05 January 2014

Focus On: Mission Bay Grade Separations

The San Francisco downtown extension, as environmentally cleared and approved, ends at the existing grade crossing at 7th and Mission Bay Drive (formerly Common Street).  Somebody apparently forgot that this will leave grade crossings at Mission Bay Drive and busy 16th Street, where it still isn't exactly clear how to cross a trolley bus line with a 25 kV electric railroad at grade.  With train traffic expected to double in the next 15 years, conditions at those intersections are expected to become unacceptable, a problem which is now being considered along with redevelopment ideas for the Mission Bay neighborhood.

A number of radical solutions were proposed by outside groups (including the city!) coloring outside the lines of the environmentally cleared and approved DTX project, thus causing significant unease at TJPA headquarters.  To regain control of this issue, the TJPA commissioned a white paper study by engineering consultants Parsons Transportation Group and Jacobs Associates (organizations with considerable tunneling expertise), entitled Proposed Alternative Tunnel Connections to the San Francisco Downtown Rail Extension.

Grade Separation Tunnels Proposed
The billion dollar PowerPoint slide,
from the city's presentation,
used as pretext for dismissing a trench

Curiously, on page 1 of the white paper, the most logical option is summarily dismissed: 
  • City SPUR Presentation Option 1 – This option consists of a trench trackway in between the I-280 freeway pilings. The presentation noted that this option may not be feasible, and therefore it is not discussed further in this white paper.
This seems like a rather weak rationale for not further discussing or studying a particular option: somebody else, presumably far less-qualified than the Parsons / Jacobs engineering team, happened to cast doubt on it.

One could cast a similarly dismissive eye on the entire study.  It boils down to considering various complex and over-wrought tunneling options (considering arcane subjects such as boring machine types, ventilation zones, and even tsunami run-up) in order to grade-separate these two streets for an estimated cost of well over a billion extra, compared to the current DTX plans.  That's right, BILLION.  Two grade crossings.  If you do the math, at least $500 million a pop works out to about ten times what a typical grade separation should cost.

The study's conclusion, not surprisingly, is to proceed with the TJPA's environmentally cleared and approved DTX plan so as not to upset or delay anything, with the understanding that nothing in the current plan precludes grade-separating in the future, if someone finds more than a billion under a rock.

Grade-separating a couple of streets doesn't need to be so ridiculously expensive or complicated, of course.  It requires an open trench built under the I-280 freeway, along the current alignment of the tracks and between the large pilings that support the freeway.  This trench would pass under Mission Bay Drive and 16th St, then climbing steeply (2.5% grade) to meet the existing portal of Caltrain Tunnel One.  No new tunnels would be built, and the freeway could later be torn down if the city so desires.

This modest 1/2 mile grade-separation trench extension would not require complex phasing; it would be built as part of the DTX tunnel, at an incremental cost far smaller than the proposed grade separation tunnels, with approval of a minor supplemental EIR.  There would be no need for fancy U-walls or tunnel stubs or temporary structures that would have to be modified at some future date with great disruption to train service.  The vertical profile would look very much like this:

Grade separation trench profile view (from west side of tracks)
There are a number of arguments raised against a trench at this location.

Constraint #1: Freeway Pilings

1987 grade separation trench concept
To see what a trench built between the I-280 freeway pilings might look like, we can set our way-back machine to the year 1987 -- more than a quarter century ago -- when Morrison-Knudsen Engineers performed a study entitled Peninsula Commute Service Interim Upgrade Study that included the I-280 diagram at right.  In 1987, the idea of building a rail trench under an elevated freeway wasn't simply dismissed without further consideration.

I-280 straddle bents at 16th Street
And suppose the trench did come too close to some I-280 piles?  The 1989 Loma Prieta earthquake triggered a massive seismic retrofit of the I-280 freeway, during which Caltrans demonstrated a surprising ability to add or move bents under an active freeway.  Even temporary solutions would be acceptable, since the city plans to tear down this section of I-280 in the short to medium term.  The biggest shame would be to spend a billion-plus dollars to avoid freeway pilings that are about to be dismantled.

Constraint #2: Major Sewers

There is already a large sewer passing under the tracks at Division Street (see profile diagram above).  Furthermore, the San Francisco Public Utilities Commission is in the planning stages of a $1 billion sewer project known as the Central Bayside System Improvements, which will involve tunneling a new 27-foot "Channel Tunnel Connection" underneath the tracks (see profile diagram).  The approved DTX project artfully avoids the sewers by surfacing before the sewer crossings, but a trench or tunnel would inevitably conflict with them.  The TJPA study assumes the new sewer tunnel forms the controlling constraint on the vertical alignment of a grade separation rail tunnel, which sends the rail tunnel so deep underground that it can no longer connect to the existing Caltrain Tunnel One.  That's what makes the proposed rail tunnels so long (all the way to Cesar Chavez!) and so expensive.

