Anybody who’s been on Caltrain at rush hour lately knows that there is a capacity crunch.  Many riders are standing, and sitting on stairs, in cars not designed for standing, with average trips 20-30 miles long. Ridership has been growing rapidly, doubling over the last decade.  To address the capacity crunch, Caltrain is in the process of purchasing extra cars from Metrolink (potentially adding a 6th car to a 5-car train).  The electrification project, scheduled to be implemented in 2019, will add extra capacity by allowing more frequent service (as well as growing ridership with the ability to stop at more stations).

If ridership continues to grow, how long will these measures alleviate the crunch? What else can Caltrain do to accommodate more riders?  We did a bit of lightweight modeling using some information from Caltrain, and some logical inferences.  The model shows that investments in capacity improvements building on basic electrification is highly likely to be needed to keep up with rider demand and changing transportation preferences, to address economic growth and regional initiatives for low-carbon transportation.

Ideas are welcome about how to improve the model and how to communicate the importance of investing in Caltrain capacity and productivity increases.

Peak hour capacity

The constraints to Caltrain capacity is at the peak times.  Also, Caltrain frequency is described in terms of trains per direction hour. So this model focuses on peak hour capacity. How many riders can Caltrain carry during peak hour? How many is it carrying now? What are potential capacity improvements, and how much would they help?

Today, there are about 650 seats in a 5 car train set (consist).  For reference, the most crowded train, the 329 NB out of San Jose, has over 850 passengers at peak season.  Ouch.

Better accommodation for some people standing. The average trip is 20-30 miles long, and many travel longer than that, so it does make sense for Caltrain’s train design to accommodate a lot of people sitting, but there are still some people with shorter trips who’d be willing to stand.  BART survey data says that about 25% of BART riders stand, so let’s model that. If Caltrain reconfigures cars for a some more comfortable standing room that would be ~845 passengers per car.

Longer trains – pre-electrification. So when Caltrain adds the Metrolink cars, perhaps adjusted for more standees, that gets up to ~5,000 passengers per hour, with a 5 train per direction per hour peak schedule.

More frequent service.  Electrification will allow Caltrain to run up to 6 trains per direction per hour.  Assume the new trains have the same capacity as the old ones, that is about 6,000 passengers per peak hour.

Extended platforms. If Caltrain made its its platforms longer after electrification, that would enable it to have longer trains. With 8 car trains, that would provide over 7,000 passengers per peak hour.

Speed and Productivity improvements

Replacing remaining diesel trains.  Caltrain plans to replace only 75% of its diesel trains travelling between San Francisco and San Jose, for budget reasons. With additional funding, Caltrain could replace those diesels, speeding travel time to attract riders, reducing operating cost.  Replacing the diesels will also improving the health and quality of life for people living near the train, especially near 4th and King and the San Jose storage yard.

Level boarding.  As Caltrain extends the platforms, upgrading the platforms to enable level boarding will make service faster and more reliable.  Where BART stops for 30-40 seconds at each station, Caltrain stops for 1 to 1.5 minutes.  A passenger needing mobility assistance can take up to two minutes on each end of the trip. This makes it very difficult for Caltrain to make timed transfers with BART, VTA, and other systems.  Level boarding will speed trips, improve transfers, and increase accessibility for the elderly and disabled.  Based on experience in other transit agencies, level boarding would increase ridership to fill the extra cars.

Even more frequent service. Many people believe that Caltrain service is capped at 6 trains per direction per hour because of the agreement with High Speed Rail.  That’s actually not quite the case.  According to Caltrain’s studies, the corridor can handle up to 8 trains per direction per hour without passing tracks, and up to 10 trains per direction per hour with additional passing tracks.  HSR’s most recent business plan shows a substantial amount of ridership on intra-Peninsula commute service. So, could Caltrain start running 7 or 8 trains per hour before HSR starts? Caltrain and HSR have confirmed that the answer is yes, and the agreement does not need to be changed to do this. With 8 trains per hour, 8 car trains, that’s over 9,000 passengers at the peak hour.

Caltrain’s studies indicate that the corridor could handle up to 10 trains per direction per hour, with the addition of passing tracks. That’s sufficiently far off in time and available budget that it’s not included in this model.

