How a traffic jam in a Los Angeles suburb made trains late up and down the West Coast

Severe traffic congestion on I-5 north of Los Angeles during 2018's busy Thanksgiving travel period (photo courtesy of The Santa Clarita Valley Signal)

Last Saturday, I took Amtrak from the Los Angeles area, where I had spent Thanksgiving, to Davis, CA, where I spent a week working remotely from my childhood home before attending tomorrow’s rescheduled Cal-Stanford football game.

California is constructing a high speed rail line to help meet the state’s north-south mobility needs. But currently, the once-a-day, circuitously-routed Coast Starlight train is the sole option for single-seat rail travel between Los Angeles and the Bay Area or Sacramento. Though the route is quite scenic, scheduled travel time between L.A. and Sacramento is nearly 14 hours.  

I opted for a more popular alternative, scheduled to be several hours faster than the Starlight: a shuttle bus from Los Angeles’s Union Station to Bakersfield, where riders can catch a northbound San Joaquins train seven times daily. Though Union Pacific refuses to allow passenger trains to continue through to L.A., citing capacity constraints on its tracks through the Tehachapi Pass, the route is one of Amtrak’s most heavily used corridors. A complex network of Amtrak Thruway intercity bus routes connects the route not just to L.A., but also to a diverse array of locations including Las Vegas, Palm Springs, and Yosemite National Park.    

The 113-mile L.A.-Bakersfield bus ride would take about two hours if vehicles on Interstate 5 operated at the road’s speed limit, but thanks to the realities of urban geometry and induced demand this is almost never the case. To account for the relentless highway congestion likely to delay the bus, Amtrak built 50 minutes of padding, plus an additional 20 minutes of layover prior to the train’s departure, into its schedule.

On this day, 70 minutes was not enough cushion. While the high-speed trains of the future will arrive at their Northern California termini less than three hours after departing Union Station, on Saturday it took that long just to get out of Los Angeles County. My bus left on time at 2pm, but was crawling past a Six Flags theme park in the northern part of Santa Clarita, an LA suburb, when it should have been approaching Bakersfield.

We did not arrive at Bakersfield until after 6pm, nearly an hour after our train was originally scheduled to depart.

Since many San Joaquins riders transfer from the shuttle buses, and my train was the last of the evening, Amtrak held the train’s departure until all the buses arrived. The last of those buses arrived at 7:20pm, about an hour behind my bus and two hours after the train’s originally scheduled departure. It took over nine hours for that bus to complete its 238-mile route from Indio, CA, a town east of Palm Springs.

Due to these car-caused bus delays, the train departed Bakersfield – and all 14 of its subsequent stops – more than two hours late. Passengers down the line were left waiting in the dark. Even projecting a real-time arrival was impossible prior to the train’s departure from Bakersfield.

Compared to my traffic-plagued bus trip between Los Angeles and Bakersfield, travel aboard Amtrak's San Joaquins train was welcome relaxation (photo by me)

Our train didn’t experience any additional delays on Burlington Northern Santa Fe’s tracks, arriving in Stockton at around 11:30pm. The café car even offered 20% discounts on food to help appease frustrated riders. 

In Stockton, I transferred to another shuttle bus headed to the Sacramento area, while the train proceeded west to the Bay Area. Though the Northern California roads were remarkably free of cars in the middle of the night, the bus had to wait for us to arrive before it could depart, and thus still was about two hours behind schedule. I rode the bus to Davis, its final stop, and since there was no local transit available at 1am I hopped on a Jump e-bike (with luggage in tow) to get home.  

Many riders on my bus transferred in Sacramento to the aforementioned Coast Starlight train, which continues north through the night from Northern California to Seattle. That train arrived in Sacramento a minute early, but like our San Joaquins train earlier that evening, Amtrak held it in the station until our delayed connecting bus arrived to ensure no one was stranded.

The Starlight departed Sacramento for its 15 remaining stops 51 minutes late, delayed by a traffic jam nearly a half-day earlier and 300 miles away.

High speed rail will save millions of riders from I-5. But California needs a solution sooner. 

