense that the most of the studies of metro expansion included a Georgetown station. This is somewhat like New Yorkers complaining about their airports - airport improvements don’t generate significant transit value by themselves, but people irrationally dislike them. Plus, in the same way that it sends a bad message if the city gateway is dingy, it sends a pretty negative message if the city’s best school is so hard to get to that it might as well be on an island.
The problem with fixing the Georgetown connection is that it basically requires a new line, so it does make sense that the WMATA took what it saw as a need for a new line and combined it with a new train to Georgetown. From both logical and PR perspective, it make sense if metro is legitimately capacity constrained. It does incur the problem that frequency on the new tunnel would be limited, if that were a problem. The problem with all of this is that WMATA has way too much train capacity to justify expansion at all.
The WMATA capacity problem
While on the one hand it definitely makes sense to build more rail capacity if rail capacity is constrained, the underlying assumption is that if rail capacity is reached, all of the trains are full and the agency is struggling to keep up. This is decisively not the case on WMATA. The subway system as it currently exists runs gigantic trains, which are mostly not full. “Unlocking capacity” as described both by the WMATA report and the Matt Yglesias piece on DC transit isn’t a rational problem to address, because the DC metro system isn’t short on actual rider capacity.
The problem instead is that WMATA is in theory a subway system and thus runs trains at subway frequencies to places with decidedly low densities. Take for an example the Orange Line. The number of people who ride from the eight Orange line exclusive stations (ie, can’t take the Silver Line) is only 8,000 people, down from 22,500 in 2019. In other words, this is a line with service to 1,000 people per station. Even at prepandemic numbers, the justification for a frequent train was sketchy, but now it simply does not exist. The orange line, in other words, gets about as much ridership as a typical Metro North or Metra Electric station (while being considerably cheaper than Metro North). The level of service that actually makes sense is something approximating every 20 minutes at midday, every 15 at rush hour, and every 30 minutes on nights and weekends. Instead, it gets a train every 10 minutes all day.
This, by the way, is not a story exclusive to the Orange Line. The section after Stadium Armory on the Blue and Silver lines garners 4,000 passengers a day (down from 10,000), the combined Silver Line phases I and II get 9,000 (down from 13,000 on just phase I). The simple reality of the Blue/Orange/Silver line is that the rails are overcrowded because it runs too many trains. The obvious solution, of course, is to simply run less trains.
Downtown Arlington, and ‘Kill the Blue Line’
To hear most of the wonk side of transit advocates tell it, the Blue Line is the original sin of the DC Metro because it constitutes a branch and reverse branch. In layman's terms, this is when a train branches off one core line in a downtown, but instead of then going off on its own, it joins a different line in the periphery. In the blue line’s case, it branches off the orange and silver lines, then merges with the yellow.
To make a long story short, most transit professionals regard this as a bad idea for two reasons. The first is reliability, and it is a large drawback of interlining generally. If the blue line becomes delayed, it can jam up the orange and silver lines, but also the yellow line. Problems on the yellow line also impact the green line and now only the red line is running normally. Killing the blue line and ending the branch-reverse branch effectively quarantines the orange and silver lines from the green and yellow lines and would improve system reliability.
This is all pretty straightforward logic, but what it doesn’t explain is why systems do interlining in the first place. The basic answer is that people don’t like switching trains. Commuting on mass transit with a one-seat ride is an experience that rivals a car, a two seat (one transfer) ride is still pretty easy, but once you get into three seat rides and beyond, transfers become pretty discouraging. If you had to change trains at every station on your commute, you’d probably seek an alternative - the main draw of transit being that you don’t have to pay attention to your commute as it happens basically dissipates.
Turning the blue line into a shuttle basically incurs all of these problems. Trips from Alexandria and Aurora to Rosslyn now require a transfer, but worse, trips to downtown Arlington become two transfers. (The alternative is a transfer at L’Enfant, which adds 15 minutes to the trip time). When the blue line was being built it could have been argued that this was an insignificant trip volume, but Rosslyn has now become a downtown core which rivals Capitol Hill (as DC’s height limit prevents any other industries being nearby). This type of trip is now worth serving, so it feels unreasonable to suggest the blue line be turned into a shuttle. Not only that, but it will slowly become even more worth doing. While prepandemic estimates are not the most reliable arguing point, it is still true that there’s a lot more construction in Virginia than in DC itself. Demand projection studies showed that demand is highest on the Yellow line crossing, as well as from Pentagon/Crystal City to Rosslyn and on to Arlington and Tysons Corners, while demand actually going to DC itself won’t be all that strained.
A Dichotomy - Metro vs Regional Rail
Running less trains is, of course, obviously not a popular solution, and in some parts of WMATA is genuinely an inappropriate solution. It has long been noted that WMATA (and its sister space-age subway, BART) are a sort of hybrid system.
In most cities, there is a subway (or metro, or u-bahn - there are many names) which serve to connect different parts of a city together. Because they stay within a city, they serve densely-developed land, and have short stop spacing. These factors generate high ridership, which justifies frequent trains which often fill, and their short average trip lengths means that focus is on capacity, not comfort - most people stand.Regional rail (or, in a peak-oriented form, “commuter rail”) is different. It serves to connect a city to other parts of its region, with long stop spacings, and fast trains with a layout focused on sitting. Because it reaches from a city out into its suburbs, it serves relatively less dense places and runs less frequently, especially off-peak.
