I'm in the copyedit phase of my book, which I've been sternly advised is my last chance to substantially change the text. So of course I'm stumbling on lots of odd little uncertainties, and I have a few questions for transit and transport experts out there.
Feel free to help me out on any of these! (Try not to say anything too immortal in the comments, as this post will be deleted when it's served its purpose.)
- Roughness. Based on usage I've heard from traffic engineers, I use this word to mean "delay in a traffic lane adjacent to a curb or parking lane caused by events such as delivery trucks and taxis stopping for customers, cars engaged in parallel parking movements, car doors being opened into your lane, slow cyclists sharing the traffic lane, and so on." But attempts to google a definition founder on the more common sense of "pavement roughness." Is roughness the right word for what I mean? What word would you use? I have my answer on this one: friction.
- Relationship of Ridership to Density. Rutherford and Spillar (1998) find that in the range of densities covering most North American urban areas, ridership's relationship to density is an upward curve. That is, if you control for service quality, if suburb A is twice as dense as suburb B it will generate much more than twice the ridership. Has anyone done or seen more recent research proving or disproving the same point?
- Deterrent Effect of Various Kinds of Delay. The Transit Capacity and Quality of Service Manual (2nd ed) presents these figures for how much different kinds of waiting time discourage ridership, compared to 1.0 for riding time. For example, it states that on average, a minute of walking time has the same deterrent effect as 2.2 minutes of riding time. These numbers are sourced, however, on a wide range of studies dating back to the 1960s. Has anyone seen anything more recent?
Delay Type: Walk Initial Wait Ride Wait for Connection Minimum 0.8 0.8 1.0 1.1 Average 2.2 2.1 1.0 2.5 Maximum 4.4 5.1 1.0 4.4
- Worker-Driver service. Do any transit agencies in the developed world run worker-driver service, where a commuter bus is driven by someone actually making the commute, who is hired as a part time employee by the transit agency? I'm aware of two outer-suburban agencies near Seattle that do this. Is it commonplace somewhere outside my awareness? UPDATE: Just to be clear, I'm not talking about vanpools, which are for particular groups of people by pre-arrangement. I'm talking about public transit vehicles running along routes, collecting fares, open to anyone.
- Driver shift start-end locations. It's universal, in my experience, that driver shifts must end where they began. The need to return drivers (on the clock) to their point of origin is a large part of the hidden cost of one-way commute express services. Are there common labor arrangements in the developed world in which a driver clocks out at a different place from where he clocked in?
A number of empirical studies have demonstrated the importance of
considering travel time variability when estimating the benefits to travellers of
time savings. There is evidence that unexpected delay should be valued
significantly different from average travel time valuations – values between 2.5 and 5 times in-vehicle travel time has been used. Small et al (1999) ‘showed that a
reduction of one minute of travel time under unexpectedly congested conditions was
valued at over 2.5 times the value of one minute of time under normal condition’.
Re: 4) Worker-Driver Service – NJ Transit tried this during the time I worked there and had some takers. Buses would be operated from a suburban garage to the Port Authority Bus Terminal in New York and left at a lot either in NYC near the terminal or in a large lot adjacent to the Lincoln Tunnel in NJ. The worker would pick up a bus for the trip back in the evening. Of course, this depended on the worker-driver having a reasonably regular work day for the evening trip to be driven.
NJ Transit has a huge number of peak period trippers into NYC; so the program was looked at as a way to cover those trippers at a low cost. I do not know if the program is still in effect.
4. I’ve known this done in the Northern Virginia suburbs of DC. Express buses from Prince William County to the Pentagon would be driven by part-time drivers whose main job was at the Pentagon. The buses would be parked at the Pentagon during the day.
I haven’t worked at the Pentagon for close to twenty years, now, so don’t know if the practice still occurs.
Re Roughness: Perhaps “turbulence” could describe the effect in traffic flow caused by transient obstacles. Admittedly, you’d be coining your own usage for the word but that’s part of a dynamic language.
For #1, we usually use the term “friction”.
Yes, I’ve always called #1 Friction, not roughness. I think that is the common North American phrase.
I have heard of 4 being used on Metra lines in Chicago.
Worker-Driver service sounds like a fantastic concept and Id like to read more about it.
Re: 2) grammar correction – you have two ‘if’s in the sentence.
