Grade Separation: The Decadal View

For the last few years of debate around the issue of high-speed rail, grade separating the peninsula rail corridor was often cast as an all-or-nothing proposition.  This extremist view clouded two important facts: first, the corridor is already mostly grade-separated (in 2013, only 40 road crossings out of 104 between San Francisco and San Jose will remain at grade); and second, grade separation is a slow and inexorable process that takes place over many decades.

If we assume the next several decades will be like the last several decades, we can take an educated guess about how and in what order grade separations will be built.  The criteria for prioritizing each project could be:

• creating long, uninterrupted stretches of grade separated right of way to enable higher train speeds without compromising safety
• creating a four-track mid-line overtake facility to increase the capacity of the corridor, to support initial HSR service
• separating crossings that rank highest in the CPUC's Section 190 Grade Separation Priority List
• delaying the most expensive and politically costly projects until last

Phase I is simply the completion of the San Bruno grade separation in 2013.  The San Mateo / San Bruno grade crossing being replaced was once rated #7 statewide on the CPUC's priority list.

Phase I: San Bruno Grade Separation

Phase II consists of four projects in San Mateo County, opening up two long stretches free of crossings by the early 2020s, including 14.8 miles free of crossings north of Burlingame, and 6.5 miles south of San Mateo.  This enables the future construction of the "short" mid-line overtake envisioned in Caltrain's corridor capacity analysis, and leaves only a few dense clusters of crossings within San Mateo County.  The new grade separations, in order of priority, are:

1. 25th Ave in San Mateo, the only grade crossing that remains between San Mateo and Redwood City.  Along with new grade separations already planned at 28th and 31st, this project enables the future 4-track mid-line overtake.
2. Broadway in Burlingame, an extremely congested crossing that has been slated for grade separation since the 1970s.  It rates #11 statewide on the latest CPUC priority list.
3. Linden Ave in South San Francisco, originally planned as part of the San Bruno project, but dropped from the final design in 2007.  Grade-separation at Linden is accompanied by the closure of Scott St, which becomes a pedestrian tunnel.
4. Center Street in Millbrae, a grade separation that will require a U-shaped elevated ramp due to the nearby BART subway tunnel box.  Such are the consequences of bad planning.

Each of these projects is independent and can be negotiated on a case-by-case basis with the four affected cities.  Starting in San Mateo County allows at least another decade for the Pacheco vs. Altamont debate to run its due course, legally and politically; these four projects are useful either way.

Phase II: San Mateo County Grade Separations

Phase III occurs mostly in Santa Clara County, creating a new stretch free of grade crossings from the southern half of Palo Alto all the way to San Jose in the late 2020s, assuming the routing of HSR over Pacheco survives as currently planned.  The last two grade crossings in San Francisco are also eliminated as part of the downtown extension project.  This phase includes the following six discrete projects:

1. Mary Ave in Sunnyvale, the corridor's busiest grade crossing in this county, with more than 25,000 daily vehicles
2. Sunnyvale Ave
3. Rengstorff in Mountain View, with about 18,000 daily vehicles
4. Castro in Mountain View, with about 9,000 daily vehicles
5. Charleston and East Meadow in Palo Alto, with a combined ~30,000 daily vehicles plus numerous pedestrians and bicyclists
6. 16th and Common in San Francisco, as part of the DTX project

Together, these six projects create a new 16-mile stretch of track that is entirely free of grade crossings.  The corridor is now left with just three dense clusters of grade crossings, in San Mateo / Burlingame, Redwood City, and PAMPA (Palo Alto - Menlo Park - Atherton), highlighted in orange in the figure below.  Note that these three dense clusters contain 27 crossings, and that to get this far, only 12 existing crossings were newly separated.

Phase III: Santa Clara County Grade Separations

Phase IV is the Great Redwood City Grade Separation.  This project, potentially for the early 2030s, would prolong the four-track mid-line overtake by three miles, by removing six grade crossings in downtown Redwood City.  Removing this cluster first makes sense from the standpoint of increased corridor capacity, the lowest number of new structures, the short mileage, and the entire project being politically simplified by virtue of its containment within Redwood City limits.

Phase V is the Great San Mateo / Burlingame Grade Separation.   This is a tougher project because it involves some of the most highly constricted portions of the corridor.  It also involves political and technical coordination between two neighboring cities, adding an additional challenge.  The sheer quantity of crossings (13 grade crossings + 4 obsolete grade separations within 2.4 miles) is also a complicating factor.

Phase VI is the Great PAMPA Grade Separation.  This project is left for last because it lies in the most expensive real estate on the corridor, involves coordination between three different cities, and is liable to cause the fiercest political and legal backlash anywhere on the peninsula.  Delaying it until last, perhaps into the late 2030s, allows the customarily long planning process to run its course without undue haste in all three affected communities.

