Friday, June 28, 2019

Building Melbourne's Useful Network: Part 9 Bentleigh East and beyond

June 18's Timetable Tuesday looked at the new Moorabbin - Chadstone Route 627 and how its trips mesh in with the long-established Route 822 that it partly duplicates or runs close to.  Trips were mostly reasonably evenly spaced from the Chadstone direction.

However there were instances where they were not. Especially on Sundays when two buses would appear at once, followed by a long gap to the next service. Not because of traffic, but because of timetabling. 

When it's busy, like when people finish shopping at Chadstone, people wait more than necessary because of this. Especially retail workers who, unlike shoppers, can't easily time their departure. When it's quiet, like Sunday evenings, we see two almost empty buses following each other. Neither situation is good for service, patronage or efficiency. 

This happened because Route 627 got introduced over the top of an existing network that had no changes to its routes or timetables. If you have one route operating every 40 minutes and the other operating every 60 minutes you can't space trips evenly.  That's the Sunday situation with the 627 and 822.

The annotated and merged PTV local area map below shows some bus network issues in the Bentleigh East, Moorabbin and Cheltenham areas. There's also spill-over into the Hampton and Sandringham area as both areas share some routes.   

Existing Useful Network

So much for broader network issues, what about the existing Useful Network? I explain the Useful Network concept here. It's those routes that are frequent enough and run over long enough hours to be useful for many trips. I've specified a 20 minute frequency on weekdays and 7 day service until 9pm. In other words the coloured lines on the Melbourne Public Transport Frequent Network map with the 20 minute frequency selected.

The map below shows the existing Useful Network (click for better clarity). 

East-west Useful Network routes in the area are on main roads, eg North Rd (Bus 630), Centre Rd (Bus 703) and South Rd (824). These are spaced about 1.6km apart and align well with train stations. 

Unfortunately the north-south Useful Network is sparse. The central part of Bentleigh East has no routes or legible combinations that meet the standard, with walks of nearly 2km to either the Frankston train line or Warrigal Rd (903 SmartBus), with only infrequent access to Chadstone. The jobs areas in Moorabbin around Chesterville Rd and Bay Rd between Sandringham and Southland also miss out. 

What if you wanted to do better? That is to give more people useful bus services without a bucket more buses or money. Today we'll try this exercise for Murrumbeena, Bentleigh East, Cheltenham and adjacent areas. 

Unlike other series I'll suggest two expanded Useful Networks. The first is a lower change/lower benefit option that keeps the new Route 627. Whereas the second is a greater change/greater benefit option that merges 627 and 822 to form a major new north-south Chadstone - Southland route. 

Expanded Useful Network 1: Bentleigh East boost with Route 627 retained

This option delivers a two route frequent service between Chadstone, Murrumbeena and Bentleigh East. It retains the new route 627 on its existing alignment and service level. Route 822 retains its service level, gets moved to East Boundary Rd and is renumbered 628. The result is a 627/628 corridor with an even 15 minute frequency on weekdays and a 20 minute service on weekends. The simpler service through the heart of Bentleigh East gives better access to Murrumbeena Station and Chadstone, as you can see on the map below.  

Route 822's shift and renumbering affects where other routes go. This reshuffling is needed to retain service near all stops currently served.   

Route 701 is less used than other local bus routes. It also overlaps Route 767 in some areas. The network above shifts it to Marlborough St to replace Route 822. By serving the busier Oakleigh station (rather than Murrumbeena) it provides connections to Monash Medical Centre from a wider area. 

Route 701's occasional Moorleigh deviation gets removed. That means consistent trips with no confusing footnotes on timetables. As compensation Moorleigh gain a regular 7-day service with a minor change to Route 767. 

This network hardly changes route kilometres. Neither are there frequency increases on individual routes. Similar annual service kilometres should mean similar operating costs to today. However the simpler direct routes and new frequency corridor should make buses more useful and boost patronage, particularly in Murrumbeena and Bentleigh East.  

Expanded Useful Network 2: Bentleigh East and beyond upgrade

Here's a more radical network. It's simpler. There's fewer routes. But they're more frequent. And it brings a Useful Network standard of service near thousands more people and jobs. If you want buses to be simple, direct and more like trams then this is your network. See it below:

The centrepiece is a new route from Chadstone to Southland Station. I call it the 620. It's basically an amalgamation of Routes 627, 822 and part of 767. Designed to get people to trains at Murrumbeena and Southland, shops at Chadstone and Southland and jobs at Moorabbin, it can serve diverse trips all day every day. Service is every 20 minutes Monday to Sunday. Extra short trips could operate from Bentleigh East to Murrumbeena Station to provide a 10 minute peak combined frequency if resources allow.  

Bay Rd is the other major Useful Network extension. It is formed by routing 828 to Sandringham via Southland Station and Bay Rd. This direct alignment serves a busy jobs area, improves access to Southland and provides a back up for when either Frankston or Sandringham line trains are disrupted. Because the route is shorter than the existing indirect alignment to Hampton, its operating cost should be less than now. Any savings should be reinvested in frequency upgrades as Route 828 is a popular service. 

How are local routes affected? Route 701 can be realigned as per Network 1, ie via Marlborough St. Route 767 is similarly changed except for running down Rowans Rd (to replace the existing 822, or Network 1's 628). 

Running 828 to Sandringham and replacing 822 with other services creates some gaps in the Hampton - Highett - Cheltenham area. A new route 707 from Hampton to Southland via Highett and Cheltenham overcomes this. It could serve all stops the 828 does until Highett. Then it runs to Cheltenham (replacing 822) before terminating at Southland. Route 707's 30 minute frequency is designed to mesh with trains at Hampton (every 15 minutes). There is also the option of an even spacing with Route 708 (also every 30 minutes) which starts at Hampton and serves a similar catchment. This is unlike now where spacings between the 708 and 828 can never be even due to unharmonised frequencies. 

