More people are speeding on the Golden Gate Bridge – here’s why

16073676100_24d1279e0a_z

The new barrier and the bridge during closure. Image by David Yu, on Flickr

When the new Golden Gate Bridge barrier opened, it was heralded as a new age of safety, but there were rumblings of problems immediately. Bridge Manager Kary Witt was quoted as saying, “I do think everyone is driving too fast. Everyone needs to slow down.” A few days later, he was rather more forceful: “We’re seeing too many drivers driving 20 to 30 miles over the speed limit. It’s completely unacceptable.” While the district seems to have been caught off-guard, this was an entirely predictable result.

Keeping roads safe is one of the most important tasks traffic engineers have. To do this, they will often try to improve a road’s safety by making it more forgiving of driver error: make the lanes wider, smooth out the curves, remove trees, and add median barriers. But this sort of improvement assumes that people drive a set way without regard to their environment.

This is not true. People drive as fast as feels safe, and they subconsciously react to visual cues to tell them what that speed is. Have you ever driven along a road at what felt like a comfortable speed, only to find that you were going 55 in a 35 zone? You were a victim of this subconscious pressure, called risk compensation.

By removing dangerous obstacles, engineers will often paradoxically make a road less safe by encouraging higher speeds. In a limited-access highway this might be okay, but on roadways that aren’t limited access – like the Golden Gate Bridge – it can create a dangerous false sense of security.

This is precisely what is happening on the bridge. Before, drivers on the bridge had a very visible cue that danger was omnipresent as traffic whizzed by in the opposite direction just a foot away. People would drive slower to ensure they had control of their vehicle and wouldn’t accidentally drift into traffic. They also had to navigate the very tight space between toll booths, slowing traffic further.

By removing the toll booths and adding the center barrier, the bridge district has lowered the perceived danger of crossing the bridge. This has encouraged drivers to drive faster, which has resulted in more crashes.

This is not a limited access roadway, either. There are driveway entrances and exits at the Toll Plaza and at Vista Point. Pedestrians and bicyclists cross the bridge on its sidewalks. This is a recipe for disaster.

It is perhaps understandable that the district would choose to spend millions on a median barrier. It was a harrowing crossing, and I know I never drove in the left lane if I could avoid it. But it is baffling that the bridge district was apparently unprepared for higher speeds as a result of this change.

There is a lesson here: traffic speed follows design. If towns and cities in Marin want to reduce speeds and increase safety, it must design roads that encourage people to follow the desired speed. The Golden Gate Bridge District has done the opposite, telling people to go one speed while silently encouraging them to go faster. If it’s serious about keeping speeds down, it won’t rely simply on enforcement to keep speeds down but will also seek design solutions.

A third lane on the Richmond Bridge is just a bandage

from MTC

from MTC

The push for a third lane to Richmond has sucked a lot of the air out of the conversation over Central Marin traffic. Cut-through drivers from San Francisco to Richmond are taking up all the space in Larkspur and causing horrific traffic. Thanks to induced demand, however, the third lane will likely fill up soon after it opens and we’ll be back to the same old story.

The most common way to think of traffic is as a gas that fills the space it’s given. No matter how much you build, there will always be traffic to fill it. This couldn’t be more apt for the situation faced by the Richmond Bridge.

Right now, for cut-through San Francisco-to-East Bay commuters, the Marin route is the fastest and cheapest way to get home. These drivers may have to deal with congestion and delays on Sir Francis Drake and 580, but it’s less than what they’d have to deal with on 80 and the Bay Bridge.

If we solve the problem and open a new lane on the Bridge, we’d reduce congestion in Marin enough that we could declare victory… until more people saw that it was a less-congested route than 80 and the Bay Bridge and switched. Either this area will return to its present levels of congestion, or the congestion will migrate to another bottleneck further south in the system, or some measure of both.

This is a much larger version of a problem faced by Los Angeles suburbs, where cut-through drivers, guided by their GPS, take surface streets to escape congestion on freeways. Delays become as bad on surface streets as on the freeway.

If congestion returns to Larkspur Landing, then the widening will simply buy us a few years of peace. If it causes another bottleneck, we’ll have bought some peace to Larkspur Landing at the cost of congestion elsewhere. If it’s both, then nobody wins.

