Setting a service standard part one – the Clifton Hill group

Something that Melbourne’s trains have got very wrong (but that our trams have got basically right) is the concept of a service standard. Riccardo has blogged on this issue a lot so I won’t go into great detail about the intricacies of the issue. Suffice to say, a service standard shouldn’t be % of trains run within 6 minutes of timetabled or % of timetabled services run – it should denote how many trains per hour should be provided on a given line, the stopping patterns and the nature of the rollingstock. Measuring how many of this trains run on time is important, but is not a service standard in itself.

So, what sort of service standard should we be looking at for Melbourne? I’ve said on numerous occasions that most of Melbourne’s lines need to be running at least 6 trains per hour all day every day. Basically we should run as many trains as the infrastructure can handle. In this post – the first of a series – I’ll be looking at the service standard for the Clifton Hill group (the Hurstbridge and Epping lines).

When deciding on a short to medium term service standard, we need to consider two main things – the underlying patronage potential of the area and the infrastructure capacity of the line. The good news is that in 2004, the DoI published a report – the Clifton Hill Group rail review phase 1 – gives some insights into these issues.

Patronage Potential

Measuring patronage potential of an area directly can be difficult. Using the distributional patterns of existing patronage on a given line can help though. Obviously, such figures aren’t perfect and don’t account for supplier induced demand. But detailed AM peak up direction figures were readily available on pages 7 and 8 of the DoI report. Here they are:


This tells us that on the Epping line, patronage is highest towards the end of the line (which is also a growth area) and that Hurstbridge has high patronage as far as Greensborough, but not much beyond there. Supplier induced demand becomes an issue with the Hurstbridge line figures, as many peak hour trains terminate at Greensborough. This makes the figures trickier to use, but given low population density of the area and the line’s incursion into green wedge, it’s fair to say there’s not all that much new patronage potential beyond Greensborough or Eltham.

Infrastructure capacity

The Hurstbridge and Epping lines are some of Melbourne’s trickiest on the infrastructure front, with long single track sections and in some sections antiquated signalling. The good news is that there is reletively little congestion in the Clifton Hill loop. The following diagram, taken from page 11 of the DoI report, shows the infrastructure constraints as well as the peak service level as of 2004.


Capacity between Clifton Hill and Jolimont will be improved somewhat with the construction of the second track between Clifton Hill and Westgargh as well as reversing the AM loop direction. The Clifton Hill-Westgarth duplication really should have had a flyover at Clifton Hill though. Eltham to Hurstbridge will benefit from the signalling upgrade that’s going on at the moment.

Setting a service standard

What the above diagram shows us is that the infrastructure can handle 6 trains per hour to Epping and 9 trains per hour to Greensborough. The Clifton Hill-Jolimont bottleneck will be eased within the next couple of years. Capacity is more constrained beyond Greensborough, but there isn’t much patronage there anyway. This shows that we could run 6 trains per hour (giving us a 5 minute average wait time and turn up and go convenience) all day as far as Greensborough and Epping without any new infrastructure spending. Eltham and Hurstbridge would receive fewer services – say 3 trains per hour for Eltham and a shuttle every 40 minutes between Eltham and Hurstbridge. Run this level of service all day every day, 5:30am to midnight. If the capacity is there, run even more trains during peak.


Now we need to look at stopping patterns. Every Eltham train would run express Clifton Hill-Jolimont. All Greensborough services would stop all stations. Every second Epping train would run express Clifton Hill-Jolimont. During peak hour, any spare capacity could be used to run extra super-express trains on the Hurstbridge line – as is the case now.

Such a plan would require more drivers – there’s no doubt about that. It’s beyond my capabilities to work out exactly how many drivers would be required. However, if you look at some of my old system wide calculations, it’s clear we’re not looking at big money.


20 Responses

  1. Very sensible suggestion and evidence-baseed too.

    And you’ve avoided “fairness” eg the idea that Eltham and Greensborough “deserve” the same service – which with today’s level of infrastructure may simply not be possible to deliver.

    Connex and their predecessors got themselves in knots trying to deliver services that just weren’t possible reliably, and therefore compromise service delivery to those parts of the system that CAN be served reliably at high frequency, and raised expectations among those that can’t.

    And you’ve independently come up with an idea I had for a permanent Eltham-Hurstbridge shuttle. This train could be a lot smaller. And why not make a virtue of it? Get the locals to own it, paint it maybe, call each railcar a different name like Wombat, Bandicoot etc. The shuttle cars would be confined to the area so what’s the problem. This principle could of course be extended to other shuttles eg Alamein, Williamstown and a shuttle beyond Upper Gully.

    And you could build a dock platform at Eltham (very short) in the middle of the island, so that all transfers are cross platform and the existing platforms are available for terminating city trains.

    something like this
    DC__/_____ \_____________Montmerency

  2. Dam, the track diagram went skew! The spaces have been dropped



  3. Epping is a funny creature. It goes through an area of high tram use (and poor traffic) as far as Preston. This area is similar to the Upfield line as far as Coburg and similarly doesn’t pick up pax

    Suggests that the trams, despite being slow, are popular because they are frequent (and I suspect have better ‘navigability’ because you see the tram tracks on the ground, the wires above, you look for the little metal square on a pole saying “Tram Stop” and you’re away – far easier to locate and use than rail)

    But it is definitely counter intuitive that these higher density areas have lower ridership.

