The short answer is that those 'outside the system' don't know. Information is scarce. Even from governments normally eager to sell the benefits of its policies. Often we know more about the number of jobs involved than what's delivered.
This is particularly so in public transport whose bosses too rarely understand and sell the power of frequency. For example, the Department of Transport stumbled last month when launching Craigieburn's upgraded more frequent bus network. And with nothing recent to go on, we can merely hope that Sunshine will get more than the current (and planned) three trains per hour at most times when Metro Tunnel opens.
On the latter we do now at least have an indication of staffing for a Day 1 Metro Tunnel service ramp-up. Today I'll run through some numbers in an attempt to understand what this could mean for service. And demonstrate that we can do a lot with a little, if we set our minds to it, before Metro Tunnel opens.
Facts and figures - what can one train driver do?
To get an idea of how significant one train driver is we need to know how many service kilometres they drive per year. We can use a number of metrics to get relevant numbers number including kilometres of service the network delivers, average speed of trains, numbers of drivers and average working hours. Try a few, do some multiplication and division and numbers should broadly tally up.
2022-23 Budget Paper No 3 (page 340) tells us that 24.9 million service kilometres were scheduled on Metro Trains. This was the actual number in 2020-21, the expected outcome in 2021-22 and the target in 2022-23. About 1% of that wasn't delivered due to cancellations but for planning purposes we should use the scheduled figure.
Metro Trains said it ran 63 705 train services in November 2019. That's about 764 000 trips per year (noting there will be variations eg special event services). Today's number would be higher due to service upgrades in 2021. The average length of a trip (24.9 million km / 764 thousand trips) is about 32 km.
How many train drivers? Metro's operations in Melbourne has around 3500 employees all up. In 2019 Metro's driver numbers were stated at more than 1100. Today's tally would be nearer 1200 given last year's upgrades. Then there's about two other staff per driver.
The average Metro driver would clock up 22 636 service kilometres per year. That's more than half way around the world! In 2019 each Metro driver did an average 694 trips per year (764 000/1100).
Now let's do a 'sense check' to see if some of the above numbers are about right. The average Melbourne Metro train goes at about 35 km/h. This means that the average driver would be driving in service for about 22636/35 or 647 hours per year. It seems low but their actual working time would be a lot more. This is because the hours figure above does not include train preparation, driving non-service trains to get into position, pilot operation (ie 2 drivers per train for route knowledge), dwell times at out stations, non-driving activities, paid leave etc.
Other things being equal, if we want 647 more hours of service per year then we are going to have to recruit one more train driver. More maintenance and support staff will also be needed as we'd be running trains harder. But, except at the margins, most station staffing and PSOs are pretty much fixed costs regardless of whether we run one train per hour or six trains per hour. With economies of scale you shouldn't need anywhere near two extra staff per driver if you're just putting on a few more trains.
How long does it take a train to the end of the lengths of suburban lines we have in Melbourne? Leaving aside short branch lines it's usually a bit under to a bit over an hour. The typical length of a main suburban line would be about 30 or 40 km, with Pakenham, at 55km, the longest.
Let's say that you add one return trip per day to the schedule of a line that needs 1 hour run time to get to the end and an hour back. You'll need 730 in-service driver hours per year (365 x 2). One driver won't be enough (unless you only ran that trip 6 days / week) while two would be inefficiently used (unless you added trips on some days). Add more trips on more lines and it gets less lumpy and more efficient.
A few trips makes a big difference
Last year I mentioned that you don't need to add many trips to dramatically reduce the longest waits for trains, eg to cut maximum waits from 30-40 to 20 minutes. The corollary is that you don't need many train drivers either, though I didn't then estimate that.
One of the examples used was Craigieburn - a busy but lowly served line with many 30 and even 40 minute maximum waits. The line abuts and serves Transport Minister Ben Carroll's seat of Niddrie. As you can see below 30 minute waits from the city start at 7:39pm on weekdays and earlier on weekends.
