by Rafael (warning: long post)

Introduction
Caltrain Firebird is a new infrastructure and operations concept I have developed to illustrate how integrated planning and modern technology could reduce HSR implementation cost in the SF peninsula while delivering the desired aggregate throughput capacity for the corridor plus improved Caltrain service, probably at a lower operating subsidy per passenger.

Note that CHSRA had previously already planned to share track with Amtrak and Metrolink in the Fullerton-Anaheim section but returned to the concept of dedicated HSR tracks in the arguably mistaken belief that FRA regulations are immutable and, that the existing dual tracks cannot deliver sufficient capacity. For now, though, let us examine the situation in the SF peninsula in more detail.

One Caltrain board member recently suggested the entire railway may have to shutter in 2012 unless it figures out a way to deliver more bang per county taxpayer buck. That will take large capital investments, but the state and federal governments will pick up the lion’s share of those – bureaucrats tend to think in terms of which pot funding comes out of, not who is putting the money into those pots in the first place. Regardless of which level of government ends up paying for them, the capital improvements have to be cost-effective, which means they ought to be driven by a financially viable strategy for future operations.

The Caltrain 2025 plan of record assumes electrification, lightweight EMUs, signaling upgrades plus access to the new Transbay Terminal will be sufficient to double line capacity and triple ridership by that year (implies 50% more seat capacity per train). However, increasing rush-hour rail traffic from 5 to 10 trains per hour each way would snarl up motor vehicle traffic at all remaining grade crossings. IMHO, it seems unlikely that Caltrain would ever get environmental approval for doubling rush hour train frequency, absent full grade separation of the entire corridor. The only available partner for that is CHSRA, so Caltrain’s plans have to change to accommodate HSR operations.

The notion that CHSRA should work around Caltrain’s already-completed plans that for the most part predate prop 1A needs to be challenged or else, the HSR starter line may never get fully funded and neither service is going to be successful. The current threat to Caltrain’s survival in the medium term makes lateral thinking more important than ever. The two organization may well prefer working independently to being joined at the hip, but the awkward fact is they need each other. This is reflected in the Peninsula Rail Program organization they’ve set up, but IMHO that body is still focused more on resolving conflicts between existing investment plans than it is with aggressively seeking out synergies to increase value per capital investment dollar spent.

Granted, it is not CHSRA’s job to sort out Caltrain’s – or any other legacy operator’s – funding issues. Nevertheless, CHSRA very much is supposed to secure full funding for HSR while also preserving “HSR feeder” systems that its own ridership forecast depends on. In passing AB3034 via prop 1A, voters explicitly reserved some 10% of available capital improvement funds for such feeder systems, with heavy rail a priority.

CHSRA has a much better chance of securing additional federal and other non-state funding if it can demonstrate that it is actively wringing cost out of its plans wherever possible. When you ask taxpayers and private investors to cough up $35 billion, there should be ZERO sacred cows.

Here is a map to help make sense of the details discussed below:

View Caltrain Firebird in a larger map

Main Line plus Long Sidings
Firebird would replace Caltrain’s existing current baby bullet, limited and local services and, the associated timetable. As a valuable fringe benefit, it would also enable HSR service in the SF peninsula without having to construct any new tracks dedicated to that.

The brand name refers to a famous ballet about a fabled beast that is both a blessing and a bringer of doom to its owner. The former would apply to funding and to rail customers, the latter might ring true for those who own property near the line – especially in the sections that will require new tracks for Caltrain.

The concept relies on a dual track main line in the SF peninsula, with siding tracks for serving selected groups of stations. The main line would be shared with HSR via the “rule of special applicability” that FRA will anyhow need to write for CHSRA. It’s common for national regulators to draft new rules in order to enable the safe operation of technologies that are new to their jurisdiction. Indeed, FRA had already begun just such an effort for Florida HSR before voters shelved the project. It’s now live again, so FRA will presumably pick up where it left off.

There would be no “Caltrain tracks” nor “HSR tracks” north of way point Lick near the Monterey Highway in south San Jose, the southern end point of PCJPB property. There would just be a main line and a number of sidings. This implies interleaved operations based on an integrated timetable and integrated traffic management, a choreography of steel and electricity.

