Transportation Elasticities Study

[Jon Bryant]

This is a must read for anyone planning our transport systems.  This is a recent study from the Victorian (Canada) Transport Policy Institute on transportation elasticities (i.e. the affect different factors have on our our transport choices).

http://www.vtpi.org/elasticities.pdf

Abstract
This report investigates the influence that prices and service quality have on travel behavior. It summarizes research on various types of transportation elasticities and describes how to use this information to predict the travel impacts of specific price reforms and management strategies.

Whilst the report says a lot and is very interesting reading it does state that

The elasticity of transit use to service expansion (e.g. routes into new parts of a community) is typically in the range of 0.6 to 1.0, meaning that each 1% of additional service (measured in vehicle-miles or vehicle-hours of service) increases ridership by 0.6-1.0%. The elasticity of transit use with respect to transit service frequency (called a headway elasticity) averages 0.5.

To me this shows that our current approach to transport planning in SEQ of assuming that Public Transport will only grow at low % is utterly flawed.  We should be setting high PT targets and building the system/service to handle the volume because people will shift out of thier cars and use the service.  Especially if combined with car parking restrictions and price increases.  

It also shows that a moderate increase in petrol prices is likely to see a mass switch to PT for which there is little or no remaining capacity to support.

[#Metro]

I think a lot of things can be determined using 'experimental' or 'trial' methods.
Sometimes you just can't predict what will/won't be popular or what works/doesn't.

Small trials and pilots help you sort out what works in theory, from what works in practice.
And they avoid costly mistakes.

[stephenk]

There are many examples where a frequency increase has resulted in larger % increase in Patronage. Croydon Tramlink being a good example. It is a no brainer that Brisbane needs to move to 15 min inner-suburban off-peak frequencies.

[#Metro]

Why not just split the trains into half, and double the frequency. (Sectorisation required)
This is something that can be done sooner rather than later.

All those guards can then become drivers. You wouldn't need guards because the train would be so short.
Trains that typically run every half hour, would run every 15 minutes. Trains that ran every 15 minutes would run every 5 minutes.

Imagine that? (Hmm. Could the network even cope?).

Trial it on a line- say the Shorncliffe line or Ferny Grove line. See what happens.

[#Metro]

This is a recent study from the Victorian (Canada) Transport Policy Institute on transportation elasticities...

Oh, Canada! 

[stephenk]

You wouldn't need guards

Oooh, controversial!

[somebody]

Real interesting, Jon B.  But the 0.5 factor for frequency increases seems a bit simplistic.  Perhaps it's explained in more detail in the report.  Obviously, increasing from 30min frequency to 15min frequency would be more effective than 15 mins to 5 mins.

Why not just split the trains into half, and double the frequency. (Sectorisation required)
This is something that can be done sooner rather than later.

All those guards can then become drivers. You wouldn't need guards because the train would be so short.
Trains that typically run every half hour, would run every 15 minutes. Trains that ran every 15 minutes would run every 5 minutes.

Imagine that? (Hmm. Could the network even cope?).

Trial it on a line- say the Shorncliffe line or Ferny Grove line. See what happens.

You'd only need to lock the back 3 cars like they do late at night.  I wouldn't want to try that on the FG line though. With no extra seats, those trains would likely become packed.

[#Metro]

Oooh, controversial!

How is that controversial? They will now sit at the front of the train, rather than at the back. Its not a big difference.

[mufreight]

So we spend hunderds of millions to revamp the signaling systen to allow trains to operate closer, and slower because station dwell time for a three car train is the same as if not longer than for a six car train and the major problem of train paths would effectively limit the numbers of services that could be operated.
Simple arithmetic ten three car trains operating at 6 minute intervals has less passenger capacity than 6 six car trains operating at ten minute intervals with slower transit times and more overcrowding by operating the three car sets as suggested.     

[Arnz]

Oooh, controversial!

How is that controversial? They will now sit at the front of the train, rather than at the back. Its not a big difference.

DDA issues too, instead of the guard helping wheelchair passengers, the driver has to get out and help instead..  Most stations on the Citytrain network are low-level platforms, and more than half of the stations go unattended after the morning peak.

There are some ways to solve this, eg raising EVERY platform on the network (very costly), or making most stations attended for the whole day (transfer some guards to station staff), or getting the driver to get out and help instead.

[#Metro]

Some good points/comments.
Thanks for pointing that out Mufreight- you are right about that point.
I was thinking along the lines of "how do you get more frequency", and as Mufreight pointed out, capacity per hour might be a more important consideration than frequency.

