Silver Bullet #2: Network Design

[#Metro]

Rail has had the capacity to carry higher crush loads and as a consequence has not seen the increased frequencies that would create higher loadings because of availability of service.

Hi mufreight, I really enjoyed what you wrote. The short answer is: IMHO the bean counters are wrong. Firstly, there seems to be mixed messages about the rail system. Some people say that it could not possibly take more people or more trains, which I deeply doubt, especially for off peak. This is more of an argument to upgrade rail and spend money on it (after decades of neglect) and more of a reason to put more services on. Peak hour trains may be full, but it would be interesting to know what the maximum number of trains possible that could run on each line at peak is for each line.

Does anyone have this information or know?

I recall that RailBOT lobbied for more train services on the Ipswich line for the Go-Between Bridge construction. IIRC there were 2 new services, and a "hair extension" of an existing one. The two new services have now become a permanent part of the timetable because they too are full; Even though works on the Go Between Bridge are now finished.

Why Brisbane has such shockingly bad off peak frequencies, and sometimes peak frequencies, I do not know. Perth and Melbourne seem to have managed to do it. My personal experience is that many rail lines are very lightly loaded in the off peak. With two trains per hour on many sections, and trains sitting idle in the depot, there should be more than enough space for more trains. And if some lines are full in the off peak, this too means that there should be more trains.

http://www.brisbanetimes.com.au/queensland/more-trains-to-help-with-coronation-drive-gridlock-20091102-ht1m.html
http://www.brisbanetimes.com.au/queensland/clogged-roads-prompt-plea-for-more-trains-20091101-hrmh.html

[#Metro]

Rail has had the capacity to carry higher crush loads and as a consequence has not seen the increased frequencies that would create higher loadings because of availability of service.

It does. But what kind of transport models are the bean counters using? Is it really true that it is cheaper to carry 1000 people by bus (16 drivers) than by rail (2 staff), including time for passengers to wait etcetera?

Do they take into account network effects from feeder buses and strategic interchanges? The Mohring effect and induced demand? Demand for off-peak or recreational travel (because not everyone wants to go to the city). If we look at the BUZ paper, we see increases in patronage between 100% and 300%. The official elasticity of demand of public transport about "0.5", I take this to mean (correct me if I am wrong) that when services double the passengers riding should increase by 50%. The fact that BUZ shows levels well above this 50% tells me that something else is at work, network effects plus frequency and that the models used might need to be seriously revised.

Even more unusual, the greatest growth is actually in the off-peak, weekends and in the evenings, when people are not going to work. Trips on 199 in the evening increased by almost 300%. This tells me that PT is being used for all day general purpose trips, not just work trips, which is a good sign.

[#Metro]

Even more telling is the situation in outer areas where buses operate on at best hourly frequencies and do not interconnect with other services (bus or rail) which rather than attract custom effectively deters PT usage and authorities then cite the low usage as the reason for failure to improve service frequencies or in some cases justification for the removal of services.

You are exactly right. Nobody is going to bother waiting for a bus that comes once an hour. Every passenger knows that.
What if you need to transfer? More waiting, and routes that cross your route are not even shown on the bus timetable!

Melbourne is managing to supply cross town travel with SmartBus services (which by the way do not go into the CBD, they are 100% cross town services), these are well patronised. Again, why Brisbane lacks this, I do not know. Route 599/598 with its random timetable (should be clock face) and poor half-hourly frequencies is a public transport abomination.

[#Metro]

An increase in passenger numbers of say 25 would require an additional bus service while with rail an increase in passenger numbers in the region of 400 would be required before the bean counters at Translink would allow QR to increase the train frequency and operate an additional service.

True, if these were walk up patrons living within 800m of the station. But what happens if we use feeder buses to collect passengers from afar?
400 divided by 65 bus seats is 6 buses. If we can get 6 full feeder buses during peak hour to arrive and terminate at 6 rail stations (or 6 feeder buses in one hour to arrive and terminate at a bus-rail interchange) this requirement is more than satisfied.

If the interchange is immediately next to or below the platform, and the connection is timed, even more people will catch the bus.
Meaning that more trains will be put on.

[Golliwog]

I would be interested to see actual figures for patronage increases in the off-peak before making big statements about stuff like that. I think the higher % increase would be simply due to the fact that origianally there were very very few people using it. If you only had one person originally and after the BUZing you had 2, that doesn't look so good but if you say you had a 100% increase in patronage it does. Don't get me wrong, I'm certainly not against increasing off-peak frequencies, I'm just saying that massive % increases may be just hiding what are actually kind of low patronage.