I haven't done detailed costing calculations. Like Network 1 the weekly service kilometres run is unlikely to be much different from now. The big difference is that unlike Network 1 there are fewer routes. And route kilometres are less. That frees up service kilometres for use on expanded Useful Network routes like the 620. 

That spreads benefits much more widely, to include people like Moorabbin residents, Southland shoppers and workers near Chesterville and Bay roads that Network 1 did not benefit. 

Oh, and what if you really wanted to keep the 627 that Network 2 scraps? A possibility could be to keep it running simultaneously with the revised network for a month or two to give time for people to try different network options. If patronage drops it could be made an hourly daytime one-bus route to better reflect demand. And if usage is very low then a review could ask the community whether they would prefer other upgrades, such as a Route 823 Nepean Hwy extension to Elsternwick and a compensatory 701 extension to Brighton via Dendy St. 

More detail

See the link to Useful Network map if you need more detail. You can also view it below after locating the area and selecting the menu (top left) to show whether you want to see existing or expanded Useful Networks. Top right opens it in another window if preferred.  


I have presented two low cost network options for buses in the Murrumbeena, Bentleigh East and Cheltenham areas. 

Network 1 uses the recently introduced Route 627 as a starting point for more useful service between Chadstone, Murrumbeena and Bentleigh East. Other areas gain simpler and more consistent services.

Network 2 also benefits Chadstone, Murrumbeena and Bentleigh East. Plus, there are further gains for people and jobs in Moorabbin, Cheltenham and Sandringham.  However, in its purest form, the network replaces the 627 with a newer, longer and more frequent route. 

What do you think? Which option is best? Or do you have other ideas on the local network?
Please comment below.

PS: Other reading. 1. Daniel Bowen's post on local buses with some history. 2. Krustylink old timetables (822 used to have a 20 min peak frequency). 

Tuesday, June 25, 2019

Timetable Tuesday #28: Where there's no return - Bus route 745's alphabet soup

We're not like Adelaide. Or Sydney.

Melbourne doesn't do letters on bus route numbers.

With one exception(*).

The 745.

With variations A, B, C, and D, it's one of the long term bus routes serving the City of Knox in Melbourne's outer east.  PTV's web timetable dates its last change as May 2008 but the routes have been running for decades before that. 

All start from or finish at Bayswater Station. Apart from that they don't have much in common. If it was anywhere else and they'd have their own unlettered route numbers. The map below shows their destinations.  

Primary schools normally have small (ideally walkable) catchments. It is unusual for regular bus routes to have them as destinations. However 745C and 745D do. 745B operates between two adjacent train stations via a residential area. 745A is (mostly) a main road service between Bayswater and the large Knox City Shopping Centre.  

The Knox network map below shows where the four 745s fit within the network. All are fairly short. In some cases they are about the only routes on the roads they travel on. That includes main roads like Mountain Hwy and Scoresby Rd. Whereas an inner suburb would have a bus every 20 minutes or better along such roads, and a middle to outer suburb every 30 or 40 minutes, Knox never got a real bus network when development started in the 1960s or in the fifty years since.   

Possibly the greatest surprise is the timetable. Each 745 variant has just one trip each.  Their purpose is unclear. All operate in the afternoon but appear too early or too late for school finish times. Even if you could take a 745 somewhere you can't return on it.

745A operates to Baywater, as you can see from the schedule below.

745B, 745C and 745D operate from Bayswater. Again it's uncertain who they're trying to cater for. While it backtracks towards the city direction, the 6:30pm departure is a potential commuter route. However there is no corresponding morning trip to Bayswater. 

There's 745 patronage numbers here (and reproduced below). 

They jump around quite a bit. There's no reason I can see for this since the timetable, dated 2008, has not changed in the whole period. Unfortunately they are not separated by A B C D variant. 

There's roughly 250 odd weekdays per year. With roughly 2500 passengers per year that amounts to 10 passengers across the four trips daily. In other words 2 or 3 people per trip. Which is low, given the 745 runs in school or commuter peak time. If its trips represent needed positioning moves for busier school or commuter trips, its running cost may be low. But dear or cheap, its limited timetable and consequent low patronage makes the 745A, B, C or D useful for very few people. 

What was the 745 like in the past? Krustylink has 1984 and 1991 timetables.  In both years 745 was an irregular service with many variations, like now. There was no 6:30pm trip. But there were morning trips. The 1984 bus timetables contained a handy guide to train frequencies at Bayswater. These remain fairly similar today, including the 30 minute off-peak frequency. 

What would you do with the 745? Is it worth upgrading at all? Or is it so useless that it should be deleted? Is there scope for a better bus network in the area, including along main roads? Please share your thoughts in the comments below.

(*) Actually there is one other, also in Melbourne's east, but that's a story for another time. 

Friday, June 21, 2019

What real network reform looks like: Regional Rail Link turns four

Today is an important anniversary. Four years since the commencement of services on the Regional Rail Link. The official opening was actually a week prior but today is the more significant anniversary.

Regional Rail Link routed Geelong trains on new tracks away from Werribee via Sunshine. This allowed trains to service the fast growth areas of Wyndham Vale and Tarneit (which never had trains before). And it freed up line capacity for more trains between Werribee and the CBD.

As for Geelong, although the line on the map appears less direct, the faster running of the trains  on its own tracks meant similar travel times to before. And off-peak weekday frequencies increased from every 60 minutes to every 20 minutes. This is a major improvement that made trains more useful for more trips including business travel. It's just a shame that the City Loop's convoluted operations (including the midday reversal) still makes getting to Geelong a lottery from the northern and eastern CBD.

Transport today would not cope without the Regional Rail Link. Although 'Regional' is in its title, the big growth in passenger numbers has come from outer suburban Wyndham Vale and Tarneit.

Five years ago Tarneit Station didn't exist. Now it's the busiest V'Line station outside Southern Cross. It and Wyndham Vale's car parks are amongst the biggest on the rail system. And well-used direct buses run from the new stations to established Werribee line stations.  This is like the pattern in the middle-eastern suburbs (eg around Mt Waverley), where north-south buses serve train lines that run east-west, providing a grid style network where both buses and trains attract good patronage numbers. When you do the same thing you tend to get similar successful results.  