A rendering of the proposed California High Speed Rail segment between Bakersfield and L.A. (photo courtesy of Bakersfield Californian)

Until the proposed high speed rail line through the Tehachapi Mountains and south to L.A. opens – which won’t happen for at least another 15 years – the Bakersfield-L.A. gap will continue to serve as an impediment to people in Southern California wishing to access rail service to points north. Thus, if the expanding system is to succeed, something must be done to ensure riders never have to tolerate the traffic congestion I experienced last Saturday.

Rerouting the buses via an alternate, inland highway may have eased the delays I experienced last weekend. But any mixed-traffic route can become a parking lot in a heartbeat, rendering an on-the-fly detour useless.

Thus, a more permanent solution is necessary. I see three main options: more effective messaging to the public regarding holiday travel options, working with freight rail companies to restore conventional train service between Bakersfield and L.A. until the high-speed line opens, and constructing dedicated bus lanes on the most chronically congested sections of I-5.

Google Maps tracks my bus as it creeps through severe traffic congestion in Santa Clarita (screenshot by me)

Multimodal messaging: Authorities and the media often instruct travelers looking to avoid the worst traffic congestion during busy periods, like the Thanksgiving weekend, to drive at off-peak times or on alternate routes. But messaging intended to convince people they can avoid traffic only encourages more driving, and thus more congestion. And if too many drivers shift their trips to alternate times, the roads can become just as congested then as they are during traditional travel peaks.

A better solution would be to educate Californians, many of whom are accustomed only to cars and planes, about bus- and rail-based options. Such options include not just the San Joaquins and Coast Starlight, but also a plethora of intercity bus options. These options typically serve central locations that, in contrast to outlying airports, are accessible via a variety of modes

A rerouted Coast Starlight train makes its way through Tehachapi Pass (photo courtesy of Clear Signal Trains via Reddit)

Rail restoration: Much of the existing L.A.-Bakersfield rail route already sees passenger trains, as Metrolink, the region’s commuter rail provider, operates its Antelope Valley Line between Union Station and Lancaster.

The primary obstacle to a full LA-Central Valley passenger rail connection is the segment through Tehachapi Pass, farther to the north. Freight trains from all over the country converge on low-speed, winding tracks owned by Union Pacific to get through the pass. The last regularly scheduled passenger train to traverse these tracks – Southern Pacific’s San Joaquin Daylight – was cut in 1971, when Amtrak took over most of America’s intercity rail system.   

UP occasionally allows the Coast Starlight to detour through the pass when track maintenance necessitates a closure of that train’s regular coastal route. But peak intercity travel periods, like Thanksgiving, arguably strain California’s transportation infrastructure much more than a disruption to the Starlight. Thus, Amtrak and the San Joaquin Joint Powers Authority (the consortium of regional governments that manages the San Joaquins route) could try working with UP to temporarily extend at least a couple daily San Joaquins to L.A. during such peak periods.

Over time, perhaps the passenger rail providers could work with the freight company on infrastructure improvements that would allow for permanent extension of the San Joaquins line, as the Capitol Corridor Joint Powers Authority (SJJPA’s sister agency that manages the busy Amtrak route connecting Sacramento to the Bay Area) is doing to increase train service to Sacramento’s suburb of Roseville.

Such improvements could even allow some high speed trains to continue through to L.A. via the existing conventional tracks once the initial San Francisco-Bakersfield line opens. Barring electrification of the existing UP line, any such trains would have to switch to diesel power in Bakersfield, as Amtrak trains traveling from the Northeast Corridor to points south of Washington, DC do in the nation’s capital. 

A bus uses a dedicated lane to pass slow-moving cars on a freeway outside of London, UK (photo courtesy of Getty Images)

Bus improvements: A possible alternative – or complement – to restoration of conventional Bakersfield-L.A. rail service would be to implement dedicated bus lanes on the most congested portions of I-5. Amtrak’s buses, as well as local transit routes in Santa Clarita and L.A., would use these lanes and avoid traffic congestion-caused delays. Riders would enjoy faster – albeit less comfortable – transit service than would be possible via the existing railroad through Tehachapi Pass.

