Wednesday, December 10, 2008

HSR across the Bay Bridge?

NOTE: We've moved! Visit us at the California High Speed Rail Blog.

In my recent post introducing rapid rail as a concept, I included a map showing the potential of running HSR trains across the Bay Bridge. This was very tentative, mostly because I did not know if the bridge could support the weight of modern, lightweight UIC-compliant passenger trains.

Early on in its service life, the Bay Bridge used to carry electric trains, e.g. this trolley car from 1938:

In 1959, the rail tracks on the lower deck were converted to road lanes. In 1998, at the urging of their mayors, voters in San Francisco, Oakland, Berkeley and Emeryville passed a non-binding declaration of policy that rail service should be restored. Quentin Kopp, then in charge of the project to build a new east span, rejected the idea out of hand. Bitterly ironic, isn't it? See also the response by Emeryville's then-mayor Bukowski.

Nevertheless, MTC did commission a rail feasibility study that was published two years later. It concluded that sacrificing road lanes would not be acceptable due to an expected 22% increase in persons crossing the bridge by 2020. This was a complete non-sequitur, since a rail track can support a lot more person crossings than a road lane. Rather, the statement should be seen as evidence of the clout of the road lobby.

The study did conclude that rail tracks could be added to the west span, either slung underneath (subject to Coast Guard approval) or, on ugly cantilevers on either side. However, the cost for getting the western span "rail ready" in this way was estimated at almost $1.33-1.45 billion, excl. the actual tracks and western approach from the new TTC. With an underground train box there, only the (cheaper) alternative of tracks slung underneath the bridge was considered possible.

The new east span was designed to support five lanes of road traffic in each direction, plus emergency and wide bike lanes. Back in 2000, the design consisted only of the two skyways. The feasibility study concluded that the design of those structures would could support light rail, but only if the inside lanes were sacrificed. The outside lanes would have been preferable in terms of tunnel construction at Yerba Buena. Modifications to add light rail tracks were estimated at $546 million, provided the design changes were made before breaking ground.

All told, MTC study estimated a cost north of $3 billion just to get entire bridge and its approaches ready for two tracks of light rail plus five active road lanes in each direction. This was considered far too expensive at the time, so all rail services were dropped from consideration. The self-supporting suspension span that was added to the new east span at a later date for aesthetic reasons was probably not designed to support light rail.

For reference, the axle load of the latest generation of European bullet trains is below 17 metric tons. That's not much higher than the maximum permitted load on the rear axles of an eighteen-wheel heavy duty truck, 34,000 lbs (15.4 metric tons). In terms of weight per linear foot of track, a measure relevant for bridge loads, the FRA-compliant Acela Express comes in at a hefty 1800 lbs/ft. However, a fully decked out Alstom AGV comes in at 1370 lbs/ft, a shade below light rail at 1390 lbs/ft!

Competitive HSR products are at similar or slightly higher values, BART at lower ones. That suggests running either BART or HSR trains across the Bay Bridge just might well have been possible at reasonable cost, provided that road lanes had been sacrificed for the purpose. Otherwise, it would make more sense to build a second transbay tube one day, e.g. under Oakland harbor via Point Alameda.

Unfortunately, with the construction of the east span and the seismic retrofit of the western approach already underway and, the Transbay Terminal project set to break ground, it seems highly unlikely that bullet trains will ever cross the Bay Bridge.


Pantograph Trolleypole said...

I've always wondered not about the bridge, but rather the approach. How would HSR get under the transbay terminal so fast?

James said...

As for giving up a bridge lane for rail, we have a data point for comparison. After the Loma Prieta quake we had to do without the entire bridge. So giving a lane to rails may just work, especially if the trains were frequent.

mrampton said...

BART has stated their plans for the next 50 years include a new transbay tube linking up with the new transbay terminal. Supposedly it would be wide enough for 4 tracks, two of which would go to bart, and the other 2 for HSR/Electrified rail). This would allow them to run another line through oakland, connect to SF, and then from there extend along SOMA, to Van Ness, and finally out along Geary...