In the SF peninsula, the main line would have a nominal speed of 90mph where possible – up from 79mph thanks to upgraded signaling and other changes. The number is not arbitrary, as the power required to overcome aerodynamic drag is proportional to the cube of velocity. Train designs for top speeds of 125mph necessarily feature longer transmission gear ratios and hence, lower acceleration performance at low speeds.

At selected locations, Caltrain would depart onto a siding at line speed, serve two to four stations and return to the main line at the speed target for that location, at a specific time. The primary advantage of this approach is that only a fraction of the corridor needs to be quad tracked, saving a lot of cost and reducing environmental impacts related to right of way width. A lower total number of miles of track also means lower infrastructure maintenance costs and, those would be shared with the HSR operator. Exploiting a main line more aggressively is the best way to cut operating overheads.

There would be a single, proven signaling system, featuring integrated traffic management and positive train control, for the whole corridor. The technology selected should be capable of headways as short as 2 minutes, something that ETCS level 2 has made possible at a line speed of 125mph in a real-world implementation in Switzerland. Other systems might also meet this spec. CHSRA must anyhow make this investment, so there is no need for Caltrain to develop its homegrown CBOSS technology any further. IMHO, absent grade crossings, that effort boils down to consultants lining their pockets and corporate welfare for UPRR.

In practice, a two-minute headway would give timetable planners capacity for up to 30 trains per hour each way. However, any emergency situation would have to be reliably detected and acted on within around 32 seconds whenever an express train is just two minutes behind a regular one, since trains are heavy and cannot stop on a dime. That’s not as hard as it sounds, thanks to signaling and positive train control. If both trains are traveling at line speed at the time of the incident, more time is available to avert disaster.

A useful analogy is a virtual conveyor belt moving at line speed, with slots at defined intervals (the analogy does not refer to propulsion, just to timing). Any train that wants to use this belt should travel at its speed and in one of the slots. It should also maintain a distance equal to an integer multiple of the headway times the line speed relative to the front of the preceding train, regardless of whether that’s a Caltrain or an HSR train.

Taktverkehr
The core operations concept would be what Germans call Taktverkehr, roughly “traffic to a beat” or “metronome traffic”. If a train departs onto a siding, the time it must spend there is equal to the length of that siding divided by the line speed plus an integer multiple of the headway and the line speed. In plain English: to permit one train behind it to overtake, a train on a siding must “lose” 2*2=4 minutes, relative to time it would have taken to cover the same distance on the main line at 90mph. For two consecutive overtakes, it’s 6, for three it’s 8 etc. Losing just 2 minutes is not enough for even a single overtake. However, a train can stop at a station on the main line and accelerate away in time to jump back on the conveyor after slipping exactly one slot if that slot is not already occupied.

Imagine you have Caltrain A running in slot #1, an HSR (or Caltrain express) train B in slot #2 and second Caltrain C in slot #3 on the main line. Train A departs onto a siding, freeing up slot #1. Train B overtakes it while A is moving on its siding and serving stations. Train C departs onto the same siding, freeing up slot #3. At the exact moment train A reaches the end of the siding, traveling at main line speed, it recycles slot #3. Train C will later recycle slot #5 etc. In this manner, you can interleave express and certain non-express service patterns without sacrificing any line throughput at all (once you reach a steady state at the end point of the conveyor belt).

However, an express HSR train running at 125mph would need several empty slots ahead of it so it can gradually “gain” on the slower conveyor belt. Alternatively, slower trains in front of it would have to move off to a siding in a timely fashion so it can overtake them without slowing down. In the above example, Train B could speed up as soon as slot #1 becomes available.

Every slot timetable planners have to leave unused to accommodate speed mismatches translates to lost line capacity. In the particular case of the SF peninsula that’s acceptable, but only up to a point. During Caltrain’s rush hour, there might not be enough empty slots to permit many – or any – express HSR trains to run. If so, the HSR operator would either run non-stop at 90mph or else, run at 125mph but stop at HSR stations on the peninsula. Note that AB3034 specifies a non-stop line haul time of 2h40m for SF Transbay Terminal to LA, but it does not require that this performance be feasible at all times of day.

