Average speed is important when identifying if a new or upgraded rail service will appeal to potential travellers. One reason so many passenger trains have vanished throughout New Zealand is because rundown trains got slower and slower, so fewer people rode them, which officials and politicians used to justify reduced timetables, so even fewer people rode them, and so they got cancelled instead of upgraded. As you might know, I’ve detailed that story extensively in Can’t Get There from Here, which happily landed on bookshelves last month (and can be ordered online here). I allude briefly to average speeds on a couple of routes in the final chapter, which looks to the future. I’d like to dig into that some more today.
What is average speed and why is it important? I’m talking about the overall end-to-end speed of a train from origin to final destination, accounting for all stops, slow curving sections, steep climbs, track in poor condition, and fast straight lines across plains. This is much more important than maximum speed: who cares if your train can do 200km/h if it can only reach top speed for about 30km of a 400km journey, and spends a solid 100km snaking through hills at 50km/h because of tight curves and poorly maintained track? A train that can only hit 150km/h but maintains that speed regularly throughout its journey is going to be much quicker overall.
In one discussion on Twitter last week, @RuralPesant noted how long a number of rail journeys in New Zealand would take if an average speed of 110km/h were possible. I chipped in a few more.
(Note: he meant “NP”, New Plymouth, but I thought by “PN” he meant he lacked the specific distance for Palmerston North–Marton, which is necessary to know if you want to calculate Wellington to Whanganui or New Plymouth)
I decided I’d take this one step further. Let’s see what sort of timings are possible at a few average speeds and talk about how realistic they are. I’ve chosen four benchmarks:
- 80km/h: a reasonable average speed for new trains with modern braking, acceleration, and tilting technology, especially if electric. This is the average speed for Queensland’s electric tilt trains from Brisbane to Bundaberg and Rockhampton, which have a maximum speed of 160km/h and operate on the same narrow-gauge track as New Zealand. Straightforward upgrades to existing track will enable this average speed (indeed, such upgrades are in many places decades overdue).
- 100km/h: a good target for new trains with a top speed of 160km/h—this is the average speed of Western Australia’s Prospector service from Perth to Kalgoorlie, and Victoria gets into the 90s km/h with VLocities on its Ballarat and Bendigo lines. This would require a significant nationwide programme of upgrades (e.g. better track, deviations to make curves gentler).
- 120km/h: this would require greater upgrades still, more expensive deviations of twisty hilly sections, etc., but would be appropriate for busier routes and maximise the potential of trains with top speeds of 160–180km/h.
- 140km/h: I included this as an example of what would be possible for trains with a top speed approaching 200km/h, although in the NZ context this likely requires unjustifiably large expenditure for most routes. Above 200km/h, new dedicated high-speed alignments are necessary rather than simply upgrading existing lines, and these are not financially viable in New Zealand.
Where possible, I have compared these benchmarks with the average speed of present or historical trains. They come from 2021 (operational routes), 1993 (services cancelled in 2001–02), or in the case of northern Taranaki and Gisborne, a 1981 timetable (cancelled 1983 and 1988 respectively). The timetables I have for services cancelled before the 1980s are not really suitable for comparison, especially as the nature of the services was so different—steam-hauled “expresses” typically stopped more frequently than modern services would.
Rail distances are sourced from the New Zealand Railway and Tramway Atlas: the most recent edition is from 1993, but little in terms of route length has changed so the distances should be close, if not exact. I have provided road speeds and distances based on searches around 1–1:30pm NZ time on Friday, 7 January 2022: the distances are rough because my searches were as crude as “Wellington to Auckland”, so they are based on where Google chooses to place the start and finish points in each city and the route it thinks is the quickest between them. Speeds vary: I suspect if you did searches at peak you will get slower times. I’ve re-run a few searches at off-peak times and in some cases my results have been 5–10mins faster or slower than what’s below—so I think they’re reasonable estimates.
Here is my overall spreadsheet for your viewing pleasure:
Or, if you prefer them ranked from shortest to longest duration:
(My thanks to @zahribeth for help when I got stuck on a formula!)