Longer-term solutions depend on which outcome occurs; let’s look at each in turn.

Congestion comes back to Larkspur Landing only

If this occurs, the only real solution is to keep traffic on the freeway as long as possible by installing a proper 101-580 interchange in San Rafael. This interchange has been proposed before, but community opposition to a towering flyover connecting westbound 580 with southbound 101 scuttled the project. If the same opposition arises again, it might be worthwhile to simply remove that aspect and only do the eastbound 580 to northbound 101 aspect.

For now, at least, Caltrans ought to remove signs at the Sir Francis Drake Boulevard exit indicating that that is the way to 580.

Congestion occurs elsewhere in the system

The most likely location for congestion to occur is south of Marin City: on the Waldo Grade, Golden Gate Bridge, Lombard, or Van Ness, all of which are good targets for transit. Both Marin commute trips and local San Francisco trips are relatively easily served by transit. The upcoming Van Ness BRT line will make a big difference to that corridor, and an extension onto Lombard would help both GGT and Muni riders. Extending the HOV lanes onto the Waldo Grade by converting one of the through-lanes would speed transit and encourage carpooling, also helping alleviate congestion.

Alas, transit sometimes functions like adding more lanes: the amount of congestion stays constant even while the transportation capacity of a road to move people increases. At least we can comfort ourselves that fewer people will experience congestion from behind the wheel.

Congestion occurs both at Larkspur Landing and elsewhere

If this occurs, then planners will need to employ both solutions: add the interchange and improve transit.

The only permanent solution

The rub, of course, is that congestion is ultimately not a solvable problem without an economic downturn. Houston, Seattle, Los Angeles – all have tried to fix congestion by increasing roadway capacity, and none have succeeded. Anthony Downs, in his 1992 book Stuck in Traffic, said that widening a freeway doesn’t work thanks to what he called a “triple convergence”:

In response, three types of convergence occur on the improved expressway: (1) many drivers who formerly used alternative routes during peak hours switch to the improved expressway (spatial convergence); (2) many drivers who formerly traveled just before or after the peak hours start traveling during those hours (time convergence); and (3) some commuters who used to take public transportation during peak hours now switch to driving, since it has become faster (modal convergence).

The only way out is to view road space like a resource, and to price it as such. Jarrett Walker describes it thus:

Fundamentally, congestion is the result of underpricing.  If you give away 500 free concert tickets to the first 500 people in line, you’ll get 500 people standing in line, some of them overnight.  These people are paying time to save money.  Current prevailing road pricing policy requires all motorists to act like these frugal concertgoers.  Motorists are required to pay for road use in time, rather than in money, even though some would rather do the opposite and our cities would be safer and more efficient if they could.  Current road pricing policy requires motorists to save money, a renewable resource, by expending time, the least renewable resource of all.

For the Bay Area, this would mean varying bridge tolls during the day so that congestion never builds up. Downs’ triple convergence would work in reverse.

With a rush hour 80 and Bay Bridge free of congestion, cut-through travel would be much less attractive for Contra Costa commuters. Those that still made the journey would likely not be enough to congest 101 at all.

But before then, we have a third lane and an interchange to try.

The 70 and 71 have issues

Last week, in covering GGT’s quarterly schedule update, I mentioned that the 70 follows the 71 too closely for some of its runs, once just 3 minutes behind. This, however, was not an accident of poor scheduling but a deliberate choice.

There are two reasons for this.

History. Years and years ago – long before the senior Marin Transit (MT) planner I spoke with came to work for the agency – GGT needed extra capacity on Route 70. They found that most passengers on the line were staying in Marin and crowded out the folks heading to San Francisco.

To beef up service, GGT asked MT to create Route 71, a Marin-only local bus. Though operated by GGT, the 71 would better fit MT’s mission to provide transit within Marin County than in GGT’s mission to provide service between Marin and neighboring counties.

At the rush hour times when the 70 closely follows the 71, then, the 71 serves as a relief bus, picking up all the local ridership and leaving the regional ridership for the closely-following 70.

