Friday, September 13, 2019

Building Melbourne's Useful Network Part 20: Fixing 302/304 crowding, Box Hill and Templestowe

Those who write to politicians advocating public transport service upgrades soon become familiar with the standard form-letter replies. The answer is generally a diplomatically-crafted “no”. The government may “have no current plans”. They may say that they “continually review services”, “service provision reflects available resources”, or that “proposals must be considered against state-wide priorities”.

The 'expectation management' reality in which these replies are drafted includes fixed funding for service kilometres. In other words a decline in public transport service per capita due to a growing population and stagnant services (graph below).

One might think that this relative decline would increase interest in getting the most from what we do have. However it's been the reverse, with less network reform occurring now than five years ago.

The paucity of reform means that department staff who answer correspondence can't say much. Regardless of their private views, the reality they inhabit rarely admits the potential of doing better by redistributing service kilometres from poorly used or duplicative bus routes to services that need it. This is despite (i) buses being easier to reform than rail networks with numerous precedents here and elsewhere, and (ii) the current network falling well short of the best possible, even assuming unchanged service kilometres.

Current official wariness to bus reform risks the perverse outcome where more departmental time is spent writing to people explaining that they can't have a network improvement than reforming the very service that is the subject of much correspondence received.

Political risks

On a crude 'political economy' calculus, not reforming can be justified on the basis that 10 voters disadvantaged by a reduced service make more noise than 1000 voters who would have gained under a reformed network.

It's asymmetrical: losses carry more weight than gains. And actual losses carry far far more weight than potential gains. Hence, despite the large gains in Transdev's aborted 2015 network (eg frequency upgrades on the busier eastern parts of the SmartBus orbitals), the then new Labor government lost no political capital in dropping the whole thing (though we have yet to see the promised 'more balanced bus network plan').

In fact Labor probably gained goodwill as Transdev's proposed network (i) had substantial cuts in some populated areas with no offsetting gains and (ii) it had little legitimacy as public consultation was very limited (Protip: Presentations to stakeholders like local and state government officials, who are generally quite privileged people who drive cars do not count!).

Before you think that politics closes off prospects for bus network reform, consider what else happened in 2015. Regional Rail Link. This was accompanied by reformed bus networks in Geelong and Werribee/Tarneit planned under the Liberals but implemented by Labor. Both networks successfully brought more service to more people and have had trips added since. Factors allowing these networks to succeed (with general public support and few problems for the government)  include (i) a multi-operator/multi-modal approach and (ii) better public consultation.

Then there's another type of political risk. The risk of doing nothing. Especially when service is known to be mediocre and existing passengers are either being delayed or left behind by full buses.

Labor paid that price in 2010 when it lost five marginal seats along the Frankston line (Frankston, Carrum, Mordialloc, Bentleigh and Prahran) and thus government. Transport service delivery was a major issue, with people no longer able to rely on trains to get them to places on time.  Because service provision lagged patronage growth crowding became chronic, further exacerbating delays (graph below). 

John Brumby and Co realised this late in their last term but for them it was too late. Even a train timetable change requires a good lead time. Train drivers can't be got off the street at a moment's notice. And we needed more trains to fix the peak hour problems.

For the rest of us though, late far beats never. A new and more frequent Frankston line timetable delivered a turn-around in performance from 2012. This improvement has been largely sustained since.

It thus confounds me why the government hasn't delivered a similar new timetable for the Ringwood train line, which, since the 2018 election, has replaced the Frankston line as the central artery of Labor's marginal suburban seats. When good politics meets good service planning, why wouldn't you do it?

The seat of Box Hill is in that Ringwood line belt. This includes the eastern ends of today's topic - the crowded and unreliable bus routes 302 and 304. As well as some major education, shopping and health destinations like some major but poorly connected hospitals. We'll try to help those out as well. 

An expanded reality

Luckily sometimes one sees replies that open rather than close possibilities to deliver services people need. 

A recent Timetable Tuesday discussed bus routes 302 and 304 in Melbourne’s inner eastern suburbs. These were the routes whose riders (sick of being left behind) established a support group on Facebook to advocate for better reliability and more service.  This advocacy included writing to local politicians like Tim Smith MP (Kew) and Paul Hamer MP (Box Hill). 