More to do

The RRL remains an unfinished project in some ways. Three road-rail grade separations at Deer Park and Sunshine West (Robinsons, Station and Fitzgerald roads) that should have happened were not done. Although we think of V/Line trains as being infrequent, these crossing are so busy at peak times that more trains pass through them than those on some Metro lines.

Timetables on both the Geelong and Werribee lines now don't meet today's demand. For example Tarneit (the busiest station mentioned before) has almost a half-hour gap in departures around 5pm. Cancelling just one critical trip would cause there to be no trains for nearly 50 minutes. And the 40 minute weekend gaps are at least twice what they should be.   

The Werribee line's peak timetable is other unfinished business. The peak timetable on it and related lines (including Laverton and Williamstown) currently have an awkward 11 and 22 minute pattern. Werribee's peak timetable does not fully exploit the capacity freed by rerouting Geelong trains. This was mentioned by the Auditor-General who found that benefits from RRL have yet to be fully realised. And the government came close to but eventually did not proceed with a radical metropolitan train timetable overhaul in 2015

New bus networks

Less heralded than the RRL infrastructure and associated train services, are the bus network improvements that also started on this day in 2015. These were not mere tinkerings. Instead, almost every bus route in Greater Geelong and Wyndham (incorporating Wyndham Vale, Tarneit, Hoppers Crossing and Werribee) got a new alignment, a new number and a new timetable. A complete redesign from a blank slate. Historically significant and quite rare.

The multimode network diagram below, though not quite accurate, showed how the more frequent services in Geelong and Melbourne's western suburbs fitted together. I did it at the time as an experiment and it never went anywhere.

You can compare sections of a reformed network (central Wyndham) with an unreformed network (central Melton) below. Notice how Wyndham has direct routes that stay on its main roads with local routes serving areas in between. Whereas Melton's routes are all half and half - a mix of main road and local street running. Its basic network structure hasn't changed for years. Like Wyndham, Melton is an outer growth area and there are large areas without service.   

The big differences aren't evident on the map. Frequency. Wyndham's bus network is two-tier. Its main road routes operate every 20 minutes during peak periods. The most important of these also run every 20 minutes off-peak. And sometimes even on weekends as well (eg 170 and 180).  Local street routes in Wyndham are typically every 40 minutes - their main purpose is to provide coverage to areas distant from the main roads. Both types of routes attract good patronage, with peak usage so high that extra trips on some will start next month.

In contrast Melton's routes are (at best) every 30 minutes on weekdays (even in peaks) and hourly on weekends. Melton's mix of unreliable V/Line trains and infrequent local bus routes makes using them unattractive for commuters seeking an alternative to parking at the station. 

Geelong also got a new bus network. It was not uncontroversial. However the only people who liked the old network were existing users. This can skew results when people are asked about network changes intended to attract new riders. Like in Wyndham, many of Geelong's routes were made more direct and upgraded to operate every 20 minutes off-peak on weekdays, harmonising with train frequencies. This is summarised in the hypothetical advertisement below.

Both revised networks represent the two biggest examples of comprehensive bus network reform we've seen for years. Despite continued strong population growth and networks that need reform, nothing bigger has been implemented since.  And where bus reforms have occurred, it's mostly been in conjunction with rail extensions such as in the lead up to and again when Mernda's line opened.

A well-known motivational quote from Norman Vincent Peale is: “Shoot for the moon. Even if you miss, you'll land among the stars.”

In the context of Melbourne public transport, this could be rewritten to say "Advocate for a train. Even if you miss you'll get a better bus network". The record shows you'll probably have more success than asking for buses directly. Whether it's trains to Doncaster, Rowville, Cranbourne East or South Morang (initially) or trams to Knox City, advocates have shamed governments into providing  improved bus services at a higher standard than almost anywhere else.

That's more than seems to have been achieved by those advocating directly for buses since about 2011. Despite a strong case, advocates for bus services haven't done very well with this government. The only bright spots have been universities (with new or improved shuttle routes) and local MPs in formerly marginal seats (who have gained local routes layered over existing unchanged dysfunctional networks). Real network change has been minimal despite the opportunities presented by road-rail grade separations to renew bus networks based on improved traffic flow and easier interchange to the rebuilt stations. 

I've veered off-topic, haven't I? But they illustrate the historical significance of today's anniversary. A day, four years ago, when real train and bus network reform was delivered with substantial community benefit. May we have many more future days like it. 

For now though, enjoy the videos I made at the time. 

Regional Rail Link Open Day

RRL construction (day)

RRL construction (early morning)


The RRL (and its associated bus network changes) have been great successes. Maybe even too successful, with services currently straining under patronage pressure. However it paved the way for bigger and even more transformative rail projects that are now under construction. 

Note: This post appears in place of Building Melbourne's Useful Network. That's back next Friday. 

Tuesday, June 18, 2019

Timetable Tuesday #27: The Two Followers - 627 and 822 from Chadstone

Bus route 822 has been running for years. It starts at Chadstone then goes south via Murrumbeena Station, Murrumbeena Rd, Southland, Cheltenham then Sandringham.

Its map is below. 

Bus route 627 (reviving an old number - but that's another story) started two days ago. It also commences at Chadstone (thoughtfully from the same interchange bay as 822), then goes south via Murrumbeena Station then Murrumbeena Rd. It then goes straight south via East Boundary Rd then a few local streets before terminating at Moorabbin Station. Pictures from its first day below.

In other words 627 duplicates the 822 for a bit before staying on the main road. But there'd still be common catchment for those living between the two routes. That's particularly handy if you're coming home from somewhere and have the choice of both routes at Chadstone or Murrumbeena Station. Hold that thought for later. 

627's map is below. 

The maps above are not to scale. You can see the relationship between the two routes on the Glen Eira network map below: 

This is 822's timetable (from Chadstone/Murrumbeena). Click for a larger view.  