Tom said...

Yeah, I think any rail (HSR or otherwise) across the Bay Bridge will likely never happen. It would be too expensive (and ugly frankly) to *add* rail capacity to the bridge (keeping the 5 auto lanes in each direction), and not politically feasible to remove auto lanes from the bridge in exchange for rail.

A couple other problems of converting car lanes to rail on the existing western/suspension span:

1) The Key rail line that ran across the bridge ran at only 35 MPH (compared to BART's max 80 MPH in the transbay tube). Would it be possible to run trains faster than 35 MPH (or even faster than 80 MPH) on the Bay Bridge? It would be hard to convince people to give up auto lanes if the trains don't even match the speed of auto traffic (non-commute times).

When the Key System ran on the lower deck of the Bay Bridge, there were 2 "lanes" of rail (one in each direction), and 3 lanes for trucks): One dedicated truck lane for each direction and the middle truck lane was a "suicide" passing lane. There were no barriers between truck lanes whatsoever. Obviously in today's more safety conscious society, that would not be allowed. So it would likely just be 1 truck lane in each direction, but that would likely slow down truck traffic.

On the upper deck, there were 3 lanes of auto traffic in each direction. Given how narrow the 5-lane configuration is today, I think 6 lanes is too narrow. Also, there was no center barrier whatsoever. With all the head-on collision concerns of the Golden Gate Bridge and the concern about how to fit a moveable barrier on the GG Bridge, 3 lanes in each direction on upper deck of the Bay Bridge would likely never work.

Original Bay Bridge lane configuration with rail (upper drawing)

So the best option would be to have 3-4 lanes of mixed auto and truck traffic plus one lane of rail on each deck. But then other issues arrise: 1) How to deal with the lack of clearance for the rail vehicle (and the catenaries) on the upper deck of the YBI tunnel? 2) Can rail run on the same side of the bridge (presumedly on the southern lane of the bridge on each deck) and have the load of the bridge remain balanced? 3) How to deal with connecting underground to the Transbay Terminal? 4) What about the westbound auto offramp from the bridge to YBI?

Anyway, I think the best bet for a transbay HSR crossing would be a tube parallel to BART's, a new higher capacity BART tube as mrampton mentioned, or even another bridge (rail and auto or rail-only) somewhere south of the Bay Bridge (though a tube would be best). Regardless, this wouldn't happen any time soon.

Anonymous said...

I did a rough sketch project for a client where we looked at a single rail across the suspension span and a new dual track bridge (and tunnel through Yerba Buena) on the east. The single track would be on the lower deck, with a reversible HOV lane on the upper deck, between the Fifth Street exit in SF and Yerba Buena.

That preserves peak direction capacity for autos, and with good scheduling can still allow 5-minute headways for rail.

Of course, the connection to TTT is a problem.

I don't think rail on the Bay Bridge will happen until auto use starts to drop.

Anonymous said...

I find the 'engineering fantasies' explored here fascinating. But help me - why would HSR run over the Bay Bridge (or even through an eventual tube)? What destinations would be served through such access to the City? Paris has four or five TGV stations - as well as bypass routes for those who want to traverse from the North to South of France. I don't see why there couldn't be eventual routes that would terminate in Oakland (from SEA/Portland? Reno?). If there are to be restored tracks over the Bay Bridge, reborn AC Transit (ex Kay) trolleys is likely to be more useful.

Anonymous said...


Agree on Bay Bridge issue. A great BART-Capitol/East Bay HSR interface in Oakland is much more helpful and needed.

Rafael said...

@ pantograph trolleypole -

the original TT had the train station on the 2nd floor and both train tracks ran on the lower deck of the bridge.

I'm not sure how they intended to go from the basement of the new TTC to tracks slung underneath the bridge, there's a massive bridge support column at I-80/Beale.

