Policy Analysis, St. Louis City, Transportation

N-S MetroLink: Devil In The Details

This past weekend we got good news (about the North-South MetroLink expansion) that quickly turned into bad news. The good news is that the expansion is moving ahead: The Bi-State Development Board of Commissioners approved a resolution authorizing the BSD team to plan, design and develop the Jefferson alignment of the extension utilizing $7.4 million in federal pandemic aid money. The bad news is that the actual initial design components of the expansion are pretty half-baked, falling closer to boondoggle than to transformative investment in public transit. The main problem arises on the second page of Bi-State’s deck, six bullet points down: “Utilizes in-street low platform light rail vehicles.”

This is the first real confirmation we have that the project will not be grade-separated like the existing MetroLink system. That’s unfortunate because grade separation is a huge reason why MetroLink punches way above its weight compared to other light rail. Here’s Uday Schultz —a person with no connections to St. Louis, just one of the big accounts on transit twitter— explaining how good MetroLink is due to its lack of street running:

But these new plans abandon that design feature in favor of the type of tramway street-running that became popular under the Obama-era DOT funding regime in places like Kansas City and Cincinnati. Because these tramways are “in-street,” they’re subject to disruptions (mostly from vehicular traffic) that grade-separated systems like the existing MetroLink or various subway systems are not. As a result, tramways have to operate at more cautious speeds which makes them less appealing for actual transportation.

Alon Levy, one of the premiere transit design experts in the world, has an excellent piece defining four different categorizations of “light-rail” based on the speed with which they travel at the city center and at the city peripheries. Speed, in this context, Levy explains, is determined entirely by grade separation. And tramways like the one Bi-State has proposed, falls into the “slow at the peripheries, slow at the center” quadrant of Levy 2*2 typology. This quadrant, he explains, can only work under a very limited set of conditions (like connecting to an excellent regional rail system) that St. Louis mostly doesn’t meet.

How slow are we talking? Berlin’s tramways max out at 12.5 mph, it seems that the Cincinnati Connector averages about 8 mph over the course of its 3.6 mile loop, and the KC streetcar maxes out at 25 mph except, as the Kansas City Star put it, “when traffic or stoplights slow things down” (good thing we don’t have any of that here!). The current speed limit on Jefferson is 35mph. That’s going to make our slow moving tramway a really tough sell to riders.

Now there are two reasons given for why we’re opting for this slower form of transit on the Jefferson extension. The first is that the existing system was constructed mostly using abandoned pre-existing grade-separated freight routes, and we, unfortunately, don’t have any to utilize along this route, so we’d have to assemble the new grade-separated route ourselves, either by tunneling or by constructing elevated tracks, both of which cost significantly more than street running.

This problem is legitimate and tough to solve, but there are legitimate ways around it. One way to achieve significant grade-seperation at reduced costs would be to use concrete el’s (thin the Chicago El’s quiet cousin) like we use near the Lansdowne station. Perhaps the most effective use of elevated tracking would be along the still to-be-determined route of the more sparsely populated second phase of the project.

The options for the routing of Phase 2 running from Fairgrounds to 270

On the other hand, if there’s a stretch where tunneling is more effective, perhaps through the central corridor, we could use the “cut and cover” method of tunneling rather than boring them with the much more expensive tunnel boring machine. This method is the one that was used to construct the Forest Park Parkway tunnel back in 2006, which the MetroLink’s Blue Line currently runs on from Forsyth to Skinker.

To achieve this on the N-S line, we’d have to shut down Jefferson for an extended period of time and open the whole street up via trench, install the tunnel, and then cover it all back up at the end. As Levy explains, this method is especially cost effective in situations where the route in question runs down a straight, wide street like we have in Jefferson. Biting the bullet and using this rarely used best-practice would mean cost savings by a factor of 1.5-2x that of traditional tunnel boring.

But the second reason given for street running suggests the exact opposite set of priorities. Bi-State says that avoiding grade separation will “minimize impacts to adjacent properties.” This reason is much more fundamentally egregious. First, it’s a cowering in the face of NIMBYs that haven’t even yet emerged. Even more importantly, though, the entire point of the extension is to impact the adjacent properties! In fact, the very next slide touts $9 billion worth of impact to the adjacent properties in the form of TOD! You can’t say something is going to make things drastically better and have minimal impacts simultaneously; it doesn’t make any sense.

A couple of weeks back I mentioned that the current zoning surrounding the southern half of this route limits how much TOD we’re going to be able to build here too. If we’re trying to minimize the impacts to surrounding properties does that mean we’re not going to alter that either? We need to make up our minds: do we want to perform “improvement” or do we want to improve?