Only about 17 extra timetabled return trips per week (34 each way per week) are needed to reduce Craigieburn line maximum waits from 30-40 to 20 minutes each day until after 9:30pm. This weekly total could be made up of two return trips per day (between approx 7:30 - 9:30pm) and three return Sunday morning trips to commence the 20 minute frequency earlier at 7:35 am rather than 9:35 am.
Benefits include reducing the maximum wait on four key train lines to 20 minutes until after 9pm 7 days per week. This makes for a much more useful network compared to now where 30 - 40 minute waits are widespread at times many wish to travel. It gets cheaper (or you could do more) if peak frequencies are slightly adjusted if it is considered that some of those who moved to working from home remain working from home. And if money is tight you'd only do Craigieburn, Mernda, Sunbury (possibly to Watergardens only), with Upfield deferred.
A higher increase (a still modest 2 - 3%) could enable upgrades right across Melbourne. Including for Belgrave/Lilydale, Hurstbridge (part), Pakenham/Cranbourne, Glen Waverley and Sandringham lines. As some of these have more service for more of the time fewer extra trips are needed to get to a 20 minute maximum wait. For example Sandringham just needs a handful of Sunday morning trips to qualify.
Conversely Belgrave/Lilydale have 30 minute gaps daytime interpeak, not just at night. Service reform here though could pay big dividends, including 10 minute Ringwood frequencies and benefits for some of Melbourne's most marginal seats.
How many drivers would a 'Getting to 20' maximum wait need? This should be closely related to the number of trips and service hours added (which will be more on longer lines). A 1% increase in the number of trips should need about 1% more drivers. Or maybe 12 - 15 up on the 1200 we currently have. That could deliver 7 day 20 minute maximum waits until around 9:30 pm on several key lines.
You could do even more with a recruitment of (say) 20 to 50 drivers. That should enable worthwhile upgrades on the rest of the network, notably the 12 lines that missed out on the 2021 upgrades.
Priority after that would be the 10 pm - midnight period. This would completely reverse 1978's evening cuts and bring service levels up to what was introduced on the upgraded Werribee and Williamstown lines last year. We could learn from cities such as Perth which has a planning culture where they add what they can today (even if it's only a little) and incrementally add service in later timetable revisions until the desired standard is reached.
The 300 new drivers and station staff in Tuesday's budget is about 10 to 20 times higher than the numbers discussed above. Hopefully their commencement can be phased in with some upgrades starting sooner.
For example Union Station on the Ringwood line will be opening in 2023 and need a new timetable to reflect its replacement of Mont Albert and Surrey Hills. Hopefully the opportunity will be taken to simplify and expand services in the new schedule. And that planning will be sufficiently advanced to promise it later this year in the election campaign given the marginal seats around there.
More broadly, let's take a punt and guess that 200 of these jobs are for drivers. Let's also assume that their driving hours and annual kilometres are the same as the average for existing train drivers. If so then something like 138 800 trips per year (694 x 200) could be added to the timetable.
A figure like 138 800 more trips per year means even less to most people than when governments spruik how they're adding X number of trips per week. (Trips per day is an improvement, or, for a Metro system, a minutes maximum wait frequency figure). Comparing it to the 764 000-odd annual trips however adds perspective in that it's roughly an 18% service boost across the network.
18% is very significant given what we could do with a well-targeted 1 - 2% service uplift as discussed before. Instead of 'Getting to 20' it could be more like 'Getting to 10' on some lines. The 2016 Metro Tunnel Business Case could give some guidance with particular gains for Craigieburn and Sandringham but less so Sunshine (as discussed here). The Network Development Plan - Metropolitan Rail from 2012 also plotted a way forward with widespread 10 minute service by now if it was followed.
'Getting to 10' on a hypothetical line
Suppose you've already 'got to 20' with a 20 minute maximum wait on a line as discussed above. A 10 minute daytime maximum wait, all week, might be the next goal.
With evenings and Sunday mornings sorted, I will assume that peaks are already every 10 minutes or better. 'Getting to 10' during the day would involve about 8 hours worth of ten minute service on weekdays (9am - 4pm and some peak shoulders) and 12 hours worth on weekends (7am - 7pm). This 10 minute day / 20 minute night service would be similar to what the upgraded Frankston line got in 2021. It also matches 2012's Rail Network Development Plan and coordination framework .
