Shutdown versus slowdown: Why the L train slowdown may be more disruptive than we think
Many applauded the announcement that the L train would remain open during repair work on the Canarsie tunnel. Instead of shutting down completely for 15 months (or more) service will be slowed down on evenings and weekends mitigating what MTA leadership has described as an inconvenience to over 250,000 daily riders. However, as soon as the new plan was announced advocates openly wondered why the MTA wouldn’t take advantage of the construction innovations to close down the line for a much shorter period of time. In this post Greg Marsden draws on lessons from infrastructure projects in the UK to reflect on the trade-offs inherent in opting for a slowdown versus a full shutdown.
If you’re planning to have major renovations done to your house, you might ask the question – will this be less painful if I move out? I’ll have the added costs of finding somewhere else to be for a couple of months but the builders can get on with the work without having to accommodate me and I don’t have to live in a house which doesn’t really work properly and each day, when I get home, is a bit different. Which room is the sink working in? Is the oven connected or not?
The answer to this question is not easy. It is essentially a set of trade-offs over the type, the amount and the duration of the inconvenience. In many ways it is the same set of decisions to be faced in thinking about the closure or partial closure of the L train line in New York City.
It is uncontroversial to say that the delivery of the upgrade will be much easier and less expensive if it is done in one big bang approach, with extensive periods of total closure. Working on infrastructure around a live railway is dangerous and so there are large periods of time lost in handing over the line from being in use to the construction company. In the UK for example, an 8 hour rail closure has typically had only 3.5 hours of productive working time (see Figure). This results from the deployment of safety procedures to hand the line over to workers and back again to passenger services alongside a buffer to avoid overrunning. Whilst Swiss Railways achieve much higher rates of productive time, the operating environment underground is more complex. NYCT president Andy Byford recently stated that the agency estimated that they could expect an average of 5 hours of “wrench time” per night but often squeezing work into small time windows is not productive. Effective spending of taxpayer money would likely result in a total closure.
However, what do we know of how travellers respond to total closures versus partial closures? We need to understand this to determine whether the social costs of a full closure for a shorter time exceed the drip, drip pain of partial closures. These are some of the questions we are seeking to answer through our proposed study of the L train. Unfortunately there is not a lot of evidence (see Shires et al., 2018). We can however infer some of the answers to this question from looking at evidence from elsewhere:
- In Edinburgh, Scotland, the Forth Road Bridge had to be closed to all but emergency traffic during December 2015 due to a serious structural fault. The new parallel bridge was not yet complete and so the road diversion was substantial. However, the rail crossing was still functioning. The closure lasted for around 4 weeks. Our research found that there was a 12% reduction in the number of days traveled to work, partly from home working and partly from more intensive shift working. The largest reduction was in people travelling to work five days a week which decreased from 63% to 51% of commuters. Some activities were able to switch destinations (e.g. some retail and leisure) but social trips were significantly reduced, in part because of longer journey times for those that drove and more congested rail services. However, in this instance, there were few alternatives for those without an origin and destination close to the train line.
- The London Olympics provided a different type of disruption but with a network similar to New York where there were many alternatives. Transport for London knew that each day would create different hotspots on their network dependent on which events were running and ran an intensive behaviour change campaign with businesses and individuals. Our evaluation found that there was a great deal of change, even amongst those that did not plan beforehand. More changes were made by those people who had actively planned in advance. The most common changes were reducing (31%) and retiming of journeys (25%), followed by rerouting (16%) for commuting.
So, there is potential for adaptation amongst the population on a daily basis. However, the event lasted four weeks and was one of the highest profile hosted in London. Similar techniques were tried with a major station closure following the Olympics and the behavioral adaptations were disappointingly small.
So, what will work best for consumers? This might depend on how messy and convoluted the partial closure is. People like stability and routine and having to readapt in different ways on a regular basis creates additional uncertainty. Not everyone who turns up at a station is prepared for what is happening irrespective of the amount of information provided. Set against this, a full closure will impact on those journeys with more limited alternatives and so it is important to provide a stable alternative. There is also the question as to whether diversionary routes and modes have capacity to accept those who re-route or re-mode. However, the evidence suggests that there will be re-timing, relocation and squeezing of activities so the additional demand burden will be unlikely to be as bad as might be imagined.
All of this suggests that, on balance, limiting service during repairs to the L line will have longer but smaller effects than a complete shutdown. That said, it will almost certainly raise the risk of major and unexpected service disruptions that will be the inevitable consequence of managing such a complex project with such narrow margins for error. This is even more likely because reopening the line every morning will be contingent on effective silica dust remediation. Recent construction-related hiccups on the line have demonstrated the disruptive consequences that relatively small issues can cause.
The closure of the L line is but one of many major infrastructure upgrades which will be necessary across the US in the coming decades. It is incumbent upon us to understand how to communicate whichever option is chosen to the population and to monitor and adjust the adaptation plans given that there is a relatively weak science of understanding how people respond to such circumstances.
Returning to the house analogy set out at the start, the main prize is a much better living (in this case travelling) experience: a shutdown with robust alternative routes weighed against a slowdown that increases uncertainty. You soon forget all the dust and inconvenience once the job is done. That’s the prize. The question is how quickly do we want the prize to be reached?