Sunday, June 06, 2010

New June 2010 Metro timetable starts today with 211 new weekly services & Coolaroo Station opens

Today was significant for Melbourne's rail network for two reasons.

New Coolaroo Station

The first was the opening of Coolaroo Station, the first new station since Roxburgh Park in 2007. Coolaroo serves three established residential areas; the 1950s housing commission suburb of Dallas to the south-east, the 1970-80s suburb of Coolaroo to the east and the 1980-90s suburb of Meadow Heights to the west.


View Large Map

Possibly due to a long-mooted rail electrification extension (opened in 2007), the station's construction is not the railway's first involvement in providing transport to the area. Victorian Railways scheduled a bus that continued where the train left off at Broadmeadows. As shown in this 1975 timetable (VR WTT), Route 903 provided a daytime feeder service between Broadmeadows and Coolaroo. Consistent with retail hours at the time, the bus ran 5 1/2 days per week.

Like the other Victorian Railways bus routes (901, 902) the service was either abandoned or renumbered and incorporated into (typically longer) conventional suburban bus routes.

New suburban timetables

The other major change today was the start of new suburban timetables. The real test will be tomorrow which will see additional services and revised stopping patterns, most notably on the busy Caulfield group. 211 new weekly services will be added to the timetable. 

In summary the changes are:

* Craigieburn line: More peak services to start and finish at Craigieburn (instead of Broadmeadows). Trains stopping at the new Coolaroo station.

* Hurstbridge line: Additional weekday evening services. Trains to Eltham will operate approximately every 20 minutes until after 9pm. This is the second line to have received enhanced early evening services; Ringwood was first a year or so back.

* Cranbourne and Pakenham lines: Simplification of peak stopping patterns and additional peak services. Off-peak weekday services change to operate express between Malvern and South Yarra, swapping with Frankston trains, which will serve these stations.

* Frankston line: Simplification of peak stopping patterns and additional peak services. Express trains to operate direct to/from Flinders Street. The increased use of direct running is a continuation of a trend started with the Werribee line, which had both peak and some off-peak services removed from the City Loop (introduced simultaneously with a doubling of off-peak weekday service frequencies). Off-peak trains will stop all stations, swapping with Cranbourne/Pakenham on weekdays and doubling weekend service frequency to stations between Malvern and South Yarra to ten minutes. Also added are some extra counter-peak train services.

A campaign of direct mail, newspaper advertisements, notices at stations, brochures, website advice and staff handing out timetables has been run to promote the changes, particularly on the Caulfield group where the changes are greatest. Because the changes represent a simplification of stopping patterns, it was possible to depict them in a simple graphical form.

As with all changes, there will be winners and losers, particularly on the Frankston line due to the loop changes. Most stations will receive the same or more services, but some will receive fewer - a price of consistency. Users of Frankston express trains who use Flinders Street will have a much quicker ride, while users of Loop stations will use their one-train ride if they want to get an express. Instead they will need to transfer at Flinders Street (recommended in the morning) or Richmond (evening).

Especially in the afternoon, it will be interesting to see the proportion of passengers for stations beyond Cheltenham (where the express service would normally be chosen) who opt for a slower stopping all stations train to avoid a transfer. If this occurs, stopping trains may well be fuller than would be expected from station usage statistics between Highett and the city, and express trains a little less full.

The choice that passengers prefer is likely to depend on boarding location. While passengers at Southern Cross can board trains that go to Flinders Street and transfer to an express service from there, passengers at Flagstaff, Melbourne Central and Parliament do not have this choice. Instead they will need to choose between boarding a stopping all stations service (which may have spare seats at Southern Cross and possibly Flagstaff, but not Melbourne Central or Parliament) and staying on to their destination, or board any train on Platform 2 and changing to an express at Richmond or Caulfield.

Several themes, some continued from the previous timetable changes are discernible from this latest timetable change. These include:

* Recognition that the City Loop (at least with current signalling) cannot handle all peak services that need to run. Hence extra services (made possible by the purchase of additional trains) have been made to run direct. In the short term this might reduce legibility (as all trains no longer run via the loop), but this can be restored by ensuring that after a transition period (such as we may now be experiencing) all trains from certain lines run direct, not just half of them.

* More consistent stopping patterns and increased frequencies on each one. This is making the network more legible and reducing distortions in patronage caused by people aiming to catch particular trains due to their super-express stopping pattern (for example). An example of an improvement on the Frankston line is that until this latest timetable the last two pm express trains were 30 minutes apart.