Station Selection
Given that sidings imply expensive quad tracking and high-speed turnouts, each candidate should serve stations that can be expected to contribute significant ridership. Once a year, Caltrain measures average weekday and weekend boardings for each station. Out of the 36,778 total for 2010, SF 4th & King contributed 8673, Palo Alto 3905, Mountain View 3264, San Jose 2698, Millbrae/SFO 2485 etc. There are plenty of stations that contribute low thousands to high hundreds each.

Past performance is not a perfect indicator of future performance, but there are 11 stations that each currently contribute (much) less than 1% of total ridership: Bayshore, South SF, San Bruno, Hayward Park, Belmont, College Park, Capitol, Blossom Hill, Morgan Hill, San Martin and Gilroy. The last 5 are served by just three trains a day (each way) via trackage rights on UPRR’s central coast line, not because there’s a hard capacity constraint but because there simply aren’t enough potential customers.

My proposal is to suspend or cancel Caltrain service at all of these 11 stations, which together comprise just 5% of total ridership on a service that is dwarfed by BART and regional rail providers elsewhere in the nation and world. If you’re going to cut costs by sharing core infrastructure with a service that depends on not stopping frequently, you have to make some compromises. Note that Caltrain has closed stations due to lack of ridership in the past, just not so many all at once. Gilroy is slated to get an HSR station and, Amtrak Capitol Corridor could theoretically be extended to either Hollister or Salinas.

Local transit advocates may howl at the prospect of anyone anywhere losing Caltrain service, but IMHO that would be letting the perfect be the enemy of the very good. There is only a limited amount of operations subsidy funding available, so the focus ought to be on serving the vast majority of customers better.

San Bruno Station and Curve
Note the special case of San Bruno, where Caltrain intends to spend a whopping $300 million to move the station in the expectation that this will boost ridership. The railway expended a lot of time and effort on coming up with a design that works for the local community. This is now shovel-ready and on the list of projects due to be implemented with ARRA and prop 1A funds. Unfortunately, the design predates prop 1A’s passage and preserves a sharp curve that really ought to be straightened so HSR trains don’t need to slow down before they pass it. Worse, current ridership in San Bruno is so low (370 per average weekday) that the station would be marginal even if it that number were to double. In addition, quad tracking is complicated by the presence of BART tracks stacked under the Caltrain rails.

Therefore, I recommend suspending the entire shovel-ready San Bruno project until the PRP figures out an operations plan for the whole peninsula that will make Caltrain financially viable while keeping total capital investment in the corridor within reason. The good burghers of San Bruno will not like this one bit, but just maybe the way to do this right is to discontinue service there altogether.

Rolling Stock Specifications
In addition to ridership, siding candidates should deliver reasonable dwell times at stations with aggressive but technically feasible acceleration specifications for Caltrain’s new EMU rolling stock. The average value for 0 to 90mph should be on the order of 0.6-0.7 m/s^2 to keep sidings short, line haul times low and, to provide sufficient reserves to make the timetable robust.

Longer single-level EMUs perform substantially better in this regard than shorter bi-level consists with comparable crash safety. In addition, they would facilitate harmonizing platform height to the high level that CHSRA prefers. Let’s not forget which agency will be funding the lion’s share of the corridor improvements.

For the purposes of my analysis, I picked off-the-shelf Stadler FLIRT trains as representative of the type of technology that will be required. This is not meant to be an endorsement of that particular product.

Right of Way Constraints
A related objective is avoiding quad tracking in sections where that would cost a minor fortune. High acceleration performance lets trains reach main line speed in a short distance and in some cases, avoid expensive construction measures to overcome severe lateral space constraints due to existing structures. The following list may not even be exhaustive:

