This video is inspired partly by a shot in the House Of Cards titles, and partly by something my dad used to tell me — that you could stand at Richmond station in the evening peak and see trains on every track coming out of the city.
He may have been exaggerating a tad, but it’s often been said that in decades past the rail system had more trains running on it than at present.
That’s true to an extent. The inner part of the network was probably more intensively used in the past, though the outer sections of the network are busier than before.
For instance in 1939 in peak, trains out to Oakleigh were every 5-10 minutes, but out to Dandenong only about every 15-20 minutes — today they’re about every 5 minutes all the way.
And there’s been substantial growth in train numbers in the past ten years.
1939 vs 2006 vs 2015
How does the network compare overall?
I compared the 1939 timetables with 2006 (just as patronage, service growth and fleet expansion began to take hold) and also with the current 2015 timetables (last revised in 2014). I used departure times at the cordon stations (Richmond/North Melbourne) in the hour 5-6pm. (Note I’ve moved the range slightly where excluding a train a minute or two outside it would have given an artificially low or inconsistent figure.)
The verdict? More trains ran in 1939 than now, but not if you discount the Port Melbourne and St Kilda lines, which were converted to tram lines in the 1980s.
- In 1939 it was 116 trains in that hour, but 15 of those were on the St Kilda and Port Melbourne lines (and another 7 were on the Inner Circle and Kew lines which no longer exist; but those trains also serve some stations that do still exist). So a reasonable figure to use is 101.
- In 2006, the number had dropped to 87.
- But by 2015, it had risen again to 109, about 8% higher than in 1939.
But the balance of trains has changed.
In separating them out into the graph, I’ve used the old line groupings, because it more clearly shows the changes:
- Northern (eg lines through North Melbourne, including western suburbs) is up, though individual lines have changed in different ways. In 1939 the Williamstown and Upfield lines had a lot more trains than at present. This is countered by huge growth in the Werribee and St Albans/Sunbury lines — a fourfold increase in both, reflecting that they now serve growth corridors.
- Clifton Hill is now about the same as in 1939 (if you exclude the Inner Circle), showing growth since 2006 primarily on the Thomastown/Epping/South Morang line, which almost doubled in peak services between 2006 and 2015, though that line is still slightly short of the 1939 number.
- Burnley lines are up, 25 to 29 — Kew trains have been replaced by other services. There’s been basically no change since 2006, which reflects that patronage has grown more slowly. Almost all the other lines on the network serve growth corridors.
- Caulfield is slightly up, though Frankston and Sandringham line numbers are about the same now as they were in 1939. The real growth is on the Dandenong line, which even just since 2006 has grown by 40%.
Update: This July 2008 paper from Paul Mees notes that in 1929, 113 trains ran in the busiest hour (including the Port Melbourne and St Kilda lines) — slightly fewer than in 1939. It was 108 in 1964 (also including those lines), and a 1969 prediction forecast 181 trains in 1985 (including the Doncaster line).
How full are the tracks?
The bottleneck in the CBD is basically the number of tracks emerging from the City Loop and direct from Southern Cross and Flinders Street.
How full are those tracks in that hour?
“Full” is hard to measure. A rail line signalled for 2-minute headways (which the City area is) could be considered full at 80% of that, eg 24 trains per hour. But loading times at stations, junctions along the way (especially flat ones), level crossing gate closures, and signal/track capacity further out all reduces that. And if we’re measuring outbound trains, then how many inbound trains can we feed in from the suburbs (given little stabling in the City area)?
- Northern direct tracks to Newport (Werribee, Williamstown lines): 11 trains, but these tracks are used by 5 Geelong trains in that hour as well. Regional Rail Link will help free them up for more suburban services. Then the problem will become the single track through Altona (often the cause of delays and bypasses), flat junctions (principally at Newport), and level crossings (at locations such as Yarraville, the gates have been known to close for almost 20 minutes at a time).
- Northern Loop (Sunbury, Craigieburn, Upfield lines): 21 trains (plus 2 more trains to Seymour mix in at North Melbourne). This shows the value of having moved the Werribee line out of the Loop back in 2007 — allowing more trains to run in peak on all these lines. But the tunnel is close to capacity — one proposal sees Upfield line trains run direct from Southern Cross instead, and longer-term, if the metro rail tunnel is built, Sunbury line trains would use that. Of course, the single track on the northern section of the Upfield line may pose problems until duplicated, though ten minute services were provided as far as Coburg during the 2006 Commonwealth Games.
