The new Coburg Station

A few pictures of the new station, visited a few days ago. 

Although the station is open for passenger service, the most direct access to the north is not yet, with long walks to buses. There is not even any wayfinding signage to buses, such as installed on a mass scale during the Metlink signage era (about 15 years ago). 

The new Coburg station looks shiny and nice but of long term significance is its poor design as a transport hub. As pointed out by the Upfield Corridor Coalition, it should have been built to straddle busy Bell St, which sees 13 buses per hour (offpeak). This would have enlarged the station's walking catchment and improved connectivity with buses by allowing people to catch buses either east or west without negotiating one of the northern suburbs' busiest roads. The diagrams below compare best and more typical design practices for stations involving elevated rail.

An emerging pattern with new stations that emerge from grade separations is that their designers do not always see the public transport system as a whole, including the need for interchange between modes (that should ideally be just a few steps). Maximising walking catchments measured in accessible population / within 10 minutes walk (including that required to cross major roads that poor designs impose on station users) should also be another key criteria when evaluating designs. 

Building Melbourne's Useful Network Part 76: Goodies in our new train timetables (450 new trips!)

Today is the 'great reveal'. We see what's in the new Metro and V/Line train timetables due to start on January 31. Significant projects that triggered them include the High Capacity Metro Trains and the Ballarat Line Upgrade. 

Even more significant is the opportunity being taken to tidy some decades-long timetable irregularities and service shortfalls. Many were planned to have been fixed in 2015 but a nervous government on a slim margin flicked the switch to infrastructure.

Though a major rationale of Regional Rail Link was to free up paths for more Werribee trains, just two were added to the timetable on Day One as a token effort. More peak trips were added in 2017 with this 2016 Age article giving the background.

Essentially it's a story of service being sidelined until very recently. Hence the five year deferral of major metropolitan train timetable changes, stagnant tram timetables and widespread metropolitan bus network reform being in abeyance. 

Last month PTV added a website section to communicate the changes. However it undersells the improvements happening, especially on Werribee and related lines. 

The habit of quoting a given number of extra trips per week is a provider-oriented mindset that inflates numbers and conceals the times at which frequency is being improved. A better account is provided on this Daniel Bowen blog post. Even so some questions remained unanswered until we saw the timetables that came out last night. 

What we already know

Some things were clear from the PTV website item. These include the following: 

* Cranbourne and Pakenham trains would enter the City Loop in an anti-clockwise direction at all times. This is a major simplification that would end that group's midday weekday-only reversal and provide a consistent direction all week. These lines would also run express full time, with Malvern through to Hawksburn being served only by Frankston trains. 

* The cross-city group would function consistently 7 days per week (well, almost). This provides a simplified pattern for Frankston, Werribee, Williamstown and Altona/Laverton trains. You will be able to board a train at Caulfield or South Yarra and be confident it will run straight through to Footscray and Newport. As part of this Newport will gain a 10 minute weekend frequency. Also the confusing weekday peak loop trains on the Frankston line will run cross-city like the rest of them. All this should simplify the service and allow departures from the same platforms all week.  Stations south of Mordialloc will also gain with some evening services extended (previously these finished at Mordialloc or Carrum). Click for better view. 

* Sandringham trains being taken out of the City Loop on weekends and will terminate at Flinders St. Hence, like the Caulfield and cross-city groups it will follow a consistent pattern all week. Additional off-peak services are added but no advice was given on what days they would run. 

* Peak services on the Craigieburn, Sunbury and Werribee line were upgraded to a 10 minute or better frequency over wider periods. This is a worthwhile boost to shoulder peak services on key lines that are only every 20 minutes at most other times. Peak Williamstown trains will be upgraded from the difficult to remember 22 minute frequency to an even 20 minutes. Upfield also gains a few  peak services though at somewhat irregular frequencies. 

All this is mostly but not entirely good. For example the post am shoulder peak has frequent inbound service extended by about 30 minutes on the Sunbury line. However immediately after that there is a 24 min hole with no trains from Sunshine, one of Melbourne's busiest suburban stations that the government was promoting as a transport hub. 