• In SF, the elevated end section of I-280 was built around the existing dual-track rail line. Its supports block the path for additional tracks near Napoleon St south of tunnel #2 and also between 22nd St and Mission Creek. This is on top of the fact that the hilly terrain already requires four short tunnels in the area.
• In South San Francisco, the Caltrain tracks run through a UPRR marshalling yard.
• Between San Bruno and Millbrae, BART has consumed so much of the originally available right of way that implementing quad tracking at grade will be difficult. If HSR trains are to dwell for an extended period of time to accommodate airline passengers with a lot of baggage or else, there is a desire to offer integrated baggage handling to maximize the relief HSR can bring to SFO (that was the whole point), it may even be necessary to construct an ugly tall aerial leading to platforms above the mezzanine level.
• In San Mateo and Menlo Park, the available right of way is too narrow for quad tracking.
• In Palo Alto, the layout of the Univ. Ave/Alma St grade separation intrudes on the railroad right of way.
• At San Antonio Rd, ramps to Central Expressway and the Caltrain platforms mean the available ROW is too narrow for quad tracking.
• In Mountain View, overpass supports, a VTA light rail track and busy frontage roads mean the right of way can only accommodate the already existing two tracks at grade between the 85 and 237 freeways. Unsurprisingly, the city would prefer that CHSRA to implement grade separation in a trench, but Stevens Creek near 85 presents a hydrological hazard well below grade and, grade level actually rises between Castro and that location (see Clem’s Focus on Mountain View). A viable solution must address both the vertical and the lateral constraints.
• In Santa Clara, the right of way at grade is constrained by the junction with UPRR’s Alviso line and a marshaling yard.
• In San Jose’s Gardner district, much like in tony Atherton further north, there has been strong opposition to quad tracking from residents near the existing line, prompting CHSRA to select the severely curved 280/87 alignment alternative, complete with iconic bridge and other cost-maximizing measures. Reversing that decision would only be conceivable with a new, less intrusive proposal.

On the plus side, Caltrain has already quad tracked a few sections of its network. Where justified by ridership considerations, those should be leveraged. Service at Bayshore should nevertheless be suspended due to low ridership. If the Brisbane Baylands development actually happens, that could easily be revisited. However, that would cost CHSRA its prime candidate for a stabling and light maintenance yard at the north end of the starter line, even though the soil there was contaminated by decades of Southern Pacific’s operations.

Siding Length and Station Count
The analysis I did yielded a mildly surprising result: the optimal siding length is 3-4 miles, with not one but two consecutive stations served for the time penalty associated with letting a single train overtake, i.e. 4 minutes with moderate acceleration requirements for the sections between the stations. This was for a target dwell time of 30 seconds, which is reasonable for level boarding of long single-level trains. Increasing the headway to 2.5-3 minutes would substantially increase dwell times, reduce speeds between stations and/or favor fewer but longer sidings containing 3-4 stations each. The latter would work out between Millbrae and Calif. Ave but less well south of Palo Alto. Additional scenarios should be analyzed to explore the trade-offs involved.

Proposed Sidings
Here are the siding sections that I propose:

1. Millbrae-Burlingame
2. San Mateo-Hillsdale
3. San Carlos-Redwood City
4. Menlo Park-Palo Alto-California Ave
5. San Antonio-Mountain View
6. Sunnyvale-Lawrence Expy

• The acceleration and deceleration sections of sidings 1 and 2 approach each other to within a couple of hundred feet. Therefore, it may make sense to also connect the the two such that timetable planners have the option of keeping selected Caltrains off the main line for a three overtake scenario in support of express HSR operations. A double overtake is theoretically feasible even if all four stations were served, but it would sharply reduce the amount of timetable padding. If planners need a double overtake at a given point in the timetable, they should have that train serve any three stations and run through the fourth.
• Siding 4 was the only case in which it made sense to group three stations rather than two. Timetable planners could decide to have a given train serve any two and achieve a single overtake scenario or else, serve all three and achieve a double overtake. Note that not all impacts can be avoided: the chicane at Univ Ave/Alma St in Palo Alto will need to go and the right of way widened a little bit in Menlo Park. However, those measures are justified by proven ridership, not supposed FRA red tape nor bureaucratic megalomania.
• The Mtn View station on siding 5 would have to move a couple of blocks toward Shoreline Blvd to give southbound trains enough distance to accelerate to a useful line speed of perhaps 80mph at the 85 underpass. This would anyhow make sense in the context of the more extensive remodeling of roads in the downtown area that would be needed if CHSRA decided to site the mid-peninsula HSR station there. San Antonio Rd will also present problems, but at least that is not a major freeway and therefore easier to reconfigure.
• Siding 6 leverages a section of existing quad track south of Fair Oaks Ave

Siding Implementation
In principle, siding tracks can run either outside or inside the main line. In the former case, all of the stations on that siding must have side platforms, in the latter all must have a single central island platform to keep slow traffic on the left. In addition, each individual siding requires four high-speed turnouts rated at 90mph for the curved section. If sidings 1 and 2 are connected for additional flexibility via an additional set of four switches, they must feature the same track order.

