SAN FRANCISCO BAY AREA: Transit/Transportation News

[Wizened Variations]

Originally Posted by mthd

the original tunnels are close to or at capacity depending on the headways, which depend on the train control system. there are plans to upgrade the control system to marginally increase capacity in the tube.

however, three of the four downtown stations - embarcadero, montgomery, and powell are capacity limited by escalator capacity and platform size. the medium-term plans for these stations include extra platforms on both sides of the trains, which sounds crazy but actually isn't that bad. however.... i'd rather they put the money towards a second line.

Thanks on the original tube update.

I think that both the extension of public platforms and the 2nd line needs to be built.

The longer I look at the picture the more I am struck by what a cool, intense subway system the combined old BART line and the new second line would make for SF particular if THE ALTERNATE CUTOFF IS ALSO BUILT. The local traffic synergies would be amazing. Off the top of my head, I could imagine 100,000 rides per day, or more, being added to BART just from that small area.

If the alternate cut off were also built, BART would have two of those traffic generating wonders called a Loop. The smaller of the two loops would lie in Oakland and would surround Lake Merritt. This loop would provide 5 stops all within short distance of one another. The synergies on this loop, while small might be surprising. The big loop would connect 13 stations- if sometime, a well designed Y track were placed at the 9th/Laney station a train could run in an endless loop which would provide world class connectivity.

This larger loop, if properly done, IMO, could add more than 250,000 additional riders to the system, I would guess.

[mthd]

Originally Posted by Wizened Variations

Thanks on the original tube update.

I think that both the extension of public platforms and the 2nd line needs to be built.

The longer I look at the picture the more I am struck by what a cool, intense subway system the combined old BART line and the new second line would make for SF particular if THE ALTERNATE CUTOFF IS ALSO BUILT. The local traffic synergies would be amazing. Off the top of my head, I could imagine 100,000 rides per day, or more, being added to BART just from that small area.

If the alternate cut off were also built, BART would have two of those traffic generating wonders called a Loop. The smaller of the two loops would lie in Oakland and would surround Lake Merritt. This loop would provide 5 stops all within short distance of one another. The synergies on this loop, while small might be surprising. The big loop would connect 13 stations- if sometime, a well designed Y track were placed at the 9th/Laney station a train could run in an endless loop which would provide world class connectivity.

This larger loop, if properly done, IMO, could add more than 250,000 additional riders to the system, I would guess.

here's what I would do before a new south-running line from downtown, if I had about 20 billion dollars to spare. construct the caltrain DTX extension with tracks for bart below the caltrain tracks. build a station at 2nd between brannan and bryant to serve ballpark/arena, Rincon hill, and for transfers to mission bay on the T-third. build a relatively deep station on second between mission and howard, with direct concourses to transbay and the montgomery bart station. then go up post or geary, making the turns under streets or low rise buildings. it would be AMAZING to loop south on van ness and to civic center station from geary/post and van ness, but from an engineering standpoint I don't see how it could be done. too many structures in the way.

[ardecila]

There's a lot of armchair speculation going on here, but I think it would be better from a regional standpoint to build a conventional rail tunnel than a BART tunnel.

Certain HSR trains could run through to Sacramento with an SF stop at 4th/King or even Transbay, and some Capital Corridor trains could be diverted to run down the Peninsula to SJ instead of through the East Bay.

There's also no reason why said tunnel couldn't have a spur down Geary using trains with a standard gauge. That service would still be part of the BART system with integrated ticketing.

[Wizened Variations]

Originally Posted by mthd

here's what I would do before a new south-running line from downtown, if I had about 20 billion dollars to spare. construct the caltrain DTX extension with tracks for bart below the caltrain tracks. build a station at 2nd between brannan and bryant to serve ballpark/arena, Rincon hill, and for transfers to mission bay on the T-third. build a relatively deep station on second between mission and howard, with direct concourses to transbay and the montgomery bart station. then go up post or geary, making the turns under streets or low rise buildings. it would be AMAZING to loop south on van ness and to civic center station from geary/post and van ness, but from an engineering standpoint I don't see how it could be done. too many structures in the way.