Such bus lanes would undoubtedly face opposition from drivers and those who make money off them. A possible compromise could involve “managed lanes” that some automobiles in addition to the buses would be able to use, or simply a redesign of the road's shoulder that would allow buses to use it during periods of severe congestion. Even this may face substantial obstacles, as a two-year, $171 million rehabilitation project of I-5 currently underway failed to include any such upgrades.

In the end, the best solution is simply to complete the high speed rail line to Los Angeles as quickly and efficiently as possible, allowing trains to fly through Tehachapi Pass via a series of tunnels and bridges completely separated from the existing freight tracks.    

During mass evacuations, people shouldn’t have to depend on cars

Residents of Paradise, CA attempting to drive away from the Camp Fire, seen burning to the right of the road, encountered heavy traffic (photo courtesy of The New York Times)

With a wildfire closing in on their homes, residents of Paradise, CA attempted to escape the flames using their automobiles. Roads leading out of town quickly turned into parking lots.

Some trapped drivers were unable to make it out in time, burned alive in traffic. Others abandoned their vehicles and fled on foot. 

Traffic jams: an icon of American disasters

Heavy traffic fleeing the Camp Fire passes a burning automobile near Chico, CA (photo courtesy of the Chico Enterprise-Record) 

Hurricanes, firestorms, and other natural disasters pack deadly, destructive force, but they don’t change the geometry of mobility. It’s simply not possible to safely and efficiently move large numbers of people out of harm’s way via a transportation mode as space-consuming as the car. 

Thus, scenes of extreme traffic congestion have long been a symbol of large-scale disaster evacuations in America. Already facing the prospect of losing everything, people must get behind the wheel and take to overwhelmed roadways, pitted for hours against other drivers who are just as stressed as they are.

Over 100 people died during the primarily car-based evacuation of Houston prior to Hurricane Rita (photo courtesy of Houston Press) 

Normally, evacuation orders come a couple days in advance of a forecasted catastrophe. In these cases, the sudden spike in automobile travel exacerbates the usual car-related problems, such as long delays, crashes, and demand for gasoline. These problems reared their ugly head during preparations for Hurricane Rita in 2005, when over 100 people died during the primarily car-based evacuation of Houston. 

Those without access to private vehicles are left behind, told to seek help from family, friends, or a vaguely-defined “local government”. Many states and cities have failed to include any non-car options in their contingency plans for emergencies. Even plans that do mention such options often assume that only disadvantaged populations, such as elderly and disabled citizens, will utilize alternatives to the automobile.

This can have deadly consequences. For example, many of the 1,300 people who died due to Hurricane Katrina in 2005 stayed behind only because they lacked viable transportation away from the storm. 

New Orleans still has a lot of room for improvement. The city cancelled an evacuation agreement with Amtrak last year even though thousands of residents rode special intercity trains to safety prior to Hurricane Gustav in 2008. The New Orleans Regional Transportation Authority’s bus system, meanwhile, still hasn’t been fully restored to 2005 levels.  

Car-based evacuations: a product of auto dependency

Residents of Orange County, CA attempt to flee the Canyon Fire in 2017. Notice the not-in-service bus stuck in the evacuation traffic (photo courtesy of CNN)

When sufficient transit is available, it can serve as a vital lifeline during emergencies, even on minimal notice. For example, Washington, DC’s Metro moved hundreds of thousands of riders out of the city center on September 11, 2001, and WMATA buses helped transport people injured at the Pentagon to hospitals. In New York, subway service was halted for over two hours following the attacks that morning due to damaged infrastructure and security concerns. But ferries arranged on the fly, as well as the decision to open bridges leading off the island to pedestrians, helped facilitate a relatively orderly evacuation of Lower Manhattan.   

But DC and New York are exceptions to the norm, as they are among the few parts of America in which multimodal options are mainstream forms of mobility. The Wikipedia article on Paradise describes the main obstacle to more efficient, multimodal evacuations in the rest of the U.S.:

“There are not many transportation options within Paradise other than driving an automobile.”

Once people believe it’s normal to have to drive everywhere, they depend on their vehicles even when it’s utterly irrational to do so. Thus, it’s no surprise that if told to evacuate, they jump in their cars on instinct, regardless of how congested or dangerous the roads are. Even if a resident of an auto-dependent place wished to evacuate using a different transportation mode, the necessary infrastructure to do so may simply not exist.   