So, how much capacity will Caltrain need? When is Caltrain likely to outgrow some of these improvements?

Peak hour ridership – current and future

To estimate Caltrain’s peak hour ridership, we started with data that Caltrain published regarding its highest ridership trains.  The most crowded direction is AM Northbound.  Here is what Caltrain reports about the most crowded trains, based on the February counts, and extrapolated by Caltrain to the highest ridership summer season, based on the typical summer seasonal peak.

Most crowded trains	Departs SJ	2013
329	8:03:00	854
323	7:45:00	793
319	7:03:00	783
217	6:57:00	716

These trains are in two different hours. So we guesstimate the ridership in the peak hour, based on available information. We know that the 3 missing trains between 7 and 8am all have fewer than 716 riders each, because the 217 was the the #5 most crowded Northbound train.  (Caltrain has the data for the missing trains – we’ll fix this model if and when we get that data). By this estimate, Caltrain is carrying over 3000 riders in the peak hour peak direction.

Peak hour AM northbound	Source	Departs SJ	2013
329	Caltrain data	8:03:00	854
227	Estimated	7:55:00	500
225	Estimated	7:50:00	525
323	Caltrain data	7:45:00	793
221	Estimated	7:18:00	690
			3362

 

So, how much will the Metrolink trains help, and how much will various other improvements help, if ridership continues to grow? This model assumes a linear growth in ridership, continuing the pattern of doubling over the last decade.  In this scenario, the Metrolink trains will help (fitting over 5000 riders), but it will be quite crowded again by 2019.

Basic electrification will support up to 6000 riders, and will be immediately crowded.  Longer platforms will support over 7000 riders in the peak hour, and will alleviate crowding through the early 2020s, according to this growth scenario.

Beyond that, the increased frequency provided by running up to 8 trains per hour will help substantially.   This is important, since the downtown extension of the Caltrain tracks to Transbay terminal, and the BART extension to downtown San Jose are expected to be implemented in the 2020s, finally connecting rail backbone service around the bay.  These connections will drive more ridership that will benefit from more frequent peak hour trains.

However, Caltrain’s studies of the blended system suggest that running 8 trains per hour in the 2020s will put a lot of pressure on high-traffic intersections. Grade separations seem likely to be needed to allow Caltrain to use that capacity.   It would be welcome to see the cities on the corridor work together – San Francisco, Peninsula cities, and San Jose – to implement the improvements to carry the riders when there is a big step up in ridership.

 

Other improvements and modeling details

There are other nuances that may affect capacity.   Incentives or congestion pricing could encourage more riders to use trains outside of the height of the peak (for example, Bay Bridge congestion pricing shifted 10-15% of riders from the Bay Bridge to BART).

This model does not look at seat turnover.  Most riders don’t ride all the way between San Jose and San Francisco.  Often, the same seat holds a different rider between San Jose and Palo Alto, and then from Palo Alto to San Francisco (for example).  The number of riders per peak hour will be greater if turnover is modeled.

Bikes on Board take up room – but also provide an extremely effective and cost-effective means of last mile connectivity, in a region where 80% of jobs are within 2-3 miles of the train, and many are further than walking distance. How effective will a bikeshare program be at helping people use a bike without bringing it on board?

What the future may hold

The ridership growth is being driven by a strong Silicon Valley economy, plus underlying changes in people choosing to drive less, companies preferring to locate in more walkable locations, and transit oriented development on the corridor.  The underlying shift in preferences toward less driving, and locating in more walkable places are generational and based on a lot of evidence likely to continue.

Transportation and land use improvements are likely to continue to drive ridership – the Central Subway is expected to be complete in 2019, connecting more riders to Powell Street BART and central SF. The Downtown Extension of the Caltrain tracks to Transbay terminal will connect to many more jobs and homes in SF and the East Bay.  San Jose is completing its Diridon Station Area Plan, and BART will connect to downtown SJ, driving more riders to and from downtown SJ.  These changes sometime in the 2020s are likely to drive demand for the next stage of capacity improvements, with more frequent service, and grade separations to keep intersections functioning with more frequent trains.

It will important to invest in the capacity and productivity improvements that allow Caltrain to continue to support. the region’s economic growth, people’s preferences to drive less, and the region’s environmental goals.

More eyes on the modeling are welcome.