Saturday, March 28, 2020

Moving the essential: Train, tram and bus timetabling for the COVID-19 emergency

Things are moving fast. A few days ago the line was public transport was an essential service and all services would run.

However patronage and fare revenue were falling. Those who could were urged to work from home. Entertainment venues closed. Restaurants offered take-away or delivery services only. The premier told everyone to stay home. 

Yesterday the RTBU stated that while public transport was an essential service there could be some short-term reductions in service. 

US transit agencies have asked for billion dollar bail-outs while UK has suspended rail franchising. Today's Age is reporting on crisis talks to keep our services going. 

Measures already taken include more frequent cleaning of buses, trains and trams, ticket refunds for those who no longer need to travel, cordoning off seats near the driver, encouraging social (ie physical) distancing, and discouraging the use of cash. 

We've so far avoided service reductions. But Radio 3AW this morning mentioned the possibility of us moving to a Saturday timetable (or modified Saturday timetable).  No official announcement has yet been made. But I wanted to tease out the implications of this or other arrangements if (or when) it happens. 

Objectives for an emergency timetable

Before I do that it's worth stepping back to consider the objectives of such an emergency timetable. A few I can think of include:

* Maintain a service so that those in essential industries can still get to and from work

* Retain mobility for people to access essential services eg food shopping and medical care

* Be operationally possible with an ability to roster, schedule and implement at short notice

* Be easy to explain and understand so that people can plan ahead and are not stranded

* Maintain a reliable service with a reduced transport workforce (some may be ill, in isolation or have caring responsibilities especially if schools shut)

* Save costs given reduced fare revenue and cut over-servicing given the dramatically reduced demand (but noting that social distancing dramatically reduces train/tram/bus vehicle capacities)

* Even if not fully justified by demand maintain a minimum network coverage, span of hours and frequency so that those who need to travel can still do so with only a little more waiting for connections.

* Maintain goodwill towards public transport so that patronage quickly recovers post-virus. 

* Keep transport workers in jobs (so that services can quickly recover once travel demand resumes) and as a form of economic stimulus to retain employment levels in tough times.

Some of these conflict. For example low frequencies are poor for those whose jobs either have fixed start/finish times or variable times workers can't easily influence. Those whose trips involve connections may be doubly inconvenienced. On the other hand excessive frequency (like three to five minute headways on major lines) is overkill. Frequencies in the 10 to 20 minute range might be a fair compromise.  

Needs by mode

Different modes have different needs. Quiet roads might mean that bus and tram timetables could be sped up. Especially for trams it might be possible to run a high frequency with fewer drivers, but only if we moved to headway timetables that did not penalise early running. Many in tram-land either have CBD white-collar jobs where there is a work at home option or are switching to cycling. This might disproportionately reduce peak tram usage. 

The same could be said for peak train usage which again is dominated by CBD workers. Students may also have online study options and not travel. Peak train frequencies on the better served lines are again probably excessive even allowing for wide social distancing. 

Buses have different issues. They provide the only public transport for the majority of Melburnians who are beyond walking distance of trains and trams. However they enable essential travel for those beyond walking distance of the supermarket or the doctors. Their usage will have also fallen but possibly not in percentage terms by as much as for other modes. And, because we've been sluggish with bus service and timetable reform during the 'good' pre-virus years their timetables still contain many quirks that may disproportionately deny service to those who can least afford it unless we are very very careful. More on this later. 

Emergency timetable options

What timetable options are there for weekdays? I've identified four main options and will discuss their pros and cons. 

1. Flat 20 minute option 

This option is based on all train, tram and major bus routes operating every 20 minutes from early morning to evening, with a possible drop to 30 minutes late at night. It harmonises headways across the network and caps waiting times at 20 minutes for connections between major routes. Local bus routes might operate every 40 minutes, potentially connecting with every second train. Operating hours would be the same as for an ordinary weekday. Therefore anyone who can make a trip now would still be able to travel, though waiting times might be longer. 