This is 627's timetable (from Chadstone/Murrumbeena). Click for a larger view. 

I'll leave you to compare the timetables. But I do wish to draw attention to one thing. Look at afternoon departure times from Chadstone (and Murrumbeena Station a few minutes later).

Weekday ex Chadstone

Route 822: 1:34 2:04 2:34 3:04 3:34 4:05 4:34 5:08 5:38 6:09 6:36 7:06 7:40 8:36 9:29 10:39
Route 627: 2:14 2:44 3:16 3:47 4:17 4:48 5:18 5:49 6:19 6:48 7:21 7:51 8:21 8:57 9:21 9:57

Saturday ex Chadstone  

Route 822: 1:49 2:29 3:09 3:51 4:31 5:11 6:11 7:11 8:04 9:00 10:00
Route 627: 1:31 2:11 2:51 3:31 4:51 5:31 6:11 6:51 7:31 8:06 8:36 9:06 9:36

Sunday ex Chadstone

Route 822: 1:08 2:08 3:08 4:08 5:08 6:08 7:08 8:00 9:00 10:00
Route 627: 1:31 2:11 2:51 3:31 4:11 4:51 5:31 6:11 6:51  7:31 8:06 9:01

The ones in red are where departures on both routes are within a few minutes of one another. On weekdays both routes run every 30 minutes. They are roughly 10 to 20 minutes apart. This isn't an even 15 minute combined service. However it harmonises with the 10 minute daytime train frequency at Murrumbeena. Weeknight frequencies diverge. 822 drops to roughly hourly while 627 retains its roughly half hourly service all day. On only one occasion (after 9pm) are there two bus departures within 10 minutes of one another.  However if you miss that it's nearly 30 minutes until the next service.

During the day on Saturday both routes operate every 40 minutes. Departures are almost perfectly staggered, providing a combined 20 minute service. That is until about 5:31pm. By then 822 has dropped back to hourly. 627 retains its 40 minute frequency until about 8pm before going up to 30 minutes after then. Presumably this is possible with the daytime number of buses due to quieter roads and faster run times.  The uneven and then clashing frequencies mean that two buses go the same way for quite a bit within a few minutes of one another after 6pm. By about 8pm Saturday trains on the Dandenong line have dropped back to 30 minutes. Hence you would have two buses often departing at once unless you altered one route never to properly connect with the train. Or it might have been possible to operate a cheaper timetable and drop 627 to only hourly at night, offsetting with 822 to meet different trains.

Sundays present buses following one another during the day as well as at night. This is because Route 822 runs every hour versus Route 627 every 40 minutes (30 min at night). This means that buses follow one another on every second Route 822 trip. From 8pm, because both buses drop to hourly and there's no staggering of times, two buses arrive pretty much at once followed by an hour wait until the next two.  What's the bet that all those trips will be carrying a lot of fresh air then?

Over to you. Would spacing evening departures provide a better overall service? Should Route 822's Sunday timetable be upgraded to provide a more even service like Saturday's. Or would that be just tinkering and would it be better to put Route 627's resources towards wider network reform with higher frequency routes? 

Sunday, June 16, 2019

A look at the recent VISTA travel survey

Exactly one month ago today randomly selected volunteers were filling in forms to document their day's travel. A few days prior they downloaded a mobile phone app. That would track their movements for a week. At the end of each day they'd review gathered positions, make any corrections and input the mode and purpose for each journey, leg by leg. 

All this was in aid of VISTA - the Victorian Integrated Survey of Travel and Activity. Data from everyone's responses helps build a picture of where, when and how people travel. It feeds into the models that transport planners use when assessing future transport needs and evaluating projects.

Data provides evidence of usage. And, to some extent, demand. Although there's limitations here as one can't measure the demand of a service that doesn't exist. However we can make a pretty good estimate by looking at demographically similar areas where they do.

What happens with the mountain of analysis done? In theory there would be public transport network strategies  developed with input from sources like VISTA. However public evidence of these existing is scant. For instance we haven't seen a substantive Victorian Transport Plan for over a decade despite it being required by the Transport Integration Act 2010.

How much is our network is shaped by evidence as opposed to running what we've always run? The answer is not much. Population growth continues but network reform has stalled. For example our train timetables can reflect 1970s service cuts more than today's needs. As opposed to Sydney's  network where a 2017 service upgrade delivered fifteen minute frequencies to most stations at nearly all times - day and night.

Ditto for dead-end bus termini and timetables that persist with relics like midday Saturday shop closings. Crowding on some routes leaves people behind while others, unreviewed for years, carry fresh air through affluent suburbs.

One could conclude, that, based on recent service reform, (ie very little happening) history is king.

And because so little is done with it, data is uninfluential and therefore unimportant.  

Except when it is.

At certain times bus services have become a political issue, thanks to community, industry and academic advocacy. Governments may be galvanised into action, as occurred for a few years from 2006 when most areas got 7 day bus service and orbital SmartBuses linked the suburbs.

Sometimes effort can be wasted on matters of marginal passenger benefit, such as operator franchising, multiple rebrandings, vanity architecture (eg Southern Cross Station) and ticketing systems that no one asked for. The late 1990s and early 2000s era was notorious for that.

Meanwhile maintenance is deferred, infrastructure is failing, patronage is building and needed services aren't being added. Once that became part of a credible story about government mismanaging growth and being unable to run basic services its interest swung from buses to trains, with major timetable improvements on lines like Frankston and Dandenong. These had large benefits with train reliability rebounding from about 2011-12.   

There may be calls for a new line to relieve a clogged system. Politicians may warm to it. But evidence is needed for support to broaden, especially if it means deferring other projects or taking money earmarked for other portfolios. There's a flurry and everything is wanted yesterday.

Then data, unloved and unused for years, may be core to making a strong case when drafting and comparing alternatives. In which case it, based on what people wrote in their VISTA forms, becomes an unsung hero.

