Central Subway: An Alternative Proposal

The Central Subway project (proposed from SOMA to Chinatown) has run into various delays and cost overruns but is still very much in San Francisco’s plans. Rescue Muni board member Howard Strassner has an alternative that would be much less expensive – rider comments please! Note that Rescue Muni’s policy now is to recommend extension of the central subway to North Beach, rather than stopping it at Clay Street as currently planned.

Update: The Examiner discussed this proposal this week.

(Details after the jump.)

Study of the 30 Stockton as BRT instead of Central Subway – 7/2007 – by Howard Strassner

San Francisco Metropolitan Transportation Agency compared the proposed Central Subway, which will run from the Caltrain Station to Clay and Stockton Street, with the existing bus service; but never compared the Subway with Bus Rapid Transit, BRT. Howard Strassner, an experienced transit advocate, produced this study to make the comparison. The study analyzes the comparison in two phases as local public approval develops.

Phase 1, partial BRT, treatment of the 30 Stockton, using methods in use in San Francisco or proposed for BRT studies, can reduce transit running time between the Caltrain Station and Clay and Stockton from the current, 21 minutes, to 15 minutes. The MTA shows that a $1.4 Billion surface and subway extension of the Third Street Light Rail, LRV, will reduce the running time to nine minutes (including an allowance for extra walking and waiting time) for the same trip. BRT will also reduce travel times for riders on the 9 and 45 Lines for a total of more than twice the number of subway riders. BRT can be provided much faster than a subway at less than 10% of cost of a subway. The question is shall we spend $1.4 Billion to save some 30 Stockton riders 12 minutes or should we spend less than $200 million to save most 9, 30 and 45 riders 6 minutes.
Phase 2, unconventional auto traffic changes can further reduce BRT running times to twelve minutes at a total of less than a fifth of the cost of a subway. Extending BRT on Stockton Street, north of Clay Street, will save riders another three minutes.

The Way it is Now:
North bound: The diamond lines help a little but there are transit delays due to: right turning cars in the diamond lane on Third Street, traffic as the Stockton bus crosses each of the major freeway like blocks south of Market; slow boarding passengers and often an extra wait for a signal to cross Market to Kearny. Transit moves well on Kearny and the bus uses its mass to facilitate the turn to Sutter and then moves quickly through the tunnel until it gets to Chinatown. Here there is only one moving lane north bound and large numbers of on and off passengers at most stops. The bus also looses time to right and left turning cars and massive amounts of double parking. Muni has provided center street bus bulbs which help a little. Department of Parking and Traffic, DPT, traffic counts indicate that there are from 121 to 196 vehicles per hour during the day at Pacific and Stockton.
South bound: Even though there is an extra lane, the Stockton bus has all of the same problems as above, north of the tunnel. South of the tunnel buses have a diamond lane to Market, but this helps little and the Muni Time Table shows that that it takes the south bound bus the same five minutes as the north bound bus to cover a much shorter distance to Market Street. South of Market the bus has the same problems as the northbound bus. The approximate run time from the Caltrain station to Clay and Stockton, where the proposed subway and light rail line would end is about 21 minutes with ten stops, in each direction. A fast walker could cover this 1.7 miles in 25 minutes with some luck with signal lights. The sidewalks, in Chinatown north and south bound, work as extensions of the shops and often goods are sold from parked trucks at the curb. This adds congestion to the already narrow sidewalks. Recently “Barnes Dance” crossings have been provided at some intersections and in time this will improve pedestrian safety. DPT traffic counts indicate from 361 to 506 vehicles per hour during the day at Pacific and Stockton.

Comparison of BRT to Subways:
General: Most of the transit speed advantages that are associated with subways can be provided by well designed surface transit whether BRT or Light Rail Vehicles, LRV.
Fast Loading: Breda cars, the San Francisco LRV, have four wide doors per side and in subways Bredas are quickly boarded from a platform at car level. Throughout the world this is achieved with low floor articulated buses with additional wider doors.
Prepayment: Subways have prepayment of fares. The LRV approaching Caltrain has fare machines for pre-payment. Many riders in SF have fast passes. Our LRVs are officially proof of payment, POP, ie riders with proof of payment can board at any door. The 30 Stockton and many busy bus lines are de facto POP as some riders board quickly and wave a transfer or fast pass at the driver, from the rear door. Stuffing a crinkled dollar bill into the fare box or picking up loose change slows the bus. BRT includes POP and ticket machines on busy loading platforms to reduce boarding time.
Right-of Way: Subways of course have no conflicts with surface traffic. BRT can have a nearly dedicated right of way with a curb separating transit from traffic. Muni, following the guidance of Prop ‘E’ is moving toward signal preemption. Studies are showing that this can reduce running time by over ten percent.
Subway Running Time: Transit in subways can run at higher speeds than on the surface but only when the distance between stations is great enough to allow time for acceleration. BART, the best that can be hoped for, shows 3 minutes from Civic Center to Montgomery which is 2 stops and 0.9 miles. The Central Subway is about 1.7 miles long, and from Caltrain to Clay there will be four stops at BART type spacing. So the fastest possible ride from Caltrain to Clay will be six minutes. However the average rider will have to walk further between stops and take time to get into and out of the subway. In addition the headway between trains, because of their larger passenger capacity, will be greater than the headway between buses. All of this increases the real average subway ride time by at least three minutes to nine minutes total.
Before we spend a Billion and a half dollars subway to nowhere and then more Billions to extend it to somewhere we should first completely consider the easy improvements and then the more difficult improvements for a BRT system.

