How many travellers use the trains?

Weekday boardings, Hurstbridge & Epping lines

The President of the Public Transport Users Association, Daniel Bowen, posted some “unofficial” stats last week on boardings at Melbourne’s railway stations in 2008-09. I’ve used these numbers to put together the accompanying exhibit showing the number of weekday boardings on the Epping and Hurstbridge lines. These two lines join into one at Clifton Hill so I’ve shown the section from there to Jolimont separately (too much effort to do any other lines!).

Daniel emphasises these numbers come with no warranty as to their accuracy but they did come from a “good source” in the Department of Transport. He reckons inflating the numbers by 11.4% will give a fair estimate of 2010-11 boardings.

I want to make a number of essentially speculative observations prompted by these numbers (for the purposes of this discussion I’ll leave the numbers as they are).

First, there seems to be no statistically significant relationship between the number of boardings and distance from the city centre i.e. from Jolimont to both Epping and Hurstbridge (admittedly my measure of “distance” is rank order of stations not kilometres, but I don’t think that matters). So while the proportion of the population resident around each station that uses the train generally declines with distance from the centre, the absolute number of boardings isn’t correlated with distance.

This may seem surprising because stations close to the centre are more proximate to the CBD’s many and various attractions and might be thought to enjoy higher dwelling densities than more distant stations. However it appears that other variables, such as the accessibility of a station to the surrounding population, are a more important determinant of the number of boardings.

Second, location on a junction of the rail network is not a guarantee of a large volume of boardings and nor does the absence of a junction mean a station will only ever have a minor role. Clifton Hill is the only station on this line that’s on a junction and has a reasonably large number of boardings, but not as many as Ivanhoe, Heidelberg or Reservoir.

Clifton Hill only ranks 36th in patronage of all stations (excluding the five loop stations) but that’s better than two other “junction” stations, Burnley (48th) and North Melbourne (90th). Eight of the 20 largest (non-loop) stations happen to be on junctions but twelve aren’t — my interpretation is being on a junction was a distinct advantage in the early days of rail and gave those eight a head start. Nowadays however the broader characteristics of centres appear to be more important drivers of boardings. For example, Ivanhoe is not a large activity centre in terms of jobs, but its station serves two large private schools and is an important pick-up point for buses serving schools in Kew. Heidelberg also has a school but more importantly has a large number of jobs in and around the Austin Hospital and has the local courthouse. Clifton Hill is disconnected from the nearby retail strip, has little nearby space for commercial or more intensive housing development, and is “in competition” with the No. 86 tram.

Third, some stations have relatively low patronage. There are thirteen with fewer than 1,000 boardings per weekday and five with around 600 or less. As discussed here (see comments), there’s probably little to gain from ceasing to serve stations like Wattle Glen (278), Rushall (478) and Darebin (530) because the extra time involved in stopping is small (the cost equation might change as the Government’s plan to staff each station at night with security officers is rolled out). There could however be a better case for servicing Hurstbridge, Wattle Glen and Diamond Creek by bus from Eltham rather than rail, although the savings would have to be pretty large to see off the ugly politics that “closure” would bring on!

Weekday boardings at stations (excludes loop stations) in descending order

An alternative interpretation is to wonder why a station like Darebin – my local – is languishing with just 530 daily boardings. There’s a wide strip of land between the station and Heidelberg Rd with some large land parcels used for low value uses (e.g. a car yard), but not much in the way of medium density housing going up. Two storey town houses are currently being constructed on one lot, but this is a location which would easily take more intensive development.

Fourth, it’s useful to put the boardings in context. Upstream from Clifton Hill, the two lines have combined daily boardings of 53,243. Tony Morton (who’s also from the PTUA) estimates that this equates to around 15-20% of the 1,320,000 weekday trips taken by residents of the four municipalities served by these lines i.e. Banyule, Darebin, Nillumbik and Whittlesea (boardings and trips aren’t strictly comparable but it’s the order of magnitude I’m wanting to convey). As Tony says, that’s a pretty healthy contribution to the total travel task of the region, yet it’s only a minority of trips and underlines why it’s important to think about all modes when planning how to improve transport as a whole.