Back to the VISTA survey.

What do survey participants have to do and what data is collected?

Participants get a pack delivered to their home. It looks like this.

It's delivered by people from Ipsos, the opinion pollster and social research company contracted to conduct the survey. They'll tell you a bit about it if you're at home. 

The first thing they ask is to download a mobile phone app called rMove. Then for the next week or so take your phone with you wherever you go. This tracks where you go in conjunction with GPS data. After each day you are asked to review data accuracy and enter why you went places and how you travelled. This includes modes and connections for public transport trips and intermediate locations if you made stops during a trip. That's the first part of the survey.

The second part of the survey is paper-based.

You complete a detailed diary of your movements on a nominated day.

The survey is household based. Thus you need to complete details not only for yourself but any others living with you. You get asked demographic information such as age, income and vehicle ownership. 

Then there's details about your travel. An example page is below. It's quite a thick book. It needs to be since you're asked about every stage of your trip and you might make several of those per day. Questions asked include its purpose, where you went and whether you travelled with others. Public transport passengers and car users had their own special questions relevant to their mode. 

You can read more about VISTA on the Department of Transport's website here.

Publications making use of the data collected can be found here.

If you ever get a travel survey like this, do participate. And complete it well. Every trip counts.

Friday, June 14, 2019

Building Melbourne's Useful Network - Part 8: St Kilda Brighton Sandringham

One of the parts of Melbourne with the most number of routes but the least service on most of them is the narrow bayside strip from St Kilda to Sandringham via Elwood and Brighton.

The area is a real demographic mix. It includes nightlife in St Kilda, young working renters in Elwood apartments, multi-millionaires in Brighton and comfortable retirees in Sandringham.  Trams serve its northern portion while the southern part is near the Sandringham train line. There used to be Victorian Railways trams in the southern part as well. These were closed down after WWII. However their legacy lives on through the long operating hours of some bus routes in the area. 

The area's increasing population density and young urban demographic (who may be more open to using public transport for a variety of trips) provide opportunities for increased patronage from a better bus network. And, as we'll discuss later, duplication of and (arguably) overservicing on some routes should make the costs of reform low.  

Existing Useful Network

I explain the Useful Network concept here. It's those routes that are frequent enough and run over long enough hours to be useful for many trips. I've specified a 20 minute frequency on weekdays and 7 day service until 9pm. In other words the coloured lines on the Melbourne Public Transport Frequent Network map with the 20 minute frequency selected.

The maps below show the existing Useful Network. It's dominated by the Sandringham train line. The northern portion also has trams.  

North-south Useful Network buses include 216/219 and 220 around St Kilda East, 605 around Caulfield South, 246 around Elwood and the multi-route 600/922/923 combination south of Sandringham.  West-east Useful Network routes include 630 from Elwood, 703 from Brighton and 828 from Hampton. The left map shows St Kilda - Brighton while on the right is Brighton - Sandringham (click for better view). 

Where the gaps in the Useful Network? Hampton St is the most conspicuous. Regular routes run near parts but there's nothing direct along its length. Given its density you might also want better service around Elwood. Four routes overlap (606 and 600/922/923) but none provide a consistently useful service. 

East-west Useful Network gaps exist near Dendy St, South Rd and Bay Rd. Reform here involves changes to out of area routes such as 811, 812, 822 and possibly 828. Changing these would affect a lot of other areas so I'll concentrate on the north-south routes today.

Notice all the loose ends in the Brighton area? This is because useful corridors often finish just short of shops and railway stations.  Network complexity also means that some corridors (eg St Kilda - Elwood) have multiple routes but with insufficient frequency for each to be useful. See map below. 

The southern termini for routes 216, 219, 220 and 605 owe more to history than being useful destinations. Routes 600, 922 and 923 are very similar but split and recombine multiple times.  Not all parts of these routes operate all day, with limited span, low frequencies and multi-hour gaps common. 922 and 923's numbering do not convey their similarity to Route 600, despite being 80% identical and knowledge of all required to successfully use them. Overall today's network is more complex than in the 1990s but gives less service.  

Service hours are also worth mentioning. It's a case of haves and have nots. Routes 246, 216/219, 220, and (to a lesser extent 600) have early starts and late finishes reflecting their heritage as ex-Met services (and before that Tramways Board). The worth of some is debatable since some of these routes run near the Sandringham train line.  In contrast operating hours are consistently shorter (9pm or earlier finish) on east-west routes such as 703 and 828 despite their unique coverage away from trains and trams. Other routes with early finishes, such as 606, 922 and 923, just make the network unnecessarily complex.  

Expanded Useful Network

An expanded Useful Network is below. Highlights include new Useful Network routes via Hampton St (new 604) and Elwood (upgraded 606). Some routes are extended to more active termini like Elsternwick and North Brighton. The overall network is simpler than what runs now, with two fewer routes.

See the link to Useful Network map if you need more detail. You can also view it below after locating the area and selecting the menu (top left) to show whether you want to see existing or expanded Useful Networks. Top right opens it in another window if preferred.  

Most or all of the resources for this network come from removing route overlaps and duplication. Routes removed include the confusing and poorly used 600/922/923 and 216. These are largely replaced by upgrades to routes like 606 and the new neighbourhood route 603.

Some resources are also freed by running some poorly used routes less frequently for less of the day. The areas that would lose are generally either close to the train or are areas with demographics unsuited to high all-day public transport usage eg Brighton. The table below (click for a better view) indicates potential resource shifts between routes.

While precise accounting has not been done, scope exists to scale down the proposed network if required. For instance the extensions of Route 220 and 605 to Brighton North are not necessary to achieve the gains in Elwood and Hampton St.  

Service priorities for expanded St Kilda, Brighton and Sandringham Useful Network

1. Terminate Route 600/922/923 to operate Sandringham - Southland only. Upgrade Route 606 operating hours and frequency to provide simpler and more regular services in the Elwood area. Make associated changes including (i) new local route 603 (to retain coverage between Brighton and Sandringham), (ii) extension of Route 630 to Elsternwick and (iii) amalgamation of Routes 216 and 219 into an upgraded 219 with minor route changes in Brighton area. 