Scenario A: (60 min / 6 minute intervals) = 10 services in one hour * 0.5 capacity unit = 5 capacity units
Scenario B: (60 min / 10 minute intervals) = 6 services in one hour * 1 capacity unit = 6 capacity units.

More overcrowding will result under high frequency (A), than under lower frequency (B) even if there is no change in passenger numbers due to a change in service frequency.

So this solution can't really help during peak hour. It might be useful when there is off peak demand though- mainly weekends, as I would imagine splitting a train after AM peak, and putting it back together again in the PM peak weekdays would be a big effort.

Access for people with DDA is important. With regards to DDA, how does Melbourne do it?
I don't recall seeing guards on their trains, and I'm sure they still meet DDA somehow.

I think that leaves other options as: extend the platform, extend the trains. (Or buy more!).
Are there any "right now" solutions that can be done in the time between now and getting more trains?

Notes:
1 capacity unit is equal to a standard 6 car train.

[ozbob]

From the Connex site http://www.connexmelbourne.com.au/index.php?id=61#wheelchair

Access to trains
All of our trains provide wheelchair access. You can board trains from our dedicated access points marked on platforms with a white or yellow rectangle. Drivers will provide a ramp for boarding. And you'll need to supply them with a note about where you want to get off.

[Jon Bryant]

Buy/build new trains as fast as possible must be a priority.  We must have the existing network operating at its maximum physcial capacity.  If these means new signalling to allow 6 car trains to operate closer togeher then that must aslo be a priority.  I train a month is just too slow.  

I do sit/walk at Roma Street of a morning and wonder if they kept the services that terminate at Roma Street and Bowens Hills running up and down the lines for a few more hours (if not all day) if this will alow people to move across town more.  Does not help peak hour but it does help.

[longboi]

It might be useful when there is off peak demand though- mainly weekends, as I would imagine splitting a train after AM peak, and putting it back together again in the PM peak weekdays would be a big effort.

You're spot on...Splitting 6-car sets into 3 would be just a huge pain in the arse for very little gain. Besides, theres no room at Mayne to do all the shunt movements that would be required.

Access for people with DDA is important. With regards to DDA, how does Melbourne do it?
I don't recall seeing guards on their trains, and I'm sure they still meet DDA somehow.

I'm pretty sure (and Somebody can confirm) that the driver handles wheelchairs in Melbourne.  Personally I think that is a bad setup because the driver should be concentrating solely on driving and if the driver is on their own, leaving the cab to assist a wheelchair passenger could make them (and the train) vunerable to all sorts of things.

[#Metro]

We'll they wouldn't get far! There's only forwards and backwards, and I think the power shuts off if you pass a red signal.
And all the cash is locked into go cards and tickets- so it would be pretty odd. 

Maybe we should just get a metro.

[longboi]

Nope...You don't lose power if you pass a red signal, at least not in the Brisbane suburban area.

Besides, I was talking more about theives, vandals etc.

[O_128]

or we could make platforms level and eliminate the problem all together. but that would require common sense.

[#Metro]

Nope...You don't lose power if you pass a red signal, at least not in the Brisbane suburban area.

Wow, that IS dangerous. I always thought they had ATP.

http://en.wikipedia.org/wiki/Automatic_Train_Protection
http://en.wikipedia.org/wiki/Balise

http://en.wikipedia.org/wiki/Train_Protection_%26_Warning_System

[Arnz]

or we could make platforms level and eliminate the problem all together. but that would require common sense.

Not to mention a lot of money too (for past mistakes).

[O_128]

or we could make platforms level and eliminate the problem all together. but that would require common sense.

Not to mention a lot of money too (for past mistakes).

my favourite would be the beenleigh line considering 7 stations were upgraded it is now going to be another 100 years before they can get upgraded

[longboi]

Nope...You don't lose power if you pass a red signal, at least not in the Brisbane suburban area.

Wow, that IS dangerous. I always thought they had ATP.

http://en.wikipedia.org/wiki/Automatic_Train_Protection
http://en.wikipedia.org/wiki/Balise

http://en.wikipedia.org/wiki/Train_Protection_%26_Warning_System

ATP doesn't kick in until past Caboolture somewhere.

[#Metro]

Why isn't there ATP or something similar in the Brisbane area? I mean, that is where all the people are...

[O_128]

Why isn't there ATP or something similar in the Brisbane area? I mean, that is where all the people are...

once again becuase that would make sense

[longboi]

Why isn't there ATP or something similar in the Brisbane area? I mean, that is where all the people are...