Nevertheless, rail performs the vital task of delivering a large number of people (mainly workers and students) to an extremely small location (the loop stations) within a very small period of time (the AM peak). Then it takes them back again. It’s hard to see how the CBD – which is vitally important to the metropolitan area as a whole – could be supported in its current form in any way other than by rail.

The second exhibit shows all 201 Melbourne rail stations (excluding the five loop stations) in descending order of boardings. That very tightly fitting curve is Logarithmic (with the loop stations included, the best fit is a Power curve (R^2=0.813)).

Update: how’s this for serendipity? Not long after I put up this post, Daniel Bowen posted a map of boardings by all stations prepared by one of his readers, Brendan Durward.

22 Comments on “How many travellers use the trains?”

  1. brisurban says:

    So what is the purpose of the curve?

    • Alan Davies says:

      I thought someone more informed than I might be able to relate the distribution to something more general or that occurs elsewhere (e.g. analogous to Zipf’s law)

    • Dave says:

      I wonder if you get the same curve on a less radially-based system, either with multiple cores or a less intense core?

  2. Natronomonas says:

    Darebin is likely languising for both the lack of density, but also that it’s very close to Ivanhoe, which has a superior service level.
    Same reason Eaglemont will be suffering – if you are in an overlapping pedshed you’ll walk further to Heidelberg or Ivanhoe for the better services (both rail and bus).

    Past Eltham could easily be run as a shuttle, having reduced impact on the rest of the line with delays on that long single-track section. I wonder if the bus could simply have some steel wheels put on it…? There’s probably some rules about that these days…!

    Daily boardings also don’t necessarily reflect the value of junction stations in their interchange role. However, I agree that modern drivers of boardings are more likely based around destinations in the stations’ pedshed (or connections to other modes that lead to destinations – eg Clifton Hill’s patronage might be improved if the tram connected, rather than competed). Some junctions don’t even have the service level you would expect; for instance, Burnley is suggested as an interchange station, but few regular passengers would use it as such, knowing that Belgrave/Lilydale trains mostly express it during the peak and shoulder.

  3. jack horner says:

    Agree that rail service beyond Eltham is extravagent.

    Why such a low figure for Nth Melbourne? Where do all the people getting off the Mel Uni shuttle bus go?

    I’d be sympathetic to closing a few inner area stations that have low patronage, no interchange role, other stations nearby and good public transport alternatives, eg West Richmond, Collingwood, Rushall, Croxton. The benefit is speeding up the trains for everyone else. Or replace Croxton and Thornbury with stations sensibly located for interchange at Normanby Ave and Miller St.

    • Russ says:

      North Melbourne has practically no walk up passengers, being isolated against the railyards/industry of West Melbourne, saves no time over the local trams getting to the city, and the 401 is the only interchange of note. Also, these figures won’t include the large number of people who get off at North Melbourne, and back on a different train without exiting the station (which is kind-of the point of a junction).

  4. Lynne says:

    You might also want to consider where the car parking is, I don’t think there is much at Darebin, whereas Ivanhoe and Heidelberg both have around 300 parks with Ivanhoe seeming to have many more commuters parked in nearby streets, something not possible at Heidelberg. They also both have bus services to La Trobe University and Northland.

  5. Jim Wells says:

    Interesting topic but be careful with the data – school children are significant users of rail and are major contra peak flow passengers.


  6. Richard Peterson says:

    I am of also one of the mere trickle of Darebin users.

    The area of surrounding railway land, plus low-grade use private land with development potential is most extensive. The station buildings themselves could have commercial use. Any development could include commuter multi-level carparking, and planning longer term, bridging across the railway cutting. Only Media House (The Age building) and Federation Square achieve this in Melbourne so far.

    For too long, every development proposal in Ivanhoe met community objections. Now some have been completed and others are proposed.

    Yet the proposed development at 40 Upper Heidelberg Road is very slow to proceed and 3 of the fine townhouses at 1094 Heidelberg Road (cnr Latham Street) remain unsold. Much more creative design, planning and marketing is needed for future multiple integrated uses over a long time-frame. The site is only 8 km from the CBD after all.

  7. Chris Curtis says:

    Wattle Glen is a tiny hamlet, so the low boarding figure is not a surprise, but it is a fair way from Hurstbridge. It’s hardly likely to close as it has just had a sealed car park built.