2. Introduce a new Useful Network route between Elsternwick and Sandringham along Hampton St (Rt 604). This brings better service to a large area lacking a strong north-south route. Some resources for this could be obtained by reducing frequency and/or span on poorly performing routes such as 216/219 and 220 in areas that are near alternative services.  

3. Extend Routes 220 and 605 to a stronger southern terminus such as North Brighton. 

4. Make minor upgrades to Route 605's operating hours and Sunday frequency. 


The St Kilda, Brighton and Sandringham area has a lot of bus routes but often poor usage on them. However there exist growth opportunities in areas with dense housing or corridors unserved by bus. 

Unlike areas where there is no 'fat' to prune, the cost of pursuing growth here is less because of the potential to simplify the network and reallocate bus hours to where they could be better used.  

What do you think? Do you have other ideas? Are there things you see are wrong with the above? Please let us know your thoughts in the comments below. 

Tuesday, June 11, 2019

Timetable Tuesday #26: Route 504 across the inner north

The most important factor that determines the usefulness of a public transport service is not whether it's a train, tram or bus but its timetable, route and stops.

On a service kilometre per head basis, Melbourne's inner-north is one of the most highly served parts of the metropolitan area. This is due to the large number of basically parallel turn-up-and-go tram lines running south to the city. Also on radial routes are the less frequent but faster trains. 

Matching the higher service per capita is the inner north's lower car ownership and use compared to elsewhere. Other distinguishing characteristics include mixed housing densities and high bicycle usage. Over the last forty of so years industry has moved out while high income city workers have moved in. 

A versatile network requires both north-south and east-west services. That makes it possible for people to reach anywhere with just one change of service. All east-west travel in the inner-north is provided by bus. We discussed the limited service bus route 506 back in February. 

Just to the south of the 506 is the 504, which is today's topic. You can see its map below. 

Like the 506 it starts at Moonee Ponds Junction. Unlike the 506 it is straighter, at least until Carlton North. Then it heads south then east to Clifton Hill Station. As you can see on the network map below it intersects with numerous tram routes. It also passes near Jewell and Rushall stations.

Mostly because of the unhelpful road layout around Royal Park and Parkville, it's the last east-west route until Route 402 along Grattan St.  This makes it a very important link. It also serves as a substitute for Route 506 when this service is not operating (eg Saturday evenings and Sundays).    

Below is Route 504's timetable. It benefited from the 'minimum standards' upgrades about 10 years ago so enjoys a 7 day service until 9pm. This makes it unlike other routes in the inner north, eg 503, 506 and 509, whose operating hours and days are shorter.

Service is close to a half-hourly headway on weekdays, with a slightly higher frequency early in the morning. This frequency does not mesh with trains (every 20 minutes off-peak) so connections are purely coincidental except at Clifton Hill which enjoys a 10 minute train frequency. 

On weekends Route 504 operates to a flat 40 minute frequency. This meshes with trains at all times except evenings, when trains drop back to every 30 minutes. A minor oddity is weekend vs weekday finish times. Even on Sundays the last bus from both Moonee Ponds and Clifton Hill are 20 or 30 minutes later than the last trip on weekdays.  

What if you lived in the large area between Route 504 and 506 and could catch either from Moonee Ponds? Is there one route whose timetable is consistently more frequent and operates over longer hours than the other? The answer is there isn't. Route 504, unlike 506, has 7 day service until 9pm. It normally comes every 30 minutes on weekdays. In contrast, Route 506 is much more frequent on weekdays and Saturday mornings. However it finishes early on Saturday and lacks Sunday and public holiday service (not conforming with the general standard on the latter).    

Route 504 is an important link in Melbourne's inner-north. Does it have greater potential? Should its frequency be improved, perhaps to the level that would make it a Useful Network route? Is there scope to extend it? Please let me know your thoughts on it in the comments below. 

Friday, June 07, 2019

Building Melbourne's Useful Network - Part 7: The Mordialloc Freeway corridor

The Useful Network series is typically about low cost ways to improve local bus networks. It normally achieves this by removing duplication between routes and increasing frequency at off-peak times when spare buses are normally available. 

Today we take a different tack. Firstly it will cost more than previous weeks' reforms, since the area has little service duplication. Secondly it covers a larger area than normal. This is because, thirdly, this network was provoked by a proposed freeway. Which we'll discuss before returning to buses.

The Mordialloc Freeway

The state government is proposing what is basically a northward extension of the Mornington Peninsula Freeway. Widely known as the Mordialloc Freeway, it was supported by both parties at the 2014 state election. Estimated to cost $375 million, it will be 9km long and include four interchanges. The Mordialloc Freeway will parallel the Eastlink toll road, about 6km to the east.

The Eastlink experience

Let's talk about Eastlink for a while.

The former Linking Melbourne website says it was Australia's largest urban road development during its construction. It runs from Mitcham to Frankston. Opening in 2008, the public private partnership project was delivered by ConnectEast, with construction by Thiess John Holland. Claimed benefits (in 2009) include faster travel (24 minutes for full trip) and 30 per cent less traffic on existing arterial roads.

While the road itself is exclusively for private cars and trucks (though some use of it may be made by dead-running buses), the project did include some public and active transport components. Their inclusion meant it could be presented as an integrated transport package for eastern suburbs.

Transport planners tell me that such packaging can increase and broaden a road project's political or economic acceptability. This is especially where decision-makers favour bold and expensive 'signature projects' over modest but more cost-effective upgrades for local roads, active transport facilities and public transport services. If things that should be done anyway (but aren't) are tacked onto a grand road or rail project then its overall benefit to cost ratio may improve (or become less worse).

Eastlink was described as Melbourne's Motorway Masterpiece in this promotion brochure.  There is other Eastlink information on this archived copy of the government's Linking Melbourne Authority website. This is the successor to the Southern and Eastern Integrated Transport Authority, the body established to plan and commission Eastlink.