The reply from the latter was better and clearer than average. It concisely mentions the two ways you can improve bus services – more funding and more efficient network design.

The 'additional funding' approach needs to go through the government's budgetary process. That's a complication in itself. Whereas the 'operational changes on other routes' might not require extra ongoing funding so in this sense is simpler if everything can be found within the transport portfolio.

On the other hand, pruning other routes adds complication and  political risk. Luckily we're three years from the next election. So upgrades could be running before then if planning starts now. Also, if you're imaginative with network reform you can spread the benefits wider than just the routes we're planning to boost for no or little extra money. For example hospitals at Box Hill or routes that enable improved connections to La Trobe University.

Addressing crowding 

Like all commuter-heavy CBD-bound routes, 302 and 304 are not cheap routes to run. A morning peak trip into town takes over one hour. Short trips from Balwyn take about 40 minutes. A single extra bus could really only do one am full length trip during a 2 hour peak window. But it could possibly do two short return trips if one is early and one is late in the peak. There's already one morning short trip from Balwyn. Depending on the loading profile maybe more short trips could be added to relieve pressure on the other trips.

Afternoon peaks are a bit longer so you might be able to have the one bus contributing frequency in the early and late parts of the peak, even on full length trips. In both the morning and evening retiming other trips would even out frequency and lessen crowding. I don't know if you'd run short trips in the evening. As people may wish to travel beyond a terminus, short-finishing trips are less user-friendly than short-starting trips. However you'd consider them if there are large loads alighting early as they may allow a more intensive service. 

I haven't gone into the detail but something like three or four more buses should provide noticeable crowding relief and capacity boost. Other things worth considering include articulated buses (which would add capacity but not frequency) and cheap evening upgrades (particularly 304) to bring services up to minimum standards (ie 9pm or later finish on both routes).

Similar comments with regards to route lengths, peak crowding and requirements for extra buses may apply to the 200 and 207. Keep this in the back of your mind as one of the potential benefits of some of the more radical network changes discussed later.

Where's the overserviced routes? 

So much for the 'give'. That's easy.

What about the 'take' to free up resources for the upgrades we want? That's harder.

And potentially controversial. But necessary if we want fast, cost-neutral gains.

Otherwise, we'd have to take a ticket and wait behind dozens of health, education, housing and welfare interests also wanting money from the budget process.

So buckle in and prepare for this quick fly over local routes and corridors. To speed things up we'll group related services together. We'll discuss their potential for reform by group.

Bad news first. There are routes we can't drain resources from. Either because they are major and busy (that if anything need more rather than fewer trips) or are already infrequent and give needed coverage.

Those with better prospects include routes that (i) have low patronage, (ii) sparse catchments and/or (iii) overlap other routes. Where they serve sparsely populated areas the number of people affected could be far smaller than those who a reformed network could benefit. And even losses can be mitigated if we extend some existing routes to new and useful destinations. It's easier to win acceptance of a service reduction if some silver linings can be shown as well.

Most routes discussed here are operated by Transdev (who also run the 302/304 that we want to boost). This makes reform easier as it means that any resource transfers between routes can be handled internally.

* Various inner east routes (284, 285, 548, 609, 612, 624): Only the first two are run by Transdev. None run very frequently. Five of the six are daytime only with little or no weekend service. Some of these routes have potential for upgrade as strong north-south routes with stronger termini, as discussed here. At the very least they need reform as part of a local area review including stronger termini and 7 day service. For this exercise I'd leave them alone unless a change is necessary to make an alteration to another route work.

* Routes 279 and 295: Both routes run north-south along Victoria St. The trunk portion of 279 enjoys a good peak service and a 15 minute off-peak frequency. It has confusing deviations (eg to Blackburn) and multiple termini (eg either Shoppingtown or Templestowe Village). Also it finishes unusually early on Saturday evening despite frequent weeknight service until midnight. Despite these handicaps it's well used. More about the 279 in Timetable Tuesday #15. A local review would simplify the route but likely retain its frequent service.

295, in contrast, is more of a local shopper type service between Shoppingtown and The Pines. Much of its catchment is not very dense (rich people with lots of cars) and patronage is, as a result, low. It probably doesn't even justify its half-hourly off-peak service.

