Friday, August 02, 2019

Building Melbourne's Useful Network Part 14: Canterbury Rd for not much money

Where are the biggest Useful Network gaps in the middle - eastern suburbs, i.e. around Box Hill, Ringwood and Glen Waverley?

The Elgar/Huntingdale Rd and Middleborough/Stephensons Rd corridors south from Box Hill are major contenders. Serving major destinations, they already have direct, well-used but infrequent routes (767 and 733). They don't require much done to surrounding routes to reform them. I discussed them in Useful Network Part 3

There are also gaps radiating from Glen Waverley to Ringwood, Knox City and Dandenong. All have existing but infrequent routes (742, 737 and 850 respectively). Proper reform involves changes to surrounding routes to avoid duplication.  It's complicated. I'll cover these in a future week.  

Finally there's Canterbury Rd. Almost alone for a main road, it lacks its own route. That's unlike its mirror-image north of the railway that has the 270 between Box Hill and Mitcham every 20 minutes. 

Instead Canterbury Rd has portions of about a half dozen routes on or near it. Travel very far along it involves convoluted bus-train-bus connections or, from some parts, a ride on a very indirect bus route.

Canterbury Road has been known for years as a missing public transport link. The government-commissioned 2010 bus service review recommended a new main route along Canterbury Rd. And the Public Transport Users Association proposed a SmartBus along it. Both proposals got nowhere. 

It's true that Canterbury Rd's western bit, between Camberwell and Surrey Hills, is close to the parallel Ringwood train line.  However the middle part is beyond rail's catchment with various second and third tier shopping centres along it. Direct travel to Ringwood or Box Hill can be quite difficult from sections of it. Then in the outer east around Bayswater there's lower density residential and industrial land use.  

Which part of Canterbury Rd?

Here I will differ from the 2010 bus review proposal for a long Canterbury Rd route from Camberwell to Heathmont. I don't think all of Canterbury Rd necessarily justifies it due to duplication with the train in the west and low density in the east. And it would miss the key centres of Box Hill and Ringwood.

But the middle part of Canterbury Rd definitely deserves a service. Preferably more frequently than the 30 minutes suggested in the bus review. Especially if the bus went to the area's biggest trip generators, like Box Hill and Ringwood. 

Serving Box Hill really is necessary. Both because of what's at Box Hill and the train connections available. Box Hill is so big and generates enough public transport trips that the bus routes that serve it are amongst the best patronage performers in Melbourne. This even goes for the 732 that has a long overlap with the frequent tram service along Burwood Hwy.

Access to Box Hill from most of Canterbury Rd is difficult on the current network. A dedicated route along it would connect Box Hill to a much larger local catchment than exists now. Even if access to some other centres is made less easy than now the better access to Box Hill could offset this.  

In contrast, the alternative destination of Camberwell is further, smaller and requires more route kilometres to serve. If you wanted a Canterbury Rd service from Camberwell one might reroute the 612 along there and reroute the 766 north to cover the existing 612 territory up Union Rd. 

The advantages for Box Hill apply for Ringwood on a smaller scale. One can quibble exactly where the route joins Canterbury Rd at either end. But apart from that the route almost draws itself. Very roughly it looks like this (I've called it the 730). The translucent line lets you see what existing routes are underneath (click for a clearer view).

There's other benefits as well. You would think that a destination near the corner of two major roads would have access to at least two bus routes to allow travel in all directions. However this is not so for the council's aquatic centre, due to the lack of a bus on Canterbury Rd. Currently this is served only by the north-south Route 902 along Springvale Rd which doesn't help most of its local catchment. A Canterbury Rd bus would provide good east-west connectivity to the centre and allow easier travel for other local trips eg Forest Hill Chase to Brentford Square.

A Canterbury Rd bus route would benefit three lower house seats. All are marginal. These include Forest Hill (Neil Angus MP), Box Hill (Paul Hamer MP) and Burwood (Will Fowles MP - on leave).

Paying for it

A Ringwood - Box Hill bus route along Canterbury Rd is about 14 km long. Buses go at about 22 km/h. So each trip is about 2/3 of an hour, or 40 minutes, each way. Two buses could provide a 40 minute service, three buses could provide a 30 minute service while four buses could provide a 20 minute service. You might want to add one or even two more buses for better reliability in peak times or a higher peak service.

You need a couple of million dollars or so to run the 20  minute service. That's assuming you're just layering it on top of what's already there. Which, regrettably, is recent but inefficient practice. Especially when we are told by the same ministers and governments that accept this inefficiency that there is no money for better bus services.

Step One - chase the overlaps

We can do better. The first step is to see whether the new 730 overlaps existing routes, and whether savings can be made by making them shorter.

The main overlaps are between Box Hill and Forest Hill Chase (zoomed in below). The most significant is with 765 which deviates via Blackburn Station. 765's indirectness shows how inefficient access between Box Hill and Forest Hill Chase currently is; something that the new 730 would fix.

Another smaller overlap involves the 703.  And some others further east. The latter would require  network changes to fix. They're small so it might not be worth doing much about them for now.