The main line should always be on the straight (or less curved) section of each switch, since as express trains need to run through at 125mph. Suitable designs do exist, but they are very long and require multiple actuators, see image below. The Firebird concept is predicated on switches that can be actuated very frequently and reliably without excessive maintenance overheads.

SF Stations and Approaches
Bypass tracks would be available at 4th & Townsend station en route to the Transbay Terminal. A high-performance design requires expensive tunneling/trenching all the way to 22nd Street, which would also separate the remaining grade crossings at 16th and Common(Berry) St. The latter is currently closed but could prove useful as Mission Bay is developed.

Crossing the Mission Creek outfall at grade, as TJPA currently assumes, would be cheaper and avoid a hydrological hazard, but the descent section toward the DTX tunnel would put the below-grade 4th & Townsend station well into that structure with just a single bypass track. This constrains the throughput of the DTX tunnel and forces planners to use 2nd instead of 3rd St for the alignment.

Instead, I propose letting the tracks cross the outfall below the culverts and excavating all of 4th & King to create a station with multiple long platform tracks plus a run-through stabling/light maintenance yard – below grade (h/t to Richard Mlynarik). This would be an expensive but inevitable consequence of putting the approach tracks below grade. My map shows four full-length active platform tracks, an optimized system design might feature a different number.

A run-through yard would be especially important if CHSRA fails to secure the Bayshore property for its stabling yard, though it is unclear how many additional trainsets Caltrain could accommodate in addition to half of its own EMU fleet. At least the HSR operator could clean the trains at 4th & King instead of at the TBT, eliminating CHSRA’s concerns about feasible dwell times at the downtown station. Note that thanks to electrification, part or all of grade level at 4th & King would become available for development or perhaps, for a new city park.

The expense of all that excavation would be partially compensated by reducing the number of tracks inside the DTX tunnel from 3 to 2, which means it can be dug using fast and cheap TBM machines. In addition, moving the run-through station out of the tunnel would mean TJPA could leverage 3rd St rather than 2nd, though that would put the DTX within earshot of the MOMA building. That in turn would permit a one-way loop track alternative to create a train box at the TBT with six dead-straight run-through tracks that would preserve main line capacity. However, those tracks would end up a little further west than is currently planned – a consideration that TJPA would have to deal with quickly because it impacts the placement of support columns for the above-ground portion of the TBT.

Note that 4th & Townsend would nominally be a Caltrain station paid for by CHSRA under the heading of ROW acquisition, subject to an easement that would let the future HSR operator leverage the platforms for a nominal fee. This is legalese sophistry but it just might get CHSRA out from under the 24 station limit in AB3034. The same – admittedly tortured – logic could perhaps apply to the mid-peninsula station.

San Jose Diridon and Approaches
In the Firebird concept, SJ Diridon would become a single-level station at grade with two main line plus four full-length platform tracks accessed via two island platforms. These would be shared by Caltrain, HSR, Amtrak and ACE, assuming platform height are compatible. Worst case, room for an additional side platform for these legacy services will have to be found. This doesn’t just translate to a cheaper station, it eliminates the need for aerials in the approaches on either side. However, it does imply that all operators, including UPRR, Amtrak and ACE, will have to install compatible PTC equipment in their locomotives and submit to corridor traffic control. This would require appropriate clauses in FRA’s aforementioned “rule of special applicability”, consistent with those of the mixed traffic waiver Caltrain has already been granted.

For reference, the in-cab equipment for ETCS level 2 costs less than $0.5 million per unit. CHSRA could easily spring for some number of those so the legacy operators don’t have to. Contrast that with the cost and disruption of needlessly pouring hundreds of millions of dollars’ worth of concrete.