Active and public transport measures were some late additions to the Eastlink project. These include a walking and cycling path parallel to the freeway.  Because there are few dense trip generators right near Eastlink these are basically for recreational rather than commuter transport purposes. Their  positive value is perhaps best compared against the negative effects that freeways and road widenings can have for access by people who could previously more conveniently walk to trains and shops.

The Eastlink project's associated public transport projects included upgrades to four railway stations and the Route 900 SmartBus between Caulfield and Stud Park (introduced 2006).

More than ten years on it is difficult to discern lasting benefits from all of the station upgrades. Noble Park needed to be rebuilt due to level crossing removals. Kananook remains bleak due to its poor surrounding environment. And, while there has been investment in the nearby Dandenong city centre, the poor day to day cleaning and maintenance of Dandenong station still make it not the nicest of places.

In contrast bus Route 900 has been highly successful. It's well used to the point that additional frequency is needed. Peak frequency was recently upgraded from every 15 to every 10 minutes. However further upgrades are desirable, especially on weekends where the current service is only every 30 minutes. Also it only goes only as far as Stud Rd. Most of Rowville's housing is east of Stud Rd where only occasional buses operate. 

One of the arguments for freeways is that they can act as a bypass and divert traffic from clogged parallel roads. In Eastlink's case, Springvale Rd and Stud Rd. Potentially one could do other things with those roads, for instance put in bus lanes, bike lanes and/or make them more pedestrian friendly. If we think in those terms we're starting to think not just about one freeway corridor and cars, but about a wider area and other transport modes. In other words what is sometimes referred to as 'balanced transport' thinking.

Can all that work in practice? It depends. The above carries certain assumptions. These include wheter decision-making for active and public transport is responsive enough to make proper use of any reallocated road space. If the freeway is tolled like Eastlink, removing car lanes from a parallel untolled road will just be seen as 'revenue raising' and be politically unpopular.

Stud Road's bus lanes - a cautionary tale

The initial decrease in traffic on parallel roads to Eastlink provided an opening for the government to introduce bus lanes on Springvale Rd and Stud Rd. However the Stud Rd lanes were not to last.

The Route 901 SmartBus between Ringwood and Dandenong via Stud Rd was introduced in 2008. It runs every 15 minutes on weekdays (including peak periods) and every 30 minutes on evenings and weekends. This was a large improvement on the previous Route 665 but short of what would run on a true turn-up-and-go bus rapid transit system. A few local routes used sections of Stud Rd but their frequency was less. The large catchment beyond walking distance of Stud Rd in the City of Knox has particularly poor service with some main road routes running just one trip per day without  significant service upgrades for decades.

Stud Rd car drivers (understandably) got upset at seeing an almost empty bus lane while their lane crawls along. Bus lanes can appear relatively empty yet still earn their keep (by carrying more people than car lanes) if a good frequent service is provided. This is difficult but necessary to communicate since bus lanes are constantly under threat by motorist interests who think that removing them would ease congestion. This is one (admittedly small) reason for buses not to have window advertising or dark tinting that makes it difficult to see in from the outside.

The bus lane infrastructure on Stud Rd was never backed up with frequent service. Infrequent service is merely inconvenient for train users, who keep their right of way. For buses though it can mean slower and less reliable travel if it leads to bus lanes being removed. An example was in 2011 when the Baillieu government removed bus lanes on Stud Rd after advocacy from local MPs.  Antipathy towards bus lanes can cut across both sides of politics, with the current government proposing bus lane removals on Fitzsimons Lane.  This only adds congestion as people swap bus travel for driving on increasingly clogged roads alongside other ex-bus users. 

The lesson is that bus lanes must be backed up by frequent and useful service along them. If this is missing, as in Stud Rd's case, then the lane risks being ceded to regular traffic. That's likely under policies that favour infrastructure over service. One can then reach the paradox where bus lanes (or other public transport improvements) are used to bolster the case for a road widening or freeway but, due to a lack of service they end up being removed with the benefit gone.

Traffic volumes - fixed or variable?

Also important to discussions about road space is the nature of traffic and whether increasing space will fix problems of congestion.  Many think of traffic as being a fixed quantity like water. At any one time there is a certain number of people on the road who need to go places. It's the transport system's job to keep them moving.

Think about a known volume of water in an elevated tank. Let's imagine it fills up each night and you wish to empty it each day. You open taps to outlet pipes to let the water flow. Add more (or thicker pipes) and the tank empties quicker. But the amount of water you start with each day remains the same.  In this analogy water represents traffic and the pipes represent roads. Hence it seems logical that if you want to speed movement then you add more or wider pipes.

A different way to look at it is to regard traffic as the result of many peoples' decisions. When you make something faster, more convenient or cheaper you attract people to it. Listen in on peoples conversations. Read advertisements. Scour social media posts. Observe traffic volumes. Available transport changes what people do.

It is undeniable that roads like Eastlink have redrawn peoples' mental maps about transport. Driving trips that people previously thought 'too far' or 'too slow' became more practical, so are being made more often. Faster goods travel shortens supply chains and increases responsiveness as people can get stuff sooner.

Employers have a wider labour market (but may lose workers to businesses further away). Specialist businesses gain wider catchments (although also face competition from further away). Eastlink has made some places relatively more accessible, and, by extension, others relatively less accessible. In short, it's changed where we go and how we travel.

While not immediate, all the above factors flow through to business decisions like where to locate and personal decisions like where to live, work, shop and enjoy. That changes the suburban structure, makes non-driving trips less viable and creates demand that simply didn't exist before the freeway was built. This means that traffic behaves less like water and more like a gas, which spreads to fit the available vessel.

No less than the Metro Rail Tunnel, roads like Eastlink have been truly city shaping projects. Decisions as to their desirability boil down to what type of city you want. You fund more of what you want more of and fund less of what you want less of. People are pragmatists who will use what works best for them based on available choices arising from infrastructure and service funding decisions.

