A view of I-94 in St. Paul from the Victoria Street bridge.

Thoughts on I-94, and Yet Another Idea

Putting rail in the I-94 trench is an idea I’ve seen pop up on occasion, whether regional rail or a metro/subway. In theory, this would provide a faster downtown-to-downtown connection for fairly cheap. I support this idea in the broadest sense possible! We lack a good downtown-to-downtown connector. The Green Line is good for local trips along University Avenue but takes far too long from end to end, and the various express buses cater mostly to commuters and thus have limited operating hours, especially on weekends.

Taking the bus from my house in northeast Minneapolis to downtown St. Paul and back takes about as much time as it does to bike. With all that said, I don’t think the I-94 corridor is the right-of-way with which to do rail or bus rapid transit (BRT).

A graphic map titled "Twin Cities Boulevard" and reading "Rethinking 94 aims to convert the freeway to an at-grade boulevard. Heavy Metro — use the trench!"
Source: @MajorBoondoggle on X, formerly known as Twitter.

In general, I’m opposed to replacing highways with rail. Not because I think highways are good — far from it — but because many of the design choices that make a “good” highway often make them difficult to retrofit for rail. Particularly, the way they tend to skirt around the downtowns of large cities instead of piercing straight through the center. This makes the conversion to rail quite difficult as, ironically, that last-mile infrastructure is now necessary.

Naturally, these are the places where land is at a premium, delivering a one-two punch of cost increases due to greater complexity and higher land values. There are also some specific issues with a potential subway line. It’s within about a quarter mile of both the Green Line and the coming B Line and — as much as I hate saying this, because I know how most people who say stuff like this think — the area around those corridors doesn’t currently have the density to support three high-frequency routes.

And Now for Something Completely Different

I’ll get back to the question of “What to do with I-94,” but first I’d like to propose an alternative downtown express connection: using the current Amtrak Empire Builder/Metro Transit Northstar route and squeezing in a dedicated passenger rail track. As I’ve discussed previously, our railroads have shrunk since their heyday but still own the same size right-of-way, so adding this capacity back is relatively easy. This would be operated as an extension of the Northstar, of course, and would still start at Target Field Station and end at Union Depot in St. Paul. As such, it would provide better connections to not only en-route destinations like the University of Minnesota, but also to points well beyond the Midway area. It would allow for one-seat rides all the way up to Big Lake and two-seat rides to St. Cloud. 

As an aside, this would work better as an extension of the Dan Patch Line to avoid turning trains at Target Field Station, but that’s a story for another time.

Google satellite image showing freeways and a proposed rail route from Target Field Station to Union Depot.
My proposed route from Target Field Station to Union Depot. Source: Google Earth

A stop at Snelling and Marshall avenues is basically a given. The A Line and future B Line connections and nearby attractions make it too good to pass up. Additional stations at the U of M near Stadium Village and one around West 7th would also be nice, but I’m worried those would cut into transit times.

I know this undermines my previous point, but if providing a fast and convenient connection is the main goal, then trip times need to stay under 20 to 25 minutes in order to compete with driving and the I-94 express bus. That means averaging around 30 to 37 mph. Historically, this is about the fastest this journey would take, with most routes in the 30– to 45minute range, and none of these trains made intermediate stops. Doable, but tricky. Local bus routes already serve these areas pretty well, and while a direct ride would be nice, a single transfer in either downtown or at Snelling probably isn’t going to dissuade too many riders.

I’m skeptical that BRT or a subway are suitable replacements for the status quo partly because of this need for higher speeds. Needing 40-plus mph speed limits on the boulevard to keep travel times competitive basically defeats the purpose of the conversion, as it will be almost as loud and hostile as the highway. Similarly, the travel times between Target Field and Cedar-Riverside (confusingly named Seward on the first graphic) and Capitol-Union Depot come out to 20 minutes, so any use of the corridor for a subway would need to average 96 mph on the remaining stretch of the corridor, nearly twice what any subway equipment is designed for. At 20 mph, roughly the average speed of the Chicago L, that’s 44 minutes downtown to downtown. That’s a lot of investment for less than 10 minutes of improvement, especially since my own legs can almost keep up with that.

The Target Field to Union Depot route has the additional benefit of being supported by the Minnesota Department of Transportation (MnDOT) and the Legislature; it also has operational precedent, as Northstar occasionally runs special trains to Union Depot. This route can use existing space along the right-of-way, as most of the route was historically double tracked but has since been consolidated as technology has advanced and traffic patterns have shifted. Not to mention that this option can actually provide service overlap — I-94 can stay open for motor vehicles while a new rail corridor is built. I doubt people would take it well if suddenly their route to work just doesn’t exist anymore. I know what’s right isn’t always what’s popular, but in a democracy the latter is more important.

A special Northstar holiday train to Union Depot.
A special Northstar holiday train to Union Depot. Source: Trains

If I had to choose between BRT and a subway, though, I would choose BRT. It has less capacity, but is less expensive, has a similar travel time and can utilize mostly existing equipment in Metro Transit’s fleet. Most subway systems use a third rail pickup — literally just a steel bar on the ground that a pickup shoe rides over — as opposed to the overhead catenary that the LRT (light rail transit) uses. Furthermore, most subway cars are single unit EMUs, which allow easy scaling of train lengths, whereas Metro Transit uses fixed three-unit sets of vehicles that bear more similarity to streetcars than subways. Mixing power and equipment types is certainly possible — a few places around the northeastern United States do so — but on a system as small as ours I think it’s best not to run two parallel but different systems.