Let’s take a closer look at existing/historical services versus current car times:
First, as they stand, many rail routes are longer than road equivalents. This was often not true historically but roads are now more direct because of heavy funding imbalances in favour of new highways and the sustained failure to upgrade, straighten, and deviate lines so that they can best serve current and future needs rather than those of the nineteenth century.
Consequently, New Zealand’s trains of the past half-century have generally not matched road for speed or journey time. Trains have averaged between 50–70km/h, hardly much to inspire the time-poor traveller. At present no medium-/long-distance train in New Zealand maintains an average speed of even 65.5km/h. Of all those to operate in the past 3 decades, only one has (just) exceeded an average speed of 70km/h, the Kaimai Express to Tauranga. It had a flat route and could race through the Kaimai Tunnel, but suffered compared to road because of its circuitous route via Hamilton (see pp.114–15 of Can’t Get There from Here for the nearly-built bypass that would’ve avoided this). No service in New Zealand has run faster end-to-end, although it is worth noting that the Southerner maintained an average speed of almost 80km/h between Christchurch and Ōamaru (the twisty coastal section between Ōamaru and Dunedin really dented the train’s overall time). Even with stops along the way, this is equivalent to how long a car takes today to do the same journey on State Highway 1 without stopping.
Notable are two Wellington commuter services. The Capital Connection from Palmerston North is not fast, at an average speed of 65.39km/h, but even though it has regular stops, is hauled by 40-year-old diesels with sluggish acceleration compared to Wellington’s electric multiple units, and the carriages are secondhand from Britain and now half a century old(!), it can match a non-stop drive—clearly this corridor deserves more than one train each way a day! As for its Wairarapa Connection counterpart, the Remutaka Tunnel ensures the rail route to the Wairarapa is much shorter than the Remutaka Hill Road. Even with similarly ageing locomotives and carriages and a slow average speed because of frequent stops for commuters, rail and road have equal journey times.
Now let’s see what’s possible at various average end-to-end speeds and, again, how these would compare to a non-stop drive:
With average speed boosted simply to 80km/h, 9 of the 20 routes above are faster than a non-stop drive and 6 more are within 20 minutes of the road time. This is a speed that our trains should be able to maintain already, a speed utterly unremarkable in any country that has had more balanced investment in transport over the past half-century. The reason I made the Australian comparisons when setting out my benchmarks is because its state railways have similar historical funding disparities and neglect, and yet trains in three separate states have at least been able to reach this benchmark. Queensland has done so on the same narrow gauge as New Zealand, and its electric tilt trains will hit even faster average speeds whenever proposals to upgrade slow, congested sections of the North Coast Line finally come to fruition. There is no good excuse for New Zealand’s trains not being able to maintain 80km/h end-to-end.
At 100km/h, an attainable average speed for most if not all of the sampled routes, only two fail to exceed the non-stop road time: Auckland–Hamilton, for which they are equal, and Auckland–New Plymouth, where the train has to take a circuitous inland route—and still is only two minutes slower than driving non-stop along State Highway 3. Auckland–Hamilton, moreover, is one of the most obvious candidates for a much more rapid rail journey: the two higher benchmarks should be eminently possible on this route (such as in the Regional Rapid Rail proposal from Greater Auckland).
The higher benchmarks are not realistic for all routes. Masterton–Wellington, for instance, is a commuter service that needs to stop frequently. It will not achieve an average end-to-end speed of 120km/h, let alone 140km/h. But, if we can lift it to 80km/h, it is half an hour quicker than the equivalent drive, and 100km/h would bring Masterton within an hour of the capital—especially if a revived western Hutt link between Melling and Manor Park allows it to be at least partly separated from all-stops Hutt Valley commuter services.
Services from Christchurch to Invercargill and from Wellington to Auckland won’t beat road’s times if we just aim for the 80km/h benchmark, but these are trunk routes with multiple major intermediate destinations and larger investment in higher speeds can be justified. They are also routes long enough where the non-stop measure does not reflect the reality for many car journeys, as travellers are likely to stop along the way for fuel, food, and toilets—but on the rails, they could sit back and feast while the train rolls on. Significantly higher rail speeds are possible on lengthy portions of these routes, balancing out slower timings through twistier sections.