Well that is the purpose of a train driver...to control the train under direction of signals 

Anyway, ATP would just be too expensive to implement and isn't really suited to stop/start suburban networks with a lot of complex train movements.

[#Metro]

Doesn't Melbourne have something like this on its network?
It only takes a slip up to put 700+ people at risk.

[longboi]

Honestly, with the amount of different controls in place, you would have to be doing something ridiculously insane. I.E. Go 100km/h around a 40km/h curve 

[somebody]

Doesn't Melbourne have something like this on its network?
It only takes a slip up to put 700+ people at risk.

Pretty sure Sydney does too, but it's not full ATP: It only controls SPADs & violations of low speed indications.  Not sure of the extent because it should have prevented the Glenbrook disaster.

[mufreight]

From memory the Sydney system uses mechanical trips at signals which apply the brakes.
More modern systems which could be progressively rolled out in the metropolotan area give a speed signal between signals and if the speed is exceeded takes control and cuts power and applies brakes, the next step up onboard signaling and from there it is only a short step to fully automated driverless trains.

[O_128]

From memory the Sydney system uses mechanical trips at signals which apply the brakes.
More modern systems which could be progressively rolled out in the metropolotan area give a speed signal between signals and if the speed is exceeded takes control and cuts power and applies brakes, the next step up onboard signaling and from there it is only a short step to fully automated driverless trains.

lets just go straight to driverless trains it amazes me that the technology is there yet we arent using it!

[#Metro]

What places use driverless trains?
I'm always worried that people will commit suicide/cross yellow lines/drive across the tracks when the boomgates are down.
How is the train going to stop if there is a fight on board?

[stephenk]

Why isn't there ATP or something similar in the Brisbane area? I mean, that is where all the people are...

Well that is the purpose of a train driver...to control the train under direction of signals 

Anyway, ATP would just be too expensive to implement and isn't really suited to stop/start suburban networks with a lot of complex train movements.

ATP is used on many suburban rail networks in varying forms. Sydney and Melbourne have ATP for starters. ATP is commonly used in Japanese suburban lines. Many European suburban lines such as Munich & Berlin S-Bahns and Paris RER have ATP.

[Jon Bryant]

I am with tramtram.  It also employees people and reduces the pressure on the support services (aka our taxes)

What technologies do we need to implement to allow trains to operate on 2-3 minute frequencies?  Is it just signalling or is ATP needed as well given the reduced separations.  Frequency of service is everthing to get more people using public transport.

[#Metro]

I wonder if headway can be reduced?
There is little use in buying more trains for more frequency if it is unsafe to have them run so close to each other.

[O_128]

What places use driverless trains?
I'm always worried that people will commit suicide/cross yellow lines/drive across the tracks when the boomgates are down.
How is the train going to stop if there is a fight on board?

1. Hong kong uses them on some lines as does korea. as do many monorails
2. A driver cant stop in any of those cases anyway and there would also be safetly tech like lucruy cars have now.
3. there would still be a guard and an emergency stop button

[dwb]

Going back to the original post on transportation elasticities don't you think this is a rather simplistic approach?

For instance, there are some major assumptions in the process. For instance, supply... that is, it doesn't matter how highly discounted an iphone may be, if it is not in stock I simply cannot buy it no matter how much I want to! (If I cannot squeeze on the bus I simply can't make the choice to buy a ticket).

The report seems to argue that fare increases will negatively affect transit ridership.  It seems to ignore things like capacity in either system to absorb that change in use.  For instance, many people currently can't catch PT so they use an alternative mode, and this may have little to do with pricing (at least in the short term).  It could be assumed that this small change in PT ridership would mean significantly worsening traffic congestion that would iteratively negatively affect the operation and quality of PT.

But for me this seems backwards and doesn't consider that changes in price rarely occur in isolation.  For instance what would happen if a network of buslanes were introduced along with peak hour fare increases, significant off peak fare decreases and major positive changes in service quality due both to the isolation of transit from traffic and secondly from the additional services that could be funded with the additional funds raised, not to mention that the services would be able to carry more passengers simply from the fact that they're more efficient?

The world is not simply some economic vacuum where maths rules.  People make highly complex value decisions on ongoing bases.  It is not simply price changes in transit that one needs to consider, but rather quality of transit (through non-price REGULATION such as buslanes), frequency (yes this is acknowledged in the report), car ownership rates, overall car pricing measures, and the availablility of the product to purchase among only a few.