Some Good News

The question is not whether this plan will offer lower quality service compared to the existing MetroLink system (it will), the question is whether or not the plan can offer useful service for its price-point and the relative opportunity cost of using the money for multiple lines of Bus Rapid Transit. And I think that’s still up in the air thanks to some details the project does indeed seem to get right.

rendering of street running N-S MetroLink pulling up to a crowded stop as a car passes

The current plan calls for the trams to “operate in a dedicated lane, separated by curb to enhance safety and travel times,” which should eliminate some of the Loop-Trolley-levels of downside risk, especially if the trams get traffic signal prioritization and feature decent frequency (we haven’t received any indications on either of those fronts yet).

Bi-State has also indicated that the N-S line will give the tramway prioritization at intersection traffic lights so that it does not face delays every time it hits a red light. This transit signal prioritization (TSP) should help make the line more competitive with vehicular traffic. Will it be enough to compete with the tramway’s overall lower speeds and need to stop for on-boarding? Probably not. But it should help to level the playing-field somewhat.

Last month, Bi-State also announced that they were cutting the number of proposed stops from 14 down to 10, eliminating the Olive, Russell, Arsenal and Natural Bridge/Parnell stops that had been proposed in an earlier iteration of the project. Obviously this move does not come without some costs in terms of catchment area, but fewer stops means faster service and lower construction costs, so this feels like a good tradeoff.

Finally, the trams will be low-platformed. In a different article, Levy explains that because trams are on tracks they can get much more precisely close to the curb than even low platformed busses can. That means the gap between passengers and the tram can be reduced to a matter of centimeters and boarding and off boarding can happen really fast compared to busses, especially for people in wheelchairs. Levy hypothesizes that this is part of the reason that trams see between 34-43% improvement in ridership as compared to busses. Even so, I’m not sure that a 40% upgrade in ridership on a single bus line is exactly what we should have in mind when we talk about a once-in-a-generation transit project.

Devil is in the Details

But while we do seem ready to clear the bar of basic tramway designs like low platforms and dedicated right-of-ways, those alone does not a good project make. A point that another comparative transit expert, Marco Chitti, makes in the thread embedded below is that for a tramway to be successful you also have to design around other modes of vehicle-transit conflict.

That means things like nixing through-traffic and left-hand turns on Jefferson. As he states down-thread: “This is why I raise an eyebrow when considerations about the effects of urban integration choices on transit performances are brushed away with shortcut arguments like “we will use TSP and will give trams/buses their own RoW so they will be fast!” because that’s not enough.”

If we want our region’s once-in-a-generation transit investment to be successful, we have to be willing to make hard calls and do it right. If they’re not willing to do that, then Pete Buttigieg and the DOT should refuse to fund it and use the money for regions who are. We only get so many bites at the apple.

7 thoughts on “N-S MetroLink: Devil In The Details”

  1. This was a flop of an article and the comparison to KC and Cincy is way off. This isn’t a streetcar, it’s surface running LRT. And they hardly became popular “during the Obama administration.” LA’s first line opened in 1990. Denver’s in 1994. Portland in 1986. Sacramento in 1987. All of these systems (and many more) have large portions of street/at-grade trackage and low floor vehicles.

    The Jefferson Alignment is basically standard fare for American LRT. It will have its own lanes, signal priority, a large transfer station with the Red and Blue lines, etc.

    Yikes…this was bad lol.

    1. Always appreciate some constructive feedback from my readers! To first address the criticism, I think the feedback to the article has definitely shifted my opinion and made me less hostile to it. It’s pretty obvious that this isn’t going to be as good as the existing metrolink typology, but maybe we can’t afford for it to be and maybe that doesn’t make it a bad project. It’s just a lower quality one, which is fine. Life is about trade offs and maybe grade separation doesn’t pass cost benefit analysis. All that being said, 1) you seem to have details on this project that aren’t public, would you mind sharing where you found the signal prioritization or the plans for the transfer station? 2) you’re being an asshole for no reason and it’s hard to convince anyone of anything if they think you’re an asshole. Bring a different energy next time man, you make it sound like you designed the system yourself.

      1. Here here. LOL, you make some good points but give the man some credit. The article brings up important points that are worth hearing about. LOL, along with critiquing the article, how about joining in with the effort to really make this happen? Be great to add to our STL mass transit.

      2. I didn’t design the system myself and have no inside knowledge. I’ve read all the same articles, seen all the same presentations, watched all the same CMT/Bi-State videos that everyone else has access to. I also be sure to really listen to the agencies leaders when they speak about it. Which is where I saw the words about a direct transfer come right out of Taulby Roach’s mouth.

        Yes, it is different than existing MetroLink and that’s because MetroLink is the exception to the rule for light rail. It’s one of the only US systems that functions closer to heavy rail instead of traditional LRT systems. In some countries MetroLink style systems are called “light-metros.”