@ James -

yes, that was my reason for exploring this possibility. Driving has to become enough of a pain that people will prefer to hop on a train. It runs counter to a deeply entrenched car culture, though, especially at Caltrans.

@ mrampton -

BART would like a lot things from Santa. I really don't see any reason for a second set of broad gauge tubes. BART could run a spur up Geary without one.

On the other hand, a set of standard gauge tubes via Point Alameda would be useful for express trains supplementing BART service in the East Bay as well as connections to Sacramento, Concord NWC and perhaps, Altamont Pass (only makes sense if express trains reach SJ Diridon from the east for rail-around-the-bay).

The caveat is that only some UIC-compliant equipment would be light enough.

@ Tom -

the MTC rail feasibility study suggest 50mph would be possible on the west span, probably more on the new east span.

The rail tracks would each get a new tunnel through Yerba Buena Island, since messing with the one for road vehicles would be too disruptive. The catenaries are less of an issue, you can mount a rigid conductor rail against a low-slung ceiling.

If rail across the bridge is ever contemplated again because gasoline had become too expensive or traffic across a permanent nightmare, it's likely that both rail tracks would have to be on the lower deck. That would mean reversing direction on one lane up top and fixing the approaches and lane divider accordingly.

Frankly, I suspect they'll not do anything until and unless an earthquake damages the western span beyond repair. Given that it's just been retrofitted, that would have to be a truly massive tremblor. I hope we never see one that bad...

Rafael said...

@ anon -

BART and light rail options were considered in that year 2000 feasibility study but simply getting the bridge ready for any type of rail service without sacrificing any road lanes was considered too expensive.

After the massive cost overruns on the new east span, SF mayor Newsom was quite adamant that he would oppose any new bay crossing of any type for the foreseeable future. That means BART will have to squeeze what capacity it can out of the existing transbay tube, principally by expanding pedestrian flow capacity away from the platforms in the downtown SF stations.

Buses will serve the TTC as well, of course, perhaps even advanced articulated types like the Huebner/Fraunhofer Autotram. New guidance electronics allow a research version to navigate unmanned, road trains of up to three articulated units are also technically possible.

I just don't see even fancy buses reducing congestion on BART. Many commuters into SF come from further afield than Oakland, i.e. Richmond/El Cerrito, Walnut Creek/Concord, Hayward and Dublin/Plasanton.

Tom said...

@ Rafael - Yes, I suppose separate tunnels could be built through YBI with a rigid conductor rail against a low-slung ceiling, and ramps could be used to so trains can go eastbound from the existing suspension span and "exit" to the new YBI rail tunnels and vice versa. But I'd imagine that this would be easier to do if one set of tracks were placed on each deck of the bridge as I suggested, instead of both on the lower deck.

Also, if both directions of trains are run on the bottom of the bridge (as was done with the Key System before 1959), you want to reverse the direction of just one lane on the upper deck? What a nightmare it would be to use that lane, especially if a car breaks down or there's an accident in that lane. And where to put a barrier? The lanes are rather narrow as it is with 5 lanes in each direction. Remember, trucks need to use the lanes too.

Tom said...

@ Rafael - Yeah, BART needs to run longer trains, which means expanding the platforms systemwide. And as expensive as that would be, I bet it would still be cheaper than expanding the bridge's capacity by adding BART to it.

Maybe BART can run the trains closer together through the tube? Though I understand that the tube's capacity is probably maxed out because the tube is a bottleneck (4 East Bay lines all using one track in the tube in a given direction).

Even if buses could help alleviate some of the congestion on BART, they'd still have to deal with a different congestion: traffic on the bridge (despite bus/HOV lanes at the approaches to the bridge).

I suppose the simplest solution would be to expand ferry service between the East Bay and SF. And maybe to get people from the far reaches of the East Bay to the ferry via bus?

Anonymous said...