We didn't arrive at today's state of grade separation (more than half) overnight.  It resulted from a slow and steady process that began in earnest in the 1940s.  The future is likely to be similar, and the peninsula rail corridor can reach a far improved state by separating just 12 more crossings over the next couple of decades, as described in Phases II and III.  This dozen should be prioritized for construction, before any of the crossings in the remaining dense clusters are touched.

Tunnel Sizes

Mountain View mayor Ronit Bryant argues that a deep tunnel for Mountain View ought to be considered because "most of the BART in our area is underground. If they managed it then, why can't they do it now?" Here's one big reason why:


The HSR tunnel requires large clearances around the train for aerodynamic reasons. Without those clearances, a high-speed train (going 1.5 to 2 times as fast as BART's maximum speed) will dump megawatts of power into swirling the air inside the tunnel, making it unbearably hot. In the diagram, the BART tunnel bore is 16 feet (5 meters), and the HSR tunnel bore is 26 feet (8 meters).

Getting Sketchy In Mountain View

The city of Mountain View recently commissioned a series of sketches depicting different grade separations options for Castro St and Rengstorff Ave. Let's take a look.

First, shockingly, the tracks are configured as separate pairs, one for HSR and another for the rather aptly-named Caltrain/Freight. As will continue to be argued on this blog, and as the preliminary alternatives analysis acknowledges, this FFSS (fast-fast-slow-slow) track arrangement would force highly inconvenient single-tracking operations during service disruptions. It would also prevent commuter train overtakes. This is the worst-possible track arrangement for Caltrain. That being said, the track arrangement isn't the point of these renderings. On to the critique.

The recently installed faux-depot, built for a couple of million dollars (a pittance compared to the HSR options now being weighed), appears in all the options for Castro Street. It is evidently a civic pride and joy, to be preserved at all possible costs--costs to state and federal taxpayers, and ongoing costs to pedestrian transit users attempting to circumnavigate it for access to the actual train station. The elevated option is even dinged for "poor relationship with depot building".

The Castro elevated alignment shows a rail viaduct with a full eighteen (18) feet of roadway clearance underneath, and gigantic 40-foot ornamental towers with "potential for civic gateway treatment" that dominate the landscape. The same urban design values as for the faux-depot (form over function!) are on full display. If you wanted to sand-bag the elevated option, this is clearly how you'd do it. A more functional and discreet plan would be (a) to reduce the clearance under the viaduct to a more reasonable 16 feet, which could still accommodate all manner of interstate trucking and military convoys, and (b) sink Castro by about 5 feet, with relatively minor impact to business frontage on the 100-block of Castro. This would not only reduce the hulking profile of the station, but also shorten access to the platforms. This exact "split-grade" plan was proposed in Focus on Mountain View, accompanied by a Google Earth 3D model, all free of charge.

The Castro trench alignment, sure to be everybody's favorite, would literally introduce a new moat between downtown and the VTA light rail station. A nice "greenway" would be wedged in between the moat and Central Expressway, on top of the HSR tracks, for a (pleasant?) stroll to Rengstorff Park. As was previously pointed out, it is not feasible to put Caltrain/Freight into a trench at Castro and still make it back up to grade over the Stevens Creek, all the while adhering to the ultra-conservative 1% grade limitation for freight. Therefore, the trench alignment clearly suffers from some basic engineering feasibility issues, unless Mountain View has plans to grade-separate the creek as well. This plan, unless Mountain View pays for it, is likely a non-starter.

The Castro at-grade alignment is a nice illustration of what happens when you try to build a full-depth road underpass on a dense city block. Street frontage is severely affected, with loss of parking and a labyrinth of ramps and steps for pedestrians, something that is unlikely to sit well with local businesses--assuming they can even survive the construction period, which involves extensive excavation. The VTA light rail station would be further isolated from downtown by a tower bridge, just another two convenient elevator rides away from all the action.

The weakness in these renderings is that grade separation is really a continuum. The elevated, at-grade and below-grade categories are artificially imposed on the problem to make it easier to understand, but end up obfuscating hybrid options known as "split" grade separations, where both the rails and road are altered (rather than just one or the other). These split grade separations can combine advantages of the different categories, to have your cake and eat it too.

It's disappointing to see renderings that show only what the train can do to Mountain View, and not what Mountain View can do with the train.

UPDATE: The city has now posted additional sketches with street-level views. Once again, the trench is made to look better than it really is; the sketch does not show the 7-foot rise at Castro Street as shown in the corresponding aerial view.

