I have advocated a more “sectorised” rail network, one in which rail lines operate more independently of each other, untangling what is currently an overly complex system. One of the downsides of this is that it will require some commuters to make a transfer from one train to another in order to complete their journey, where currently they do not have to do so. For example, someone going from Schofields to Central would have to take two trains instead of one as they currently can, as would someone going from Rhodes to Town Hall.
Ray made the following observation in the comments section (reposted in full):
“I don’t think Northern or Richmond Line commuters would ever accept being denied direct access to CBD stations. Why should they be singled out while every other line maintains direct access to CBD stations beyond Central? Why can’t some Western Line services also be terminated at Central? In the case of the Northern Line, which north of the Parramatta River is Liberal heartland territory, there would be hell pay for the local members of Parliament, for their inability to stand up in support of their constituents. There may be acceptance for some Northern Line services to terminate at Central, so long as the option of direct services through the CBD to North Sydney is maintained. Epping and Eastwood, which are both interchange stations, already have the benefit of express services from Newcastle and the Central Coast to Central terminus.
Whilst the transport bureaucrats might push for further sectorisation along the lines previously canvassed, the political reality is that ALL lines must have direct access to the CBD via either the link to the North Shore or the City Circle and they have to find a way to accommodate this operating pattern.” – Ray (8 April 2013)
He rightly outlines the reasons against such changes, particularly the fear that many people will refuse to transfer from one train to another. This is, all other things equal, a much worse outcome than a direct service. But it is the assumption that all else remains equal which is not correct here, and this is where I disagree with Ray’s assessment.
The current system is designed on the basis of direct trips, one train or one bus from your origin to your destination, and this means having services from everywhere to everywhere else. In reality, this is not possible, so what we have instead is services from some places to some other places, but still so many routes and lines that frequencies are often as low as every 15 minutes in the peak and every 30 or 60 minutes in the off-peak (if at all). This makes transfers difficult. Even on the rail network, only 49% of stations on the Cityrail network have a train to the CBD every 15 minutes or less all day (and this is being generous with some stations).
NOTE: When a network is not designed to use connections, then you need to focus on frequencies of the direct journeys. A network designed around connections facilitates transfers in a way that a network designed around direct services does not, which is why this frequency map excludes stations on the Upper Northern and Epping to Chatswood Lines, which have 15 minute frequencies, but with every second train terminating at Chatswood. If frequencies from Chatswood to the CBD were higher, then it would allow for easy transfers, and effective mobility for users of these stations.
A network based on connections, on the other hand, purposely limits the routes and lines available to you. But in doing so it is able to run more frequent services directly to an interchange, where you can then take another service to your destination. It is both predictable and reliable. Unlike the previous model, it is far more likely to provide an everywhere to everywhere service. The diagram below shows a generic example, taken from a Brisbane Government transport policy document.
In this example, the average trip is cut from 15 minutes to 9 minutes. This is due to 3 factors:
- Higher frequencies for local services, meaning a shorter initial wait time. Here the local service has gone from 30 minute to 10 minute frequencies, meaning an average wait time of 15 and 5 minutes respectively. While it’s true that you can use the timetable to minimise that wait time, it is still possible for you to be delayed, for your service to arrive early, or for that service to be cancelled. In each of those cases, you must now wait the full period (i.e. 30 minutes or 10 minutes). That is why the average wait time is still important.
- The main trunk service (red in the example above) can be converted into a high capacity, high frequency, and (optionally) high speed service. This again means less waiting time, which is important for the reasons mentioned above.
- The transfer must be simple, preferably cross platform where possible, minimising the additional total travel time.
Taken together, these result in a reduction in total travel time. The key here is higher frequencies, which allow transport based on connections to occur.
Using the proposed changes to the Richmond Line shows the actual improvements possible there:
Right now, there are 6 trains per hour on the Richmond Line (from Schofields onwards) in the AM peak – although there is a maximum 15 minute gap between trains. This would be increased to 8 trains per hour, but shifted to the Cumberland Line, and so no longer continue through to Central. In the off peak, there are currently 2 trains an hour, which would be raised to 4. The Western Line would also see 20 express trains per hour to Central in the peak, and 8 trains per hour in the off-peak. Travel times are 55 minutes from Schofields to Central in the peak, 62 minutes from Schofields to Central in the off-peak, 26 minutes from Schofields to Westmead, and 31 minutes from Westmead to Central.
Implementing these changes would result in an average time saving of 1 minute (maximum of 5 minutes) during the morning peak, and an average of 12 minutes (maximum of 15 minutes) during the off peak.
This would not be possible to do effectively by just adding more trains with the existing network. In the AM peak there just aren’t the additional slots available into the city centre, and creating them would make the network even more vulnerable to delays and cancellations than is already the case. In the off-peak, adding more trains is possible, but they would either have to be all stations in order to maintain a high coverage or be express and give a low level of coverage. Either way, neither option provides the frequencies necessary to allow for easy connections that allow the sorts of time savings provided by the proposed changes.
The lesson to take from this is that connections are not something that we should be afraid of. Not only will they provide faster overall journeys, but more reliable ones. And when they occur entirely within the rail network, they happen without the sorts of financial penalties that currently exist when you transfer from a bus to a train or from one bus to another bus because fares are not integrated.