Nice illustration.

The best way to make a loop is to make it tight, and, buried.

This would involved a minimum radii circle- Ok, if you want a slight bigger subway U-turn put another station at the far end of the circle. This gives you both a loop and an additional stop, which might be a sales point.

This loop can be a bit bigger, but, my point is that it would have to be subway which would provide at least 1 additional stop in a an area for which the stop would be a good sell.

Loops are the best way to radially increase ridership. Well placed loops exist in the busiest five subways in NA (Montreal, Mexico City, NYC, Washington DC, and Toronto). Great subway systems like the Moscow, Tokyo, Seoul, Shanghai, Beijing, Paris, the Tube, etc., all use a mix of continuous loops, 1 stop transfer loops, and, 2 stop transfer loops.

Efficient way of turning around.

With a loop there is not need to change the engineer from A to B, rather the train continues to go forward. Small change? Nope- if the internal between trains is 2:00 the turn around time is exactly 2:00 (or 4:00 minutes .

Back up the loop with 2 X switches, and, during emergencies you have almost a km of track which theoretically could store 50+ cars on two tracks.

In your nice case, I could consider a subway loop with one station on it. Maybe 300 or 350 meters out.

Don't mean to get loopy.

EDIT: I couldn't resist the possibilities here. Assuming all is electrified (Come on Caltrain!) and you run 6 car trains caring 80 per car at 2:00 headways on two tracks, seems to me like you get 480 people per minute combining both tracks and having continuous run through. That just shy of 29,000 people per hour total for full utilization of both tracks.

Now, you have the first piece of that will result in higher rider usage.

Now the Bay Loop simply would be world class. Put continuous running trains in each direction and divert all other traffic to each end of the loop and you get, across the loop

60 trains per hour x 6 cars x 80(?) people car per 2 tracks, 2x for 4 tracks with a couple miles serving multiple stations. Ok, in each of the two tunnel sets- with two tracks running 2:00 per train per track
separated by 1000 feet-servicing 6 stations bordering each side of a 10 or 11 km^2 rectangle! Talking central Paris here! World class, my friends, world class!

EDIT 2+:

Oakland would get a great deal, too.

[CharlesCO]

This is armchair speculation, but I really would want to see how existing station upgrades might improve service before constructing a completely new tube. I don't remember the exact number of trains per hour, but the current Transbay Tube is quite surrprisingly below its maximum capacity mostly because it can't board and deboard passengers at Embarcadero and Montgomery fast enough on the other end. Upgrading those stations, the new fleet, and a new computer system could have a huge effect on reducing headways through the tube. It's too bad they're not looking at rebuilding Powell as well, though.

If they ever do build a second tube, I would prefer the Geary alignment as shown over the one seen in the Chronicle, which would mostly duplicate T-Third and Caltrain service in my opinion.

[Wizened Variations]

Originally Posted by CharlesCO

This is armchair speculation, but I really would want to see how existing station upgrades might improve service before constructing a completely new tube. I don't remember the exact number of trains per hour, but the current Transbay Tube is quite surrprisingly below its maximum capacity mostly because it can't board and deboard passengers at Embarcadero and Montgomery fast enough on the other end. Upgrading those stations, the new fleet, and a new computer system could have a huge effect on reducing headways through the tube. It's too bad they're not looking at rebuilding Powell as well, though.

If they ever do build a second tube, I would prefer the Geary alignment as shown over the one seen in the Chronicle, which would mostly duplicate T-Third and Caltrain service in my opinion.

IMO, a double tube loop would be the most effective, and, actually the only way to get the system to work at full potential.

Granted it would be extremely expensive, but, if done correctly- i.e., as an endless loop the entire system would benefit.

Every spoke of the system would be accessed via the loop with 1 transfer. Every spoke of the system would meet a line with very high frequencies. For example, a person transferring off of a spoke would have a 4 minute wait max at the loop station for trains in either direction.