This leads to an important point. You might look at these timings and think “how on earth can Auckland to Wellington or New Plymouth maintain these speeds when they pass through such rugged country in the central North Island?” The costs would be eye-watering for the sorts of work required in the heart of the island to permit sustained speeds much above 100km/h. This is, though, exactly why I am drawing attention to average end-to-end speed. With sufficiently fast speeds in flatter and more densely populated areas, where investment is a) cheaper and b) benefits multiple services, these longer routes can race through with minimal stops. If you can average 130–140km/h for half or two-thirds of the journey, being stuck on 80–100km/h for the other half or one-third is less of a big deal and still permits competitive end-to-end times. Wellington–Auckland trains, for instance, might not stop between Paraparaumu and Palmerston North so that they can sit on their top speed across the plains of Horowhenua, with the Capital Connection providing connections for destinations in between.
The selection of stops is crucial, as the time delays of slowing, waiting, and accelerating need to be balanced with serving adequate markets. In recent decades, one of the rare few steps taken in New Zealand to improve overall journey times has been the least constructive: cutting stops. When there are no connecting services, cutting stops is detrimental to patronage by removing potential journeys. In an integrated network, people from smaller stations catch local trains and connect at hubs with long-distance expresses. But in New Zealand, for example, the Overlander ceased to call at Te Kauwhata, Huntly, Te Awamutu, Waiōuru, Taihape, Ōtaki, and Waikanae in April 2005. At the time, no other trains served any of these towns except for the Capital Connection to the last two, and that train didn’t connect with the Overlander at all. Even with those stops cut, the Overlander remained a slow trundler, so this change was the worst of both worlds: potential journeys were eliminated without improving the average speed enough to make the train more attractive to end-to-end travellers!
One other great thing about upgrading the network for passenger trains with average speeds of 100km/h is that this will deliver network-wide gains to all users. Railways in New Zealand (and similar countries like Australia) often struggle to secure necessary public investment in additional passing loops, signalling upgrades, better quality track, etc. Politicians love to turn up to cut the ribbon at a new station—it’s easy PR, a facility everyone in the community can use. Standing in a paddock to point at new, heavier rails is rather less sexy, even if those rails permit longer goods trains to run faster and more often, attracting lucrative freight contracts. Because faster passenger trains in New Zealand will use the existing network, the necessary but less visible upgrades that these trains require—better signalling, heavier rails, concrete sleepers, eased curves, improved drainage, and so on—will deliver major benefits for goods, and politicians still get their moment in front of the camera cutting ribbons at reopened stations. Never ever let anyone tell you the demands of freight militate against passenger rail: upgrading the network for one should also benefit the other!
Faster trains are not simply plausible but necessary. Genuine high-speed rail—as in bullet trains—is not viable for New Zealand but we don’t need that: we can have excellent transport simply by upgrading our existing conventional rail network. Passenger trains with timetables that outdo non-stop car journeys are so realistic that we should have them already. If we strive for intercity and regional services to average 100–120km/h, they will be very attractive to the travelling public: the most time-sensitive will still fly, and some journeys will always require automobility, but rail will become the most appealing option for many journeys. New Zealand’s national mobility and accessibility will be transformed, travellers will have wider choices and flexibility, and we will make big strides to meet climate and environmental goals. The sooner these journeys are a reality, the better.
Update, 13/1/20: Stu Donovan has made a very salient point that I ought to have included in this entry: “increased speed not only directly benefits passengers through reduced travel-times but can also reduce the turn-around times for rail rolling stock—enabling the latter to make more runs per day and yielding indirect benefits in the form of higher frequencies.” One of the key factors that I outline in Can’t Get There from Here is how poor frequencies drove away patronage in the past. Frequency, after all, is freedom, while a slow train that runs only two or three times a week suits a very limited market indeed.
Dr André Brett 