It would read better as “That is, controlling for service quality, if suburb A is twice as dense as suburb B it will generate much more than twice the ridership.”
on density and transit try this (if you don’t know it already):
Paul Mees / DENSITY AND SUSTAINABLE TRANSPORT IN US,
CANADIAN AND AUSTRALIAN CITIES:ANOTHER LOOK AT THE DATA
#4 is basically the way vanpooling operates in the US, and transit agencies are the vanpool service providers in some areas (definitely in Washington state) and the vans bear the agency logos and such. In other areas, transit agencies or other organizations contract with a private vanpool service provider, such as VPSI.
Here’s Wisconsin’s vanpool program. Not sure it really counts for #4, because the driver isn’t really employed by a transit agency. He/she just gets limited use of the van and reduced rates.
#2, Does the study also control for number of routes (or total daily trips) serving the suburb? I would think that would matter more than “service quality”, unless controlling for this implicitly includes controlling for number of trips/routes.
On density and transit ridership it seems very important to point out how hotly contested this issue is, and how different methodologies produce starkly differing results.
In Transport for Suburbia (2010) Paul Mees has done a fine job of pointing out how weak the correlation often is, and how hard it is to study density in a way that produces consistent and meaningful results. He also point out that some claims about the correlation are based on errors that have since been corrected in later publications.
False certainty is a dangerous thing in transportation planning.
My take is that other factors such as service frequency and reliability are often far more important than density, and these other factors can be changed much more quickly than density.
#4 I remember from my days at VTA, there was a service that ran from Stockton to the high tech companies around the Lockheed Martin light rail station and that ran with a “volunteer” driver. Not sure if it is still being done now [that was back in 2001] nor whether the driver was compensated as an employee. One example of compensation might be that the driver doesn’t have to pay the fare if he/she offers to drive.
#5 You can relieve the driver anywhere but eventually the driver has to get back to his/her car or his/her car has to get to the operator. As a light rail operator for UTA, I was asked once if I wouldn’t mind driving an operator’s car down to Salt Lake Central. Usually relief operators would ride the train. So the operator got off at SLCen and went on his merry way. He/she would get compensated for the time whether he/she rode back to the yard or not. So yes I suppose it can be done.
Metro in Los Angeles subsidizes vanpool service for employers. My hospital has a few parking spaces up front reserved for the vans. Some come from up to 40 miles away. It’s basically a cheaper carpool, since the vehicle costs are being paid.
#5 At TriMet in Portland, Oregon, bus and rail operators commonly end their shifts at different locations from the start. They are not paid travel time, but instead are paid a nominal “road relief allowance” that ranges from US $1.50 to $4.50 for the most common points, and up to $15.00 for the most remote rail relief.
Shifts that start and end at the same non-garage relief point do not pay any allowance. The amount of the allowance is subject to collective bargainng. All reliefs must permit transit travel to or from the garage in less than an hour, but operators are not in pay status nor are they required to travel to or from the garage.
Long Beach (CA) Transit has pieces that start and end at different locations. Like presumably TriMet, unless you are taking a bus out or pulling one into the garage you never have to actually go to the garage. Drivers park their car on the street close to one of their relief points, and then take the bus back. Occasionally management tries to have a more normal arrangement where you would drive a relief car from the garage to the relief point, but the union doesn’t want that.
5) At UMass Transit some shifts begin at the garage and end at the driver change spot about a mile away (or vise versa). Given the nature of our structure many drivers will ride the bus to that change point, or take the driver change van, but that doesn’t count as paid time.
The main conclusion one should draw from Mees is that standard urban area density means very little, and instead people should look at perceived density, or alternatively at standard density in a very small area.
I notice 4 and 5 are pretty tightly related.
Not exactly what you asked, but I think relevant to the intent of the question:
I have noticed that our local bus system (Hampton Roads Transit) doesn’t make the bus return to the garage at the end of a driver’s shift. Instead, the relief driver takes a car to one of the stops along that route, and when the bus arrives he’ll take over the bus, and the driver ending his shift will take the car back to the garage to clock out.
So, while the driver still clocks out where he clocked in, the transit vehicle remains in the field and doesn’t return until the end of its shift.
I find this system slightly amusing, as if the transit agency itself is admitting that it would take too long to reach the places where the shift will change by transit, and that anybody who can afford a car would use one instead. (Note I do not intend this statement to describe transit in general, but rather the state of our local system. If you build service so bad that nobody would ride it if they have a choice, ….)
