San Francisco’s Central Subway: Part I – The Roots of Tunneling

As a native New Yorker I spent years riding the subway before moving to San Francisco. Although San Franciscans may assume that this experience is not one I would want to repeat, the city’s lack of underground transit has strengthened my conviction that such a service is an essential component of a livable city. So I look forward to the completion of the Central Subway (CS); it will be a welcome option for people who live and work along the city’s north/south axis.

Map of the Central Subway route. Click to see the complete map of the existing T Third St. Line and this Central Subway extension.

The CS route will connect to the existing “T” – Third Street Line near the CALTRAIN Station, south of Market Street to Chinatown and in the future, I hope, extend the connection to North Beach and Fisherman’s Wharf. Perhaps this is asking too much, however, a subway system’s evolution is not predictable.  Consider London’s underground system–the world’s oldest–has been a work-in-progress for 150 years and where would London be without it today? For that matter where would San Francisco be without its current underground rail?

Granted, subway construction is expensive and risky by definition. However, we are fortunate that since the pioneering days of subway building in the Victorian era, the civil engineering and construction community has accumulated a lot of practical experience.  Although the builders of that time would recognize the concepts used by those in charge today, they would be astounded by the advances in material science and information technology that have turned their dreams into reality and expanded the scope of what is possible.

Figure 1: Underground Path of the Central Subway Tunnels Legend: Proposed Stations are, 1) Chinatown; 2) Union Square Market St.; 3) Moscone, Existing Tunnels A) Stockton Street B) BART/MUNI

Figure 1, depicted in exaggerated scale, shows the path of the CS Tunnels through the subsurface geology.  Keep in mind that there will be two tunnels, each carrying one track.  Tunneling will begin at 4thStreet under the Freeway (right hand side of the graphic) and continue north, through soil known as the Colma Formation.  “Colma” is made up mainly of dense sandy material deposited on bedrock 80 to 120 thousand years ago. It can be seen exposed along our city’s pacific coast. Moscone Station (MS) and the Union Square Market Street Station (UMS) will reside in this formation.  Between MS and UMS the tunnel will pass under BART/MUNI Market Street Tunnels.  After UMS the tunnel will run under the Stockton Street Tunnel (Opened 1914) through the Franciscan Complex.  This is a geologic formation that passes for bedrock in our seismically active patch of land. “Complex” refers to the fact that this is crumpled up prehistoric sea floor hardened over time—which for the tunnelers amounts to a mess of rocks of variable hardness. When the tunnel reaches the Chinatown, it reenters the Colma formation.  These varying conditions are challenging and special tools will be required to do the job.  The primary tunneling “tool” will be an Earth Pressure Balance Tunnel Boring Machine (EPB-TBM). CS builders will employ two machines to dig the tunnels in parallel, each 20’-8” in diameter.  EPB-TBMs belong to the TBM family of equipment.  A TBM is essentially a steel cylindrically shaped shield that is fitted with a cutting face at its leading end and an open end at the back.  The back end houses the machinery that conveys away the soil that was cut away at the face and the equipment that installs the precast concrete elements that make up the circular tunnel lining. TBMs can have different cutting faces.

The face types vary according to the soil conditions to be tunneled through.

Figure 2: Earth Pressure Balance Tunnel Boring Machine Legend: A) Archimedean Screw B) Jack C) Tunnel Lining

As shown in figure 2, the EPB-TBM is engineered to balance the pressure of the soil in front of the steel cylinder with the soil that was removed by the cutting face. The steel cylinder advances, bit by bit, pushing forward (with hydraulic jacks) off the lining that was placed behind the shield. The soil that is cut passes through the cutting face, into a chamber where it is then extracted by an Archimedean screw (a.k.a. “extraction worm”) that delivers the spoil to a conveyor belt.  The screw’s orientation and rate of removal is controlled to compensate for the pressure difference between the cutting face and atmospheric pressure.  This compensation or “balance” enables tunnels to be dug in unstable variable ground conditions.

Tunneling lore has it that EPB-TBM itself is the modern incarnation of an idea inspired by the lowly shipworm (a.k.a. Teredo Navalis), which is actually a long mollusk (same family as a clam) with a shell only at its head (see cartoon below).  As the shipworm munches its way through wood, and advances, it secretes a calcareous “lining” around its body.  Its tail end has siphons that act as a snorkel of sorts as well providing waste disposal.  Teredo bores to eat the wood and bacteria contained in it.