In addition, trains particularly on the Oakland side can be dedicated to run from one end of a given spoke to the end of another, avoiding the tube, slow downs in SF, etc. On the SF side, Civic Station and Central Subway/Caltrain would become terminal stations for traffic from and to the south where linking traffic going in both directions would arrive and depart every 4 minutes.

Increase the length of platforms on the Bay Loop, run 8 car trains every 3 minutes (done in Tokyo, Moscow, Seoul, Shanghai, Beijing, etc.) and suddenly the figures get staggering.

8 cars x 150 (make more people stand) people per car x 20 trains per hour x2 directions equals 48,000 people per hour EVERYWHERE on the loop, and between any station set.

As the Loop covers 13 stations there will naturally be a lot of user traffic that does not need to use the tube(s). While feeder traffic has a lower frequency, each feeder rider would only have a 3 minute wait to travel on the loop to board a train going to 5 SF stations.

IMO, the short hops might even approach the tube ridership in SF. On the Oakland side, any rider on a feeder line would have a 3 minute wait at the platform to travel to 6 stations on the Oakland side.

Taking a huge amount of liberties, say that 30,000 people per hour use the loop and stay on their side of the Bay (mostly SF side including transfers from Caltrain .) Say 40,000 riders per hour head to SF via the two tubes and 20,000 riders per hour go to Oakland via the tubes. Consider 50,000 per hour transferring from the Loop to a BART feeder on both sides of the Bay

We are running about 140,000 people per hour for all portions of the loop. Consider 2 rush hours with a total of 4 hours under the "curve" and we get about 560,000 people per day during rush.

Add 25% for non-rush hours, and, we get 700,000 riders per day on the loop.

I have no doubt that synergies would kick in and the rest of the system could easily carry an addition 250,000 to 300,000 riders. Remember in both SF and Oakland transfers are quick from the feeder lines so the commuter's comfort zone would increase. Run trains on end to end runs on the Oakland side, for example, at 10 minute intervals (remember no feeder lines use the tubes) and the capacity of feeder lines doubles. Similar the feeder lines from the south could easily double their frequency to 10 minutes, rather than twenty. Do this effectively doubles system capacity outside the loop.

I see little reason why such a system could not carry 1,000,000 riders per day.

[mthd]

Originally Posted by Wizened Variations

...
8 cars x 150 (make more people stand) people per car x 20 trains per hour x2 directions equals 48,000 people per hour EVERYWHERE on the loop, and between any station set.

...

I see little reason why such a system could not carry 1,000,000 riders per day.

bart actually runs 10 car trains during peak periods, and assuming the system is integrated, all the platforms would need to be built to that standard. the crush capacity of one of those 10 car sets is over 2,000, and they actually run like that through the tube during rush hour, on headways of just a couple minutes.

the idea of a "bay loop" might be the true solution to san francisco's affordability crisis. consider berkeley, oakland, san leandro, etc the brooklyn, bronx, queens, etc of the bay area. with more robust connections between them, better last mile connections to bart san jose, and less reliance on bridges and freeways, more housing could be accomodated by workers going both directions. (west and south)...

[Rational Plan3]

The Chronicles route makes no sense.

A new line needs to take it's riders as close to where they want to go. So really that means a line that is close to the existing one, certainly in the downtown area. The route chosen should not take longer than the current one, otherwise you'll have a lot of angry riders.

Also a loop will reduce the number of trains that can come from the south.

[OhioGuy]

I thought this piece from The Transport Politic website from several years ago was a good read: Crossing the Bay Again -- But Not Necessarily With Bart

[ardecila]

Yep, that's basically what I was parroting. It was Yonah's idea, all credit to him - but I think he's correct.

[Wizened Variations]

Originally Posted by mthd

bart actually runs 10 car trains during peak periods, and assuming the system is integrated, all the platforms would need to be built to that standard. the crush capacity of one of those 10 car sets is over 2,000, and they actually run like that through the tube during rush hour, on headways of just a couple minutes.

the idea of a "bay loop" might be the true solution to san francisco's affordability crisis. consider berkeley, oakland, san leandro, etc the brooklyn, bronx, queens, etc of the bay area. with more robust connections between them, better last mile connections to bart san jose, and less reliance on bridges and freeways, more housing could be accomodated by workers going both directions. (west and south)...