* Spreading the peak. On the Frankston line this timetable introduces additional early morning 'early bird' services. On the Hurstbridge line this timetable introduces additional evening services. Nevertheless this remains a critical area on the Caulfield group; for example unlike the Burnley group, where peak frequencies and express trains continue for around 3 hours in the morning and evening, the Frankston line's peak service span remains about one third this in both the morning and evening.

* Removing 'holes' in timetables. It was not long ago that services from Frankston dropped to a 30 minute frequency shortly after 6pm on weekdays. Ex-city morning counter-peak services were sometimes less frequent than off-peak services. A previous timetable change introduced additional services that filled some of these gaps. This latest timetable continues this work, with early evenings from Frankston gaining additional services.

* The importance of frequency. This and simpler stopping patterns is supported by the current train operator. We have already seen more frequent off-peak services introduced by the previous operator on the Werribee line, along with evening improvements on the Ringwood, Pakenham and Cranbourne lines. Today's timetable changes stopping patterns to provide a ten minute weekend service to all stations between South Yarra and Caulfield. In addition there are frequency improvements at some other times, as discussed above.

The pace of train timetable change has quickened in the past couple of years, after years of relative stagnation. This is appropriate since timetables are key to a service's usefulness to passengers. They also play a large part in the efficient utilisation of track capacity and the service's reliability. Change driven by how to most efficiently serve booming patronage, allows a constructive reappraisal of our use of infrastructure (including the City Loop) for maximum capacity, reliability and legibility. Although reliability is currently below standard on parts of the network (most notably the Caulfield and Northern group), it is the increasing discussion about how the network is planned and run that I think offers cause for optimism in the future.

This is Phase 1 of the 2010 timetable change. It concentrates on Pakenham, Cranbourne and Frankston. Phase 2 of the 2010 timetable change (later in the year) will concentrate on Werribee, Sydenham and Craigieburn. 

Archived Metro Trains Website item on these changes: https://web.archive.org/web/20100727160606/http://junetimetable.metrotrains.com.au/

Sunday, May 30, 2010

How much does it cost to get around in the city: bikes, public transport and taxis compared

Melbourne's public bike share scheme starts tomorrow.

Each mode has its own advantages and peculiarities. For instance public transport isn't always door to door, waiting times for taxis varies, and the bike hire scheme requires a $300 security deposit (daily and weekly members only) and BYO helmet.

The charging structures for each mode varies, and this will be the topic here.

Taxi: Fares are based on a $3.20 flagfall plus $1.617 per km or 56.6c / minute time charge (the time charge applies if the speed is under 21 km/h). A 20% surcharge applies late at night or on public holidays. Taxi fares are several times dearer than public transport fares or bike hire charges, to reflect the personal door to door service provided.

Public transport: Fares depend on ticket type, validity period and passenger. Discounts apply for bulk-purchase, periodical and weekend tickets. A single-trip City Saver costs $2.80. 2 hours unlimited travel in Zone 1 (which includes the CBD and inner suburbs) costs $3.70, while a day costs $6.80. Concession fares are a bit over half these amounts. Bulk purchase and periodical tickets can offer savings of 25% or more.

Bike Share: Charges are based on a $2.50 daily flagfall (or membership charge) plus a rising charge per 30 minute block. The first half-hour is included in the flagfall. The next half-hour costs $2.00. Above then per half-hour charges rise; to $5 for 61 - 90 minutes and then $10 for each subsequent half hour.

Comparisons: Imagine a graph with time along the x-axis and cost along the y-axis.

The taxi curve is very steep, shooting up to nearly $10.00 by around 10 minutes of travel.

Public transport and bike share are both cheaper than taxis, but they swap places depending on the amount of travel time required.

Up to 30 minutes bike share is almost always cheaper, at $2.50 (much less for weekly and annual members). However users of City Saver x 10 pay $2.18 per trip - this is the only case where full fares are lower than bike share charges for up to 30 minutes of use. While the City Saver x 10 is a bulk-purchase product, there are no restrictions on when it must be used up by, unlike periodical Metcards or weekly bike share memberships.

The costs of bike share and public transport cross over in the 30 to 60 minute region. 60 minutes of bike share costs $2.50 flag fall + $2.00 usage charge, or a $4.50 total. In contrast a 2 hour full fare Zone 1 ticket costs $3.70. Hence for the casual pay-as-you-go user, public transport is always cheaper for an hour's travel.