This raises an interesting question of underground right-of-way: should the rail tunnel yield to the sewer tunnel, or the other way around?  Without very much expertise in the art of pumping enormous flows of sewage, one might guess that sewage is more flexible than trains when it comes to changes in vertical profile.  The SFPUC and TJPA need to be prodded to come to a more realistic agreement on this issue, rather than flushing well over a billion of scarce rail dollars down that new sewer.

Constraint #3: Track Gradient

There isn't a lot of distance between the portal of Tunnel One and the 16th Street crossing.  If 16th Street is left where it is, the tracks need to dive about 28 feet down in that short distance, resulting in a rather steep grade of 2.5%, after accounting for vertical curves.  The TJPA study points out:
  • insufficient length is available to transition from the tunnel below 16th Street to at-grade at Tunnel One while complying with vertical grade and curve requirements of DTX, Caltrain design criteria, or accepted railroad engineering practice.
Tunnel One North Portal
2.5% is not an impossibly steep grade for EMU electric rolling stock of the sort contemplated by Caltrain and the high-speed rail authority; portions of the high-speed rail system are being engineered with grades approaching 3.5%.  If there is a problem here, it is clearly with overly conservative DTX vertical grade requirements and Caltrain design criteria.  American "accepted railroad engineering practice" is totally irrelevant here, being firmly rooted in freight train operations.

If anybody tells you a 2.5% grade is impossible, ask for a second opinion, preferably from a European engineering consultant.  And there are ways to reduce the required dive: 25 kV overhead electrification clearances could be reduced by several feet by using rigid overhead conductor rails (instead of wires), or 16th Street could be raised by a few feet.  It sure is tight, but it most definitely can be done.

Constraint #4: Keep Caltrain Operating

Shoo-fly concept
Old wye track viewed from
under Mariposa, since removed
Pair of yard tracks to the east of the
main tracks, looking north from 16th
Keeping Caltrain operating during construction of the DTX connection is obviously a major concern.  It would be necessary to construct a temporary shoo-fly track on the east side of the existing tracks, to enable the trench to be constructed where the current tracks exist.  There are already some siding tracks to the east under I-280, but the challenge is to dodge out of the forest of freeway bents at Mariposa Street without violating side clearances between trains and concrete, along the path of an old wye track that was recently removed.  This can most likely be done as shown in the diagram at right; it is a puzzle that requires all the ingenuity our transportation industrial complex can muster.  If bents must be moved, then move them.  If clearances are truly insufficient, CBOSS can be used to enforce single-occupancy of the narrow section.

The TJPA study describes how the DTX project would be staged to allow tying in to a future grade separation tunnel.  The staging concept described would require disconnecting the Transbay Transit Center for several weeks or months while the new connection is made, with the 4th and King terminal used on an interim basis.  This staging concept is surely more time-consuming and disruptive than the trench tie-in work implied in the shoo-fly diagram above, which could probably be pulled off over a three-day holiday weekend.  Far more complex operations were pulled off on that time scale at the Bay Bridge.


Dismissing the trench grade separation alternative is premature and poorly justified.  Boundaries should be pushed and "requirements" should be questioned.  Coordination between the city, the TJPA, Caltrans and the SFPUC should be improved with the help of the Mayor's Office, and if necessary the Governor.  Most importantly, the small sliver of land between Mariposa and Sixteenth should be temporarily preserved from development until the DTX is completed, so that a shoo-fly can be built there.

Without a doubt, this trench would be a tough grade separation to engineer and construct, but when the only alternative is a billion-plus bored tunnel, there are suddenly a billion-plus reasons to take another look to make it work.  With the extensive means of excavation to be mobilized for the DTX project, the incremental cost may be small enough that Mission Bay and Sixteenth can be grade-separated concurrently with the DTX, without delaying the project.

The TJPA and its consultants should be sent back to the drawing board to study the feasibility and cost of the trench option in much greater detail.


  1. Excellent post, but I can't help but think that even though it might be possible to snake the trench and the shoo-fly tracks through the freeway pillars, the presence of the freeway is an unnecessary complication to the task of extending the DTX to Tunnel #1. Accelerate the removal of I-280 and issues #1 and #4 are now dramatically easier to deal with, and also introduce the option of creating a 16th St/Mission Bay Caltrain station. Given that the funding isn't in place for the DTX yet anyway, the extra SEIR time to include the freeway removal shouldn't be a show stopper, providing the agencies involved can line up behind the solution.

  2. Agreed. SF has the political will to remove the freeway, and the plan doesn't seem to have encountered much opposition. Removing the freeway first would make the grade separation both easier and cheaper.

    What I see here is that the PCJPB and the city really aren't communicating on this issue.