But why not just use a larger bus for the 70? Because of the next reason:

Fare media. MT issues a raft of bus passes and discount cards for students, seniors, the disabled, and low-income riders that are accepted on all MT routes. Because GGT operates most of MT’s system, the most common place to use these is on a GGT-operated bus, and it would seem logical that GGT would also accept these on its routes – the commute and inter-county routes – for trips starting and ending within Marin. Alas, it is not so.

Transit-dependent riders equipped with MT-issued passes will therefore avoid the 70, as they will need to pay full price for their trip. Instead, they’ll board the 71, which does take the passes.

As a result, the 70 and 71 serve different ridership pools, both of which need to make the same transfers at San Rafael.

One planner I spoke with said that “everyone knows” the 70 and 71 don’t make sense. Given the reasons above, it’s hard to argue with him. GGT ought to take MT’s fare media for intra-Marin travel but either can’t or won’t, and so the 71 provides 101 corridor service to MT’s poorest riders. It’s an exceptionally inefficient and wasteful workaround to what should be a simple problem to fix. (But, given transit law in California, I’d wager GGT can’t take MT’s fare media without a waver of some kind.)

Lack of seats is a concern, but the benefits of all-day 15 minute frequency on 101 would far outweigh the negatives. Larger buses, such as the MCI commuter buses or double-deckers like the Google shuttle, could also provide some of the necessary capacity.

New GGT schedule cuts Route 80, but more could be done

In GGT’s quarterly schedule adjustment, the agency will do as it planned this past summer and cut the long-suffering Route 80, a plodding local bus route that stopped at every bus stop from San Francisco to Santa Rosa. However, Highway 101 service could use more tweaks to better integrate Route 71 and other lines south of San Rafael.

New 101 bus service will look like so. Image by the author.

New 101 bus service will look like this. Image by the author. Routes 4, 10, 27, and 101X excluded for clarity.

The new basic San Francisco-Santa Rosa service will operate like the image to the right. Timed transfers between the 101 and 71 at Novato will ensure easy access to any stop between Santa Rosa and Marin City, while the express/local service pattern will better serve the larger long-distance market to San Francisco.

This is similar to Clem Tillier’s proposed Caltrain schedule, with a Silicon Valley Express and San Mateo Local.

Unfortunately, the timed transfers aren’t ideal. Before 8:55am, transfers between the 71 and 101 are 6 minutes. After 8:55, they extend to 9 minutes, a hefty chunk of time for riders.

As well, the 70 often follows the 71 too closely, once as close as 3 minutes behind. Though GGT is providing four local buses in a given hour, or one on average every 15 minutes, it instead has a 3 minute wait, followed by a 30 minute wait, another 30 minute wait, then another 3 minute wait. Though GGT is running enough buses to provide show-up-and-go service, the agency effectively continuing its two-bus-per-hour baseline schedule. It is, quite honestly, a waste of money.

Other tweaks to the schedule are quite good.

  • A new ferry shuttle, the 37, will run between Smith Ranch Road/Lucas Valley to Larkspur Ferry Terminal in the morning and evening. Though the parking situation is still horrid, it’s good to see the agency is continuing to expand its successful shuttle program.
  • Rerouting the 56 to the Broadway Tunnel instead of North Beach will save time – about 8 minutes per run – which will let commuters sleep in just a little bit longer, get home a little bit sooner, and save the agency about $50,000 in annual operating cost.
  • The popular routes 4 and 54 will receive more runs in both directions.
  • The 35 and 36 will become more consistent, with no service to Andersen Drive. Before, some runs would start on Andersen.

The largest concern is whether this new adjustment will lead to personnel shortages, which have plagued GGT for the past two quarters. Schedules should meet, not exceed, the agency’s capacity, and adjustments are when any personnel shortages can be smoothed out.

While the adjusted Highway 101 service isn’t as smooth as it ought to be, with long transfers and weirdly inconsistent headways between the 70 and 71, the new schedule is overall positive and uncharacteristically bold. As long as this doesn’t lead to more personnel shortages, the new schedule will be a success.

UPDATE: To clarify, the updates to the 35 and 36 are at the direction of Marin Transit, which contracts out their operation to GGT. Any change to the 71, which is under the same contract, would need to first start with Marin Transit. A fuller post on that mess is forthcoming.