Focus on: Mountain View

Mountain View, home to some of the biggest names in Silicon Valley, is the third busiest Caltrain stop behind San Francisco and Palo Alto, and serves as a transfer point between Caltrain and the VTA light rail system. (photo by LazyTom.) The city has so far adopted a conciliatory attitude towards the high-speed rail project, unlike several cities to the north. In its scoping comments (p. 90), Mountain View first and foremost asks to be considered as a candidate for a peninsula HSR stop, citing its strategic location, transit connections, and easily accessible freeway network. The CHSRA's preliminary alternatives analysis for the first time featured Mountain View as a potential peninsula HSR station.

The railroad right-of-way through Mountain View parallels Central Expressway, a four-lane divided highway that is part of Santa Clara County's expressway network. This inherently gives the peninsula rail corridor a less residential character than in communities to the north, although several new housing developments have been built near the tracks in recent years, with more planned in the vicinity of the San Antonio Caltrain station. The latter station, opened in 1999, replaced the primitive stop formerly located at Rengstorff Avenue with what was billed as a model transit-oriented development.

Vertical Profile

The vertical profile of the existing Caltrain tracks is shown in the diagram below, created from Caltrain track survey data. The tracks slope gently, with the southern end of town a full 50 feet higher than the north. The slope is steepest (a bit over 0.6%) in the section between Castro St. and the Rte. 85 overpass, near the downtown station.

There are just two grade crossings remaining in Mountain View: Rengstorff Avenue and Castro Street.

The city has been planning a grade separation at Rengstorff Avenue since 2002, which is currently entering the environmental review process independently of high-speed rail. The city's preferred alternative (see feasibility study in PDF Attachment 2) calls for Rengstorff and the nearby intersection with Central Expressway to be depressed under the tracks. The $45 million plan already includes a corridor expansion to four tracks, as shown in the figure at right. Other design alternatives were eliminated: trenching the tracks at this location would have been complicated by the nearby Permanente Creek, and residents were concerned about the visual impact of elevated tracks. Nevertheless, the HSR scoping comments submitted by the city (see p. 95) ask that all vertical alignment alternatives be re-examined for the Rengstorff grade separation. Since the technical and community constraints have not changed at this location, the HSR alternatives analysis is likely to be consistent with the city's prior study.

That leaves the crossing at Castro Street in the downtown area, the most challenging design puzzle in Mountain View, where only one vertical alignment solution stands out as realistically feasible. The options that will probably not be practical are:

• A deep bore tunnel. While technically feasible, such a tunnel would be prohibitively expensive and difficult to build because the entire Mountain View train station would need to be sunk below grade, including 1/4 mile platforms if the city is selected as an HSR stop.


• A cut-and-cover tunnel or trench. The rails would need to be sunk 25 to 30 feet deep to pass under Castro Street, but the natural slope of the terrain to the south would require a steep 3% rise back to grade level at the Stevens Creek, potentially compromising the ability of freight trains to use the corridor. (This assumes re-routing the creek is not practical.)


• An at-grade alignment. The high train speeds would require the permanent closure of Castro Street, or the construction of a bridge over the tracks with extensive impacts to the 100 block of the historic downtown and businesses on Moffett Blvd. Either way, Castro would no longer connect to Central Expressway.

That leaves just one reasonable option: a split grade separation with the tracks elevated about 15 feet above grade and Castro Street sunk by about 5 feet. Sidewalks could stay at the current grade level, and the station would be elevated, not unlike the existing design at Belmont. To the north, the tracks would require a 2% grade to duck under the bridge at Shoreline Blvd. To the south, the tracks could run level (which is of benefit for building a new train station) to meet the rising natural grade, returning to ground level at the Stevens Creek. The resulting vertical profile is shown in the diagram below, in green; compare to the infeasible below-grade alignment shown in red.

While Caltrain and high-speed rail would comfortably handle this vertical profile, a rigid adherence to a 1% grade limit for the benefit of freight trains would likely require Castro Street be closed entirely to road traffic.

Horizontal Alignment

The Caltrain right of way is quite wide throughout Mountain View, with 75 - 100 feet available to build out the corridor to four tracks. The relatively sharp curve at San Antonio (the #8 worst curve on the peninsula) just barely allows 125 mph operation, although it could be straightened within the confines of the right of way. Does that mean Mountain View will easily accommodate an expanded, four-track, high-speed corridor? Sadly, no.

In the late 1990s, Caltrain agreed to let VTA and the city construct a light rail extension on its right of way. The light rail system now runs alongside the Caltrain tracks for about one mile, ending at a 50-foot wide terminal with expansive storage tracks near Castro Street (photo by LazyTom). On the western side of the Caltrain tracks, the city built a "transit center" to replace the dingy asphalt strip that previously served as a Caltrain platform. This $20 million intermodal facility was completed in 2002 with a new plaza and modern-day replica of Mountain View's old train depot, which might have earned a spot in a register of historic places had it not been razed in 1959. All this recent construction consumed a large portion of the available right of way, where ample space to build four tracks existed as recently as 1998. Not surprisingly, the city clings to this new civic infrastructure and suggests among other options that HSR be routed via the Central Expressway median--never mind that $20 million is a relative pittance in the context of the peninsula HSR project.