[mufreight]

It is a point worth consideration that the BUZ services were an initiative of BT not those in George Street whichever end of the street one considers and they were initiated by BT to meet a perceived commercial need to service the business community, and they have been more than satisfactory in that respect, that the increased frequencies provided additional capacity and improved frequencies has increased the numbers of commuters is undeniable proof positive that such an approach works, a point obviously to this time lost on those incumbent at both ends of George Street.

[#Metro]

BUZ paper: http://www.thredbo.itls.usyd.edu.au/downloads/thredbo10_papers/thredbo10-themeA-Warren.pdf

I think the higher % increase would be simply due to the fact that origianally there were very very few people using it. If you only had one person originally and after the BUZing you had 2, that doesn't look so good but if you say you had a 100% increase in patronage it does. Don't get me wrong, I'm certainly not against increasing off-peak frequencies, I'm just saying that massive % increases may be just hiding what are actually kind of low patronage.

This is a property of all the BUZ routes BUZzed so far. Yes, the actual patronage is listed there for all the BUZ routes studied, in a bar chart. On many bus routes, the patronage on Sunday in 2007 was higher than patronage on a working weekday in 2003.

The increases are also not explained by population increases, the bus network overall increasing by 37% patronage (line graphs) compared to increases in the BUZ routes which are all at least 2.5x this amount.

[#Metro]

It is a point worth consideration that the BUZ services were an initiative of BT not those in George Street whichever end of the street one considers and they were initiated by BT to meet a perceived commercial need to service the business community, and they have been more than satisfactory in that respect, that the increased frequencies provided additional capacity and improved frequencies has increased the numbers of commuters is undeniable proof positive that such an approach works, a point obviously to this time lost on those incumbent at both ends of George Street.

Can you imagine what would happen to rail patronage is this was applied to rail?

[longboi]

It is a point worth consideration that the BUZ services were an initiative of BT not those in George Street whichever end of the street one considers and they were initiated by BT to meet a perceived commercial need to service the business community, and they have been more than satisfactory in that respect, that the increased frequencies provided additional capacity and improved frequencies has increased the numbers of commuters is undeniable proof positive that such an approach works, a point obviously to this time lost on those incumbent at both ends of George Street.

Can you imagine what would happen to rail patronage is this was applied to rail?

Yeah but we don't have the rollingstock or the track capacity for the universal BUZification that you're proposing. QR obviously are already working towards improving frequencies but they need the right level of rollingstock, staff and other other associated infrastructure before they can do it properly.

[#Metro]

Yeah but we don't have the rollingstock or the track capacity for the universal BUZification that you're proposing. QR obviously are already working towards improving frequencies but they need the right level of rollingstock, staff and other other associated infrastructure before they can do it properly.

Why not? Why can't we get more trains, or use the ones that are sitting in the depot during off-peak times? What is the "right level" of rollingstock and staff? Understandably finding staff is hard, but for a train you need a guard and a driver per 1000 pax on a train, for a bus you need to find 16 people to drive the buses. Surely finding 2 people is easier than finding 16.

I'm not convinced that all the lines have no more spare capacity- especially for the off-peak. The Ferny Grove line should have some capacity, as StephenK's signature line shows that there have been cuts. Could these services be re-instated? And during off-peak, with 2 trains per hour, there seems to be room. There may be bottlenecks during peak hour re: the merivale bridge, but CRR will fix that, and I don't believe that these bottlenecks are anywhere as bad during the off-peak.

I would be very very interested in finding out what the maximum number of trains each individual line could accommodate in trains per hour. If anyone has this information, please post it.

This problem cannot be solved using more buses, because they will get stuck in traffic and congest the city centre.

[somebody]

I would be very very interested in finding out what the maximum number of trains each individual line could accommodate in trains per hour. If anyone has this information, please post it.

Stephenk has posted on several occasions that the mains can get 19tph through the CBD, and the suburbans 23-25tph according to the ICRCS.

Yeah but we don't have the rollingstock or the track capacity for the universal BUZification that you're proposing. QR obviously are already working towards improving frequencies but they need the right level of rollingstock, staff and other other associated infrastructure before they can do it properly.

Sigh.  After all the time you have been reading this forum do you still believe this?  Only Doomben/Gold Coast/Nambour and the outer Beenleigh and Cleveland lines have this problem.  All other lines have this problem:

It is a point worth consideration that the BUZ services were an initiative of BT not those in George Street whichever end of the street one considers and they were initiated by BT to meet a perceived commercial need to service the business community, and they have been more than satisfactory in that respect, that the increased frequencies provided additional capacity and improved frequencies has increased the numbers of commuters is undeniable proof positive that such an approach works, a point obviously to this time lost on those incumbent at both ends of George Street.