How we allocate street space

A painting that has been going around the internet for the past week dramatizes the pedestrian view of our streets and how we allocate space.

Image by Karl Jilg

Image by Karl Jilg

The drawing, by Swedish artist Karl Jilg and commissioned by the Swedish Road Administration, was forwarded to me by a friend (who found it on Vox). Streets, it says, are yawning chasms of death, no-go zones with only thin strips along the edges where we can exist as people on foot rather than people sitting behind the wheel.

It makes sense, then, that a barrier of parallel-parked cars would make us feel safer (as Jeff Speck notes in his book Walkable Cities), and that very wide streets with a lot of road frontage taken over by car space would feel very unwelcoming.

New on the IJ: Housing close to transit hubs is a time-tested model

Last week, IJ columnist Dick Spotswood wrote that he had a revelation: The best ways to provide new homes in Marin are to add housing to downtowns, emulate downtown forms, and add second units.

It may have been a revelation to him, but it’s not news to the Coalition for a Livable Marin — CALM. We’ve been advocating for just such an approach since we were founded.

Spotswood wrote the foreword to Bob Silvestri’s pro-sprawl manifesto, but he’s starting to understand the wisdom of Marin’s small, dense, rail-oriented downtowns.

Up until the 1940s, Marin was built to maximize ridership on our old light-rail system, the Interurban. Planners put high-density commercial and residential buildings right up next to stations and less-dense homes farther out.

The layout was deliberate. While people today often drive from parking space to parking space on their way home to run errands, yesteryear’s Marinites would walk from shop to shop to run errands on the walk home.

People taking Golden Gate Transit can often still do that, especially at one of the downtown hubs. Take the 27 from the Financial District to San Anselmo, pop into Andronico’s or Comfort’s for the night’s dinner, then walk home.

Most wonderful about this sort of development is how it’s used when people aren’t commuting. Kids can stop by the doughnut shop on a Saturday, parents can watch the street from the coffee shop, and seniors can live their days seeing neighbors and family without ever setting foot in a car.

Marin ought to encourage people to live in places like this, not just for the sake of affordable housing or greenhouse gas emissions but for the health of the town.

Continue reading on MarinIJ.com

RVSD board member pushes parking in Larkspur

According to a source trusted by The Greater Marin, Ross Valley Sanitary District (RVSD) Board Member Frank Egger wants to convert RVSD-owned land at Larkspur Landing from a 120-unit residential area to a 600-space parking lot, presumably to serve the Larkspur Ferry Terminal (LFT). To make this possible, Egger is pressing Larkspur officials to change the zoning on the RVSD parcel.

While it’s no great secret that LFT has a parking shortage, Egger’s idea is exceptionally foolish from almost any angle.

District Finances – RVSD has a crushing maintenance backlog and huge financial problems. Developing its residentially-zoned property for either sale or lease would provide a much-needed cash influx to the agency. While parking fees would generate some income to the district, it is significantly less than what 120 homes could bring in. Parking simply doesn’t generate the income development does, and that’s not to mention sales or property taxes to Larkspur.

Traffic – A new lot of 600 parking spaces would generate as many as 600 extra rush hour car trips. A development of 120 new homes would generate, at most, 240 rush hour car trips, though likely much less given the proximity to the ferry. If Egger is concerned about traffic, parking will be worse than homes.

Ferry Ridership – The new lot would generate up to 600 peak-period, peak-direction ferry trips, precisely the sort of trip the system has little capacity to accommodate. The 120 homes would generate up to 120 (more likely less than 100) similar trips. However, if built to attract visitors, the homes could complement future new development that would attract reverse-peak and off-peak trips. Golden Gate Ferry desperately needs people to commute from San Francisco to Larkspur Landing for the service’s sustainability.

Frank Egger has led the charge against the Fairfax Housing Element, especially against allowing the downtown to expand into areas now dominated by parking (an idea that even Dick Spotswood endorsed). From his perch on the RVSD Board, Egger is continuing to push a cars-first ideology. He wouldn’t phrase it this way, but it’s clear he’d rather build homes for cars instead of homes for actual people.

Follow

Get every new post delivered to your Inbox.

Join 1,238 other followers