  4. I like the idea of running Eltham-Hurstbridge as a shuttle. Realistically it can hardly justify it’s current rail service, let alone boosting it to 30min or more as some Railpagers insist on. If it was not electrified back in the 1920s I would be amazed if it was still in existance today.

    On both the Epping and Hurstbridge lines, one of my main observations is that most patronage seems to come from the outer areas of the line (as Riccardo said with Epping). Clifton Hill-Preston and Heidelberg/Ivanhoe are both fairly dead sections patronage-wise in the off-peak periods, whilst Preston-Epping and Heidelberg-Greensborough/Eltham have a number of busy stops.

    Anyway, quite good ideas there Phin.

  5. Thanks Riccardo and Somebody.

    The Eltham-Hurstbridge section is a strange (and lovely) piece of railway infrastructure. I heard somewhere (can’t remember where) that the primary reason for electrification was that it was cheaper for the VR than to station a steam locomotive at Eltham for the shuttle. So it’s really a quirk of history that the line survives at all. Ironically, if the St. Andrews extension went ahead all those years ago there probably wouldn’t be anything past Eltham now.

    But Eltham-Hurstbridge sees a tiny amount of patronage and really doesn’t warrant the increase in service levels that Greensborough demands. I also like the idea of getting the locals more involved. Giving people a social or economic connection with the train means it’s more likely to used and taken care of.

    The lack of patronage on the Epping line between Rushall and Reservoir is also interesting. I’d guess that the frequent 86 and 112 trams are a big part of the reason for this, as well as poor integration with High St. Make the train as frequent (or close to) as the trams and I think that line has the potential for lots more growth in the inner stretches. Encouraging train use through the construction of a train/tram interchange at Thornbury would also help.

  6. I’m pretty sure that frequent trams are a major, major competitor for rail. Witness Upfield (vs 19) and the Epping stretches mentioned. Improving interchanges and frequency are key, while signage could be improved at most stations. Signs should exist up to 500m (at least) away, with an arrow and approx distance. And not like street signs, proper signs. Although they probably shouldn’t mention that frequency won’t be there offpeak (even though the trains seem pretty crowded to me offpeak, suggesting there is existing demand, and significant untapped demand (as Phin has previously blogged on).

    With regard to the Hurstbridge line, the train provides a useful service along the major Heidelberg Rd section, with metro-spaced stops, but without metro frequency. Thus, the patronage mainly boarding earlier on the line rather than later is not surprising, as one falls back to Riccardo’s premise (Pack’em in or whatever) that those further away will plan their trips around timetables, while those closer expect a frequent service, and, not always getting it, choose the quick drive instead. The short trips basically lose out to getting people from 10km+ into the city instead, and this shows in timetables and patronage patterns. Chicken/egg perhaps, except sometimes it takes innovation to change those patterns, not simply cater for the status quo.

    The Eltham-Hurstbridge shuttle though is a reasonable suggestion – the area is such that it might take more readily to ‘ownership’ of a train than say Alamein. And it is a reflection of demand patterns, and surely would simplify running schedules. Watching people disembark from the Alamein shuttles pretty darn often, I can’t really see the problem with transfers, provided it is cross-platforn (no stairs, underpasses or other craptastic transfer mechanisms). Inter-platform is OK, but only when you’re changing lines, not from spurs (which Elth/Hurs should be treated as – it’s got log trestles!!)

  7. Dave, one observation about the Clifton Hill-Heidelberg section is that, unlike Epping and Upfield is that there is no tram or bus that takes away patronage, minus the even less frequent weekday-only route 546. Methinks everybody must drive in those areas – most people would be affluent enough to afford a car in those areas.

    For Eltham-Hurstbridge, I am not sure exactly what would you would run on a shuttle. Running 6car trains out there for passenger loads that would fit in 1 is definately a waste. I have only once been on a train that had a full 6car load on that stretch. Be interesting to see a modern equivalent of a double ended Tait!

  8. Thanks Somebody. I agree that there is no tram/bus to take away patronage – but I think this has the effect of suppressing interest/demand in PT, whereas those in a tram/train corridor will consider both, then choose the tram if they don’t know the timetable (frequency) or the train if they do (faster travel speed).

    On the Clifton Hill-Heidelberg section, there is reasonable development along the length, so frequent services could enable that kind of travel (eg 2-5 stops, going to shops/schools etc) but these are prohibited by 20min frequencies which would be longer than the journey.

    Certainly when I used to live near there, planning was required for train use, which was fine for regular trips, and annoying for one-offs. And completely dissuades the casual user. Thus, yes, cars were/are a necessity around there, pretty much from Northcote (Westgarth/Dennis) out, where you lose the trams from.