* DART routes (905, 906, 907, 908): These are well used routes, operating to Melbourne CBD, at least in peak times. They have had crowding problems of their own (and some may still do). However, unlike 200/207 and 302/304 some have recently had service upgrades. I'll leave these alone for this exercise though some may justify further service upgrades if cuts (discussed later) release sufficient resources. I'd also improve off-peak service on busier routes like 907, including 15 min weekend frequencies and Sunday service until midnight.

* High St corridor (Templestowe): Parts are served by routes 281 (limited service), 309 (even more limited service) and 908 (SmartBus). I talked about economical upgrades to this corridor back in July (UN 12). Again there's nothing to take from this for 302/304 but there may be interactions with other network changes. An upgrade here would be a byproduct of giving the hospitals at Box Hill a better 7-day service to Doncaster and Templestowe and make other network changes more palatable.

* 280/282 ("Manningham Mover"): This is a very large circular route whose aim is to provide coverage of many areas between the main routes. However it connects to no train stations, can be indirect and sometimes overlaps other routes, even in sparsely populated areas. Operating hours are limited, with start and finish times not quite suiting commuter hours. In brief it is a dog of a route that is everything that buses shouldn't be.

There's some history here. The Manningham Mover only started in 2008 as a result of local advocacy and a state government, unwilling to build Doncaster Rail and not yet ready to roll out the DART SmartBus routes, wanting to be seen to be doing something for transport in Manningham.  While it gave some areas improved coverage, the route's length (near enough to 90 minutes) makes it expensive to run (6 buses for a 30 minute frequency). Unfortunately it was not reformed when DART and orbital SmartBuses were introduced two years later. The result is a lot of duplication and extremely low passenger boardings per kilometre. Reform of this route may free up resources to boost more heavily used routes, extend operating hours and potentially provide some asked-for connections.

* Route 309: This route runs from The Pines Shopping Centre to the CBD via Reynolds Rd and Foote St. It is largely a peak only route though a few off-peak weekday trips also operate. It has little unique coverage. Almost all of it has overlaps with multiple routes including the 280/282, 281 and 901 SmartBus. Route 309 has no evening or weekend service.

* Route 901: This is a very long orbital route from Frankston to Melbourne Airport.  Of most interest is the portion between The Pines Shopping Centre, Foote St, Fitzsimons La, Greensborough and South Morang.  As a SmartBus service levels are relatively high - every 15 minutes on weekdays and 30 minutes on weekends. There is also service until midnight Monday to Saturday. The route has little unique coverage and goes through very low density suburbia on Reynolds Rd (map above) and rural areas around Yarrambat. Consequently patronage is low, especially for its high service level.

While governments have been reluctant to break up the SmartBus orbitals (Transdev tried in its unsuccessful proposed 2015 network), the north-east part of the 901 traverses areas that would hardly justify a regular route, let alone a premium service operating until midnight. How many buses are we talking about? I get about six.

Six buses is too many to ignore if we're on the hunt for a few of them to help the people left behind on the likes of the 302 and 304. While changes to Route 901 are complicated they are almost certainly necessary if you wish to deliver cost-neutral service upgrades to the Manningham area.

* School services: Depending on the operator's peak bus fleet usage some low-cost means to provide service with existing buses may be possible. For instance 6-7am and 5-7 pm commuter trips may be possible by adding regular route shifts to school runs. There may also be potential to revise school routes in conjunction with local network reviews if the latter develops routes that better serves schools.   

In 'robbing Peter to pay Paul' exercises like these you need to be careful to prevent new problems being created. You need to provide enough 'goodies' that the overwhelming majority of people get a better service. And, as with any change, we need to be better at explaining why network reform is necessary, especially for those whose travel would change. But there's huge possibilities given we're often not currently running service where the people are or where needs are greatest.

What one might do

* Manningham Mover reform. Split 280 and 282 to two simpler linear routes mostly along existing alignments but extending to Heidelberg Station.

Southern portion (282) could start at Heidelberg Station, operating to The Pines via Shoppingtown.  Span could be extended to better suit commuter times. Off-peak frequency could drop to every 40 or 60 minutes to reflect usage and harmonise with trains at Heidelberg.