The timetable below shows the peak direction in the evening. This also gets a broader shoulder. However the frequency upgrades are quite small (eg from 20 to 17-18 min around 7pm) diminishing to nothing after 8pm (where the schedule remains half-hourly).  A similar pattern applies for Craigieburn, where the 30 minute weeknight service cuts in at 7:45pm (currently 7:47pm). 

* An improved evening frequency with the Frankston line improving from every 30 to every 20 minutes late on weeknights and after 7pm on weekends. This was stated but not very strongly. This upgrade is one of the most important restorations of service since the 1978 cuts that saw evening frequencies cut from every 20 to every 30 minutes across the network. The upgrade slashes maximum waits on the Frankston line from 30 to 20 minutes 7 days per week. Frankston has now overtaken Sandringham as the most generally frequent line on the network at all times. This sets a standard that longer and busier lines should follow. More on the recent tendency to roll out more 20 minute service here. 

* Ballarat V/Line trains to be every 40 minutes off-peak on weekdays. This is a substantial upgrade on the current hourly service. Weekend service is hourly.

* There will be cheaper off-peak fares for interpeak and after 7pm weekday trips for trips including Fare Zone 1. This might help spread the peak and encourage off-peak travel. No fare cut for Zone 2 only trips though. 

What PTV didn't make clear (but the new timetables reveal) 

I've long argued that PTV hasn't been good at communicating and selling service improvements. The early advice for this timetable is not an exception. It has more good things than PTV are telling people about (as of today). Instead they rely on people to compare timetable frequencies for themselves. Few will. Let's hope they realise this and ramp up information nearer the timetable's start date. 

Here are the main undersold improvements: 

* The peak frequency of Altona trains. Like Williamstown these currently run at difficult to remember 22 minute intervals. You would think that if Williamstown is improved then Altona would as well. However the PTV communication did not make this clear. 

Altona passengers rightly complain that they are neglected by PTV communication. This is due to the latter's ignorance of the network (Laverton being the only station that on weekdays can be reached via two rail routes from Melbourne, one of which is via Altona).  The new timetable confirms that Altona peak trains will run every 20 minutes, receiving the same upgrade as Williamstown. This is good news and restores full Useful Network service to Altona that was lost in 2011 (though I took liberties and retained it on my frequency maps). 

* Late weeknight and after 7pm evening frequency on the Werribee and Williamstown lines. There will be a major upgrade with the maximum wait dropping from 30 to 20 minutes. That's like the Frankston line, however unlike Frankston it's not communicated at all. The new timetable confirms that weeknight and weekend trains will run every 20 minutes until midnight to Werribee and Williamstown, a major service upgrade. 

* Cranbourne and Pakenham weeknight service. Previously every 30 min after approximately 10:30pm. Now improved to every 20 min until midnight. 

* Weekend 10 minute service to Newport via 7-day cross-city group. Currently Werribee trains run via  the City Loop and Altona while Williamstown trains operate as shuttles on weekends. The 7 day cross-city pattern will make arrangements more like weekdays. But not completely. As an example, Newport on weekdays offpeak sees 9 trains per hour, with direct services to Werribee, Laverton (via Altona) and Williamstown.

On weekends Newport service will increase from 3 to 6 trains per hour, with Werribee trains operating via Altona (unlike on weekdays where they don't). Hence this will remain as a minor weekday/weekend inconsistency. Williamstown will also be inconsistent on weekends because passengers will gain direct trains to the CBD between approximately 11am and 7pm but will need to change at Newport outside those hours. 

* Major Sunday morning improvements. This has long been a sore point amongst train travellers especially in northern and western Melbourne. Services are typically every 40 minutes despite high travel demand. In contrast most eastern lines get 30 minute Sunday morning frequencies while Sydney and even Perth enjoy 15 minute intervals. In a major but (again) not promoted upgrade the new timetable improves Sunday morning frequency to every 20 minutes from Werribee, Williamstown and Frankston.  

Frankston's gain is less due to its higher existing service (every 30 min) but there are some oddities I'll mention later. Lines remaining with 40 minute Sunday morning service after this change are Sunbury, Craigieburn, Upfield, Mernda, Hurstbridge, Sandringham and Pakenham. 

* Melton line service. It's been well communicated that Ballarat will get trains every 40 minutes. What hasn't been communicated well is what happens at major closer in stations like Melton, Caroline Springs and Deer Park. Currently these have two trains per hour at uneven intervals. 

This new timetable increases service to Melton and Caroline Springs to three trains per hour and an even 20 minute frequency. This is good and expands the weekday Useful Network to Melton township for the first time. Deer Park and Ardeer also gain, with the former benefiting from Geelong trains. My Useful Network maps will be revised when the new timetable comes into effect.

For places like Melton, this timetable looks a lot like the first (2015) Regional Rail Link timetable. That increased weekday service from hourly to every 20 minute but left weekend service at every hour. Crowding later forced an upgrade to every 40 min, though 20 min would be desirable, at least to Wyndham Vale. The decision to leave Melton and surrounding stations with a 60 minute weekend service (despite a 20 minute weekday frequency) will likely underwhelm many residents who were hoping for more. With weekdays boosted, a weekend frequency upgrade is now the main thing standing between Melton and the full suburban level train service it needs. 

Another key need, along with the planned grade separation and station rebuild, is a bus interchange at Deer Park Station to recognise its role as a major connection point in an otherwise underserviced area. Although Sunshine is publicised as the area's main transport hub, the potential of Deer Park for Geelong - Ballarat line and more local interchange (including bus) should not be overlooked, especially given its increased train frequency.   

Oddities and loose ends

I haven't had time to look at too many oddities or pay too much attention to stopping patterns. I have not seen the increased trips that PTV vaguely promised for the Sandringham line but didn't say what time of day. 

Timetable reform is supposed to make services simpler to use. There are a couple of cases where this timetable only half-heartedly does this. For example the Werribee line has inconsistencies as to whether trains run via Laverton or not. That's normally a weekday/weekend thing (apart from weeknights). Then there is whether Williamstown passengers must change at Newport. That's mostly a day/night thing (with times varying by day). If you wanted to get a train from say Frankston to Williamstown or Altona it will remain quite confusing if or where you need to change.   

Observers were hopeful that the new timetable would resolve the confusion around the City Loop and its weekday midday reversal. So far they have not got off to a good start with inconsistent start locations and up to 80 minute gaps from Loop Stations on this Cranbourne line timetable. Passengers at loop stations may still need to change trains and build buffers into their itinerary to be assured of catching a particular train (such as might be required for a connection to an hourly bus). 

If they can't schedule, communicate and run loop services consistently and reliably, concepts to smash the loop open (eg rebuild it to enable Richmond - Parliament - Melbourne Central - Flagstaff - North Melbourne through running) could be worth another thought. While seen by some as heresy (as not all trains would run via Flinders St) this more legible arrangement would would increase core capacity at a relatively low cost. It is not a crazy idea and was officially proposed under Stage 4 of 2012's Network Development Plan - Metropolitan Rail.  More recently it has received the backing of the Rail Futures Institute in their Melbourne Rail Plan 2020-2050.  

I also noted a potential issue with the Frankston line concerning the interface between Night Network trips and regular Sunday morning trips. This arises because although the Night Network solved the issue of the Sunday timetable's late start of service on Sundays, it does not on other days where a Sunday timetable is effective but Night Network isn't. This happens on Good Friday every year and Christmas on most years. 

Before Night Network started the first trains that arrived at Frankston arrived at approximately 9am on Sunday. When Night Network was introduced the first non-Night Network train was made a bit later. That remains the case today, with the first arrival at 9:08am. This is important for Frankston because a number of very long and sometimes high profile bus routes depart shortly after 9am Sundays on their first trip. If you don't make them you face an hour wait.  

Assuming that Night Network is defined as when the hourly services are running, the January 31 timetable makes the first non-Night network arrival at Frankston significantly later, at 9:24am - after all the buses have gone. That likely represents as service span cut early on Good Friday and (mostly) Christmas morning. These are likely to be quiet times for travel but it is a cut nevertheless unless the previous one or two Night Network trips can be made part of the regular timetable. This should be done across all lines anyway so that 7 or 8am trips become part of the core daytime timetable on all days. A 30 or even a 20 minute frequency should also apply then to benefit early Sunday morning regional and airport connections.

The other issue is the lumpy frequency - from every 60 to every 20 minutes - on early Sunday mornings for trains towards Frankston. The new timetable adds more after 9am trips but exacerbates existing oddities before 9am.   

Where possible one should jiggle services around these transition periods to avoid sudden frequency changes. In this case it is most unlikely that travel demand at 8am is just one-third of what it is at 9am. And bringing forward the 9:24am arrival to 9:00am approx would allow a smoother frequency transition and assist with connections to buses leaving shortly after 9am. 

Also the Stony Point line remains unusual in that its last Sunday trip from Frankston remains nearly an hour later than the last Monday - Thursday evening trip. This early finish makes the Stony Point line difficult for even CBD office commuters to use. Stony Point's timetable has not been significantly changed in this timetable revision. 

Sandringham's timetable needs a lot of study to find substantial improvements. However the new timetable lengthens waits around 7am Saturday mornings (from 20 to 30 minutes). And there's a curious inbound express trip on weekdays at 6:05pm from Sandringham. That trip is in the current timetable but, like the other trips, stopped all stations. Because Sunday mornings were not fixed it retains the oddity where trains run half as frequently around 8am on a Sunday morning (every 40 min) as at midnight Sunday evening (every 20 min). 

A recurring theme of this blog is that there is little relationship between the service a line (or route) gets and objective measures like patronage levels or needs. This timetable change continues this theme with the busiest lines getting less than some quieter lines. For example the cross-city WWF lines (Werribee, Williamstown and Frankston) plus Sandringham are each quieter than Dandenong. However they get upgraded to a 20 minute maximum wait, unlike Dandenong that retains its 30 minutes. Future timetable upgrades need to adopt 20 minute maximum waits at key stations such as Ringwood, Watergardens, Craigieburn and Mernda to spread benefits more widely and boost patronage.   

Buses

There are some new bus timetables online. For example some in the Caroline Springs area eg 460. Craigieburn area timetables like 533 have finish dates of 30 January 2021 so it is likely that these too will have revised times. It is understood that bus timetable changes will occur in dribs and drabs over the coming months. 

Conclusion

These rail service improvements will make catching trains easier and cut waiting at times people need to travel. 

With a minor exception of early Sunday morning, when the last couple of Night Network trips are running, you will now have a maximum wait of 20 minutes between 6am and midnight on the Frankston, Werribee and Williamstown lines. Hence they join the club that the Sandringham line was the sole member of after its upgrade in the early 1990s. Yes, train timetable reform in Melbourne can have time-lines that equal the biggest infrastructure projects, purely due to a past lack of political will. We still trail Sydney and Perth (with their widespread 7 day 15 minute service) but not by quite as much. 

Long-neglected Melton finally got some love, with it joining the Useful Network. No longer will  middle-suburban Ardeer have its sparse trains on weekdays, though the Melton line's weekend offering will remain slim. And there's some small but worthwhile shoulder peak additions on busy but historically neglected lines including to Watergardens and Craigieburn.  

These are the sort of relatively low cost service upgrades we should see more of. It is encouraging that after several inactive years that we are once again seeing service boosts. This should be the start of a program of upgrades across all rail groups so that by the time of the 2022 state election the service on most lines, especially those serving marginal seats, is substantially upgraded compared to now. There is no time to waste.

 

See other Building Melbourne's Useful Network items here

Building Melbourne's Useful Network Part 75: Should we have a new Keilor East Station?

A few days ago we read that Melbourne's proposed Airport train will use elevated rail via Melbourne's north western suburbs including Sunshine North, Keilor East and Airport West. This was not unexpected; if you do not have at-grade space then elevated rail is much cheaper than tunnelling. 

The Age article on the same topic mentioned that Moonee Valley Council was campaigning for a new  Keilor East/Airport station as part of the project. It's running an online campaign at stillwaiting.com.au  complete with cheesy video. No doubt the state transport minister, whose seat (Niddrie) any new station will be in, is following with interest. 

The station would add heavy rail coverage to an area that only has buses and the slow (although frequent) 59 tram. That's a substantial  benefit, especially if the area's bus network was revised to feed the train. However a new station would add to project costs and may mean that the much claimed 'under 30 minute' airport - CBD travel time would not be delivered. Although even a 32 or 33 minute time could still be described as being 'about half an hour'.  

Styles of airport rail service

Before we go into the specific merits of a new station, let's look at the generalities of airport rail. I needn't say much as the principles are covered in this Human Transit item. 

You may eventually have different types of rail serving the airport, as indeed large cities like London do. But if you're doing it sequentially the first thing you'd do is to connect it to your city's main electrified rail system. That's the subway, the tube, the underground or, over here, the Metro Trains network. That would provide connectivity to the CBD and important interchange points along the way, like Sunshine and Footscray. The Sunshine connection, in particular, helps not only Melbourne's west but also Geelong, Ballarat and potentially Bendigo (if their trains can be persuaded to stop there). This is the proposal that got funded in the November 2020 state budget. Completion is expected in 2029 but there's scope to bring forward many of its benefits, as explained here.  

The second stage might be to route one or more regional lines via the airport. That would be highly desirable where you've got large regional cities with high passenger flows to the airport. And there may be some wider network benefits, for instance if a via airport routing allows the separation of fast regional trains from slower metropolitan trains. A possible example for Melbourne is this Rail Futures Institute proposal to divert Bendigo and Seymour trains via Melbourne Airport on new lines. It's a bold idea but given the state's population distribution (concentrated in Melbourne with a substantial inner-regional component that has easy rail access to Sunshine) I can see why it's not been the first priority here. Also you'd want a higher CBD - Airport frequency than the regional lines by themselves would provide (unless both Bendigo and Seymour went to every 20 min each all day).  

There is also the 'fast rail' concept. This places CBD to Airport journey speed above all else with a dedicated line, no intermediate stops and expensive to build tunneling. It may be backed by a private consortium and be operationally separate from the wider rail network. An example of this was the AirRail Melbourne proposal that the state government rejected. Sometimes these types of projects are justified on the basis of providing a dedicated corridor for longer distance high speed rail (which may or may not ever be built). However if we zoom in to the metropolitan and even inner-regional scale the network benefit is less than the first two options. A focus on CBD/business travellers may also be too narrow when other airport rail options could better serve a wider passenger base including leisure/suburban travellers and even airport workers at a lower cost. And given all the time-wasting that happens at airports, I'm not sure if a 15 minute (maximum) saving in travel time that dedicated rail would offer is such a big deal anyway. The Human Transit article is critical of this type of service at Toronto's Pearson International Airport. 

Airport only or other stations?

As well as the type of service, there is also the issue of whether it serves stations on the way to the airport and/or goes beyond it. Topologically it's better if it does, provided there is a large populated area beyond it to serve. The maps below (click for better view) show all Australian cities that either have or are building airport rail links. 

Sydney, assisted by its airport's near-CBD location, has good rail access not only from the CBD direction but also to the outer south and south west via rail-rail connection points like Wolli Creek and Glenfield. Along with their all-day frequent service, the large number of non-CBD train-train connection points is what makes the Sydney rail network more useful than any other in Australia. 

In second rank is what Perth is building. There's no outer rail-rail connections. However the High Wycombe station will provide a good parking or feeder bus option for currently poorly served foothills commuters. Also Redcliffe station will provide some new bus connections and development opportunities in a handy area just minutes from both the CBD and the airport. Hence more than just airport users will benefit from the line. 

The same cannot be said for Brisbane's Airtrain. It's a stub line. It serves no station apart from the airport terminals. The city's only orbital bus route skirts around the area, so there are no bus connections either. Hence its airport rail is for one purpose only. Even for that its effectiveness is mixed with an uneven timetable sometimes featuring 30 minute waits between daytime trains. 

Which way will Melbourne go? We know that its frequency will beat Brisbane's, with a ten minute service planned.  However the map above indicates, like Brisbane, the airport line is a stub with no unique stations beyond or closer in. That will restrict its benefits unless we can do other things. 

Possible usefulness of short stub lines 

(skip this bit if you just want to hear about a new station)

Just because a new line adds just one unique station doesn't mean that it cannot be used to provide a wider network benefit. Below is a diagram of a hypothetical established railway (dark blue) that has had a new, shorter line (light blue) added from Station Z to Station E. Contrary to the impression  that may be gained from the diagram (which shows a possible service pattern, not track configuration) there is no extra track between stations E and I.

The established line is busy while the newer line is expected to be quieter. You might want both lines to run into the CBD (right of diagram) but you want to do it in a way that there isn't overcrowding on some trains while others are quiet. In other words you want to balance the loads between trains. And, because the dark blue line is long you are mindful about travel time and want express running if possible. 

One option is to modify the dark blue line trains so that instead of stopping all stations they express through F and G. To retain service all trips on new light blue line would stop all stations. This spreads benefits over the whole line. For example dark blue line crowding would be less as it would no longer have two station's worth of patronage. Instead of having 7 station's worth of passengers at station H it would only have 4.7 station's worth (as some passengers from station E would board the light blue service). Meanwhile the light blue line would benefit not only new station Z but would provide less crowded conditions for people at stations F and G, for whom it would become their unique service.  

The expressing would also speed travel by two minutes for dark blue line commuters. That might not seem much but the psychological effect of fewer stops makes the saving seem more. There are however capacity trade-offs as discussed later. 

What about those going from the dark blue line to stations F or G which it would now skip? Passengers would need to change trains at station E. However it would be the best possible change, assuming both lines share the same platform. And it's only a three minute wait since the light blue stopper has been scheduled to depart just after the dark blue's arrival. 

I have so far assumed each station has about the same walk-up patronage. However in practice it tends to be uneven. E, as a junction station may be a major centre with significant surrounding density. Because the service is frequent (12 trains per hour) people are likely to treat the service as 'turn up and go'. If that is the case then people will turn up at random times. Given this even distribution of arrivals versus the uneven train departures (with 3 and 7 minute gaps) those trains that arrive after the 7 minute gaps are likely to be more heavily loaded than those that arrive after a three minute gap. If you look at the timetable above you will see that it's the (already busier) dark blue line trains that get the higher loading. Making them 2 minutes later could ameliorate this slightly and provide an even 5 minute service. However waiting for those needing to change trains is increased from 3 to 5 minutes. If F and G aren't major destinations then this change is probably acceptable. 

Capacity was mentioned before. If you look at gaps between trains in the above timetable at Station E you can see that there are 7 minute gaps around 10:07, 10:17, 10:27 etc that you could shoehorn another train in every 10 minutes. However because this gap is between a stopper and express the gap narrows to 5 minutes later. You might be able to squeeze a train into that gap but this depends on the capabilty of your safeworking and signalling systems. And there may be greater knock on effects if trains are delayed. Eventually you get to a point where the uneven gaps caused by expressing lessens train throughput capacity.

Ways around that are passenger-unfriendly 'skip stop' arrangements or adjusting timetables so that expresses are no faster than stoppers to create perfectly parallel lines on the train graph and maybe squeeze in extra trips. Or you could have all peak trains stop all stations with expressing off-peak only. While that sounds counter-intuitive it does at least maximise peak capacity while providing a small incentive to travel during the shoulder peaks due to express trains.  

Managing capacity is a larger discussion I won't go further on. The key point is that even if a new line adds just one station (like Melbourne's airport is proposed to do) it may have wider network benefits beyond that one added station if services are scheduled carefully. 

A real life example is Perth's Thornlie line, where despite adding just one station, can justify itself by allowing daytime express running on the longer and busier Armadale line Monday to Saturday (with only the limited Sunday timetable restricting its frequency and headway evenness then). Similar comments apply to Adelaide's extended Flinders line, though benefits are less due to its lower frequency than the Perth example.   

If the new line has a wide spacing between stations Z and E there might be a case for an intermediate station. Especially if it fills a network hole where there's only infrequent buses or (at best) slow trams nearby. As well as the faster travel wider network benefits could include local redevelopment opportunities, relief of patronage pressure at surrounding stations, more balanced patronage between the dark and light blue lines and a logical terminus for local buses.  

A station for Airport West/Keilor East

Last year the City of Moonee Valley set up the stillwaiting.com.au website to campaign for a station at where its material terms Airport West. Campaign efforts have included a petition, video advert, billboards and cinema advertising. 

And there is little doubt that the area isn't that well served by rail compared to similarly distanced areas 15km north, east and south of the CBD. The local map below shows the geometry. There is almost 90 degrees between the line to Sunshine/Sunbury and that to Craigieburn. 

The Airport West tram helps but is quite slow. And it heads north, missing a substantial area to the west near Keilor East. Buses 406 and 465 provide substantial coverage but have weak northern termini and generally inferior frequencies and operating hours compared to trams and trains. And some local trips, such as Keilor to the CBD or almost anywhere to the airport are quite slow or require multiple changes. 

Similar comments apply to Highpoint Shopping Centre. It has a tram but it's the 82, which is the runt of the pack with regards to operating frequency. Traffic is clogged and the area is rapidly adding residential density. Bus services are limited and haven't been comprehensively reviewed for decades. It almost got a SmartBus but missed out when the Blue Orbital was scrapped. 

Hence, amongst the many dozens of airport rail concepts, some have sought to fill the 90 degree gap with a north-western line including stations at locations like Highpoint, Keilor East and Airport West on the way to Melbourne Airport. That way the airport line would have far wider benefits than just to flyers. An extension of the special event Showgrounds line has also been suggested. You can read a history of various proposals for airport rail in this parliamentary paper. 

In the end everything, including that which was supported in 2018 and funded in the 2020 state budget, keeps coming back to a Sunshine alignment. From there trains would run via the Albion - Jacana alignment before branching north around Airport West to head to the airport. No intermediate stations were included in the announced plans though in the current public consultation many have suggested one (or more). 

It's something like 10km between Albion and Tullamarine. That's a long spacing between stations for a metropolitan railway, especially one that traverses a suburban residential area. Moonee Valley Council claims that a station at an intermediate point would benefit a catchment of over 25 000 residents. Hence their campaign. 

Before I go further, I want to clear up a naming matter. The council describes the station site as being 'Airport West'. Which is technically true given the Keilor East / Airport West border is south of Calder Fwy and all of Roberts Rd is to the north. 

However Airport West has one clear and established centre. That is the area near Westfield, the 59  tram terminus and several bus routes including the 902 SmartBus orbital. A station named Airport West would cause untold confusion given the established centre and transit hub is more than 2 kilometres away with (possibly) no public transport link between the two. Plus it might confuse some airport passengers, something to be avoided at all costs. 

Instead, despite it being technically slightly outside the suburb, any new station might better be called 'Keilor East'. That would lessen confusion with the shopping centre, tram and buses. And it would be a more legible interchange with nearby buses from suburbs with names like 'Keilor Downs', 'Keilor Lodge', 'Keilor Park', 'Keilor East' and Keilor itself. Not to mention the avoidance of confusion for airport users. Its only disadvantage is that there is a Keilor Plains station on the Sunbury line. But apart from that there is no contest and I will refer to the station site as 'Keilor East' from here on. 