Figure 3: “Shipworm” (Teredo Navalis), Legend: A) Shell, B) Calcareous “Lining”

Teredo Navalis was (and remains) a problem for wooden structures in sea water. French-Anglo-American Civil Engineer Marc Isambard Brunel observed the shipworm boring its way through ship timbers during his navy service at the Chatham dockyards.  Close observation led Brunel to conclude the creature had its own natural “tunneling shield”.   Thus inspired, he patented a device, namely a shield, for ‘Forming Drifts and Tunnels Underground’ in 1818.  Excerpts from the original patent are included below (from Copperthwaite, 1909).

Figure 4: Brunel's Shield, from drawing attached to Specification of Patent No. 4204 of 1818 (Copperthwaite, 1906)

Figure 5: Brunel's Shield, from drawing attached to Specification of Patent No. 4204 of 1818 (Copperthwaite, 1906)

The basic idea was to create a shield that would protect workers digging a tunnel in soft soil; and would enable one to dig tunnels under waterways or terrain more common in coastal areas, where Cities were growing.

His patent was ambitious, “it is hardly too much to say, (it) covers every subsequent development in the construction and working of tunnel shields” (W.C. Copperthwaite, 1906).  As you can see from the patent graphics the original idea was for a cylindrical tunnel shield. However, when planning the tunnel under the Thames Brunel opted for a rectangular shield.  Brunel needed a very large shield. The Thames tunnel was to have two parallel carriage ways, resulting in a cutting face roughly 880 sq. ft. in area (CS’s twin tunnels add up to about 640 sq. ft.).  After considering the tools available, Brunel and his team judged that the technology of his day was not up to the task to build a cylindrical shield.

Figure 5: Brunel's Shield used to build the Thames Tunnel Legend: A) Cast Iron Shield Frame with adjustable (with poles) timber face B) Screw Jack C) Brick Lining

Instead he constructed a rectangular shield 37’-6” wide, 22’-3” tall and about 9’-0” long.  The cartoon above shows the general arrangement.  The shield pushed off the lining behind it using screw jacks.  Digging the tunnel was a difficult job on all fronts and the work was delayed many times for long periods. It took two generations of this shield to complete the digging, first between1825 and1828, and then an improved version from 1835 to1843.  In the end, this pioneering effort did not meet with financial success and shield tunneling did not set the world on fire.

6 Responses

  1. Cris Subrizi says:

    Some answers to some of the questions posed in the commentary:

    RE: TBM cutting rate – per the MTA fact sheet they are planning on 40-ft per day (see:, for more details), RE: Hwy 1 tunnel at Devils Slide and the Waldo Tunnels – these are examples of mined tunnels without using a TBM. The Stockton Street Tunnel (above the Central Subway) was also mined without a TBM.

  2. Such a good explanation of the soil conditions and drilling procedures. After much exposure to this project, I finally understand the physical challenges. Hopefully what isn’t covered will be so in a future piece, including how this under grounding compares with that done for BART in the 70’s and what constraints all the underground infrastructure in the lower part of the financial district may impose. The bigger questions are how the costs justify the benefits. A comparison with the Manhattan subway misses the huge differences in densities and traffic, both vehicular and pedestrian. Surface shuttles and streetcars might be a much more cost effective way to link the TransBay Terminal area with Chinatown and North Beach. Curious to know what the transportation planners may have to say.
    Still, a very informative and interesting article. Thanks.

  3. John Parman says:

    This is a really good explanation of the project. Thanks for posting it!

  4. Really fascinating read. I look forward to more installments of this how-dunit. What is the timetable — how many feet per day do you estimate the TBM will carve? Is this how the new tunnel for Highway 1 below Pacifica has been dug? Ditto for the other major tunnels, Waldo, etc…?

  5. Cris Subrizi says:

    Thanks Robin, Glad you enjoyed it. Yes, Marc Brunel was quite an innovator. After escaping the turmoil of the French Revolution, he settled in New York (he was Chief Engineer for a time) where he proposed, and convinced leading merchants, to fund one of the first great canals of North America, connecting the Hudson River to Lake Champlain. Brunel left NY for England to pursue a commercial idea for the Royal Navy, which brought the shipworm into his life. While the worms are said to have sparked the idea, it was a project to cross the Neva river in Russia that firmed up his thinking about soft ground tunneling, which culminated in his 1818 patent. As a famous vulcan would say; he lived long and prospered.

  6. Robin Chiang says:

    Thank you, Cris, for this fun to read and fascinating glimpse into engineering the underworld. I enjoyed being brought back to the Industrial Revolution and look forward to learning more about tunneling methods. I hope to remember that Marc Isambard Brunel was an innovator in his own right (not to be confused with his famous engineer son). And Teredo Navilus recalls a Star Trek TV episode and makes me wonder if Gene Roddenberry shared an aha moment similar to Brunel.

Leave a Reply

Your email address will not be published. Required fields are marked *