  9. Phin – challenge for you. Come up with a effective journey time performance measure

    being the sum of

    navigation point A to navigation point B (navigation point being the point where a consumer expects to find the public transport ie you expect to find a bus in a neighbouring street, but expect to find planes at an airport)

    comprised of: journey time rail station A to B PLUS (1 of bus journey navigation point A-rail station A or walk navigation point A-railstation A if walk journey less than 400m) PLUS average timetabled waiting time for a continuum of (turn up and go-await scheduled departure, with the average weighted towards turn up and go as frequency approaches 3 minutely all day) PLUS average delay at Navigation Point A plus connection time at rail station A plus average delay at rail station A.

    I would be interested in what would happen if you built that as a spreadsheet, superimposed the grid coordinates for the rail stations on the system, major bus stops and the respective areas that are within walking distance of both, then assigned timetable values, to work out what the TRUE commuter-experienced PT journey times in Melbourne are. And use the spreadsheet to calculate the best investments.

  10. Hi Phin,

    I wasn’t sure how else to contact you so this will have to do. After you are done with Riccardo’s latest challenge, perhaps you might want to blog on the economics of electrification. I do have some perhaps outdated information that may be of use to you – send me an email if you are interested.

  11. Phin and others

    This is a diagram for my second part – a basic 3 minute frequency FSS-Richmond-Pakenham where every second train runs express to the fastest Connex table – SY-Caul-Oak-Clay

    As you can see the fast overtakes slow exactly at 10.6 km, Caulfield. Fast leaves 4 min after slow, gets to Clay 2 minutes before following and stays 2 minutes ahead until end.

    Interested in people’s thoughts on this.

  12. Thanks for the comments everyone.

    Somebody and Dave, my personal experience with the Hurstbridge line between Heidelberg and the city basically mirrors the patronage data. Between Ivanhoe and Clifton Hill there’s not a huge amount of patronage, with the notable exception of Fairfield. There’s plenty of patronage potential there though. I have seen ABS data on public transport use in that are, and I’ll try and chase it up.

    Riccardo, that is an interesting challenge indeed. I’ve actually been considering a similar idea for the tram system with a friend of mine. He’s looking at programming something in C# which would show the difference in travel time between peak and off-peak periods, and displaying it on a map in which the length of a line is determined by travel time. We’re also hoping to add average wait time to the mix.

    If he gets it to work, the same idea could be readily adapted to the train system. It might take a while to do, but the results could be very interesting.

    Your timetabling plan for the Pakenham line is a good way of running a long line with long double track sections. I have two questions though:
    1. How would you fit in the V/line trains?
    2. How would you change the culture in both Connex and the DoI so that government is serious about getting it to work?

    I see the biggest challenge being a cultural rather than technical one.

    Drwaddles – thanks for the offer – email sent!

  13. A lot of the stations westgarth to Ivanhoe seem to be ‘walk-ups’ ie no connecting bus feeders, and their car parks aren’t huge. They are a long series of ‘laburnums’ from what I can see. Fairfield, as noted, the exception.

  14. […] posted on a service standard for the Clifton Hill group a while back, advocating 6tph to Epping, with every second train running express Clifton Hill – […]

  15. Regarding the Clifton Hill to Heidelberg stretch you should consider car parks – the reason that Fairfield has the most patronage of any of those stops is a direct correlative of the size of its carpark, and Heidelberg has in addition to being the first Zone 1 stop. But people from further up the line from Heidelberg may be driving away from the city (and that particular stretch) just to get a car park, or a super-express..

    Similarly if I lived in Alphington etc I might drive to Clifton Hill where the frequency is much greater…

  16. P.S. I’m not advocating car parks. In fact if the car parks were made smaller it might increase patronage at nearby stations.. but the implications of the of car/train combinations shouldn’t be overlooked.. People will take the path of least resistance, so if it is raining or they’re running late they’ll drive to a station with a large carpark rather than walk 15 minutes to their closest one.

  17. Inner suburban park-n-rides – yucky yuck yuck.

    There is *some* justification for park-n-rides, such as at the rural end of a line, but definitely not in an inner-middle suburban area with houses as far as you can see.

  18. Agree Somebody

    Some locations where park n ride MIGHT work:

    -East Pakenham (end of the wire where the Fwy crosses)
    -General Motors overpass
    -Moorooduc level crossing

    In each case the aim is to pull people out of surrounding acreage properties onto the rail line. Services would need to be carefully structured as the demand would never be enough to sustain turn up and go. In each case I would suggest commuter style, maybe 3 services in, 3 services out and 1 at lunch time for those who have to go home early.

  19. Agree that inner suburban car parking is generally a waste of time, space and money. But agree they might be ok for some outer suburban infrequent commuter services.

    Andrew, car parking would skew the figures somewhat. For example, Heidelberg has 300 spaces, and if we assume one person per car using the space all day – AM peak to PM peak, that accounts for around 1/2 of 2004 peak boardings at the station. Mind you, Fairfield has over 600 AM peak boardings, and only 79 car spaces, indicating that the station is getting mainly walkup/bus patronage. Likewise, Greensborough has 230 spaces and over 900 AM peak boardings.

  20. […] six are good candidates for commuter type services – I’ve left Hurstbridge out as it’s quiet past Greensborough and dead beyond Eltham. On the remaining six lines, two tier services are the optimal solution, with metro trains […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: