Pampered parking. Starved buses. The story of station access.

Politicians love promising and building car parks near suburban train stations. 

Why?  

Perceived good politics..

In their world the politics stacks up for at least ten reasons, as follows. 

a. Station parking is seen as a vote winner popular with busy suburban working families that are otherwise not politically engaged and could vote either way. 

b. Where there's spare land an expanded car park can be promised, built and opened in one electoral cycle. 

c. Individually they are not big projects. So you could build them in your 20 most marginal seats. 

d. They are visible, so you can see the steamrollers, smell the tar and read the sign proclaiming which government funded it. 

e. Success can be seen each day as people observe how full they are. 

f. They are scaleable, so you can promise another expansion in 3 or 4 years if your government is having a rough time. 

g. Station parking is seen as a 'public transport' project with environmental and 'congestion busting' benefits. 

h. It's something they can officially open with signs, ceremonies and selfies. Boosting buses, from two per hour to four per hour, is not seen as being as tangible, despite their potential to carry more people. 

i. They are seen to address a common anxiety in modern life, that is the commute to work. Their construction shows that our representatives (however self-serving or stupid they may appear) have the common touch and are determined to deliver for their communities. However they give rise to another anxiety - "parking rage". 

j. Station car parks are largely seen as capital projects. That appeals greatly to federal governments who, in transport, would never fund services but may fund capital infrastructure. It also sits well with the current Victorian state government which will borrow billions for infrastructure but balks at recurrent funding for improved bus service (with declining service per capita being recorded). Hence both governments have lavish station car park programs. Federal's is the Urban Congestion Fund while the state offers Car parks for Commuters. 

..but (mostly) bad policy

The problem is that with uncharged parking at stations the politics works better than the projects in the more populated suburban areas. Station parking does become more appropriate further out. For example one could justify a car park at an exurban commuter station in a low land value area too sparsely populated for walking, bikes and buses. If you want to redevelop a town centre station precinct you might redistribute its parking spaces to stations either side.   

But once we enter suburbia, with higher land values and greater population density, things change. Station parking becomes hugely cost-ineffective per train passenger potentially gained. Car parks can cost tens of thousands of dollars per space. You'd think that a slab of asphalt in the suburbs would be about the simplest thing a politician can promise but even that can fail as the Herald-Sun reported on Wednesday. 

Using vacant rail-side land is one thing but per user costs skyrocket if land needs to be purchased and/or multi-storey structures built. And each space is occupied, all day, by a passenger paying only a few thousand per year in fares. Their patronage of the trains is good but at what cost? And could it be won through more cost-effective means with wider benefits?

Station car parks can crowd-out more productive land uses. Even short-term parking would be more locally beneficial if there is retail nearby. Or it could be used for dense retail, jobs and housing that would attract a greater economic return and encourage transit usage. Having such higher density at stations could also allow suburban residential character in neighbourhoods further away from them to be preserved.  

Station parking means that a very small number of people is getting a large public subsidy as we too rarely charge for it (unlike Toronto and San Francisco). That's something to remember. Parking lots look impressive but their space inefficiency means that relatively few directly benefit. A rail network needs efficient access to its stations to get high usage. This involves most access being via walking, cycling and buses whose space impacts are much smaller per passenger. For example a bus can carry 50 people every few minutes to a waiting area the size of a few car parking spaces. And in purely capital terms you can buy several new buses for the cost of adding a few dozen parking spaces with no doubt as to which option can move more people each day.

The customer is always right. Most people are pragmatists. They will use whichever access mode works best for them. Upgrade the space-efficient modes and people will use them. That will free scarce parking spaces for those who lack the other choices. Even the complaint of there being 'not enough parking' can be mitigated with intelligent demand-based parking charging. 

Part of the clamour for station parking is that alternatives just aren't very good at the moment. Especially in outer areas where stations are widely spaced and roads are hostile for those not in cars. For people in the critical 1 - 5 km distance from a station, walking is too slow, cycling is unsafe while buses are unreliable, infrequent and need recurrent investment to improve. 

Opportunities  

None of this is necessarily inevitable. Walking can be sped with more crossings, roundabouts removed and new station platform entrances. Cycling can be safened with connected off-road paths. And buses can be made better. Imagine if all stations had buses like the 465 in transport minister Ben Carroll's seat. This doesn't look much on a map and it goes to few significant places but it feeds Essendon station well and, unlike any similar route, runs every 7 minutes or so in the peak. Williams Landing has been another feeder bus success story from almost the day it opened. A dollar spent on bus services will create continuing jobs, as opposed to the one-off 'sugar hit' of parking construction (with the land acquisition component creating no jobs except for property lawyers and agents).

Are we doing much of this? For each station that is part of a government car park program I'll briefly summarise the state of current bus services, whether there have been recent improvements and any known plans for improvements. 

I'll list stations in roughly clockwise order around the bay. Those in red are part of a state government car park program, those in blue are part of the federal government program while those in black are in both. 

Werribee

On average every 20 - 25 min in peak. Major network revamp 2015. Minor upgrades since. New route 182 planned. 