Compared to regular weekday timetables this option would reduce tram frequencies the most. Train timetables would be very similar to what currently runs on Saturdays except that the lines in the eastern suburbs that run every 10 minutes wouldn't. Some main bus routes (eg those that currently run every 30 minutes off-peak like the popular 733, 737 and 767) might even get an overall increase. SmartBus would get a cut from every 15 to every 20 minutes - milder than other options. Extra trips could be slotted in on routes that serve major hospitals and essential services if demand warrants. 

The word that best describes this option is 'egalitarian'. It delivers a similar basic service to many suburbs including those without trains. However it involves a rescheduling and rerostering of almost the entire network. Communication to passengers might also be difficult, especially for local bus routes. Whatever its merits the chance of implementation at short notice is so close to zero that it need not be discussed further. 

2. Sunday timetable (without early morning Night Network services)

This is what Wellington did a few days ago. Running Sunday timetables on weekdays is a bad option for Melbourne. Here's some reasons why: 

a. More than 50 Melbourne bus routes serving significant residential areas do not run at all on Sunday. The actual number without Sunday service is 102 out of a total of 349 but nearly half are special services eg peak-only, industrial or university shuttles. Many streets would no longer have service, removing the ability for people to make necessary trips. Areas particularly hard hit include Campbellfield, Thomastown/Reservoir, Dandenong/Springvale, Templestowe, Croydon/ Lilydale, Belgrave, Bayswater/Scoresby, Frankston South and more. Service to job areas with no other nearby service, such as Port Melbourne, Laverton North and Dandenong South would not run. And a Sunday timetable wouldn't include shuttles like the 401 between North Melbourne and the hospitals at Parkville. While all these areas get cuts rail corridors like the Frankston line would keep their midday trains every 10 minutes with almost no one on them.   

b. The trains would start too late and be too infrequent. Essential workers, even if they start at 'normal' business times like 8 or 9 am would be unable to get to work. This is because before Night Network our trains started late on Sunday mornings. For example the first train to arrive at somewhere like Frankston would get there after 9am. And even if people could start a bit later, the widespread 40 to 60 minute frequencies until about 10 or 11am is unsatisfactory.  

c. Limited bus operating hours. Even if your bus runs on Sunday, you would not necessarily be able to get it for work trips. This is because 'minimum standards' bus routes are only just starting by the time people need to be at work (ie 8 or 9 am). That particular affects early starters who on weekdays can board buses from around 6am. Finishing times are also an issue. Regular routes, particularly in the Ringwood - Doncaster area often finish around 10pm on weeknights but only 6pm on Sundays. SmartBuses would also lose service, with the implementation of Sunday timetables deleting their 9pm to midnight service. 

d. Lower tram frequencies. Possibly less significant than the above three are reduced tram frequencies. These include reductions to half-hourly in the mornings (when people need to get to work) and evenings on most routes. 

There are certain operational benefits of a Sunday timetable. It's the cheapest of these options to run, at least for train and bus. It's easy to roster for since all operators (except for Moonee Valley Bus Co which doesn't operate on Sundays) have timetables proved to work. Communication is also easier too. However it would severely reduce the network's capabilities for many essential work and other trips that are currently easily possible.  

3. Saturday timetable (without late evening Night Network service) 

This is what was talked about on the radio this morning. It is far better than the Sunday timetable option. There are two big reasons why:

a. Trains, trams and some major bus routes have similar operating hours on Saturday as they do on weekdays. This means that most people would still be able to do trips that the weekday timetable enables (though usually with extra waiting times) if a Saturday timetable was in force. The main difference would be at night on most train lines where the Saturday timetable drops abruptly to every 30 minutes from approximately 7pm. Also on Saturdays most SmartBus routes drop to every 30 minutes as opposed to the 15 minute weekday frequency.

b. Many more bus routes run on Saturdays than they do on Sundays. This allows more people distant from trains to make essential work, shopping and medical trips. Out of Melbourne's 349 bus routes 301 run on Saturdays, leaving 48 that do not. The majority of those that do not are special routes like peak only, university shuttle and industrial area routes. 

Like with Sunday timetables, the Saturday timetable is a proven arrangement known to work with regards to rostering. It is also easy to communicate. For instance all bus and tram stops already have Saturday times posted. Train, tram and bus operators are much better equipped to run a Saturday timetable at short notice than a custom timetable as per Option 1. 

There are still disadvantages with the Saturday timetable in some areas served by buses. It's basically similar issues to Sunday but more localised. These include:

a. Some weekday bus routes (48 out of 349) do not operate on Saturdays. These fall into the following categories:

i. University shuttle routes. In all cases except the 401 (which also serves hospitals) these have less frequent non-express overlapping routes that should suffice. 

ii. Peak only/weekday only express commuter routes. These almost all serve areas that have local bus options. Patronage on these routes is likely dramatically reduced (by a greater proportion than regular bus routes). Not running these (as would happen with a Saturday timetable) should not cause undue hardship. 

iii. Industrial area routes that only run weekdays with no or limited Saturday service. Examples include the 235 in Fishermans Bend, the 417 in Laverton North and the 857 in Dandenong South. These routes would either not run or be rendered unusable (eg noon finish) if a Saturday timetable operated. 

iv. Late starting times for local bus routes. Most Saturday routes start around 7 to 8 am but this may still be too late for those with early starts at jobs that were accessible when buses started at 6am or earlier as they do on weekday. More than 100 bus routes fall into this category. 

v. Residential area routes that have no or limited Saturday service. There are not as many of them as  areas lacking Sunday service. But it's still an issue. 531 in Campbellfield, 680 near Lilydale, 802 in Dandenong North are examples of neighbourhoods that would lose all service if a Saturday timetable ran.  Also significant are routes with Saturday timetables that finish around noon (or at best) early afternoon. Areas affected include Campbellfield (Route 538), Reservoir (Route 558), Thomastown (Route 559), Noble Park (Route 800), Springvale South (Route 814), Doveton (Route 844), Patterson Lakes (Route 857) and more. Many are low income neighbourhoods where buses provide vital links to essential services. Work travel is also made harder if the last bus ran five hours before your shift finished.  

It's not all about underservicing. Some train services may greatly exceed demand in the middle of the day but not be enough for social distancing earlier in the day. For example the Saturday timetable for Ringwood, Dandenong and Frankston feature a 10 min midday service but a 20 minute service during the am peak when more may be travelling. Similar issues may apply with some tram routes. 

There's also some silver linings. Tram 82 actually has a better weekend frequency than weekday frequency interpeak. It might actually get more trips under the virus timetable. The same may also apply for Belgrave/Lilydale trains, where weekend frequency is 20 minutes (10 min to Ringwood) versus 30 minutes (15 min to Ringwood) on the weekday timetable. If we go to a Saturday timetable for Ringwood due to the virus ideally the interpeak portion of it should continue permanently.   

Something else we need to watch is the adequacy of V/Line service, especially from areas like Melton, Caroline Springs, Wyndham Vale and Tarneit. Their weekend services drop greatly in frequency compared to weekday. And it may be that some lower income outer suburbs contain more people in jobs that don't have a work at home option, and/or people cannot afford not to work. So demand may not drop by an even amount on all lines or routes. 

Overall using the Saturday timetable as the basis for an emergency weekday is the best option so far.  It is easy to implement and easy to understand. However it still has substantial problems with buses since Saturday service is not universal and where it does exist operating hours may not suit work starts and finishes. And we may still be overservicing trains and trams in the middle of the day. 

4. Modified Saturday timetable (without late evening Night Network service) 

If we completed minimum standards upgrades and reviewed more area's bus networks we might not need this option as straight Saturday timetables would be more adequate in more areas. But we didn't. So we may need to consider the possibility of a 'modified Saturday' timetable option, especially if virus timetables are likely to persist for weeks and months. 