Below shows what one could do with the 765 and 703. The new Canterbury Rd route replaces the 765 to Box Hill. 765 can instead start from Blackburn and be made simpler. Admittedly it means that a few lose a one seat ride to Box Hill, but it was very slow anyway. Those near Canterbury Rd get a much better service, and other areas are near the 732, 733 and 735. There seems no reason why it couldn't be run up Main St full time to make it simpler. 703 can then also become simpler, running the full length of Blackburn Rd. It might be marginally faster, which helps reliability.

Pruning the 765 can help with the financing of the new Canterbury Rd route. The extract of the timetable has it taking more than fifteen minutes between Box Hill and Blackburn. That's slightly over one bus for its 30 minute off peak frequency, and nearer to two buses during peak periods (where more service runs). 

The other potential saving involves the 703. Not running it via Forest Hill Chase saves a few route kilometres, provides much more direct access to Blackburn Station and gives a consistent all-day service. However peak trips already skip the shopping centre, so you might not be able to free up any buses.  But the extra time might help reschedule the 703 so it's an even 15 minute off-peak (which meshes better with trains at Blackburn) rather than the current weekday interpeak gaps of up to 20 minutes. And it might aid punctuality.

If you could find your three or so buses required (after the savings from the 765) you could start the Canterbury Rd route right away without further network changes. 

Need to find more savings to make it happen? Read on. 

Step Two - A dead end

Now it gets more complex. Like unravelling a ball of string. A change to one route in one area might have implications for other routes. And you can sometimes run into dead ends that turn you off the whole idea. Here's an example. 

South of and parallel to Canterbury Rd is Burwood Hwy. It has a frequent tram. And a bus. The 732. It runs to Box Hill approximately every 20 minutes. The idea was that you could flip the 732 to run to Box Hill via Canterbury Rd instead of Burwood Hwy, as per the map below. 

It seems logical to remove duplication to provide better service elsewhere. And, in this case, Box Hill, the end point and a major destination, remains unchanged. However the 732 also passes fairly near other trip generators including Deakin University. Rerouting it would cause more forced transfers at Vermont South. And Forest Hill Chase is likely to be a far weaker trip generator for those coming from Knox than Deakin University. Besides, despite the overlap, the 732's patronage is reasonably high. 

Other shortcomings include the Canterbury Rd route not continuing to Ringwood and the north-south overlap with the 736. That's not very useful. Unlike if the overlap went further north to Mitcham Station where it would provide an improved train feeder. But then the route would be less direct to Box Hill (if it continued there at all). 

Though interesting (and cheap), an approach involving the 732 had too many cons to be continued with. It wasn't good for existing 732 passengers and it didn't fully fix Canterbury Rd. 

Step Three - A puzzle

Having given up on the 732, the next possibility was whether one could reconfigure other existing routes to free up a bus or two and make the Canterbury Rd route more affordable.

Since the Box Hill - Blackburn part of the 765 had been replaced by the straighter Canterbury Rd route, would it be possible to use other routes to replace the remaining portion? 

There are several ways to do it. It's inevitable that some trips that are a single bus now would involve a change on a revised network. Conversely other trips would become easier and the network would become simpler, especially for those travelling along Canterbury Rd. 

The concept below is based on several big ideas. These are as follows: 

* A direct Box Hill - Ringwood Canterbury Rd route is very good and would help many local trips.  
* There is enough gap between Canterbury Rd and Burwood Hwy (1.6km) for there to be an east-west bus route roughly parallel to these. This should start at a popular destination (eg Box Hill) and end at a station to the east. It also needs a mid-route trip generator (even if the route has to kink a bit) to generate all day usage. 
* Thanks to road/rail grade separations the Ringwood train line doesn't have to be the big barrier to buses that it once was. The SmartBus orbitals opened things up but there is no reason why local routes cannot also cross the railway. 

The map below shows what such a network incorporating those ideas might look like.   

Most important is the 735 / 765 amalgamation. The western bit of the 735 is unchanged. As is the eastern part of the 765. But in the middle, around Forest Hill Chase, they are changed, joined and become one route.  It starts at Box Hill and runs as directly east as possible to Mitcham while serving Forest Hill Chase and Brentford Square. Directness is better at the cost of not serving Burwood One.

How many buses does merging 735 and 765 free? First we need to see how many they currently use. A return trip on the 735 takes roughly 90 minutes. That's 3 buses for its 30 minute off-peak frequency. Peak traffic slows return trip times to around 100 minutes. That affects the timetable, with a 30 - 40 min peak service offered.

Route 765 is a bit longer, taking about 100 minutes off-peak and about 10 minutes more during peak times. It's every 30 to 40 minutes off-peak, making it again three buses. But unlike 735 its peak frequency increases to roughly 20 minutes, increasing its requirement to about five buses (noting that there might be scheduling complications like sharing buses between routes as frequencies are irregular).

To summarise, both routes together require about six buses on weekdays and eight during peak periods. The revised 735 is longer than either 735 or 765 on its own. How much longer? I get around 17% longer than the existing 765. Four instead of three buses should do the trick off-peak.

