Last week, in the closed session of the AT board meeting there was an item about running pattern of the rail network after completion of the City Rail Link. It sparked a bit of debate about the merits of the purple line. With this post I thought I’d look at why I don’t think it’s a good idea.
The basic version of the future operating pattern this is below.
A lot of people go to Grafton and Newmarket
The CRL will be benefit a lot of people and none more so than those on the Western Line. But the one group that will be the most negatively impacted will also be on the Western Line, those who travel to Grafton and Newmarket.
To start I thought it would be good to quantify just how many trips that is.
So, I turned to the rail station data that Auckland Transport have provided me in the past. It shows for the year to 30 June 2017 how many trips were made using HOP from each station to every other station. In total it represents about 17.6 million trips, or about 90% of all trips for that year.
The data shows that there were just over 3 million trips boarded at stations between Swanson and Mt Eden heading towards the city and a similar number of trips in reverse.
|To/from stations west Mt Eden||854,009||28%||852,737||28%|
There are a interesting observations from this:
- More people on the western line are travelling to other stations on the Western Line than to Grafton and Newmarket. This is a good thing regardless of what happens with the purple line as most can benefit from the improved frequencies the CRL can provide.
- You will notice there’s over a 100k difference between boarding and alighting at both Grafton and Britomart. We believe this is due to what we call ‘down hilling’. In other words, they’re getting off at Grafton in the morning and walking, cycling or catching a bus down to Uni but at the end of the day, instead of going back to Grafton, they keep heading downhill to Britomart. Post-CRL I assume these will mostly shift to the Aotea station.
This suggests that in total, about 21% of Western Line trips are genuinely to or from Grafton or Newmarket, or about 8% of all trips. Given western line trips to those two stations are already well served, we should expect even if the numbers stay about the same, the percentage of trips affected will drop as many more people will be travelling to the city.
However, we also need to consider that proposed purple line only goes as far as Henderson. Up to 14% of those trips are to or from stations west of Henderson. All of those passengers would need to transfer services anyway.
What are the travel time differences
Western line services to Grafton and Newmarket, at peak at least, are currently pretty good in the Auckland context, being every 10-minutes. That’s in large part because those stations are ‘on the way’ to Britomart. But given the numbers above, and infrastructure constraints, it’s hard to justify that level of service in the future. Even if you could find enough money to run the purple line at 10-minute frequencies, there would almost certainly be something better on the PT network to spend it on. The same goes for track capacity and even the trains themselves. As such, the purple line is only expected to run every 20-minutes, that’s outside Auckland Transport’s own definition for a frequent service and certainly not something you’d class as ‘turn up and go’. So, what does this mean for travel times?
For this part I’m going to use travel between Henderson and Grafton as my example. Trains are currently scheduled to take 32-minutes between Henderson and Grafton (if the train turned up just as you got to the platform). If however you just missed the train, at peak times you’ve got to wait 10-minutes for the next one for a total travel time of 42-minutes. How does this change after the CRL
In this scenario, turning up right on time would net you the same travel time, just miss a train though and you’ve got a 20-minute wait giving you a total time of 52-minutes. That’s a wide variation. Instead of waiting, more likely you’d just catch the next CRL bound train and just transferring at Karangahape Rd. Taking an average wait time would mean 42-minutes.
CRL at opening
AT opening, the CRL is said to have the capacity for 15 trains per hour per direction. More detailed versions of the plan above show a base pattern of 6tph on the red and blue lines with additional peak only services, three per hour from Henderson and three per hour from Papakura but travelling via the Eastern Line (which is the purpose of that peak only dotted line at Manukau above). So 9 trains per hour from Henderson through the CRL. That’s a train less than every 7-minutes.
CRL documents suggest trains would take about 3-minutes from Mt Eden to Karangahape Rd so about a 33-minute travel time all up. From Karangahape Rd it would take about 4-minutes back to Grafton. This gives us a best-case scenario of 37-minutes, 5-minutes slower than the similar scenario for the Purple Line. However, the worst-case scenario of just missing your train at both Henderson and Karangahape Rd, at 51-minutes, would still be faster than the worst case for the Purple Line. The average would be about 44-minutes.
CRL in future
It is expected that with further upgrades, the CRL would be able to move up to 24 trains per hour per direction. Assuming they’re evenly split across the lines that represents 12 per hour from Henderson. In this situation, the average drops to 42-minutes and the worst case to 47-minutes.
With faster trains too
We know that AT are planning on improving travel times on existing trains. We don’t know just how much time they’ll save but using the calculations I wrote about recently, travel times for Henderson to Karangahape Rd could see up to a 3-minute saving. That would benefit all of these options relatively equally but I thought it was useful to include to compare to the current state. It would mean the scenario above, with 12 trains per hour would a range of 34-44-minutes, only two-minutes slower than what we have now.
I’ve highlighted these travel times in the graph below. The line through the middle is the average travel time.
On top of the travel times, there’s also something to be said for a simple, easy to understand network. To be able to turn up at a station and catch the next train that comes along without having to think if that’s the one that will get you to your destination the fastest or have weird gaps in the timetable because one train every 20-minutes goes somewhere else.
A better use for the resource?
Given some of the travel times above, the purple line seems to be more about running trains to satisfy a model than running a simple, frequent network. As such, it goes against the very principles of the new network AT are rolling out right now.
In my view, the purple line is something we should only consider if we had spare money, infrastructure or trains available, something I don’t see happening. Even running the purple line as proposed isn’t going to be cheap. It effectively represents using about 1/3rd of the entire capacity of the Western Line to move what is already less than 1/3rd of the trips, with that percentage about to plummet as people flock to trains to the city.
At 28.6km between Henderson and Otahuhu, each run covers about 4km more distance than a service from Swanson to Britomart (post-CRL). That means that for the cost of the purple line, you could extend the initially planned peak frequency to cover the entire day. That would benefit everyone, not just those going west-south. It turns our rail network into much more of a metro style operation.
Another issue I have with the current plans, it’s absurd to spend $3+ billion to, among other things, shift the west up to 20-minutes closer to the centre of town and then not improve services to almost 20% of the Western Line. That’s because there is no plan to increase frequencies past Henderson with both the extra peak services and the purple line starting from there. As such, another alternative is that you use the purple line resource to send those peak-only services past Henderson. There’s probably enough operational resource to do both better off-peak services and sending them all to Swanson.
It’s not just running trains either, for example we will need approximately six trains to run such a service. At close to $9 million per 3-car train, that’s about $54 million in trains before you run anything. While we have trains now and more on the way, we’re going to need all of them just to service our existing network and the extra frequencies that will be needed in the future.
By now I think it’s pretty clear that I’m not convinced of the merits of the purple line.