Multiple loops feed on each other, particularly if each loop intersects the other at more than 1 station. Two or four would be better.

There are caveats: for example a loop with a low number of spokes that takes more than a certain time to get around fails to produce ridership synergies. Where loops shine, IMO, is when stops are less than 5 minutes apart and all or most stops intersect spokes. *

A big loop might work differently, and, be designed more as a primary line where riders board and exit from the loop as part of a long commute. The South Bay loop has the advantage of connecting 3 or 4 major population centers each of which provides commuters to other stops on the loop.

In my opinion, particularly with your pointing out 2000 per train capacities, that if two loops could share both Bay tunnels, or that one tunnel could have both loops. In this case, say 6 minute headways on loop one on tunnel 1 and 3 minute headways on tunnel 2 sharing both loops (loop 2 has 6 minute headways also).

Here are links to successful systems with loops that are highlighted. Granted these are huge systems but the principal works the same. These maps are idealized for visualization not geographic accuracy.

London
http://speedymole.com/Tubes/London/l...-tube-map.html

Moscow
http://speedymole.com/Tubes/Moscow/m...map-metro.html

Seoul
http://wallpoppy.blogspot.com/2011/04/colourblind.html





*The Bay makes this different so a loop across the Bay should only affect the time required to cross the Bay, not the other stops.

[electricron]

Originally Posted by Wizened Variations

Multiple loops feed on each other, particularly if each loop intersects the other at more than 1 station. Two or four would be better.

Here are links to successful systems with loops that are highlighted. Granted these are huge systems but the principal works the same. These maps are idealized for visualization not geographic accuracy.

London
http://speedymole.com/Tubes/London/l...-tube-map.html

Moscow
http://speedymole.com/Tubes/Moscow/m...map-metro.html

Seoul
http://wallpoppy.blogspot.com/2011/04/colourblind.html

London, Moscow, and Seoul have rivers bisecting them less than a mile wide. The Bay Area has San Francisco Bay bisecting it, which can be 4 (Bay Bridge) to 7 (San Mateo Bridge) miles wide and 33 miles long (Bay Bridge to Milpitas). Putting those distances into prospective, 33 miles in the London area stretches from M25 west of the city to M25 east of the city as a bird flies along the Thames. The English Channel between Dover and Calais is just 25 miles wide. Yes, just the south Bay is longer than the crossing.
The Bay stretches further north, and just by itself (uninhabitable swamp and bay) is larger in area than central London.
I don't think the Bay Area is geographically the same as the three cities you mentioned, and I highly doubt circular transit patterns will work. The Bay Area is set up for linear transit lines, not just by the Bay coastline but by the surrounding hills/mountains as well.
Ignore the landscape and distances involved at your peril. I don't think you would advocate building a circular tunnel under the channel, so why are you suggesting building a circular transit lines around the south Bay?

[Wizened Variations]

Originally Posted by electricron

London, Moscow, and Seoul have rivers bisecting them less than a mile wide. The Bay Area has San Francisco Bay bisecting it, which can be 4 (Bay Bridge) to 7 (San Mateo Bridge) miles wide and 33 miles long (Bay Bridge to Milpitas). Putting those distances into prospective, 33 miles in the London area stretches from M25 west of the city to M25 east of the city as a bird flies along the Thames. The English Channel between Dover and Calais is just 25 miles wide. Yes, just the south Bay is longer than the crossing.
The Bay stretches further north, and just by itself (uninhabitable swamp and bay) is larger in area than central London.
I don't think the Bay Area is geographically the same as the three cities you mentioned, and I highly doubt circular transit patterns will work. The Bay Area is set up for linear transit lines, not just by the Bay coastline but by the surrounding hills/mountains as well.
Ignore the landscape and distances involved at your peril. I don't think you would advocate building a circular tunnel under the channel, so why are you suggesting building a circular transit lines around the south Bay?