So should we have a station at Keilor East on the Airport line as the council advocates? Yes we should. The benefits outweigh the disadvantages. The only penalty is a slightly slower trip for airport travellers. Instead of being 29 minutes it will be a little over 30 minutes. That's pretty small in the context of (a) the time people waste waiting at airports anyway, and (b) the infrequency that most people would use the airport train. 

On the other hand a station at Keilor East, especially with well-connected buses, improved access over a wide area for diverse trips. Not only for CBD commuters, but also for trips to destinations including Sunshine and Melbourne Airport. Some journey times could be cut by as much as 20 or 30 minutes if not more.  Plus it could provide an impetus for a more efficient bus network for reasons discussed next.

You could argue about the location though. It's highly unwalkable and doesn't strike me as being the ideal spot for a transit oriented development. Like Airport West (which already has a major shopping centre, a tram and a SmartBus) could be. On the other hand the line to the airport heads north before Airport West on the map presented so it's likely not an option without a substantial rethink.    

Benefits for buses and the wider network

A major problem with the Keilor East area is that, without it having a station, there is no major local destination where buses could be terminated at. As an example key routes like the 406 and 465 have weak termini. And though they are well used the weak terminus means they have unidirectional patronage. Whereas a strong terminus would increase usage with largely the existing bus fleet as both mornings and evenings would attract more balanced bidirectional patronage via stations at both ends of the route. 

One feeder bus network concept is shown below. Thicker routes are more frequent, typically 20 minutes off-peak and every 10-15 minutes peak. There will be some extra costs compared to the current network but usage with this network is likely to be higher with improved connectivity between major destinations. Click for clarity. 

  

To summarise: Route 406 extends directly north to Keilor East Station to allow airport access with just one change. Route 465 provides the same facility from Buckley St. Some CBD or Sunshine passengers will prefer to ride it west rather than east for a new Keilor East station connection. Route 476 is unchanged except for it being routed via the new station. That would greatly CBD reduce travel time from Keilor and provide a good Watergardens/Keilor Downs airport connection. Potential exists to increase its peak frequency to compensate for the removal of Route 465 from Keilor East (whose residents will generally benefit from the dramatically shorter distance to their nearest station). 

Some less frequent neighbourhood routes might also be extended to the new station. These include the upgraded 407 that replaces the 406 loop in parts of Keilor East. Depending on the home locations of industrial workers, the 482 industrial route might run to Keilor East Station rather than Airport West. A basic Airport West to Keilor East connection could be provided by extending the (small) 490 demand responsive route via the Sharps Rd industrial area and Keilor Park Reserve (which would gain a service) to the new station as a fixed route on that portion. Or if that's not practical potential may exist to split the new 469 at Airport West to do likewise. 

If a frequent and fast connection between Airport West and Keilor East is desired that may be doable with something along the Ring Road. The problem here is the lack of intermediate stops. I've drawn in a red route that could continue west then south across the Maribyrnong to Sunshine (as per the current 903) or to St Albans to replace the 419 (which could be routed south to Sunshine to provide an eastern version of the successful 420 west of the Sunbury line).  A cheaper option could be to use the 406 for this purpose (as discussed in previous weeks) for less overlapping but also reduced connectivity to the new station.  

This is just one network concept. Like any other there are pros and cons. However given the number of 'dead end' bus routes in the area a new station is likely to lead to connectivity improvements. And, as I've mentioned before, a new station is often a catalyst to bus network reviews, which although they have merit in themselves, rarely happen unless there is an associated rail project.  

Conclusion

One might hesitate to build a dedicated line, but given that one is going to be built for another purpose, there's a good case for a Keilor East station to serve a currently poorly served residential area within 15 km of the CBD. Benefits include improved airport and CBD access plus local connectivity improvements in an area that has a discordant mix of old and new, connected and disconnected. Just avoid confusion and don't call it Airport West!  

See other Building Melbourne's Useful Network items here