Aircraft

No significant bus connections to station. 

Wyndham Vale

Major network revamp 2015. Little change since. 

Tarneit

On average every 20 - 25 min in peak. Major network revamp 2015. Minor upgrades since. New routes 152 and 182 planned. 

Melton

On average every 30 - 60 min in peak. One route added (454) but otherwise no major change for many years. 

Cobblebank

Every 40 min approx. Route 454 added when station opened. 

Rockbank

Every 30 - 40 min approx. Route 444 added a few years ago to Aintree. 

Caroline Springs

Every 20 min in peak approx. Recent extension of 462 and improvement of 460. 

Deer Park

Every 20 min in peak approx. Last significant upgrade in 2014 with new Brimbank network. 

Sunbury 

Every 20 - 30 min peak approx. Last significant upgrade a few years ago. 

Watergardens

Every 20 min in peak approx. Last significant upgrade in 2014 with new Brimbank network. 

Donnybrook

Every 40 min approx in peak. Served by new Route 524. 

Craigieburn

Every 15 - 30 min in peak approx. Recent and planned upgrades but lagging growth. 

Merlynston

Every 30 min in peak approx. No significant network reform or upgrades for >10 years. 

Epping

Every 20-25 min in peak approx. No significant network reform or upgrades for >10 years except for Epping North/Wollert. Local routes often unharmonised or irregular frequency. 

Hurstbridge

Only bus is 343 which parallels/duplicates rail line. No feeder buses from surrounding area. 

Wattle Glen

Only bus is 343 which parallels/duplicates rail line. No feeder buses from surrounding area. 

Eltham

Every 15 - 30 min. No significant local network reforms for >10 years. Complex routes with generally poor usage. Significant areas (eg Bolton St) without coverage. 

Montmorency

Buses do not operate right to station. Nearest route (293) every 30 min with limited operating hours/days. No significant network reforms for >10 years.

Greensborough

Every 15 - 30 min. Some recent network changes but patronage performance of some newish routes (eg 343, 385) is poor. Established routes eg 513, 518 and 566 are very complex. 

Watsonia

Every 20 - 30 min. Both routes (533 and 566) very complex. No significant network reforms or service upgrades for >10 years.

Mooroolbark

Every 30 - 60 min approx. All routes have limited operating hours. No significant network reforms for >10 years.

Croydon

Every 15 - 30 min with no recent service upgrades. Served by complex 380 loop. 

Ringwood East

Every 20-30 min served by 380 loop whose introduction made buses more complex. Large area along Eastfield Rd has no bus to station.  

Belgrave

Every 20-40 min approx in peak. Complex network with no significant reforms for >10 years.

Upper Ferntree Gully

Every 20-30 min approx in peak. Complex network with no significant reforms for >10 years.

Ferntree Gully

Every 20-30 min approx in peak. Complex network with no significant reforms for >10 years.

Boronia

Every 20-40 min approx in peak. Complex network with no significant reforms for >10 years.

Heathmont

One bus route (679) every 30 min approx in peak. No significant network reforms for >10 years.

Ringwood

Every 15 - 30 min approx in peak. No significant network reforms for >10 years except for recent 742 extension. 

Heatherdale

Every 20 - 30 min approx in peak via Route 742. No significant service upgrade in >10 years except for recent extension to Ringwood. 

Mitcham

Every 20-30 min approx in peak. Some network reforms in Mitcham north area in 2014. 

Surrey Hills

Every 30 min approx in peak. No significant network reforms for >10 years. Note: Station is being merged with Mont Albert as part of grade separation. 

Canterbury

One bus route every 30 min approx in peak and limited operating hours (285). No significant network reforms for >10 years. 

Camberwell

Two bus routes every 30 min in peak (285 & 612) with limited operating hours.  No significant network reforms for >10 years. 

Glenferrie

Tram only. No significant timetable upgrades to these in last >10 years.

Jordanville

Every 20 min average during peaks (single route). No significant recent network reform or upgrades. 

Pakenham

Every 60 min average during peaks. No significant recent network reform or upgrades. 

Cardinia Rd

Every 60 min average during peaks. Route 925 stops about 800m short of station. No significant recent network reform or upgrades. 

Beaconsfield

Every 40 min average during peaks. Little substantial network reform or upgrades for >8 years. 

Berwick

Every 20 - 60 min average in peak. Recent addition of Routes 888 & 889 but infrequent. 

Narre Warren

Every 20 - 40 min average during peaks. Little substantial network reform or upgrades for >8 years. 

Cranbourne

Every 20 min average during peaks. Significant network revamp in 2016 but more coverage is needed. 

Lynbrook

Every 20 - 40 min during peaks. 

Dandenong

Every 30 - 60 min average during peaks. Many routes finish ~7pm. No significant network reform or upgrades for >10 years except for Endeavour Hills.

Yarraman

Each route is every 60 min in peak. No network reform or upgrades for >10 years. 

Sandown Park

No buses to station. 

Frankston 

Every 30 - 60 min peak on average with limited operating hours on many. No significant network reform or upgrades for >10 years. Terminus station with town centre but poor buses intensifies parking pressures. 