There simply is no operational need for an expensive bi-level design, this is a run-through station with short dwell times even for HSR. The inside platform tracks should be laid out such that Caltrain can reverse direction without crossing the main line. The only other proviso is that an additional yard for mid-day stabling of Caltrain and ACE trains must be secured. I’d recommend looking for that either north of the station (river/Ryland/87/UPRR Milpitas Line, partially above the river) or else at the southern end point of the PCJPB right of way. Please consult the map above for details. The land I have in mind currently belongs to a business called Concrete Ready Mix and is apparently used to store cement, which may or may not be shipped in by UPRR. The property might have to be taken by eminent domain and an existing access road off Hillsdale Ave paved over.

The remaining secondary grade crossings at Auzerais Ave and W Virginia St would be closed, perhaps with ped/bike underpasses. The right of way in the Gardner district would not be widened and no additional grade separation works constructed. There would “only” be additional rail traffic. The curves already there impose a speed limit of around 55mph for acceptble passenger comfort , yet even that may only be feasible with gauntlet tracks to address the superelevation mismatch with UPRR trains. Transparent sound walls plus other noise-reducing measures such as triple glazing should be offered for properties abutting the right of way.

Keep Caltrain/HSR through San Jose as straightforward as possible please, the BART extension to Santa Clara is going to be nose-bleed expensive enough. The city needs to focus on delivering overall public transportation functionality.

Other Retained Stations
SF 22nd St, Santa Clara and Tamien would all be retained as isolated stations on the main line without bypass tracks, which means serving them would reduce line capacity. If and when timetable planners can no longer afford to waste slots on these stations, they will need to be closed or else bypass tracks constructed at that time. In the latter case, dwell times would be on the order of 2.5 minutes for a single overtake scenario.

Impact on Freight Service
Given the 90mph line speed and the need to keep track geometry maintained to a level that is safe even at 125mph, there will be a need for greater superelevation in curves and, for banning super-heavy FRA-compliant locomotives. Perhaps UPRR and its customers could live with restrictions such as night-only operations, an axle load limit of 22.5 metric tonnes and vertical transition gradient of up to, say, 2%. Installing and maintaining a gauntlet track all the way to Santa Clara just for UPRR would not make economic sense and also greatly complicate the design, construction and operations of the aforementioned high-speed turnouts on the main line. It matters little that freight spurs could still be served using the regular kind.

If rail freight business models no longer make sense if restrictions are imposed, PCJPB has the legal right to unilaterally cancel freight service by invoking clause 8.3(c) of its 1991 contract with Southern Pacific. That would, however, trigger federal abandonment proceedings, i.e. compensation claims. Whether there would be the political will to take such action is another matter entirely. Businesses that currently depend on rail freight might well have to relocate or liquidate – at the expense of local jobs – or else, switch to trucking on congested peninsula freeways. CHSRA is still hoping to change UPRR’s mind on ROW acquisition in the Central and Antelope valleys, if not south San Jose to Gilroy, so it’s being cautious. PCJPB might not care as much if it anyhow decided to terminate service south of Tamien.

In addition, planners would also have to weigh the cost of financial compensation against the overall cost savings associated with remodeling the corridor for passenger trains only – free of constraints related to high axle loads, limited transition gradients and access to freight spurs. Also, the small but useful So SF yard would become available for stabling or light maintenance.

Conclusions
Except for the arguably justified extra cost of re-framing the DTX tunnel complex such that it effectively extends out to 22nd St, a lot of cost can be saved by redefining Caltrain service such that it can continue to serve 95% of its current ridership well into the future. There would be some operations constraints on the future HSR operator, including an integrated timetable and limited scope for express trains at certain times of weekdays, but those are a small price to pay at this point considering the HSR starter line is nowhere near fully funded.

Timetable planners would gain a great deal of flexibility in defining 2-3 Caltrain service patterns – albeit based on groups of stations rather than individual ones – while preserving a sufficient number of slots for HSR service. Analysis of current Caltrain ridership between station pairs should guide optimization. Overall, the focus should be on delivering 90+% of the functionality at perhaps 60% of the cost.