That's roughly 64 hours of better service per week. I'll assume an existing 20 minute frequency of three trains per hour. This is 192 trips each way per week. Or, doubled and then multiplied by 52, about 20 000 trips annually currently. As we're just doubling the service that's the number you'd add to get from a 20 to a 10 minute frequency.
How many drivers? With 694 annual trips per train driver that's 29 more needed for that line. Very roughly as line lengths vary and longer lines will need more drivers.
What about more lines?
A comprehensive upgrade would deliver such 10 minute service on at least parts of five major lines currently with 20 minute interpeak service. These include Werribee, Sunbury, Craigieburn, Mernda and Hurstbridge (noting infrastructure constraints on sections).
Now you're up to nearer 150 drivers. And that doesn't include the number needed to 'get to 20' as needed on five of the six.
Two further lines (Glen Waverley and Sandringham) are also in the mix. However they are shorter and start from a higher base on weekdays with a 15 minute interpeak frequency applying. That makes them cheaper to upgrade with fewer drivers needed compared to the 20 minute lines.
I've left branches like Williamstown, Altona, Belgrave, Lilydale, Pakenham (Pakenham East?) and Cranbourne (Clyde?) at 20 minute frequencies. However given their growth the last two might justify 10 minute service in their own right with a 5 minute frequency from Dandenong in.
I've also omitted Upfield due to my emphasis on off-peak uplifts that don't increase the peak train requirement. Adding a 10 minute off-peak service would look odd if we didn't also boost peak service. But the Upfield line serves a dense catchment that justifies improvement. And it would be a local transport game changer, relieving pressure on the 19 tram that people prefer due to its frequency (despite the slowness). Frequent service on the Upfield line has constraints but it's possible on its inner part.
Within a few years of the Metro Tunnel opening it will be time to prepare for the 10 minute service that will run from Melbourne Airport.
Conclusion
The boosted numbers of new train drivers and station staff earmarked in the state budget is both necessary and beneficial. My 'back of the envelope' estimates indicate that it is transformative, enabling widespread 10 minute frequencies across Melbourne's rail network. This would transform how we see and use trains in Melbourne with knock-on benefits for tram and bus connectivity.
Provided it has a large contingent of drivers put to work as soon as trained, you could rate it as amongst the most important items in the whole budget as regards transport. I'd even rank a system-wide 'Getting to 10' service policy as rivalling the Metro Tunnel itself in network importance. This is because when you make something better more people will use it. A large part of that 'better' in public transport is frequency.
Having said that we should also not ignore what you can do with 1% (and especially 2 - 3%) more train service. This could enable many smaller 'Getting to 20' upgrades with waits of no more than 20 minutes until much later at night across the network. Again vastly better than now with widespread 30 - 40 minute waits on all but three lines.
Consistent with best practice elsewhere, extra frequency should be scheduled as soon as sufficient train drivers become available. This could enable a staged uplift (eg by rail group and/or only to 9:30pm initially) in the years leading up to the Metro Tunnel's opening and bring forward its benefits.
See other Building Melbourne's Useful Network items here.
2 comments:
Well, they are meant to be running 'more trains, more often' with that extra inner cross-city capacity!
It will be interesting to see what happens both sides of the Metro Tunnel - unless it's a white elephant, the West Footscray turnback would mean lower relative frequencies on the western side - at least until Airport rail opens (which will only increase western frequency at Sunshine and Footscray).
Before Wyndham Vale and Melton electrification happens, could frequency in the Sunshine corridor be increased by extra deisel services that were basically suburban (even if they only stopped at Sunshine and Footscray)?
Ultimately Sunshine would have four suburban branches, so a 10-minute frequency on all four would see 24 trains per hour. A far cry from the current!
Update: The Big Build website says that 'more than 100 train drivers' will be recruited.
https://bigbuild.vic.gov.au/news/metro-tunnel-project/train-drivers-wanted-for-the-metro-tunnel
Post a Comment