BART is addressing capacity through the Tube by getting their next generation of trains with three pairs of doors per side and fewer seats. The capacity issue is dwell time at Embarcadero and Montgomery. They are considering platform doors at Embarcadero, due to the narrowness of the platform. They are also looking for ways to add more vertical circulation to the stations.

Transbay buses compliment BART capacity by providing transit for people in the inner East Bay in many of the neighborhoods that aren't served well by BART. Especially on the Concord line, trains are SRO by the time they get to Rockridge. At the further reaches of the Richmond-Fremont line, Transbay buses beat train travel times.

Rafael said...

@ Michael Kiesling -

BART is now implementing a crossover section in CC county to improve things.

In the long run, the best way to address the bottleneck would be to give SF financial district businesses incentives to set up back offices in CC county, Hayward, Berkeley etc. That would reduce congestion and possibly even use some rush hour capacity in the opposite direction.

James said...

How about special BART cars with three sets of doors and no fixed seats. Only fold-down seats and provision for wheelchairs. Would have to have poles and hand-holds. Without seats the special cars would have extra room for standing and there would be more room to maneuver in and out of the doors past the other people. The fold-down seats would be prioity for elderly and disabled. The rest of the car can be packed by able-bodied and willing commuters. One special per consist and always in the same order of the consist. Would also make room for bikes? The standard external stripe would be of a noticable style to let people know which car. Oh and make the windows higher so the people standing have a better view. Maybe some overhead windows since so many of the people are up there anyway.

James said...

BART special alternate configurations, (all include the wheelchair provision, just variations on how much open floorspace is available):

- Fixed seats at each end and no fixed seats between the doors. Still may have to inclue some fold down seats.

- Row of fixed seats on each side facing the center.

Mid bulkheads may be required if they are part of some mechanical system or part of the shell structure. Even so without the fixed seat there would be more standing room. Bulkheads may also help with crash safety to give people sometihng to push on. Bulkheads could have a shape thay people could lean on to take a rest from standing. This is mainly a commuter car. Would need ventilation arranged to account for densly packed, standing passengers. Maybe a recessed hooks that would not be a safety hazard but could hold a backpack or shopping bag.

Anonymous said...

@Rafael --

My response is why???

I would much rather see a connection to Marin. This would enable the north bay to continue the SMART train direct to SF.

I don't think such a connection is easily possible but I see no advantage to a Bay Bridge connection. A Dumbarton Bridge connection is both easier, cheaper, and offers the equivalent benefits of connection to Sacramento and the East Bay.

Anonymous said...

Worried about capital costs?

More ferries.

Rafael said...

@ Pat Moore -

running lightweight standard gauge trains across the Bay Bridge make Oakland, Oakland airport and the entire East Bay down to SJ available for express service alongside BART's inherently local trains. It would also facilitate connections to Sacramento, Concord NWC (if rail infrastructure right next to BART North Concord is already re-used), the Delta counties as well as Napa, Vallejo and before long Sonoma, Novato and Santa Rosa (NWP already owns the ROW between Novato and American Canyon and, there are still tracks and/or ROW up to Sonoma town).

Basically, getting standard gauge trains running between SF and Oakland as directly as possible would be the cornerstone of any effort to develop a decent passenger rail network throughout northern California. So that's why.


I have also looked into a connection to Marin. Between the strong tidal currents near the Golden Gate, the hilly terrain and the loss of the old ROW through Mill Valley to cyclists and encroaching homes, it's basically impossible to get a direct train connection of any kind.

At best, it might be possible to extend SMART south to Tiburon.

A new rail bridge between Richmond and San Rafael would be technically feasible but expensive. However, that only gets you to the East Bay, not to SF. Without any tracks on the Bay Bridge, you still have to change trains to sloooow BART.

Franco Marciano said...

If BART is to run longer trains, especially trips from the East Bay, the Livermore extension is a vital part to that becoming possible. After attending a BART environmental scoping meeting earlier this year, it was clear the extension would provide enough storage facilities to handle the extra cars.