The good news is that trains or trams would be largely a Saturday timetable. Unless you needed to swap some midday and morning service levels to better meet demand as suggested before. Most bus routes would also have a Saturday timetable. 

But you might wish to keep the Saturday frequency but extend operating hours to match weekdays. For most local routes (that got the minimum standards upgrades) this would mean a couple of extra trips to give an earlier start. Some routes, especially popular weekday routes with a limited (eg 40 - 60 min) Saturday timetable might have extra peak trips slotted in. 

Routes with Saturday morning service only might have services extended both earlier and later to match their weekday span. 

Some routes that don't run Saturday (or have very limited services) might keep their weekday timetables going. Or have a special timetable drawn up. 

Then there are routes that have a more frequent Saturday morning service, dropping down to about half on Saturday afternoons. Like Reservoir's 552. That made sense 40 or 50 years ago but not now. Neither would it make sense as part of a weekday virus timetable. It will be interesting to see what happens to this and other routes that have more gentle frequency drops in their Saturday afternoon timetables. Although there are routes that drop back to as infrequently as every two hours on Saturday afternoon (like the important Route 800 past the premier's electorate office) that you'd keep at least its morning frequency going all day with weekday operating hours applying. 

To summarise, this 'modified Saturday' option is messy. Necessarily so because our Saturday bus timetables are messy. It is time-consuming to roster and explain. So it might not be implemented straight away.  However it is worth considering if the reduced timetables look like they'll be in for some time. And the extra cost could be kept down with some adjustments to train and tram frequencies where warranted.  


They may still cause hardships but Option 3 (Saturday timetable) and when possible Option 4 (modified Saturday timetable) are best and most practical at preserving service for those who need it most.  That includes those whose jobs are in essential industries and those who depend on public transport for essential shopping and medical travel. 

This is a hastily written post and there may be some errors or omissions. But I thought it was important to address this topical issue. If there are other things we should consider please leave them in the comments below. 

You might enjoy these well-regarded books on transport topics

Better Buses, Better Cities: How to Plan, Run, and Win the Fight for Effective Transit Steven Higashide 

Breaking Point: The Future of Australian Cities Peter Seamer

The Public City: Essays in honour of Paul Mees Gleeson & Beza

A Political Economy of Access: Infrastructure, Networks, Cities, Institutions (Access Quintet Book 4) David Levinson

Human Transit: How Clearer Thinking about Public Transit Can Enrich Our Communities and Our Lives Jarrett Walker

Transport for Suburbia: Beyond the Automobile Age Paul Mees

(Sales links: I get a small commission if you buy via the above - no extra cost to you)


Daniel said...

Yeah, this is complicated, and rostering can be very complex especially if you consider that the virus may eventually (unfortunately) spread through the PT workforce, and it may be uneven, affecting some depots much worse than others, for instance.

I suspect the answer is some kind of hybrid, and an analysis needs to be done of how the demand is likely to track (to ensure social distancing) and the logical break-up of rostering.

For instance you might switch trams to a straight Saturday timetable, Metro trains to Saturday plus some morning extras, and then look at buses depot by depot or operator by operator - with the aim of leaving the infrequent suburban 30 (weekday)/60 (weekend) buses on their weekday frequencies, but slimming down other routes a bit. V/Line might be different again.

No easy answers, and additionally complicated by the autumn construction blitz which has closed some sections of rail lines anyway.

Rob said...

I think the simplest and easiest solution would be Saturday timetables for trains and trams (with additional morning peak services to match commuter demand) and leave buses on weekday timetables.

Weekend bus timetables are too poor to be implemented across the board

Yellow over Green said...

Why bother with a weekend timetable when you could just run a weekday timetable and remove some of the peak trains. This will give express trains for people who need to get to work but maybe just fewer of them, the buses still run (some still don't run on Sundays), the connections should work through the interpeak because it's the same frequency as a weekday interpeak. When this is all over everything will won't just go back to how it was. With a weekday timetable you can just reinstate cancelled peak trips to match demand until there's a full weekday timetable again.