Hi Jarett, I’m presuming the Worker/Driver programs that you’re aware of are Mason County Transportation Authority and Kitsap Transit.. Interestingly they’re both for going into the same destinations Puget Sound Naval Shipyard and Naval Station Bremerton which are five miles or so from each other…
Although in my mind I don’t quite count Bremerton as a suburb of Seattle, but I’m a local. Bremerton is a good hour ferry ride or drive away from Seattle…
#4 Tranz Metro Wellington NZ do this with train crews on long distance commuter services to/from the Wairarapa. Conductors ride the commuter trains and collect tickets, and then go to their day job in the city, and then collect fares again on the return in the evening. Basically a way of getting paid for your commute time instead of paying to commute. Quite a good option when you have a 1.5 hour commute.
Almost 10 years ago the TTC published the Rapid Transit Expansion Study (RTES), which included a discussion on density and modal split in the vicinity of subway stations. (At the time of the analyses, the most recent data didn’t include the Sheppard subway, so those stations are omitted.)
In particular, see the graphic on page 21 of the presentation. It ranks all the stations from sparsest to densest catchment areas (pop + jobs per hectare) and then shows the modal split, and yes, it generally increases as density increases. The TTC uses these data to say that successful rapid transit stations must have a minimum density of 100 residents and/or jobs per hectare in the vicinity of the station.
Inserting the RTES density and modal split data in an Excel table, a logarithmic trend line has an R-squared value of 0.74.
Interestingly, though, you can get a (slightly) better linear correlation with something much more straightforward: subway travel time to Yonge and Queen (R-squared = 0.76).
In practice, there are a lot of factors that influence mode choice. Certainly density is one factor, but likely more because greater density means more riders to support the service, which means more service can be provided, which attracts more riders — and because high levels of density can better support services within walking distance, further reducing the need to use a car. Other factors include travel time, urban form, parking availability and price, socio-economics and income, etc.
I would also argue land use contributes significantly; data in Toronto show that a suburban residential high-rise next to the subway can attract a high modal split much more easily than an office tower; the residential tower attracts residents that are more inclined to choose to live there due to proximity to the subway, often for a downtown commute, whereas the office tower contains employees from a variety of areas and backgrounds, many of which will not be commuting from transit-friendly areas or even be anywhere near the subway.
@ Alon Levy
Suggests that “density in a very small area” is more important than larger areas. I agree, but this becomes harder to study especially since density is often not the dominant factor (and may actually be less important than several other factors in the area in question).
Mees has a rather practical take on the question of planning in the absence of density as THE dominant factor and suggests:
“Urban planners can make the task of adapting urban transport to cities easier, by discouraging scattered fringe development, clustering higher-density housing and major travel destinations in centres along major trunk transit corridors, and designing neighbourhoods in ways that foster walking, cycling and efficient bus operation. But none of these measures can substitute for a genuine public transit network that offers convenient ‘anywhere to anywhere’ travel.”
In the absence of false certainty backed by dubious statistics you have to resort to common sense. Not the worst outcome, even for a recovering engineer like me.
#4 Part-time drivers is at the heart of a concept I am promoting to use shared cars for commuting, thus getting this exciting service into the suburbs to be used for daily commutes in a ridesharing mode. I call it “trans-seat.”
This allows any driver to do the piloting, rather than require special licenses for the larger vehicles used by transit authorities.
This will accomplish something else: provide cars at (isolated) suburban business parks that can be used for regular carsharing trips during the day, and also for family trips (replacing a second car) in (equally isolated) residential neighbourhoods in the evenings and weekends. At the workplace, their availability overcomes a major reason commuters give for being their own car, even their need for a car during the workday is only infrequent (but often with little advanced notice).
Sticking with regular cars might reduce the number of seats-per-driver, but it makes the question of where to park the vehicle during the day much easier, and reduces a kind of stationary dead-head inefficiency (which makes it a natural for car-sharing organizations to run — no subsidies needed).
I don’t know if it has changed, but at least a few years ago, in Milwaukee (WI), an operator could start his/her shift at the garage, drive the bus all day, get off at a relief point, and be off the clock at that point. So, in that case, no, the driver wouldn’t have to start and end his/her day at the same location, and there was no paid travel allowance to return to the garage.
Chicago CTA pays operators to/from the garage at the start and end of the shift, but if they have a swing (split AM/PM) run, and get relieved on the road somewhere, they are not paid any travel allowance for the in-between portion. They are off the clock when they finish their first piece, and aren’t on the clock again until they start their second piece.