The issue is frequency of travel and switching from feeder loops.

Both should occur, but the most important is the Bay Loop.

The time taken to travel under the Bay is rather short- from the West Oakland Station to the Embarcadero Station takes 7 minutes.

http://www.bart.gov/schedules/byline...ate=02/18/2014

Each tunnel should be close together.

The model I am using was posted by Itsmotorsport, 021614, #1017 on this blog.

What the Bay Loop would do:

A) Directly connect 13 stations.
B) Directly connect (with the critical Alernate Cut-Offz) 5 feeder lines if one includes Caltrain.
C) Provide a very well served 6 station core (6 stations) directly in SF, from which rider synergies would radically increase. These six stations would be no more than 1 km or so from the one of the tube lines.
D) Within 1 transfer (a quick Alternate Cut-off ride) an additional 4 or more stations- say 10 minutes out from either Civic Center or Central/Subway forms a dense station net.
E) The same situation would exist on the East Side, with a tight cluster of stations (proposed Alameda, Lakeshore/Grand, and existing McArthur (need to complete the loop her) 19th,12, and West Oakland Stations.
F) Eliminate delays caused by feed lines.
G) Provide backup for if one Tube needs to be repaired.
H) Provide very short headways for trains.

You did make me think, though, at a far cheaper solution

I would interested if something like this just used the existing tunnel. The same tight loop could be made by collapsing the loop, something like a figure "8" with a line segment expanding the center cross over. This certainly would be straight forward on the Bay side, but on the Bay side a circular connector would have to made further south, built in less developed land. Ball Park Station would then lie on a feed line, i.e., would then be 1 transfer away from Oakland.

The geometry still works and with 2000 person capacity trains at 3 minute headways we are still taking about 40,000 rider per hour capacity in each direction.

Again the key to the Oakland/SF end of the Bay is a very frequency fast loop with easy, quick transfers. And, this does not have to be done with two tunnels (although 160,000 riders per hour ranks with the worlds best.)

Thanks for getting me on the 'right track.'

*************************

The south loop is rather nebulous, because, as you point out, much of the Bay patterns are linear. In addition, we have variables like integrating the Caltrain train which has standard track gauge (BART does not) and a loop of additional track layer to go around the south end of the Bay. Based on your input, I think any second loop would have to involve two transfers in the loop (at each end of a segment on the Caltrain line) and would not provide a good rider increase per cost.

Incidentally, as has been discussed by OhioGuy, this blog #1029, others have suggested extending Caltrain and a new (standard gauged) tunnel under the Bay and forming a 2nd loop formed by extending Caltrain through Jack Lon and Hayward and possibly completing the loop either across the Bay or just to to the south (the cross bay would be new) and the most of the rest mixed use).

IMO there is little reason to do this short term. I think that the key to this possible two transfer loop would be the speed of BART from Hayward around the south side of the Bay. A possible improvement on the BART portion would be 4 track passing stations from Fremont south through San Jose to Santa Clara.

The keys to the South Bay transit, IMO, boil down to the speed and frequency of the Caltrain, and, the speed on BART south of Fremont.

[electricron]

Originally Posted by Wizened Variations

The keys to the South Bay transit, IMO, boil down to the speed and frequency of the Caltrain, and, the speed on BART south of Fremont.

The keys to South Bay transit isn't more regional trains and/or lines but more local trains, i.e. more light rail lines like in Santa Clara and San Francisco. There's no reason why each large city around the South Bay shouldn't have its own light rail or streetcar lines. There would be far less congestion on the regional trains if there were more local trains.

[Wizened Variations]

Originally Posted by electricron

The keys to South Bay transit isn't more regional trains and/or lines but more local trains, i.e. more light rail lines like in Santa Clara and San Francisco. There's no reason why each large city around the South Bay shouldn't have its own light rail or streetcar lines. There would be far less congestion on the regional trains if there were more local trains.

IMO. the quickest way to move people is via feeder lines and loops and feeders. In addition to the simple mathematics, such route arrangements enable chaos caused irregularity scheduled changes in arrival times to be buffered by the constant 3 minute headway (5 max) of the loop, i.e., the distortion is confined to the feeder line.