WMATA has parts of both. Some stations on the red line are literally miles apart and serve only suburban houses and parking garages, while other stations, especially downtown, serve dense combinations of office, retail and apartments. When regional rail and metro combine, the solution is usually to combine many regional rail branches (up to 5) to achieve a combination. The issue is that DC’s system is billed as a subway with no separate commuter rail branding, so riders (and voters) expect subway-level service in places that really don’t justify it. The result is the WMATA we have - mostly empty trains which trip over each other
There is a reason that combined urban-suburban subways like WMATA are rare - they are expensive. Subways are separated into their own regulatory category from other trains, because they do not have to share tracks with freight trains. This allows them to be much lighter and less durable in the event of a collision. This separation is pretty necessary anyways, because subways (as they serve short trips) must get people almost exactly where they need to go. This justifies tunneling under streets. Regional rail, on the other hand, does not need to do this because trips on it are longer. It can instead feed into last mile transit, which allows regional rail builders to use pre-existing freight tracks even if they sometimes pass through industrial areas nobody lives or works in. While it may make sense to build or extend a regional rail line that can share tracks with a freight line that already exists, it almost never makes sense to build a whole new right of way to capture that (likely very low) ridership.
Operating Deficits
Almost a direct follow-on from the above discussion about WMATA’s construction choices making it an expensive system is the 2024 operating deficit which comes to $750 million. While recovery at core stations has been somewhat approximate to other relatively successful systems (around 70%), ridership recovery at the peripheral suburban stations I’ve been discussing has been abysmal, (in the 20-30% range). I’m not here to discuss what WMATA should do about that deficit, but I do think that simply ignoring it would be pretty irresponsible, and as such any extensions in the long term should probably try to avoid a severe imposition onto the operating budget, and should definitely focus on dense areas within DC which are most likely to capture a revenue-neutral amount of riders.
The People Want The Georgetown Subway
My first reaction to the Rosslyn capacity problem is to do nothing. Quite frankly the trains are not full enough to worry about a capacity crunch, and even with the forecasted 30% increase in demand by 2040 I still don’t think the load levels will be unmanageable (although at 30% more ridership they will need to think again about doing something at that point. My second idea is to just move to automated trains on modern CBTC signals. This would allow more trains to be operated, while also saving on operator cost through automation, and it would fix the capacity problem for a few more decades. However, the purpose of the Rosslyn capacity study wasn’t to find nothing, it was to find the next WMATA expansion so I suppose that at this point I should probably offer some ideas on what should be done.
First I think I should say that any new subway tunnels which get built should go to Georgetown. There are a few factors for this. Firstly, the press seemed to think that the Metro to Georgetown was the only expansion worth reporting on (not even National Harbor got an exclusive headline). By preference as revealed by media market pressures, this is where the interest is. Second, Georgetown is the focus of DC’s education sector, which is an important part of its economy still remarkably disconnected from the rest of its economy. Finally, Georgetown managed to develop the rare combination of relatively dense and rich, which is a pretty golden combination for high trip generation and even better for revenue potential. For extra bonus points, two universities nearby are fantastic ridership opportunities, since students (especially Georgetown) don’t use cars. Even usually-progressive twitter seems to agree that this sort of extension should be prioritized.
The second point in favor is that while the rationale for keeping the Blue Line around is relatively strong (as I elaborated on earlier), my reasoning doesn’t exactly cover a frequent blue line (after all, this isn’t a huge volume of riders being serviced). Coincidentally, I also don’t think the branch to Vienna deserves much service either, given that it is just three highway median Park-and-Rides with low ridership. Together, this makes for a pretty intuitive fix - plug the low-demand branches together. Of course, this does beg the question of where the Orange Line trains from New Carrollton go, which is exactly where the Georgetown Subway comes into play. The only concern is issues at Rosslyn with transfers to the silver line. However, I don’t think this blue line can fill more than a four-car train every fifteen minutes, I think that the silver line (6x the capacity) will probably be fine.
I do want to draw attention to a couple of features about this map. Firstly, it is pretty much operation-cost neutral. It operates the same amount of lines at roughly the same frequency at roughly the same runtime, which together consume roughly similar levels of labor. There are only two key differences, both advantageous from a budgets perspective. First, the subway to Georgetown runs through more dense areas than the highway medians the old orange line runs to, meaning that it gets more ridership per revenue train hour. On net, this saves money. Second, by right-sizing supply to demand on Vienna-bound trains (the new blue line cannot justify high frequency), this plan constitutes an indirect service cut. The plan allows for more area to be operated on the map, without significant damage to the underlying operating deficit, which I feel is relatively important.
It’s worth noting that the Georgetown subway in this scenario is relatively unique in being justifiable purely by attention. Very few other cities have a missed connection (that isn’t an airport) which is glaring in its absence. In the next blog post I think it’s worth discussing the selection and criteria for metro expansion, especially given how expensive these expansions have become, because in general I think that metro expansion is over-discussed: urban heavy rail rapid transit is incredibly expensive relative to the fairly small benefits in speed that it provides.