    There is no secondary school in Hurstbridge, so some of those boarding there may be students going to Diamond Valley Secondary College at Diamond Creek. Hurstbridge, as the end of the line, services Panton Hill, St Andrews and a few outer places that have no station – because they have no railway line. On the other hand, it is in the green wedge and so development is limited; thus population is limited; thus potential train travellers are limited.

    The dual track ends at Greensborough – though there are passing lines at Eltham and Diamond Creek. This reduces the services further out.

    The distance between Eltham and Diamond Creek stations is probably the longest on the whole line, meaning that many people would go to Eltham to catch the train; i.e., it has a large catchment area.

    Inner city residents, such as those in North Richmond and Jolimont, have trams as well.

    These are juts a few random thoughts to illustrate the complexity of transport systems.

  8. Andrew Rodda says:

    [Rant mode on]
    The issue of Hurstbridge vs Epping Line has been peeving me for years.
    The disparity of service has never been explained.

    Epping line trains get minimal express services compared to Hurstbridge. Have a look at the timetables….Almost every Hurstbridge line is express Clifton Hill to Jolimont, at a minimum, whereas almost every Epping line stops all stations.

    Train Frequency is also massively unbalanced. A typical train sequence through the loop in the afternoon would be… epping – hurstbridge – eltham – greensborough – epping.
    Uhmmm who are they fooling! 3 of the 5 are the same line, and all of these would be express.

    16 Train stops from Melb. Central to Reservoir Station (my stop). Where 30% of the carriage empties. Almost the equivalent number of stations as Box Hill, which often has limited express services directly to it.

    I don’t expect red carpet service, just parity. On a crowded platform we will see 3 trains go through with diminishing numbers of occupants and empty seats (Eltham, etc… line), only to have a already packed epping line train pull up, it gets a little much.

    Then there is the carriage type. Epping gets the cattle car, where the central aisle has been widened to allow for more standing room. I have never seen this carriage on the other line.

    Little wonder why I ride my bike as often as I can.

    [Rant mode off]

    • suze says:

      I think you only see with the one eye. I’m on the Hurstbridge line and I regularly get those “cattle car” trains you refer to. I had one this morning as a matter of fact. Though I did observe that while it’s easier to make your way down the aisle, the lack of handholds (that can be reached by people under 6′) mean the aisle section of the train actually holds LESS people than if there were the extra seats on the row (one person holding the handle as opposed to two sitting). And that train came in at platform 14 this morning.

      It seems to me that it’s only during peak periods that there are extra trains on the Hurstbridge line, which is only right as there are 4500 extra people on that line to clear.

      I think the express through to Clifton Hill is because the Hurstbridge train travels further and takes longer to complete its journey. Also, after 6:30pm the trains stop at those stops in Richmond and Collingwood, though I don’t understand why these patrons suddenly need twice as many trains as the rest of us get. Especially as they have trams and buses serving them that the rest of us don’t get.

      I understand your frustration though. I think it’s fair the way it is, but if I was on the Epping line, it would seem unfair and it would shit me too.

      • Andrew Rodda says:

        Actually, I sometimes catch the train from Dennis Station in the mornings, (Hurstbridge Line), Trains from 8:45 – 9:15, and they have seats available, not quite the same for the people squeezing onto the trains at the same time in Rushall or Merri….
        It won’t be fair until the timetables reflect patronage numbers and carriage densities. Epping will only get worse when South Morang Station comes on stream.

        • I lived for years right near Thornbury Station and used to feel the same way. Now I live right near Fairfield Station.

          There are pros and cons to each line, but neither is much better than the other.

          Express services may be good for some, but are the opposite for people that use the stations that get expressed. If I miss the 5.18 Greensborough for example I have to wait for 3 trains to pass before the next train that stops at Fairfield comes past, 2 of those trains are Epping services.

          As for the “cattle cars” we get plenty of them. Seems to be about 1/3 of my morning trains. Personally I prefer them, the seats are far less crammed, even if there’s less of them and they’re easier (and faster) to get in and out of. If people didn’t have to squeeze past 5 people sitting with their knees touching, then squeeze down a very narrow aisle past people that have no where else to go before they’re even at the absurdly crowded door area, then dwell times at stations would be reduced considerably.

          It dumbfounds me how poorly those trains are designed and that the DoT bought not one, but two classes (more?) of train with this same layout.