They are tight fisted.  Also, it seems that QR have been reluctant to upgrade the frequency also to make it easier for freight.  Witness the termination of the additional daytime weekday Ipswich line trains at Corinda when Darra would be far better due to serving the 6th and 9th busiest stations on the network by AM peak boardings.

[stephenk]

I'm not convinced that all the lines have no more spare capacity- especially for the off-peak. The Ferny Grove line should have some capacity, as StephenK's signature line shows that there have been cuts. Could these services be re-instated?

The Ferny Grove Line's timetable has roughly the same am peak capacity as pre-March 2008 at 7tph, although the pm peak is slightly worse (between 4:30 and 5:30pm) at 6tph. Due to the pointless token expresses being extended from Enoggera to Mitchelton, then Enoggera and Gaythorne had a significant reduction in pm peak services from 7tph to 4tph. Overall line capacity is now unchanged to pre March 2008. Other stations did see a service improvement in the March 2008 timetable, but doesn't help passengers who use stations where the timetable became significantly worse.

Excluding the express issues, there is still a silly all stations gap of 21+mins in the early am peak, and if the timetable hadn't been corrected 2 days into the March 2008 timetable (when QR realised they had seriously stuffed up) there would have been an all stations  mid am-peak gap of 20mins. I cannot understand how QR did not see that the original March 2008 timetable was so seriously flawed. Even though the am peak timetable was quickly corrected, but unfortunately commuters are still living with the effects of a somewhat bizarre pm peak timetable.

Yeah but we don't have the rollingstock or the track capacity for the universal BUZification that you're proposing. QR obviously are already working towards improving frequencies but they need the right level of rollingstock, staff and other other associated infrastructure before they can do it properly.

We have sufficient rolling stock, but not staffing. Track capacity is also an issue for 15min off-peak services on some parts of the system, in particular the Beenleigh/Gold Coast Lines.

[#Metro]

We have sufficient rolling stock, but not staffing. Track capacity is also an issue for 15min off-peak services on some parts of the system, in particular the Beenleigh/Gold Coast Lines.

Is this because the Gold Coast expresses catch up with the slower Beenleigh trains?
One solution might be to run a shorter service from Kuraby or some other station.
It won't give everyone on the line 15 minutes, but it is better than nothing.

[stephenk]

We have sufficient rolling stock, but not staffing. Track capacity is also an issue for 15min off-peak services on some parts of the system, in particular the Beenleigh/Gold Coast Lines.

Is this because the Gold Coast expresses catch up with the slower Beenleigh trains?
One solution might be to run a shorter service from Kuraby or some other station.
It won't give everyone on the line 15 minutes, but it is better than nothing.

Yes.

Reversing 2tph at Kuraby would still slow down Gold Coast trains by approx 4mins (and result in minimal operating margin). However this is more cost effective than spending $500m on quadrupling a recently triplicated track - don't you just love Queensland's half-baked infrastructure projects!

[#Metro]

Reversing 2tph at Kuraby would still slow down Gold Coast trains by approx 4mins (and result in minimal operating margin). However this is more cost effective than spending $500m on quadrupling a recently triplicated track - don't you just love Queensland's half-baked infrastructure projects!

I always wonder why they don't build everything in one go. It's so wasteful.

[somebody]

Reversing 2tph at Kuraby would still slow down Gold Coast trains by approx 4mins (and result in minimal operating margin). However this is more cost effective than spending $500m on quadrupling a recently triplicated track - don't you just love Queensland's half-baked infrastructure projects!

The quadding will probably be required in the 2010s though.

There are other alternatives such as using the Rocklea siding and turning towards Tennyson after Yeerongpilly.

I think a Sunnybank turnback makes a lot of sense also.

[stephenk]

Reversing 2tph at Kuraby would still slow down Gold Coast trains by approx 4mins (and result in minimal operating margin). However this is more cost effective than spending $500m on quadrupling a recently triplicated track - don't you just love Queensland's half-baked infrastructure projects!

The quadding will probably be required in the 2010s though.

There are other alternatives such as using the Rocklea siding and turning towards Tennyson after Yeerongpilly.

I think a Sunnybank turnback makes a lot of sense also.

I would hope that the quadrupling as far out as Banoon (as recommended in the ICRCS for 2010) is constructed by the time CRR is opened.

[colinw]

We have sufficient rolling stock, but not staffing. Track capacity is also an issue for 15min off-peak services on some parts of the system, in particular the Beenleigh/Gold Coast Lines.