The northern part of Manningham Mover (we'll call this the 280) would also start from Heidelberg and terminate at The Pines. However it would go via Templestowe Village, Foote St and Reynolds Rd. Service levels could be as above for the 282.

The Heidelberg terminus will be beneficial for both routes due to access to a train station, hospitals and, via a bus connection, La Trobe University. This extension could also compensate Reynolds Rd for the shortening of Route 901 (described later). Overall this change increases rather than reduces the route kilometres of 280 and 282. However adjusted off-peak frequency may reduce service kilometres. And the Heidelberg extension may make other network reforms, such as we'll discuss with 901 acceptable.

* High St corridor (Templestowe): A low cost reform to provide a new 15 minute weekday corridor with 7 day service through a consolidation of Route 281 and 293. This is low cost, greatly simplifies the network and provides a major boost to 7-day service to the hospitals at Box Hill. It would provide a service boost for Elgar Rd passengers near the eastern ends of the 302/304 as well as those on parts of the 309 (discussed later). A map is below with more detail at Useful Network 12.

* Route 901: Reform to this SmartBus route is key to whether you can get cost-effective improvements or not.  The more you do here the more crowd-busting improvements you can buy on 302/304 and potentially other routes.

'Do something' options (in declining order of 'radicalness') include:

1. Delete Route 901 between South Morang and The Pines completely. No change to existing network.

2. Delete Route 901 between South Morang and The Pines. Changes to existing routes to compensate populated areas considered to be underserviced by this change.

3. Delete 901 between South Morang and The Pines. Add a new route, operating approximately every 30 minutes weekday/60 min weekend to connect with every second 901 at both ends (reduced service level would better reflect catchment though is arguably still excessive).

4. Only delete Route 901 for a smaller section - ie either South Morang - Greensborough or Greensborough - The Pines (both due to low population density of overlap with other routes). 

My bias is toward the second option. The compensatory measures in themselves could deliver benefits. For example Foote St/Reynolds Rd would lose the 901 but gain the 280 extended to Heidelberg with better operating hours than now. Similarly the current Greensborough - Epping connection  could be retained if the 566 is extended the short distance to Epping to compensate for the 901's removal. These and other measures are summarised below.

* Route 309: The High St portion of this route would gain a large service increase with buses every 15 minutes between Templestowe Village, Doncaster and Box Hill thanks to the 281 upgrade mentioned before. Foote St/Reynolds Rd would gain an alternative (though somewhat indirect) connection to the CBD via the 280 extended to Heidelberg Station. Given Foote St's low population density this route might drop back to a peak-only service if off-peak usage is low.


I have sought network efficiencies so that routes like 302 and 304 can be upgraded for low or no expense. As mentioned before this was made difficult because close to an operator's full available fleet would already be deployed during peak periods.

Even though the Manningham Mover uses a lot of buses, I wasn't able to squeeze much from it. One might free a couple of buses by cutting peak frequency to (say) 45 minutes. However the Heidelberg extensions add kilometres. Also the Manningham Mover is one of the routes that would compensate  some areas affected by the 901 being split and removed.

Which gets to the nub of the matter. Apart from minor economies that might be possible from routes like the 295 or even 309, it seems to me that reforming the 901 orbital is key to delivering bus reform in Manningham and nearby areas.

Am I right, or are there other potential savings not mentioned? And if there are do you think they would be publicly acceptable?

Please leave your comments below. And, even more importantly, due to the current political interest, let local MPs know your views if you use these routes.


The above has kept service reform to a fairly tight area. But if you wanted to free up bus resources to provide additional peak service boosts for crowded routes one might wish to cast the net wider. Eg 906 has all morning trips starting at Warrandyte bridge, with a 5 min frequency operating. Could some of those be shortened to start at The Pines to free up a bus or two? 216/219/220's southern section and 232 are other quieter routes you might be able to find spare buses on.

PS: An index to other Building Melbourne's Useful Network items is here.

1 comment:

Tom said...

Some counter-peak runs of the 302/304 could run express city-Doncaster/Box Hill, via faster routes, to help increase peak direction capacity. More capacious buses, either bendy or double-decker, may also be useful on some busy routes such as potentially these (Box Hill bus interchange may present an issue for the 302, 901 etc.).