Phase 1, Low Cost Stockton BRT Improvements:
1) Even though studies show that signal preempts can save 10% of running time, this bus route crosses many other bus lines and priority for one line will slow another line. Therefore preempts may only save one minute.
2) Quicker boarding time saves 20 seconds per stop or three minutes with 10 stops. This includes ticket machines at stops to reduce standing time.. This full improvement will take new low floor cars or special platforms which will take time and/or major funding when buses must be replaced, so it will not be available for some time. However, broader use of POP and using all doors can immediately save half of this time at nearly no cost.
3) Phase 1, requires no traffic changes in the tunnel because the BRT, like the current bus, will move as fast as auto traffic.
4) South bound on Stockton between Market and the Tunnel there is a diamond lane which should be converted to dedicated transit lane with six inch curbs on each side. DPT didn’t have traffic data for south of the tunnel but this will leave a moving lane and a right or left turn lane for general traffic. This represents only a small reduction in auto traffic capacity because auto traffic is not supposed to use the diamond lane anyway. At Market the diamond lane disappears because DPT needs a lightly used right turn lane to the garage on Ellis. This lane can be removed extending the dedicated lane for Muni and provide an additional lane for all traffic to cross Market to Fourth. This treatment will save at least a minute. Dedicated lanes will not be necessary northbound on Kearny.
6) South of Market Section: During the peak hour, 8:00 to 9:00 A.M. the 9A and B San Bruno and the 30 Stockton each have 12 runs on Fourth. The 45 Union-Stockton has about 8 runs per hour. Thus, there is a bus about every two minutes. Based on Muni maximum load standards there will be 3,000 people on these buses. This should easily justify a nearly absolutely dedicated lane because the transit lane will carry much more than twice as many people as an auto lane, with 1.4 people per car. This dedicated transit lane should be located on a center lane of the roadway to eliminate any impact from right turning or parking cars. Similar to Third Street the curbs will not be continuous to allow for fire truck movement and backing out of trucking where required. There should safe loading islands adjacent to the transit lane with ticket machines. The safety islands would also serve as pedestrian safety medians. This treatment will reduce running time by two minutes on both North bound Third and south bound Fourth. Muni service planning notes that the recently added unenforced diamond lanes already save a minute (not currently reflected on the schedule).

These Phase 1, easy improvements, reduce running time by a total of seven minutes to 14 minutes for 1.7 miles and the current 10 stops. To be conservative say running time is 15 minutes.

Comparison of BRT vs. Subway service: Some of the BRT running time improvement will accrue to riders on the 45 and 9 lines while their bus uses the Stockton BRT improvements; Buses provide more opportunities for turnbacks and limited service expresses than a subway; BRT capital cost is lower and faster construction will have less disruption to neighbors and merchants. A subway ride is more comfortable than BRT but BRT running in a dedicated right-of-way will be smoother than a bus because the driver will not have to swerve to avoid traffic or accelerate to pull out of bus zone; BRT will have a street view; Subway riders with North Beach or Marina destinations will have to leave the subway and transfer to get their final destination and this means additional walking and waiting; BRT has higher expense for driver salaries because one subway driver can handle more than one car; Partial BRT can easily be provided for the Stockton Line north of Clay Street, saving riders more time; Subway have longer headways which effectively lengthens run time. Subways cover the route faster than BRT.

Phase 2: Stockton Street in Chinatown: Dedicated center lanes with loading platforms are also possible in Chinatown but this will require non conventional thinking. As a comparison to the subway plan this will only involve one block and save one minute; but it could be part a low cost extension north of Clay Street.

Imagine the 68 feet wide Stockton Street, between building faces, with no sidewalks and two BRT lanes in the center. This will leave about 23 feet on each side to be shared by shop keepers, pedestrians and/or an occasional merchant’s vehicle moving at a pedestrian pace. Boarding platforms will be provided alongside the BRT lanes for every stop. The community should decide how to use the space alongside the BRT where there are no boarding islands. One way, could use the space for: pedestrians, benches, trees or outdoor restaurant tables. This would require truck parking spaces on the side streets and have no auto or truck traffic on Stockton Street. Another way, could allow trucks and commercial vehicles to park along each side of the transit lanes. Truck drivers will have to exit away from the bus lane for safety. All vehicles, except Muni, will have to turn at every corner. This will be enforced by opposing traffic which will also have to turn. As an example: Washington is one way east bound so that traffic between Clay and Washington will go northbound east of the transit lanes and southbound west of the transit lanes. Between Jackson and Washington this will be reversed and traffic east of transit will go south while traffic west of transit will go north. There may be other uses and maybe each street could decide how to use the space.

This longer project (two miles) will save over three minutes through the most congested part of Chinatown. Except for the quickly moving surface transit, this new Stockton Street will perform as an ideal pedestrian market street. This will improve pedestrian ambiance and the merchants may prefer this or other alternatives that improve both transit and shopping. Northbound this change will have no traffic impacts on the nearby parallel and crossing streets because Stockton traffic is so low. Southbound the redistributed traffic impacts will be more noticeable but still low.

Ultimate 30 Stockton Running Time: The ultimate includes a low cost extension on Stockton as a market street to Columbus and low floor buses when new buses are purchased. Ultimately, before a subway can begin operation, the current 23 minute running time to Columbus can be reduced to 15 minutes compared to eleven minutes for a subway, which will require a new expensive extension.
For a reality check on this time we should look at The ‘K’ LRV line from St. Francis Circle to West Portal Station, 0.65 miles and two stops – Muni says four minutes and I think that this is accurate – but this section does not have easy boarding or signal preempts and has two stop signs plus a slow entry to the station. West Portal does have the equivalent of some enforcement of diamond lanes because autos rarely impose major slow-downs to transit. This indicates a running time for 2.0 miles of 12.3 minutes with less then perfect BRT. Extra time is necessary to cross Market Street but 15 minutes total should be enough.