To make the changes more attractive, with peak upgrades for 735 passengers and off-peak gains for 765 passengers you would want a 20 minute peak/30 minute off-peak frequency for the merged route.  That looks like around 4 buses off-peak and 6 buses peak. A 40 minute frequency could apply on weekends - generally better than now and harmonised with trains.

Compared to what runs now, the amalgamated route frees up about two buses. That compares with the one bus released when we just made the 765 shorter. Yes the amalgamated route misses Burwood One but it does restore the one-seat ride to Box Hill that the 765 shortening option removed. Directness to Forest Hill Chase from the west is improved and there are some frequency gains.  

Getting back to the ball of string, this creates some loose ends. Merging 735 with 765 at Forest Hill Chase means there is no longer a direct bus to Nunawading Station. And merging 765 with 735 takes a bus off Main St, Blackburn. 

Main St can be looked after by rerouting 736 there instead of going via the lake. That compensates for the removed part of the 765 and gives a handy Glen Waverley connection. The two remaining problems are the area just east of Blackburn Lake (which loses the 736) and Nunawading Station which loses the 735 connection.  

 What appears to be a solution lies to the north. The 273. It's not a well used route. The picture below explains why - much of its catchment is very low density. Nevertheless it still runs every half hour on weekdays, although its operating hours aren't very good.  

It runs south to Nunawading Station but not before a dogleg through Blackburn North (shown below) that would reduce its appeal as a feeder. All of the roads have other buses (270, 271, 303) so removing this would not leave any stops without a service. 

In our circumstance the dogleg represents potential energy that could be put to use. In this case running beyond Nunawading to Forest Hill Chase. Routing east of the lake replaces most of the coverage that the rerouted 736 removed. And it makes the Ringwood line less of a barrier for those travelling to Forest Hill Chase. A few stops near the lake wouldn't be served but the passenger catchment is low and the area is generally walkable to stations. 273 didn't have the dogleg and used to run to Forest Hill Chase back in the 1990s, so an extension would restore directness and a previous destination.

How would you compensate those in north Nunawading who lose the 273 'dogleg'? Springfield Rd has already gained, with buses added in the 2014 revisions to Route 270. While Junction Rd could be looked after by introducing Sunday service on Route 271 (neither route now on it runs Sundays). 

This step is more radical reform than putting in the Canterbury Rd route and making only minor changes as per Step 1. Some will like it while others will not. However, by keeping the number of routes in the area the same (deleting 765 and introducing 730) it lowers the cost. The benefits of the frequency gain and the Box Hill connection probably outweighs smaller minuses elsewhere on this network. 

Step Four - what frequency?

How frequently should the Canterbury Rd route run? I like 20 minutes. This qualifies it as a Useful Network route. And it would only need two extra buses if 735 and 765 were merged. While 20 minutes does not harmonise with trains (15 min base frequency), Box Hill gets better frequency than that for most of the day. And it's not an issue for the many who will be starting their trip at Box Hill.

15 minutes is another good choice if additional resources were available. Both 15 and 20 minutes would make it a higher tier route, something which doesn't exist in the area apart from the north-south SmartBuses.  In fact the only east-west routes with this frequency (or close to it) in the area are the 75 tram, the 732 bus (which overlaps the tram) and the train. That's it. North of the railway (with its government bus heritage) is better served with more frequent east-west routes, with the 901 between Blackburn and Ringwood and the 270 between Box Hill and Mitcham. 

30 minutes matches trains but is not ideal for what should become a well used main road route. Even  back street routes in the area run at this service level. Although parallel to the train line, Canterbury Rd is sufficiently distant for it to have its own catchment to support a good service. Nevertheless, if this cannot be achieved, even a 30 minute frequency Canterbury Rd route would be an improvement on today where too many trips are either indirect or require an unnecessary bus/train change. And it would be cheap, possibly needing just one extra bus if the 735 and 765 are amalgamated.


We have tackled Canterbury Rd, a missing link for buses in Melbourne's eastern suburbs. While it lacks the large trip generators that Burwood or even Maroondah Hwy has, it still has some medium sized ones like Forest Hill Chase and locally significant destinations such as the aquatic centre. Plus it forms a convenient corridor to pick up passengers going to the much larger centres of Box Hill and Ringwood. 

A new Canterbury Rd route could be layered over the existing route (at highest expense), added with small changes to existing routes (moderate expense) or added with larger changes to existing routes (lowest expense). The latter is more efficient and deliver wider benefits including frequency improvements. But it would need political guts to implement. 

What do you think? Do you like the concept of a strong direct route along Canterbury Rd? Would it get much use? And should the surrounding network be changed if it does get put in?

PS: Want to see expanded Useful Networks in other areas? An index to them all is here.

1 comment:

Tom said...

I agree that Box Hill and Ringwood are the better destinations, given their importance.

The 612 via Canterbury Rd means of serving the Box Hill-Camberwell section of Canterbury Rd, with covering 766 deviation, is possibly the best way of serving that section.

The bus review Camberwell-Heathmont proposal likely did not do it self any favours by describing it as a rail alternative for the Belgrave line, when access to local shopping centres is its main purpose.