          Mind you the “cattle cars” aren’t much better designed. As stated above the hand-straps are unusable to anyone less than about 5″9 and they’re above the seats rather than the standing area! Clearly not designed by someone that rides trains much.

  9. mich says:

    If 4% of the daily journeys undertaken by the residents of Banyule, Darebin etc are made by boarding a train there, then another 4% of journey’s are made by taking a train to come home again, which gets counted as boarding the train at Flinders Street, or wherever. So really, its 8% on the train.

    • Alan Davies says:

      Fair point, although on reflection, would those return trips be counted as originating in the City of Melbourne rather than within the four municipalities served by the Epping and Hurstbridge lines?

    • mk says:

      I also see this as a failure in terms of adequate public transport in the NE of Melbourne. We get very limited bus services, considering that unlike the East and North sections of Melbourne are served by a number of frequent tram lines, incentive to use public transport for whole journey is quite low, especially when getting to stations.

  10. mk says:

    Also I think the article should consider Heidelberg being a high use station due simple to the fact that it’s on the edge of zone 1 and 2. Many people living in Rosanna will travel to Heidelberg to simply save money on tickets going into the city, not to mention that it is a premium station and therefore is perceived as safer at night and ‘nicer’ to depart and arrive from.

  11. Tony says:

    An interesting post Alan. Your observed lack of relationship between station boardings and distance from the CBD is quite consistent with other evidence, that the majority of train passengers are people who live within walking distance of the station. This is probably due to the fact that in Melbourne we’re hopeless at coordinating the train network with the bus and tram networks to enlarge the catchment of the rail system. Students are supposed to be more predisposed to using public transport than the general population, but just try getting from Latrobe Uni or RMIT Bundoora to any railway station.

    Regarding the figures you quoted for weekday trips: I presume they’re from VISTA? In which case these are for unlinked trips: so for example if someone drives to a railway station and catches a train, the car trip and train trip are counted separately. In fact if the survey participants complete their travel diary correctly, the trips from home to the station and back home again – and from the remote station to and from the final destination – will all be counted even if they’re done on foot. And Mich above is right: unless there’s been some additional processing done on your data, they will reflect the complete travel diaries of all residents of those municipalities, so will include travel done outside those municipalities as well.

    So in principle, each individual train boarding actually corresponds to 3 or 6 ‘travel instances’, depending whether the train journey was entirely within the relevant municipalities or not. Indicating that somewhere around 15-20% of linked trips in Banyule, Darebin, Nillumbik and Whittlesea involve a train component: a pretty healthy figure when you consider that half of all linked trips are confined to the local area, and beyond the ambit of the train system.

    Another point worth noting is that it’s only in the last couple of years that we’ve finally restored the Epping line to the level of service it had in 1960 (when it only went as far as Lalor, and many trains still only ran to Reservoir). The Hurstbridge line meanwhile is still operating *worse* service than in 1960 – there were 9 trains an hour to the city in the morning peak in 1960, compared with 8 an hour now. By no stretch of the imagination is the catchment population for these trains any smaller today than in 1960: the problem lies in planners’ complete unwillingness to compete with car travel, which no amount of high-density urban development will fix by itself. Building apartment towers without attention to multimodal transport planning doesn’t lead to sustainable transport nirvana. What it leads to is Los Angeles: 27 people per hectare and 4% public transport mode share.

    • Alan Davies says:

      Thanks Tony, fair cop. I’ve removed my mistaken figure (4%) and replaced it with your better estimate (approx 15-20%) so that new readers won’t be mislead.

  12. Campbell Wright says:

    Another point that should be considered is how uncompetitive for inner/middle suburbs the travel times of the Epping/Hurstbridge lines are compared to the tram and car. From Flinders Street to Thornbury takes 27 minutes by train, despite it being only 6-7km out of the CBD by car, or 9km by train – an excruciating 20km/h average. Even the 112 tram – with much better frequencies – only takes 24 minutes from the Melbourne Town Hall to Hutton Street in Thornbury. Admittedly, travel time is reduced somewhat through Jolimont-Clifton Hill express services, and most people boarding in the City Loop. The mobility offered is poor at best, and for all but trips to/from the outer suburbs, and for people who live closer to the station than the tram stop – the train is thoroughly uncompetitive.

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 )

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