Existing public transport near the Mordialloc Freeway corridor

The map below shows the public transport network in the area surrounding the Mordialloc Freeway corridor.  It is pieced together from PTV local area maps. All routes are shown equally, whether they operate every 15 minutes all day or just run occasionally. Hence it overestimates the number of routes likely to be useful for a wide variety of trips. Click for a bigger view.

A better representation is from the Useful Network map. It shows 7 day routes that feature service until at least 9pm and a 20 minute (or better) frequency on weekdays.

Both the Frankston and Dandenong train lines exceed this standard. As do some bus routes. These include: the north-south red, green and yellow SmartBus routes (903, 902, 901) and the east-west bus route 828 (shown in blue). 

However there are some long distances between them. Residential areas beyond walking distance of Useful Network service include most of Carrum Downs, Patterson Lakes, Chelsea Heights, Aspendale Gardens, Parkdale, Mordialloc and Mentone. Similar comments apply to the light industrial and job areas of Braeside, Mordialloc, Chelsea Heights and Carrum Downs.

The distance that most people are from good service limits the existing public transport network's usefulness for all types of travel. That includes trips where public transport should have an advantage, such as trips to the city, especially during peak times where competition for road space and parking at train stations is highest.

As well as their limited frequency, the layout of existing bus routes can limit access to local jobs. For example some trips require multiple changes between infrequent services. Or bus routes to employment areas may terminate short of train stations.  We'll try to deal with those issues next. 

Future public transport near the Mordialloc Freeway corridor

What if you wanted to give those in the Mordialloc Freeway corridor an alternative to driving?

Mapped below are some public transport network upgrades that would dramatically improve access to trains, education and jobs. They will require new buses to be bought as they involve more frequent service on existing routes and extensions to better serve employment and education destinations. With one exception all operate every 20 minutes through the day, making them suitable for a diverse range of employment and education commutes. 

Click this interactive map for more detail, including notes on each route. Note that the map shows only Useful Network routes - ie those that would operate every 20 minutes or better on weekdays. 

Main points of revised network

* New Mordialloc - Monash University Clayton route. Formed from an extension of Route 705, this makes  access to Monash University's Clayton campus much from the southern suburbs much easier with good frequency and directness. The route also connects people to jobs at Braeside and improves service in the Clayton South area. This route along Boundary Rd is the public transport option that most matches the Mordialloc Freeway alignment.

* Route 708 upgrade. An all-day frequency upgrade greatly improves connections to trains from Parkdale, Aspendale Gardens and Chelsea Heights, relieving pressure on station parking at Mentone,  Parkdale, Mordialloc and Carrum. Access to Monash University Clayton is also provided via a connection to the extended Route 705 (above).

* Route 760 and 778 upgrades in Carrum Downs. Both routes are upgraded to operate every 20 minutes to Seaford Station to provide an improved direct train feeder service. This should relieve  commuter parking pressures at Seaford. In addition there is better access to industrial areas in Seaford and Carrum Downs from both Seaford and Cranbourne Stations.

* Route 770 frequency upgrade in Karingal. Current routes are indirect loop services that are difficult to understand. Route reform could involve operating Route 770 and 771 as straight there-and-back routes. Route 770 has been shown as being upgraded to operate every 20 minutes to improve access to Frankston and relieve parking pressure at its station. 

* Route 788 upgrade on the Mornington Peninsula. The peninsula's main bus route is the 788. This gets overloaded at peak times and on warm weekends. Frequency upgrades are proposed to cater for demand and improve connectivity with trains at Frankston. 

* SmartBus Route 902 frequency upgrade. Route 902 operates along Springvale Rd, which like Boundary Rd is parallel with the Mordialloc Freeway alignment. It is often crowded during peak periods and even weekends, with maximum patronage usually reached between Glen Waverley and Springvale South. Adding extra buses over this section would lessen crowding, improve punctuality and make the bus a better train feeder at Edithvale, Springvale and Glen Waverley stations. 

* Upgrades to train services. The Dandenong and Frankston lines enjoy a good turn-up-and-go frequency most of the time. However there are still times when trains are 30 minutes apart. Particularly after 7pm on weekends and on Sunday mornings. A frequency upgrade to every 20 minutes would make services more useful, particularly for those who work shift hours or catch early flights. 

* Other measures. Some aspects are not covered above. These include: (i) poor access between Frankston line train  stations and  jobs in Moorabbin east, largely because some routes stop short of stations, (ii) Frankston South, including access to Monash University, due to an existing complicated local network, (iii) coverage issues on the Mornington Peninsula, and (iv) access to jobs in Dandenong South. All these require wider local network reviews. 


The Mordialloc Freeway corridor is currently poorly served with public transport. While most people have a local bus to a station for CBD travel, its frequency (typically every 30 minutes) is often too low to be attractive. Melbourne experience is that feeder buses usually need to be at least every 20 minutes (preferably better) to start to appeal to commuters.  In addition current routes don't serve trips to major destinations like Monash University Clayton and surrounding industrial areas such as Clayton, Moorabbin, Seaford and Carrum Downs very well from most areas, with three lots of waiting and two changes often required. 

The outline of an improved local bus network based on routes operating at more useful frequencies is presented above. The higher frequencies are likely to encourage use of them by city commuters to reach stations. This supports usage of the rail network and relieves parking pressure at stations. In addition, the extended routes would improve bus travel to destinations such as Monash University more practical.  

While I've tried to minimise overlaps with existing routes, the higher frequencies and longer routes mean that money will be required to buy more buses, maintain them, fuel them and pay more drivers. However, as we see each day on Route 902, a full bus is an extremely space-effective, cost-effective  and environmentally-friendly form of transport. And, partly because most roads are untolled, most drivers aren't necessarily exposed to the full cost of their own travel. 

Comments on this network are welcome below. Is it too much or does it not go far enough? And should there be connections elsewhere, or are some ones suggested above unnecessary.