Further south, one southbound lane of Central Expressway will likely be taken to route the VTA track east of the bridge support pillars at Rte. 85 and Whisman, freeing up space for the four-track peninsula corridor.

The light rail situation is further complicated by (a) VTA's inexplicable ambition to double-track the short section of single track that encroaches on the Caltrain corridor, (b) the possible need to maintain freight train access to the Moffett Drill Track (see docket FRA-1999-6254 ; the switch that connected the VTA tracks to the peninsula corridor was recently dismantled, but the Moffett spur is not formally abandoned), and (c) the difficulty of relocating the Evelyn VTA station--incidentally second-to-last in ridership on the entire light rail system, with roughly 60 daily passengers. All three of these factors need to be examined with a critical eye towards who will pay and who will benefit.

Further north, another pinch point exists near the San Antonio station, where the San Antonio Road overpass (a 1960s structure that Caltrans lists as a deficient bridge) does not provide sufficient horizontal clearance for four tracks. The nearby pedestrian underpass would also require modification, likely as part of its extension across Central Expressway to the planned residential redevelopment of Hewlett Packard's former Mayfield campus. It goes without saying that the San Antonio platforms will need to be rebuilt, just like anywhere else on the peninsula.

Downtown Done Right

To visualize downtown "done right," download Mountain View 3D Model (2.2 Mb) for Google Earth. All the illustrations in this section are taken directly from this model, built by Richard Mlynarik.

While Mountain View's transit center is billed as an intermodal, accessible facility, there is vast room for improvement in passenger circulation. The challenge of designing a rail alignment that overcomes the twin obstacles of Castro Street and the light rail tracks presents an opportunity to improve Mountain View's transit center by creating a modern, functioning gateway into downtown. While the construction impact would indeed be frustrating, especially so soon after the area was rebuilt, this would simply reflect the haphazard manner in which the existing facility was planned, with little regard to future HSR requirements or effective pedestrian circulation.

Currently, the light rail tracks end on an unpleasant concrete island hemmed in by Central Expressway and the Caltrain tracks, requiring a circuitous route for pedestrians to access the trains. Centennial Plaza, the faux-depot and the bus loop cut off the Caltrain platforms from downtown. All passengers wishing to transfer between buses and trains must use crosswalks.

There is surely a better way. One possible redesign of the Mountain View Transit Center is described below. It is an exercise to achieve the best possible transportation functionality, providing these specific benefits:

• Keeps Castro open to all traffic
• Provides direct, elevator-free access from Caltrain/HSR platforms to sidewalks on both sides of Castro, without circuitous detours
• Places the VTA light rail terminus in Centennial Plaza, where it meshes intimately with the pedestrian fabric of downtown
• Allows direct platform-to-platform, passenger-friendly transfers between light rail, Caltrain locals and expresses, HSR, buses and employee shuttles--without the need for an umbrella during the rainy season
• Locates station amenities such as ticketing, bathrooms, snacks, etc. at the crossroads of pedestrian traffic
• Provides station parking under the tracks, reducing the need for unsightly parking structures
• Does not force bus passengers to use a crosswalk

The resulting configuration is a four-track, two-platform Caltrain / HSR station, elevated over Castro Street. The light rail station slants diagonally underneath the elevated tracks, with the trolleys pulling right up to the corner of Castro and Evelyn in a sunken Centennial plaza. (Light rail requires the same 16-foot vertical clearance as road traffic on Castro). The twin Caltrain / HSR island platforms are extended north over Castro, enabling direct access to sidewalks on both sides and a platform-to-platform transfer between all transit modes.

Centennial Plaza looking east	Evelyn looking north to Castro	Castro & Central looking south
View from northbound Central Expressway

For additional views and detail, download the 3D model for Google Earth.

What happened to the retro-faux-depot? Remember, this design is an exercise in form following transportation function. In 1895, the depot would have housed important functions such as mail and baggage handling, signaling staff, and a telegraph operator. Today, those functions are obsolete, so the depot building has no place or purpose. It's gone, for the greater good of functional 21st-century transportation. Architecturally, there are far more exciting possibilities that enhance rather than impede functionality.

Regardless of the details of how exactly HSR is integrated with Mountain View's transit center, one thing is clear: the high-speed rail authority's charter is to build HSR, period. Ensuring that the intermodal connections at Mountain View can live up to their potential will largely be up to the city and its citizens, who ought to take strong initiative to ensure it's done right--without letting their judgment be unduly clouded by the $20 million they spent this past decade.

NOTE: This post will be updated continuously, as warranted by additional information or new events relating to Mountain View.