Losing the Plot

This could actually make a good test route for trolley buses. They use a very similar OLE (overhead line electrification) system to the LRT, to the point where most of the infrastructure, like substations and catenary poles, could be shared. Likewise, the current trolley bus offerings, like the Newflyer Xcelsior, are based on some of the same buses Metro Transit uses. That said, I think the B or D Line would be better suited for this; the D Line is running and the B Line is about a year from completion, so future LRT infrastructure could leech off of them more easily.

Speaking of buses, have you ever noticed we have only about a dozen routes that operate between the two cities? If we don’t count commuter buses and routes that don’t go near a downtown (like the A Line or the 30), we only have five routes connecting the central cities: the Green Line, 3, 21, 61 and 67. Improving connectivity elsewhere will have knock-on effects with respect to this project.

Back to the Matter at Hand

So, then, what to do with the I-94 trench? I think it’s a prime spot to experiment with off-arterial density. People don’t like living near busy streets with lots of traffic, so why put them there? Or rather, why put in a six-lane road where we don’t need it? Keep St. Anthony and Concordia avenues more or less where and what they are. Add another lane if it’s really necessary, and local bus service, of course. But much more than that and I fear it will cut into the usable space too much. A six-lane road plus bike paths and bus lanes leaves little room for anything else.

A cross section illustrating a possible Twin Cities Boulevard configuration.
A cross section of a possible Twin Cities Boulevard configuration. Source: Our Streets Minneapolis

Furthermore, how much of an improvement will this be for livability, really? Sure, the speed limit will drop to somewhere around 30 to 40 mph (I assume), but now the sound barriers are gone and traffic is at-grade, making both sights and sounds much harder to avoid. Moving across is theoretically easier with all the streets reconnected, but the existing problems with driver conflicts will persist and now pedestrians will have more lanes to traverse, meaning more conflict points. Looking to the past for inspiration — as urbanists are wont to do — many cities actually set major developments away from large arterial roads to avoid these problems. Let all the other transit options — both current and proposed here — bear that burden and give people some peace and quiet.

Black and white bird's-eye photo of a typical prewar Chicago neighborhood.
A typical prewar Chicago neighborhood. Source: @carlcicahn1.

That said, I don’t think the 94 corridor is completely useless from a transportation perspective. In fact, I think it can fill perhaps the second greatest missing link in our transportation system: the Midtown Greenway extension. It’s both shorter and flatter than the proposed Ayd Mill Trail routing, as well as the current options for biking along Como and Summit avenues. This is a much safer routing, too, as large segments of the existing routes offer little protection from motorists — usually just a stripe or two of paint — and even that is questionable at points. Measures such as raised crossings and improved sight lines not only make this route safer for those using it, but also for those trying to get across the former trench.

A brand-new pedestrian mall would be completely isolated from vehicle traffic and allow us to get things right the first time. A pedestrian mall here wouldn’t need to be as wide as Nicollet Mall, since there are no buses and (probably) fewer people. Something more akin to Milwaukee Avenue in the Seward neighborhood of Minneapolis is what I have in mind. At about 40 feet, this is smaller than the inner two lanes of 94, including the center median, which leaves plenty of room for redevelopment. Density-wise, I’m thinking something similar to Exchange Place in New Orleans, but make the mall about twice as wide.

Too Little, Too Late?

In a classic case of comedic timing, I started this article before the release of MnDOT’s corridor alternatives, and now have to face the question of “Am I wasting my time?” With alternatives selected, is any of this actually worth debating? Well, MnDOT has only just started taking feedback on these alternatives (its survey is open “through the fall”) so I suppose it’s still something. If I’m wrong, then please indulge me as I shout into the void.

Of all the options, the only one that I think is workable is the At-Grade–A and At-Grade–B, with personal preference being A. Center-running bus lanes are superior as they don’t have to deal with right-turning vehicles and are generally safer for pedestrians. It also doesn’t have as much greenwashing by way of extra grass strips like option B. That said, it suffers the same problem of feeling like a waste of space.

I know MnDOT says not to interpret this as the final design, but lanes take up a set amount of space, probably four 11-foot vehicle lanes, two 12-foot bus lanes and two 8-foot sidewalks for the bus stops. That adds up to about 84 feet, or two-thirds the width of I-94 currently, and that doesn’t include sidewalks or bike paths. Streets like University Avenue can get away with this since they’re plotted as part of the grid, but trying to squeeze all of this inside an existing block leaves little room for anything else.

Granted, MnDOT mentions that the frontage roads may be removed as part of this, in which case I view these as acceptable. All the other proposals are still car sewers and offer little if any improvement, and a few of them are actually worse.

Cross-section and diagram of the "At-Grade — A" proposal from MnDOT.

The Elephant in the Room

When this is all said and done, and if the end result is something that isn’t a highway, then what do we do with Highway 280? It exists mostly as a shortcut between I-94 and I-35W, so getting rid of 94 makes the smaller highway borderline useless. I haven’t thought as hard about what should replace it, so I’m going with something inoffensive: a four-lane road, two of them bus lanes, plus a two-way bike route. Shift Route 87 over onto it and County Road B, have the 227 take the old route and maybe start an express route from the northern suburbs down to the Westgate or Raymond LRT station.

And finally, while I have issues with some of these ideas, I do appreciate that people are thinking so far outside the box on how to improve these Twin Cities, and I still prefer any of them to the status quo. Even better, some of the ideas are actually being taken seriously, not just by other nerds but by those determining the corridor’s future. The latter may not be dreaming as large as we are, but if these proposals came forward even less than five years ago, I feel like none of them would have been considered.

We can always do better, but for the moment, at least, I think we could also do worse.

Ian Gaida

About Ian Gaida

Pronouns: he/him

Foamer, Maker, Outdoorsman. bsky: @trainsfan Youtube: Midwest Urbanist