    1. It's not clear that the PCJPB communicates with anyone.

      The City should just rip down the freeway ASAP and leave Caltrain with a fait accompli.

  3. Also agreed with the above two commenters. If the removal of I-280 north of 16th street is likely to happen in the short-to-medium term, is the Caltrain DTX expected to be completed much before that? Probably considering them as a single project would yield significant benefits in terms of neighborhood disruption and construction phasing - the area would only have to be torn up once.

  4. Some more gems from the draft study:

    1) a 10-minute "turnover time" i.e. platform re-occupancy time is assumed from when one HSR trainset leaves the TTC until another can occupy the same platform. This is because of "additional clearance time required by the 1312-foot trains to traverse the CP Second Street interlocking"... self-inflicted by the TJPA and its consultants, who curiously point out elsewhere in the same study that a "shorter interlocking allows trains to clear the interlocking sooner and accelerate as soon as the rear end clears the last turnout". One wonders why they are still incapable after all these years of shortening CP Second Street to a more efficient length. What's frustrating about this is they understand the problem perfectly well but don't think it's important.

    2) "Although trains using the straighter Mission Bay Tunnel Route alternative may have a
    1.5 minutes shorter trip time than the Approved DTX Route Plus Grade Separation Tunnel, this time savings is negligible for the commuter rail and intercity rail service involved.
    " This statement, although made in the context of discrediting an alternative that deviates from the approved and environmentally cleared DTX design, betrays a cavalier attitude about run times that is especially toxic when it comes to HSR. 1.5 minutes is a very significant difference. A minute here, a minute there, and pretty soon it isn't HSR anymore. Similarly, daily commuters do indeed care about small handfuls of minutes. If they didn't we would be talking about electrifying Caltrain. With attitudes like this, one is left wondering whether our transit industrial complex is even constitutionally capable of engineering operationally efficient solutions!

    1. Clem, please could you link to or post a copy of the white paper?

  5. The Realtors....I mean the City, is pushing for the end of 280 to be cut back to before 16th Street, as you mention above. This would then transition to a surface street that would follow the current rail right-of-way and curve around to connect to King Street. (That is actually an interesting and related subject, because a grouping of apartments on Berry Street are directly in the way of the best road alignment. Sigh.) My question is this- could the trench you propose be covered later by this new road? What would be the costs and extra engineering required to make that plan workable?

    1. Covering a trench makes it a tunnel, which invites all sorts of complications with fire safety, ventilation, etc. I wouldn't really recommend that. An open trench allows the street grid to be re-formed over the tracks, which is the main point of sinking the tracks... not so much to gain a few extra slivers of an acre. In fact, the entire Mission Bay station should have been in an open trench, but this is America and a train station isn't a proper train station if it doesn't have a mezzanine level.

    2. Cut-and-cover isn't really so hard to do -- the fire safety issues are real, but were solved back in the 19th century (step one: provide uncovered sections fairly often), while the ventilation issue is nonexistent if you have electrification.

    3. "... the ventilation issue is nonexistent"
      Not so. Fire and evacution, again.

      If your think "uncovered sections fairly often" is on anybody's table at Caltrain/CHSRA/TJPA/SFPlanning, you're wrong.

      If you think "19th century" life safety design (or even 1980s or 1990s requirements) are in force, you're wrong.

      Always think "project cost maximization" and you'll be heading in the right direction.

  6. Anyone who thinks the 280 tear-down idea is "likely to happen" and hasn't engendered "much opposition" doesn't read anything but streetsblog. Every MSM article on the topic is plastered with commenters opposed to the idea, and if it ever comes to a vote (and in San Francisco, it will), it'll be voted down in a heartbeat.

    1. There is some big developer money involved... and either way, the tracks should go in a trench to reconnect the street grid at grade. Whether that occurs under the freeway or under the sky remains to be seen.

    2. It seems like SF's big political powers are on board with the idea. It's not easy to judge opposition from blog comments, but people haven't exactly been railing against the plan publicly. I don't think it's worth writing off.

  7. Of interest, the SF Planning Department just published the scope of work for a proposed Railyard Alternatives and I-280 Boulevard Feasibility Study, to include five topic areas:

    - I-280 tear-down
    - 4th & King rail yard relocation
    - 4th & King redevelopment
    - DTX redesign
    - DTX loop track (it's baaaaaack...)

    Phase 1 = $350k for 6-9 months (~3 EP), Phase 2 $1.1 million for 12-15 months (~5 EP).

    1. Nice work if you can get it.
      And America's Finest Transportation Planning Professionals surely get it.

  8. And there goes the public Right of Way which was need for the shoo-fly while the simplest, cheapest, most direct grade separation was constructed.

    Well done, San Francisco. Well planned. The City That Knows How, indeed.