For reference, a Firebird could serve the TBT plus 7-8 stations in the peninsula in the same line haul time today’s baby bullets need for 4th & King plus 4-5 stops (albeit different ones). A 16-stop “local” would make the SFTBT-SJ trip in about 70 minutes vs. 90 to 4th & King today. My results are within a few minutes of those returned by the spreadsheet Clem Tillier provided in his post The Tao of Timetables, after plugging in the relevant parameters. However, he did not explicitly take HSR traffic into account.

Recommendations

• The Firebird concept would reduce the number of miles of new tracks between SF and way point Lick from over 50 to a more manageable 20 or so, while improving service to 95% of Caltrain customers.

  Against that, the concept implies a large number of high-speed turnouts that must be switched frequently and reliably and, the closure of 11 tertiary stations.

  While some difficult section would remain (e.g. Transbay Terminal to 22nd St in San Francisco, Millbrae, San Mateo, Palo Alto), others would be left alone (e.g. 85 to 237 in Mountain View) or “rightsized” (e.g. SJ Diridon station and approaches).

• Before appropriating any more funds, members of Congress, state lawmakers and peninsula officials should all require CHSRA and Caltrain to deliver an optimized integrated operations plan for the SF peninsula corridor, because that should be driving an integrated capital investment program. This includes important technical details such as electrification design, signaling and traffic management, EMU acceleration performance and platform height harmonization.

  That plan will need to be more detailed and may well be substantially different from the concept described above, which I put together in a matter of days. My point is there are no concrete plans for actually sharing any tracks other than in downtown SF at all right now, because the PRP planners and their consultants are not focused like a laser beam on wringing cost out of the corridor upgrade project. Someone should force them to increase value per taxpayer dollar and ease the environmental impacts.

• The state of California should ask FRA to urgently station a senior regulator on site in Sacramento, with prop 1A funds used to reimburse it for the additional headcount. This person’s office should be in close proximity to that of his/her counterpart at the California Public Utilities Commission (CPUC), which may also need funds for an additional headcount. Two headcounts is peanuts, absolutely peanuts, compared to potential cost savings related to cutting through red tape as early as possible.
• Do not actually break ground on the San Bruno project until there is clarity regarding the viability of service there in the context of a Caltrain operations plan for the entire corridor that the peninsula counties are willing to subsidize at the level required. I realize there’s an urgent need for construction jobs and a deadline on the ARRA funds, but that would be a poor excuse for constructing a white elephant.
• Revisit the design of the 4th & Townsend station, 4th & King yard and the DTX tunnel to eliminate capacity bottlenecks, even if doing so actually increases cost in this section. The long-term success of both Caltrain and HSR operations depends on high throughput at the Transbay Terminal. Planning mistakes made now could come back to haunt both operators and the city decades from now.
• Give the acquisition of stabling yards for Caltrain/HSR in SF and Caltrain/ACE in San Jose much higher priority.
• PCJPB should require Caltrain engineering to leverage investments CHSRA will anyhow have to make, especially regarding big-ticket subsystems like electrification and signaling. If need be, CHSRA and PCJPB should transact certain planning prerogatives under the heading of ROW preservation, at a price that will plug the hole in Caltrain’s operating budget for e.g. the next three years. For starters, PCJPB could hand over decision control over the remaining $37 million in prop 1A funds for capital investments in Caltrain, but the counties would still have to reprogram more of their future MTC and other capital investment dollars to the corridor upgrade project. On the plus side, CHSRA could claim those as non-state matching funds in the context of the upcoming request for prop 1A bond appropriations.
• Strongly consider Mountain View for the mid-peninsula HSR station, since quad tracking between 85 and 237 would be far more expensive than moving the station in the context of a larger remodeling project. Of the three candidates, Mtn View may have the highest ridership potential due to the proximity of 3 major freeways plus Central Expressway, in addition to VTA light rail, buses and bicycle infrastructure.

  A future post will examine the nitty-gritty of the Mountain View-San Antonio section, including especially a concept for managing traffic and connecting transit consequences while upgrading the downtown business district with its popular eateries.