- Clifton Hill Loop (South Morang, Hurstbridge lines): 15 trains. Some trains run express Jolimont to Clifton Hill, which reduces capacity somewhat, though the flat junction at Clifton Hill also makes it difficult to run trains through it at a consistently high frequency (outbound Hurstbridge trains may have to wait for inbound South Morang trains, and vice versa). Ditto single track on the Hurstbridge line.
- Burnley Loop (Lilydale, Belgrave, Glen Waverley lines): 20 trains. Mostly expresses to Glenferrie, Camberwell, Box Hill and then out beyond Ringwood, but also 6 Glen Waverley trains. This Loop is also therefore close to capacity. Single track on the outer ends of the Belgrave and Lilydale lines can lead to delays.
- Burnley direct (Alamein line, Blackburn and Ringwood stopping trains): 9 trains. It’s not hard to see why in the long term, the plan seems to be to move Glen Waverley trains out of the Burnley Loop, and allowing more trains to run both to there (direct from Flinders Street) and also to Lilydale and Belgrave.
- Caulfield Loop (Dandenong line, Frankston stopping trains): 20 trains. Approaching capacity. 2 V/Line trains mix in with the Dandenong line trains at Richmond. The Dandenong line upgrade should help resolve level crossing issues, though duplication of the Cranbourne line is not currently in scope.
- Frankston direct (Frankston express trains): 5 trains. Clearly scope to move more Frankston trains to run direct, replacing them with Dandenong line trains, but it would be a fine balancing act to ensure large numbers of Frankston line passengers wanting the Loop were handled well.
- Sandringham direct: 8 trains. It’s not a growth corridor, but peak patronage does continue to grow, probably reflective of demographics (lots of CBD-based white-collar jobs). So there is scope there for an increase in services, though at some point the single platform at Sandringham becomes an issue. The old solution of terminating some trains at Brighton Beach might be the solution, unless Sandringham is upgraded with a second platform.
Other things to bear in mind
Trains in 1939 didn’t have the same capacity as those running now. 6-car trains now are slightly longer than the old 7-car Tait trains (which necessitated minor platform extensions as newer trains were introduced) and have more standing space, so the overall capacity of each train would be higher now.
Old-timers sometimes say that automation and modern technology has reduced capacity: for instance, they say that Tait trains with 9 (small) doors per side could load and unload quicker than modern trains with automatic locking doors, and blokes throwing switches for signals and points could respond more quickly as trains went past than the computers now controlling the infrastructure.
Required capacity on each line is reflective not only of population growth, but also of the number of people employed or attending education or other activities in the inner city and CBD, and needing to travel during peak hour.
As noted above, capacity of individual tracks has many factors, including signalling, dwell times at stations (which worsens as crowding gets bad), stopping patterns (consistent = good), train acceleration/deceleration, level crossings, junctions (preferably grade-separated, and the fewer the better), and even less visible factors such as capacity of the power supply.
In some ways the City Loop didn’t add greatly to CBD rail capacity. It helped distribute passengers around to more stations, and reduced the need to reverse trains. And associated upgrades (such as the “new” viaduct from Flinders Street to Southern Cross) did expand capacity. But ultimately the number of tracks out of the city stations to North Melbourne, Jolimont and Richmond is what determines track capacity in the City area — which is part of why Regional Rail Link went ahead, and why the metro rail tunnel is being pushed.
Average trip lengths are now longer than they were in the past. This means more demand for express trains (which burns up capacity if provided), as well as a bigger fleet and more staff needed to run the same frequency of service. More longer trips may also emerge via the fare cuts that took effect in January.
Where to from here?
With expansion of the CBD and Docklands, and a strong and growing economy (particularly the “knowledge economy“), demand for train travel into the congested core of the network is likely to continue to grow into the future.
PTV’s Network Development Plan’s proposals show the way forward, in terms of expansion of signal upgrades to High Capacity Signalling, high capacity trains, and re-organisation of lines through the city area to form dedicated high-frequency lines (including capacity expansion such as the metro rail tunnel).
All of which is expensive, but it’s got to be done — more than ever the rail network underpins Melbourne’s economic growth.
And remember that expanding evening, weekend and inter-peak into a 10 minute all-day service can be done far more easily, by making use of track and fleet capacity already available. This can help spread peak loads, by providing a much more usable service outside peak times, and helps to grow patronage when there is more spare capacity on the network.
(What have I missed? Leave a comment!)