In April I cited research saying that service provision was lagging apartment builds.
May saw a comparison with Sydney. The two cities are going in opposite trends. We're adding population faster than they are while they're adding public transport service faster than us. The gap is so wide is that their waits for public transport are now often half ours, especially at night. It will widen further if there is not a transport service agenda commensurate with our housing targets.
And earlier this month the Climate Council report on PT services documented the big differences between Sydney and Melbourne in their population's access to frequent public transport.
How do we compare with other cities and how have they trended over time?
While snamuts.com is most known for its maps showing relative public transport service and connectivity within metropolitan areas, it also has some handy comparisons between the bigger cities in Australia and New Zealand. These include metrics like public transport service per capita and the proportion of people and jobs that are walkable to public transport that exceeds a defined service standard (every 20 minutes weekdays, 30 minutes weekends). Furthermore these numbers are recorded at various times over the last 15-20 years so you can get an idea of trends.
Graphing service trends
The first thing I did was to make separate graphs of six cities (Adelaide, Auckland, Perth, Brisbane, Melbourne and Sydney). These have years across the bottom (noting that not all cities have data for all years).
The lower line (blue) is service intensity of service kilometres operated per 100 000 residents. This does not discriminate between modes, so all are weighted the same. The source numbers appear at the ends of the SNAMUTS service intensity bar graphs for each city and year. A city that throws a lot of service kilometres across the metropolitan area rates highly here. However a high service intensity does not mean that the service kilometres are efficiently deployed (although if they are not then the scope for cost-effective network reform is highest).
The red line comes from the network coverage metric. As noted above this is based on a frequency standard set by SNAMUTS that is considerably higher than the 'minimum standards' that transit agencies might set (eg service every 60 minutes or better). Transit agency records might thus rate a network's coverage as being 90% of an urbanised population while SNAMUTS numbers for the same network might only be 40% due to its higher threshold.
Inspect the graphs below. Click for the clearer full size version. Discussion to follow below.
Adelaide has been relatively steady. Partly due to its lower population and service growth than the other cities. However its service intensity (blue line) is relatively high, again much assisted by its low population growth. This is because faster growing cities have to add a lot more service each year just to stay still. That's not always happened (eg Melbourne, Brisbane, Perth).
Where cities have added growth area bus service on their expanding fringes these are typically routes whose frequency does not meet minimum SNAMUTS standards (red line). So again it's a struggle for cities with a lot of suburban growth to hold their own unless (a) they add frequent service in growth suburbs and/or (b) they add people and jobs in established well-served areas.
Auckland is in complete contrast to Adelaide. It has added population AND service massively. Not only that but, largely due to bus reform that added a 15 minute frequent network, it has increased the proportion of people and jobs near a minimum service from the lowest to higher than all but Australia's two largest cities.
Both Melbourne and Perth have somewhat similar patterns but Perth's percentage growth in service has been higher, starting at a low base. It's a bit like Auckland in this regard. It has overtaken Brisbane in service intensity but not (as of 2021) on the minimum standard criterion. However since 2021 Perth has opened a new railway to Airport/High Wycombe, reformed its eastern suburbs buses and is about to open its Yanchep line extension (again with a new bus network) so may now be equal if not ahead of Brisbane on this measure too.
Melbourne came off a very low service base in 2006, largely due to its (then) terrible bus network. Major improvements had been made by 2011, with more in some subsequent years. However its population was growing strongly, leading to a service per capita decline. Growth in population and jobs near SNAMUTS minimum standard services can be largely attributed to SmartBus roll-out (in 2006-2011) and CBD jobs growth along with intensification around trams in the period to 2016. Since 2016 service growth has been insufficient to keep up with jobs and population growth due to Allanism's favouring of big infrastructure builds over service.
Sydney had a drop in service per capita and then a rebound. It has consistently been higher than Melbourne. This is attributable to a strongly pro-service state government and lower population growth than Melbourne. The population near minimum standard of service is high and rising due to both these factors and network extensions including the Metro.
On the one graph
Below is the same data (plus a bit more) overlapped for easier comparison. This allows you to better compare the service performance of various cities. The bracketed number is the number of public transport trips per capita for the latest available year. Again click graph to enlarge.
Though there's only two data points, Auckland's progress is conspicuous. Perth and Melbourne also made significant strides between 2006 and 2011. As noted before Melbourne and Sydney have gone in opposite directions since 2016, with Sydney adding service per capita and Melbourne reducing it. Melbourne's continued (slow) growth in the percentage with minimum service since 2016 can likely be attributed to apartment building in the CBD and near tram lines more than actual service growth.
Very roughly the higher the service intensity the more likely people and jobs will have better than minimum service levels. You might think that having a high proportion of the latter creates the conditions required to increase per capita usage given that increasing frequency generally does increase patronage. However it's not a straightforward relationship as service intensity tells us nothing about how efficiently a network is planned or whether modes are optimally used.
Service intensity versus trips per capita
Of note is that Adelaide scores highly on both service intensity and the reach of minimum service (assisted greatly by its Go Zone buses). However its average resident takes 42 public transport trips per year, which is lower than bigger cities.
Sydney and Melbourne have less service intensity but has almost double or more in trips per capita. OK you might say that driving conditions are poorer in the larger capitals, leading to lower car ownership, but look at Perth. Much less reach of minimum service (its non-CBD light industrial areas are notoriously poorly served by PT and it has huge sprawl with typically hourly buses) yet it attracts higher trips per capita than Adelaide.
Can network structure affect usage?
This may be due to something that's harder to quantify by numbers like these - while Perth's urban form and land use patterns aren't very good for public transport, its network structure is sufficiently better than Adelaide and SE Queensland to attract a higher trips per capita.
Most notable is that, in terms of service levels, Perth has a 'big city' rail network with well-planned buses that also cater for local trips. Whereas both Adelaide and Brisbane have less frequent underperforming rail networks and a lot of buses that parallel radial rail. Neither have the circumferential bus routes that Perth, Melbourne and Sydney have. An excessively radial system limits the potential for public transport to cater for diverse suburban trips.
And it can be inefficient with service kilometres that do not necessarily lead to good patronage outcomes. Even though Perth's network is more rail oriented than either Brisbane or Adelaide's, the attraction of its trains and their ability to relieve buses for other purposes is sufficient to give Perth a higher bus usage per capita than either of the other cities.
Conclusion
Service per capita is an important (but underappreciated) metric of investment in public transport. Service kilometres needs to keep up with population growth just to remain constant. Unfortunately it hasn't in too many cities.
Service per capita is a quantitative measure. The way that service is deployed, a qualitative measure, also affects patronage. Cities that throw a lot of service on excessively radial or duplicative bus networks to the exclusion of radial trains and circumferential bus routes get a network unsuitable for any trip but towards or away from the CBD. The networks most like that, ie Adelaide and SE Queensland, had the lowest patronage per capita. The pandemic has only increased the gap between the types of trips such legacy public transport networks are best at and the trips that people actually make.
The performance of Perth (and even more so Auckland) gives encouragement that good network design can lead to higher patronage outcomes. Brisbane and Adelaide should be following Perth and Auckland on this. Auckland has potential to do even better by boosting train frequency from 20 to 15 min to match buses. And Melbourne should be following Sydney with consistent 7 day frequent trains and better buses given the latter's improvements in services and strong ridership numbers in recent times.
Index to Building Melbourne's Useful Network items
No comments:
Post a Comment