Multimodal evacuations happen all the time – in theoretical models.

A Staten Island Ferry departs Manhattan's Whitehall Ferry Terminal on September 11, 2001 (photo courtesy of CNN)

With both natural and man-made disasters consistently omnipresent in the news, numerous researchers have developed models in their efforts to optimize the speed and efficiency of mass evacuations. Though some of this research incorporates multimodal options and acknowledges the superior efficiency these options could provide if effectively implemented, it’s not possible to fully account for all the variables of any particular disaster until it happens.

In addition to the basic infrastructural shortages inherent to auto-dependent areas, there are a number of practical challenges to developing a multimodal, high-efficiency evacuation plan. For one, transit employees can’t be put in harm’s way, and may need to be with their families and evacuate the affected area themselves.

Also, since almost all riders would be heading in the same direction, the transit provider would face an extreme version of a common challenge when moving commuters to job centers during rush hour – trains and buses would be packed in one direction, necessitating high-frequency operation, but would then have to return empty to pick up more passengers and fulfill demand. Buses would also need dedicated lanes to avoid becoming bogged down in car traffic, as even roads that don’t normally experience severe congestion may become clogged in emergencies, especially if primary highways and arterials are obstructed.

Hub and spoke models – in which local service transports evacuees to “staging points” from which longer routes provide fast connections to safe locations – may address some of these problems. For standard operations, transit providers are moving away from this model, embracing grid- or route-and-branch based approaches centered on frequent networks as they pursue bus system redesigns to keep up with shifting job centers and development patterns.

But in the unique case of a regional evacuation, temporary implementation of a well-distributed hub-and-spoke system would ensure passengers are riding in a variety of directions, to and from different hubs. Instead of buses running empty in one direction and full in the other, vehicles would originate at one hub and slowly fill as they move towards a different one. Riders would then transfer to a smaller number of very-high-frequency routes out of town, reducing the number of temporary bus lanes necessary to ensure these routes aren’t delayed by car traffic. 

As service would be oriented toward getting people away from the threatened area, all transit vehicles – and the employees operating them – would wind up in safe zones once the evacuation process is complete.

What if there’s no notice prior to a disaster?

An emergency vehicle passes abandoned cars burned during the Camp Fire (photo courtesy of The Mercury News)

A well-planned transit-based evacuation plan could work excellently – assuming there’s time to execute it. Such plans would be a strong fit for hurricanes, as preparations for such storms normally begin a few days in advance.

But all too often, disasters strike without notice. This occurred with the firestorm in Paradise, and also is typical of earthquakes, military or terrorist strikes, and tsunamis. Not only would it be incredibly challenging for a transit provider to switch from standard operation to evacuation-oriented service in an instant, but damage from the event could impede operation on major routes.

In such cases, it could be up to good Samaritans to maximize transportation efficiency on the fly and keep riders safe, as happened in New York on 9/11.  

We will never be able to eliminate the risk of an unexpected disaster. But by cutting back on our car-based travel, we can help reduce emissions-caused climate impacts, oil-motivated foreign policy decisions, and suburban sprawl-necessitated paving over of areas unfit for development – all of which have contributed to increases in the frequency and severity of these catastrophic events.     

WMATA’s proposed service increase may not be perfect. But the proposal has exposed transit opponents lurking in the shadows, and will force elected officials to show their true colors too.

After 7pm on weekdays, when rush hour-level service currently ends, this is what WMATA's train arrivals often look like. The agency's proposed FY 2020 budget would extend rush hour service through 8:30pm. (Photo courtesy of Greater Greater Washington)  

Earlier this week, WMATA General Manager Paul Wiedefeld unveiled his agency’s proposed FY 2020 budget. The budget leaves many of the DC area’s transit shortcomings left unsolved – track work will continue disrupting rail service on weekends, bus routes and late night rail service cut over the last couple of years won’t return, and most buses will continue mixed operation with automobile traffic at inadequate frequencies.

But for the first time in years, tangible transit improvements are within our region’s grasp. The proposed service increases, which include rush hour-level frequencies later into the morning and evening, restoration of Yellow Line service to stations that currently are badly underserved, and complementary access to the bus system for rail monthly pass holders, will greatly increase the utility of our existing transit infrastructure.