While irregularities in a loop (and G* bless the redundancy of the double or 4 track loop) does affect the feeders to the loop, the total distortion of the system cancels out due to all spokes sharing the same deviation from scheduling (talking actually about 2 track loops) simultaneously (the same thing said two different ways, )

(Logical inverses)

For this reason, frequency density can be maximized.

You triggered by my thinking here, thanks.

[Wizened Variations]

Originally Posted by electricron

The keys to South Bay transit isn't more regional trains and/or lines but more local trains, i.e. more light rail lines like in Santa Clara and San Francisco. There's no reason why each large city around the South Bay shouldn't have its own light rail or streetcar lines. There would be far less congestion on the regional trains if there were more local trains.

2nd Reply

Loops thrive on high volume. Caltrain, and, regional trains should be feeders. Likewise, local jurisdictions need light rail, street car, and, BRT 2nd order feeder lines to feed the 1st order feeder lines running along each side of the Bay.

The key, IMO, is that each local municipality need to have their lines intersect more than one 1st order feeder line station. Perhaps a zig-zag pattern crossing the 1st order feeder multiple times. Possibly, a light rail with a loop intersecting the 1st order feed line, as well as spoke lines.

1st Order feeder lines need express and local service along their routes, something we have not built into any of our passenger rail systems since WWII. This means 4 track stations for the same direction passing of trains, and, for the few 2 track stations, platform fences and gates to permit high speed travel through stations.

[CharlesCO]

This discussion is getting to be too theoretical. Loops may work in theory, and they do work for some of the largest transit systems, but I don't really see how this discussion is at all relevant to Bay Area transit. We need to first identify specifically which problems we're trying to fix. The way I see it:

1) We need to fix capacity constraints of the Transbay Tube. This can be fixed relatively easily in the next decade with a more efficient Embarcadero, a new fleet, and a new computer system, and in the long term, another tube, ideally serving SoMa and the Geary corridor, may be a good solution, no matter whether it may be traditional BART technology or a shared tunnel that includes a new standard gauge BART.

2) As I just mentioned, the Richmond District is in dire need of a better transit solution. A below grade heavy rail solution (traditional BART or a standard gauge BART) is tempting, but I could also see how this might be really difficult to fund. Maybe a new Muni Metro line (either at grade or below grade or a mixture of both) may be a nice complement to the rest of the system.

3) An AC Transit streetcar/LRT through Oakland/Berkeley would be really cool.

4) County Connection sucks. At least I now live within walking distance of BART.

[Wizened Variations]

Originally Posted by CharlesCO

This discussion is getting to be too theoretical. Loops may work in theory, and they do work for some of the largest transit systems, but I don't really see how this discussion is at all relevant to Bay Area transit. We need to first identify specifically which problems we're trying to fix. The way I see it:

1) We need to fix capacity constraints of the Transbay Tube. This can be fixed relatively easily in the next decade with a more efficient Embarcadero, a new fleet, and a new computer system, and in the long term, another tube, ideally serving SoMa and the Geary corridor, may be a good solution, no matter whether it may be traditional BART technology or a shared tunnel that includes a new standard gauge BART.

2) As I just mentioned, the Richmond District is in dire need of a better transit solution. A below grade heavy rail solution (traditional BART or a standard gauge BART) is tempting, but I could also see how this might be really difficult to fund. Maybe a new Muni Metro line (either at grade or below grade or a mixture of both) may be a nice complement to the rest of the system.

3) An AC Transit streetcar/LRT through Oakland/Berkeley would be really cool.

4) County Connection sucks. At least I now live within walking distance of BART.

Just get the one tube we have to be a single loop. Speed frequency up to a 2000 rider train running every six minutes, then feed all the traffic into it you can.

People in the Bay Area still do not know what great subways can do.

One has to start somewhere, and, the key to the entire system is the Oakland SF connection as the original designers knew. The problem now is simply that there are so many feeder lines running into the tunnel that either the entire system is centrally controlled, or the tunnel is separated from the rest of the system.