        I also know how the original alignment that opened between the Airport and 5th & Missouri in 1993 came into existence. The St. Louis region put up very little actual money for the alignment, instead the majority of the “local match” came from the old, unused railroad ROWs that litter the region. It was really an ingenious move. It essentially allowed the region to build 13 miles or so of quality LRT mostly on the federal government’s dime. After that happened, the Feds made ROW considerations unacceptable for local matches. Fast forward to the early 2000s and Metro tackles the Cross-County extension, however without the ability to use the old Central Belt and Forsyth Branch ROWs (where the Blue line was eventually built) as a local match, they ended up receiving no federal dollars and funded it entirely with local money. It ended up going over budget, lawsuits were filed and the 100% local debt load from that extension pretty much killed any LRT extensions for the next decade and half or so. Which is about how long it took for St. Louis County’s Prop A funds (passed in 2010) to recover enough to start providing enough for local matches for projects (hence the ongoing North County Connector study). The City’s expansion tax for this (and the previous 2018 N/S proposal) wasn’t passed until 2016.

        Never mind there is no solid old railroad ROW that could run N/S like existing MetroLink. IT’s Tucker Tunnel was filled in years ago, but even then that only extended south to Washington Avenue and dead-ended at a brick wall. North of Tucker, IT had street level tracks along Hadley that took to the Iron Horse Trestle. That still exists but is owned by GRG to be converted into an elevated park. Besides, the Iron Horse would be an insanely stupid route to run MetroLink to the North Side, it wouldn’t service any of the residential neighborhoods but instead waterfront industry and low density warehouses. Union Pacific’s Desoto Subdivision would be decent for the Southside. But it doesn’t break off the UP mainline until west of Vandeventer and it’s still in use by UP and Amtrak’s Texas Eagle service. St. Louis doesn’t have the density or traffic levels to support tunneling under city streets that are already significantly wide and overbuilt like Jefferson. In fact part of the reason Cross County went over budget was the demand of wealthy Clayton and U-City residents that the train and stations be tunnels or below-grade on much of the Forest Park Parkway alignment.

        Well…that was a lot.

        TL;DR. Previous MetroLink lines were built using old ROWs with the initial alignment essentially being partially paid by them. The Feds no longer allow that. Besides old ROWs don’t exist for MetroLink in North and South St. Louis, the ones that do are either in use or poorly located, or both. Jefferson is more than wide enough to support existing car traffic and dedicated LRT facilities (bike facilities have moved to 20th as part of the mobility project underway in that corridor). In fact the ongoing rebuild is being built with N/S in mind. Jefferson is also surrounded by some large institutions (Harris Stowe, South City Hospital, Doorways, several schools just south of the old PI site) companies (Wells Fargo and NGA), stadiums and event centers (MLS, Union Station and Fairgrounds Park), and most importantly residential areas. Some census tracts in South St. Louis along Jefferson can number greater than 10,000 people/square mile. Well past the sweet spot for LRT construction which is between 4,000 and 5,000 per square mile.

        All I said was your article was a flop and was bad. Which…it was. If you’re going to publish publicly then learn to accept the criticism. I even pointed out multiple existing systems to prove my point. My energy is fine. No hard feelings.

        1. “All I said was your article was a flop and was bad. Which…it was. If you’re going to publish publicly then learn to accept the criticism. I even pointed out multiple existing systems to prove my point. My energy is fine. No hard feelings.”

          I did it without name calling too.

        2. This is really informative context, and I’m really glad to have it under this article, so thank you! My point is that the systems you point to notoriously underperform in terms of ridership per mile compared to systems with more heavy rail qualities (like the existing metrolink). Now, as you cite in your more detailed comment, perhaps recreating some of those heavy-rail-like qualities is infeasible given our economic constraints, and obviously, much of that underperformance of the systems you compare us to has to do with land use policy in those cities (as I link to in the article, I think St. Louis needs to revise its land use policies ahead of N-S: https://kirkwoodgadfly.com/to-unlock-promise-of-metrolink-tod-is-key/. Still though, the point of the article was to convey two things to a much less engaged audience: 1) the system we will be getting with N-S will be of substantially lower quality than the existing lines, and 2) there are relatively cheap ways to do a grade separated system if you’re willing to bear additional political costs. I stand by both of those points.

          I appreciate the additional engagement. You’re obviously free to comment whatever you want, but I just want to encourage you to take a shot at actually persuading people (both me and the people who stumble across this post and read your comment). You clearly have the knowledge needed to do so. Persuading people is a lot harder than saying “Yikes this was bad… lol,” and perhaps it’s not as fun, but it also has a much better shot at improving things, which I think is why we’re both here.

Leave a Reply