Is this because the Gold Coast expresses catch up with the slower Beenleigh trains?
One solution might be to run a shorter service from Kuraby or some other station.
It won't give everyone on the line 15 minutes, but it is better than nothing.

Yes.

Reversing 2tph at Kuraby would still slow down Gold Coast trains by approx 4mins (and result in minimal operating margin). However this is more cost effective than spending $500m on quadrupling a recently triplicated track - don't you just love Queensland's half-baked infrastructure projects!

Speaking as someone who lives in Kuraby, other than the nice stations & full height platforms the upgrade has been a real disappointment.  The rhetoric, delivered in glossy form to our mailboxes throughout the project, was that this upgrade would see a significant improvement in service standard.

The reality is that our rail service is unchanged from the days of dual track, most services to the city are 2-3 minutes slower than they were 10 years ago, and in a couple of cases gaps in peak are longer.

If Salisbury to Kuraby is an example of how Queensland goes about "upgrading" a suburban rail service, then we are in big trouble.  It seems bizarre that we spend hundreds of millions building great new stations & additional track, for a net effect of no improvement in service at all.

About the only real benefit I can see is that the Gold Coast service is now more reliable.

[somebody]

If Salisbury to Kuraby is an example of how Queensland goes about "upgrading" a suburban rail service, then we are in big trouble.  

I'm afraid it is an example.  Other examples are the Northern Busway, Beerburrum duplication, Mitchelton-Keperra duplication.

[Golliwog]

In fairness though, I believe some of the reason it gets broken up is because the gov't simply doesn't have the money for it all at the same time. Although rail vs. road (and other) spending is a completely different argument, which I'm not looking to get into.

[cartoonbirdhaus]

Reversing 2tph at Kuraby would still slow down Gold Coast trains by approx 4mins (and result in minimal operating margin). However this is more cost effective than spending $500m on quadrupling a recently triplicated track ... It seems bizarre that we spend hundreds of millions building great new stations & additional track, for a net effect of no improvement in service at all.

At least if they hadn't insisted on the stations having three platforms, there may have been room for passing loops for the bi-di track.

[#Metro]

In fairness though, I believe some of the reason it gets broken up is because the gov't simply doesn't have the money for it all at the same time. Although rail vs. road (and other) spending is a completely different argument, which I'm not looking to get into.

IMHO stop/start can add extra costs- extra costs for the government in the form of getting crews together. There are also costs which are not seen- extra waiting time for trains to pass each other at stations and lack of services is a loss which falls on commuters.

IMHO Keperra-Ferny grove in one hit would be good.
Again Perth has no problem building 72km of double track all in one go.

[colinw]

To be fair, WA finances were significantly better than QLD's at the time that happened.

[somebody]

In fairness though, I believe some of the reason it gets broken up is because the gov't simply doesn't have the money for it all at the same time. Although rail vs. road (and other) spending is a completely different argument, which I'm not looking to get into.

I do not believe that is at all "In fairness".  The current approach is a bit of a knee jerk reaction that makes the finances worse in the long term.

[Golliwog]

I agree its not perfect, but would you prefer it to be a bit more expensive, and get bits and pieces as they are available, or have nothing for a while until they can afford to put the whole thing in in one go?

[longboi]

I would be very very interested in finding out what the maximum number of trains each individual line could accommodate in trains per hour. If anyone has this information, please post it.

Stephenk has posted on several occasions that the mains can get 19tph through the CBD, and the suburbans 23-25tph according to the ICRCS.

Yeah but we don't have the rollingstock or the track capacity for the universal BUZification that you're proposing. QR obviously are already working towards improving frequencies but they need the right level of rollingstock, staff and other other associated infrastructure before they can do it properly.

Sigh.  After all the time you have been reading this forum do you still believe this?  Only Doomben/Gold Coast/Nambour and the outer Beenleigh and Cleveland lines have this problem.  All other lines have this problem:

It is a point worth consideration that the BUZ services were an initiative of BT not those in George Street whichever end of the street one considers and they were initiated by BT to meet a perceived commercial need to service the business community, and they have been more than satisfactory in that respect, that the increased frequencies provided additional capacity and improved frequencies has increased the numbers of commuters is undeniable proof positive that such an approach works, a point obviously to this time lost on those incumbent at both ends of George Street.

They are tight fisted.  Also, it seems that QR have been reluctant to upgrade the frequency also to make it easier for freight.  Witness the termination of the additional daytime weekday Ipswich line trains at Corinda when Darra would be far better due to serving the 6th and 9th busiest stations on the network by AM peak boardings.