The scheduling problems in large part- contrary IMO to many that state the problem is platform size and computer controlling- is due to having to cushion for the irregularities of all the lines that feed into it.

For example, how would the tunnel handle 2x or 4x the train frequency the East side lines? What infrastructural improvements would need to be made on the SF side to handle the increased frequency (assuming the tunnel was handling the increased load)?

Again the solution, IMO, must involve decoupling BART into 3 separate parts that each can accept schedule variations. This would be cheapest way to radically improve system capacity and reliability.

[CharlesCO]

Have you ever been on the Embarcadero or Montgomery platforms anytime between 5-7 pm? There are people everywhere, and while BART trains are the longest in North America, they still can't get people on and off trains fast enough— even when trains are running almost every 90 seconds. Giving it the Spanish Solution (like the DIA people mover) where exiting passengers alight onto another platform is expensive (nearly a $.5B project) but will have a somewhat large impact on headways. And building spurs to the line is not as easy as it is in Denver— not only is it below grade, but it is a prohibitively expensive technology to build, especially today. I also think that more irregular peak hour trains that only run part of the way (e.g. Concord-Montgomery/Daly City in the morning, Concord-24th Street in the afternoon) also are a good way to go once they are able to run more trains through the tunnel.

[Wizened Variations]

Originally Posted by CharlesCO

Have you ever been on the Embarcadero or Montgomery platforms anytime between 5-7 pm? There are people everywhere, and while BART trains are the longest in North America, they still can't get people on and off trains fast enough— even when trains are running almost every 90 seconds. Giving it the Spanish Solution (like the DIA people mover) where exiting passengers alight onto another platform is expensive (nearly a $.5B project) but will have a somewhat large impact on headways. And building spurs to the line is not as easy as it is in Denver— not only is it below grade, but it is a prohibitively expensive technology to build, especially today. I also think that more irregular peak hour trains that only run part of the way (e.g. Concord-Montgomery/Daly City in the morning, Concord-24th Street in the afternoon) also are a good way to go once they are able to run more trains through the tunnel.

Denver's light rail and commuter system is being built in recently was, empty land. The system is largely a real estate development scheme, IMO, where too much money has been side barred for infrastructure improvement- sewer, water, electricity, sidewalks, paved roads, etc., at the expense of capacity, speed, and, efficiency. The downtown node has had to fit around very recently drawn property lines in a very small space, and have it's north and south rail lines stub lines separated by 300 meters, etc.

SF's original BART layout was far superior even with municipalities pulling out of the "Grand Vision" than Denver's system will be for the foreseeable future.

*********************

What about putting one of those Japanese style platform edges on downtown stations and using run through trains? Express A might stop at 1, 4, 7, B at 3, and 6 and train C at 2, and 5? Rather than build a huge facility downtown how about building 3 track stations (center track is express and is only used during rush hour in the busiest direction). This might help short term.

****************

SF's problems are needing to be remedied in a highly built up area where traffic loads are high.

I suppose that something has to be done with the tunnel/subway interface, but, the question in my mind would be how far south must this interface be and what can be done to increase tunnel capacity.

The tunnel has to accept all BART trains or be separated from all BART trains on feeder lines. There is no other alternative besides splitting the system in some fashion.

I think what you say about building an interface at the heart of SF is spot on, so unless SF (and the Country) gets a flood of money a return loop can never be built there.

The question is where can this be done? Where can a continuous return line be built for the tunnel on the Peninsula and also be able to act as a stub line for the south portion of the BART line? How might all of this interface with Caltrains, and, HSR when and if it is ever built?

The issue is not design but the cost of the footprint needed to handle perhaps 12 or 16 tracks (including HSR and Caltrain) on multiple levels with several hundred meters of switching track on either side.

Regardless, if something like this was done on both sides of the Bay, the issue would no longer be adjusting schedules to fit precisely into time slots for all BART trains to go under the Bay.

As one might say: "Sorry, folks, no same seat boarding if you want to cross the Bay on BART.