These service increases may not generate the excitement of a glistening new rail line. But if the WMATA board approves the proposed budget – including the funding needed to implement the improvements – the transit system will be a step closer to becoming a serious way people get around the DC region, rather than just fodder for happy hour jokes.

The budget proposal’s first accomplishment: exposing DC’s anti-transit Tweeters for what they are

Residents of the DC region aren’t happy with existing, inadequate transit options, and have been forced to find ways to healthily vent their frustration. Anonymous Twitter feeds focused on WMATA’s shortcomings have been among riders’ most popular outlets for this.

For years, the Tweeters managed to play the region like a Ressikan flute, helping shape how people use, perceive, and discuss transportation. Be it at work, bars, or parties, everyone seemed to repeat the same manufactured taglines and buzzwords when talking transit, making productive public discussion incredibly difficult.

But today, the tide has turned, and the flabbergasted Tweeters’ reaction to the proposed service increase may seal their fate.

Unsuck DC Metro, the most infamous of the bunch, reacted to the news with a libertarian tweetstorm, doing his best to echo Randal O’Toole. However, he was unable to maintain the laid-back demeanor of the Cato Institute senior fellow, referring to WMATA as a “mugger” and calling the Washington Post a “shill for Metro” because the paper bothered to publish a story describing the proposal.

The operator of another account, called 20 TRAINS PER HOUR, ALL OF THEM 8 CAR TRAINS (yes, all caps; the handle for this account is @HeadwaysMatter), ripped WMATA for proposing to run more…eight car trains. (WMATA’s trains range from 6 to 8 cars in length and the shorter consists, which also tend to be made up of older railcars, can get uncomfortably crowded). 

Update (11/2/2018): shortly after I posted this article, @HeadwaysMatter changed their Twitter name to "100% 8 car trains running on 3 minute headways." The name is no longer all caps, and may change again in the future.   

While these accounts may still have a following in anti-transit circles, people who want DC’s transportation system to improve are on to them. A long thread started on October 30 in the Columbian Memes for Congressionally Disenfranchised Teens Facebook group, a DC-area spinoff of New Urbanist Memes for Transit-Oriented Teens (NUMTOT), demonstrates how Unsuck has lost the transportation community.

One group member, Simone Rai, stated in the thread that “I don’t know how I ever liked unsuck DC metro [sic], but now I’m proudly blocked by them because they can’t take people liking tweets that are mildly critical of them.”

“I think I stopped following [Unsuck] because they were so negative,” member Laura Labedz wrote. “I’m glad I did.”

No one came to Unsuck’s defense.

        

Next to be tested: elected officials

Exposing the Tweeters for what they are may feel gratifying, but they’re not the decision-makers who will determine whether or not the proposed transit improvements become reality. Rather, this will come down to the members of the WMATA board, who will do exactly as those who appointed them – our elected officials – tell them to do.

WMATA’s battle for dedicated capital funding, which wasn’t won until earlier this year, demonstrated how challenging it can be to convince the leaders of DC, Virginia, and Maryland to back transit and work together. But in the end, the funding passed with bipartisan support in all three jurisdictions.  

The dedicated capital funding was essentially a do-or-die for mobility in the region – without it, it would have been extremely challenging for WMATA to develop a long-term maintenance plan, and our transit infrastructure would have continued to crumble. Legislation associated with the dedicated funding also bolstered oversight of WMATA, helping the agency better manage its finances.

While the dedicated capital funding is helping make the transit system safer, more reliable, and more sustainable, it didn’t come with any promises to expand what riders consider most when choosing their transportation mode for any given trip: service frequency and extent.

Wiedefeld’s proposal puts some of the service increases riders need on the table but, like WMATA’s capital projects, will require financial commitment from the three jurisdictions.

This time, the very survival of usable transit in our region is not at stake. Instead, the decision our leaders make will signal whether they’re serious about making multimodal transportation viable for all, or instead are satisfied with perpetuating today’s mediocrity.

No matter what they decide, they’ll be required to own up to their choice, just like the Tweeters.