After working in QR and speaking directly with people who develop day-to-day timetables I know we do no have the overall capacity or rollingstock yet to comfortably upgrade frequencies all lines.

[#Metro]

After working in QR and speaking directly with people who develop day-to-day timetables I know we do no have the overall capacity or rollingstock yet to comfortably upgrade frequencies all lines.

Not even in the off-peak? 

[longboi]

After working in QR and speaking directly with people who develop day-to-day timetables I know we do no have the overall capacity or rollingstock yet to comfortably upgrade frequencies all lines.

Not even in the off-peak? 

You could probably provide a very random, inconsistant melange of services on most lines (Shorncliffe, Doomben, Sunshine Coast, Rosewood and Outer Beenleigh/Cleveland excluded). With freight and empty train movements you could probably achieve 15-20min frequency but then you would have little to no standby rollingstock in case of breakdowns and it would only ever be second-rate.

It makes much more sense to invest in a system to bring it up to standard first instead of trying to cut corners to squeeze in more services.

[#Metro]

You could probably provide a very random, inconsistant melange of services on most lines (Shorncliffe, Doomben, Sunshine Coast, Rosewood and Outer Beenleigh/Cleveland excluded). With freight and empty train movements you could probably achieve 15-20min frequency but then you would have little to no standby rollingstock in case of breakdowns and it would only ever be second-rate.

It makes much more sense to invest in a system to bring it up to standard first instead of trying to cut corners to squeeze in more services.

Melange does not sound good.  

It makes much more sense to invest in a system to bring it up to standard first instead of trying to cut corners to squeeze in more services.

This sounds bad. If the system needs investment, bean counters will want to see how much use it is getting- if it isn't getting much use off-peak (due to the said non-investment) they may justify low patronage as a reason not to invest in rail (and maybe put that money into buses) - which will perpetrate a self-fulfilling prophecy.

If rolling stock is the main issue, buying more trains could ease that one.

[stephenk]

Reversing 2tph at Kuraby would still slow down Gold Coast trains by approx 4mins (and result in minimal operating margin). However this is more cost effective than spending $500m on quadrupling a recently triplicated track ... It seems bizarre that we spend hundreds of millions building great new stations & additional track, for a net effect of no improvement in service at all.

At least if they hadn't insisted on the stations having three platforms, there may have been room for passing loops for the bi-di track.

Apparently there is space available for adding the 4th track, but in some locations I can't work out where the 4th track could be added without demolishing newly built structures.

[somebody]

After working in QR and speaking directly with people who develop day-to-day timetables I know we do no have the overall capacity or rollingstock yet to comfortably upgrade frequencies all lines.

I think the operative word there is "comfortably".  For example: using the excuse that we can't upgrade the Cleveland line's off peak frequency due to the single track sections is unnacceptable.  You can easily upgrade as far as Manly.

[#Metro]

My proposition is that the train line is extended from Cleveland, going south through Thornlands, Victoria Point, Redland Bay, Mt Cotton and then through to Logan/Beenleigh. Even if it means building a subway, this train line is desperately needed and would remove thousands of cars from our local roads.


The buses here are an absolute joke - they run on a Public Holiday timetable during peak hour. It's no wonder they are hardly used.

The roads struggle to carry the exponentially increasing amounts of traffic.

If real action isn't taken to address this, the southern Redlands area will only become more congested and polluted and the lifestyle we have come to appreciate here will disappear.

-Mick  | Redlands - July 19, 2010, 6:02PM

http://www.brisbanetimes.com.au/federal-election/magnificent-seven-sweeteners-will-swing-votes-seq-mayors-20100719-10his.html

[stephenk]

My proposition is that the train line is extended from Cleveland, going south through Thornlands, Victoria Point, Redland Bay, Mt Cotton and then through to Logan/Beenleigh. Even if it means building a subway, this train line is desperately needed and would remove thousands of cars from our local roads.


The buses here are an absolute joke - they run on a Public Holiday timetable during peak hour. It's no wonder they are hardly used.

The roads struggle to carry the exponentially increasing amounts of traffic.

If real action isn't taken to address this, the southern Redlands area will only become more congested and polluted and the lifestyle we have come to appreciate here will disappear.

-Mick  | Redlands - July 19, 2010, 6:02PM

http://www.brisbanetimes.com.au/federal-election/magnificent-seven-sweeteners-will-swing-votes-seq-mayors-20100719-10his.html

Are you "mick" in disguise? 

[#Metro]

Are you "mick" in disguise?

Nonsense Stephenk. As if I would live out at Cleveland!
I would need a car then. Unthinkable!