It is in the 60 minutes or under usage level that the benefits of long-term bike share memberships are greatest. A weekly member riding up to an hour 5 days per week pays $1.60/day flagfall + $2.00, or a total of $3.60 per day. This is comparable to casual Metcard ticket purchase (for a 2 hour ticket) but dearer than a 10 x 2 hour Metcard. Yearly bike share membership is only 20 times the daily rate, so yearly members who cycle weekly or more would pay just cents for flagfall each time. Yearly members could bike share for an hour a day and pay less than public transport users, even those using periodical tickets. Around the 30 - 60 minute region bike share and public transport costs are roughly the same, with public transport generally cheaper, especially when compared against costs for casual bike share users.

Beyond an hour the cost differences between bike share and public transport diverge markedly. Because public transport tickets are mostly time-based, and the shortest is 2 hours, it is no dearer to take transit for 2 hours than it is for 1 hour. There is also a relatively gentle rise from 2 hour to daily fares, with less than a doubling for an increase in travel time allowed by over five times.

Bike share charges behave in the opposite manner, rise steeply with time of use. Charges (excluding flagfall) increase from $0 (30 minutes), $2 (31 - 60 minutes), $7 (61 - 90 minutes) to $17 (91 - 120 minutes). At this 2 hour level casual bike hire is $19.50, ie approximately 5 times dearer than a 2-hour public transport ticket. This almost exponential rise is presumably to encourage use of the scheme for short trips in the CBD region only.

Conclusion: Of the three modes of travel compared, public transport is the most cost-effective for CBD travel. This is especially for visitors from the suburbs, who in many cases will have arrived on public transport with a daily or higher ticket, so would face a marginal cost of further CBD travel of zero. The bike share scheme occasionally works out cheaper, but only for those frequently making short trips.

Sunday, May 09, 2010

Effective Frequency - a measure of timetable co-ordination

One of this blog's hobbyhorses, that of timetable co-ordination, made it into today's Age. Not only that but it even got a one-liner on Gold 104 News (heard while travelling on the 901 SmartBus if anyone asks).

The above link has a spreadsheet prepared by the PTUA. There is also a web application showing average connectivity by route and station (note: timetable data from last year).

More buses and timetable co-ordination were cited as ways to improve connectivity. Though the article focused on buses, train frequency can also be part of the solution, especially in outer areas where hourly minimum standards buses cannot connect with trains every 40 minutes and middle suburbs where SmartBus and train frequencies don't always match.

Average connection times provide a rough guide, but as they say, 'your connection may vary'. Provided headways are fairly constant (which they are for most services), a measurement method I like is the concept of 'effective connected service frequency', most commonly for a suburban bus feeding a city train.

Let's take Route 436 at Werribee, a local route that scored well for connectivity. As the service is headway harmonised with trains throughout the day and connections are good, the effective connected service frequency of a train+bus trip is the same as the bus route's individual frequency (in this case 40 minutes) seven days per week. This is how it should be for an effective network.

A similar suburb to Werribee, but in Melbourne's north, is Craigieburn. However its buses operate at frequencies incompatible with trains. Therefore passengers need to be choosy about what services to catch to minimise waiting.

A sample of Saturday train/bus times ex city are as follows:

Train arrives 9:52am

Bus 528 departs 9:57am

Train arrives 10:12am

Bus 528 departs 10:27am

Train arrives 10:32am

Train arrives 10:52am

Bus 528 departs 10:57am

Train arrives 11:12am

Bus 528 departs 11:27am

Train arrives 11:32am

Train arrives 11:52am

Bus 528 departs 11:57am

With a 10 minute maximum waiting time allowance, the effective connected service frequency is only 60 minutes (connecting services shown in bold), despite each part of the network operating more frequently. Even though Werribee's local buses are less frequent than Craigieburn's, Werribee's effective frequency is higher due to better connectivity. Werribee passengers get a good connection every second train, but for Craigieburn it's every third train (this is on weekends, on weekdays Werribee has a higher train frequency).

Pakenham on weekends has similarities to Craigieburn. The difference here is that the frequencies are halved - trains every 40 minutes and buses every 60 minutes. The effective connected service frequency ends up being two hours because of unharmonised scheduling. Not exactly attractive for your discretionary passenger, and inferior to the planned connectivity that hourly minimum standards routes were intended to give.

Everyone knows that frequency is key to attracting patronage. At the same time buses cost money to run, so it is not always possible to deliver the highest frequency.

However effective service planning requires that the effective connected service frequency is no worse than feeder bus frequencies to maximise network connectivity and thus patronage. Not doing so means that taxpayers are paying for a higher service frequency than can be realised and passengers are getting a lower overall frequency than ought to be possible given the number of buses on the road.