Kananook

Every 15 - 60 min on average during peaks. No significant recent network reform or upgrades in >10 years.  

Seaford

Every 40 - 60 min on average during peaks. Recent introduction of Route 760 but need for other Carrum Downs area connections.  

Bentleigh

Every 15 - 30 min peak on average. No wholesale bus network reform in area (added Route 627 does not serve this station). 

Sandringham (project under threat!)

Every 20 - 40 min peak on average. Complex local network with no significant recent reform or upgrades in >10 years.  

Hampton

Every 20 - 30 min peak on average. No significant recent network reform or upgrades in >10 years.  

Brighton Beach (project under threat!)

Every 20 min peak on average. Complex local network with no significant recent reform. Lacks simple  feeder route along South Rd. 

North Brighton

Every 20-40 min peak on average. Complex local network with no significant recent reform or upgrades in >10 years.  

Elsternwick

Trams and buses every 10-30 min peak on average. Some reforms in last 10 years.   

Balaclava (project under threat!)

Served by trams only. No significant timetable upgrades to these in last >10 years. 

This is a long list but the service offering is fairly consistent. The average station that's getting a car park upgrade has a couple of bus routes about every half hour, even in peaks. It's had no recent service upgrades and none are planned. 

What works for station bus access

In one word it's frequency. And most stations now have about one third or one half what it should be.

Feeder buses are potentially lower cost and more space efficient than car driving and parking. But they can only widely appeal if the service is good enough. Bad interchange and network design can make changing and waiting a hassle. This is particularly in the afternoon where people can't as easily pick their finish times and unreliable trains can make aiming for timed bus connections a gamble. The gold standard then is a service where even if a bus has just been missed the next is not far away. 

How much peak service do you need? Maximum waits between buses cannot exceed 20 minutes. If they do then commuters won't use them as feeder services. Once feeder bus frequency hits 20 minutes then you start to get work commuters (as opposed to school commuters) using them. You can tell that because of rising bus patronage in the 5 - 7am and 5 - 7pm time slots, ie before and after the school peaks. See the ATRF paper on the 2014 revised Brimbank bus network for evidence of this.

To keep peoples' confidence services levels need to be retained and where possible improved. They cannot be allowed to atrophy through lack of vigilance. An example of the latter happened when the Department of Transport sloppily allowed peak direction headways on a major Werribee - Tarneit route (180) to blow out to 28 minutes on a nominally 20 minute service as part of the January 31 2021 bus recoordination. 

20 minute waits are still not fantastic if a bus has just been missed. 15 minute waits (like standard on most parts of the SmartBus orbitals) can even be marginal. But when you get to 10 minute feeder bus frequencies then the service becomes highly saleable and people use it in large numbers. Even in outer suburbs. Melbourne's problem is that we have very few feeder routes that operate at that frequency, even in peak times. 

What happened with bus services is similar to earlier trends with metropolitan trains. The network got bigger to serve the growing outer suburbs post WWII but frequency in established areas fell. Amalgamated routes were often less frequent than their shorter predecessors. Local buses in the 1980s often ran every 15 or 20 minutes during peaks but may now only be every 30 minutes, ending their practical usefulness as rail feeders. The Hampton example below represents a halving of peak bus service since the 1980s. It's significant as Hampton is one of the stations chosen for additional parking under the federal program.   

See the full 638 timetable at Krustylink

As you saw from the above list, 30 minute peak frequencies for feeder buses is common.  This is not necessarily true in other places. It comes down to service standards. Our minimum standard for local bus routes is every 60 minutes, with no allowance for an increased peak service (though most routes are usually a bit better). Whereas in Perth even neighbourhood feeder bus routes have a 20 minute peak service with busier routes every 10 minutes. 

The example below shows a middle-suburban feeder bus route with such a service and the efficient interchange it connects to. The latter involves smart infrastructure design. Though we've built a lot of that lately it's been hit and miss with regards to bus connectivity. For example it is regrettable that we didn't get Perth-style bus bridge/interchanges for the likes of Williams Landing where walking distances to the furthest bus is too long and the indirectness introduced is unhelpful for any through-routing. 

Conclusion

The typical Melbourne suburban station has two or three bus routes operating about every half hour in peak. Some services have been added, mainly to extend coverage. However frequency upgrades are rare in areas (ie the majority) that have not had local bus reviews implemented that could have delivered simpler networks. There is thus significant scope for station access improvements with buses. 20 minutes should be an absolute minimum with 10 min the preferred peak frequency on feeder routes within 70 km of Melbourne CBD. And bus wormholes should provide congestion-free last-mile rapid access in selected dense areas to key destinations. 

More buses undoubtedly need to be bought and operated to deliver upgrades to all stations listed. However their capital cost per passenger carried is much lower than that of station parking. Plus you are getting fringe benefits such as reduced traffic congestion, better land use near stations and a more useful and versatile transport network. And household financial benefits accrue where people do not need to add another car as there are good buses that they can use for many trips. 

Index to Building Melbourne's Useful Network items