Executive Summary: A look at the legacy of CL Fouvy, a transport engineer who identified the need for SRL-style orbital rail lines as early as 1975 and designed a massive suburban rail grid to double public transport travel speeds by 2025.
It's fair to say that this State Government has had some difficulties with the Suburban Rail Loop project in the last few weeks.
New Premier Ben Carroll regards it as a lower priority than his predecessors did. Construction will proceed but with cost reductions that lessen connectivity to its stations. The Auditor-General found the project was unlikely to be completed on schedule. And that some funding will come from a fares levy that was not disclosed to the public. The premier said he'll not continue with the levy but that would leave a gap that still needs to be filled somehow.
SRL East, due to its entirely tunnelled nature, substantial stations, construction during an inflationary period and the special difficulties the Anglosphere has with containing construction costs on major projects, is an expensive project. And the State Government appears to be having difficulty in making its funding formula (1/3 state, 1/3 federal, 1/3 value capture) work. Especially as it is being built simultaneously with other high cost projects such as North-East Link.
However imagine a city where the SRL was not exceptional or particularly expensive, but just one of an extensive grid of rail systems under (and sometimes above) Melbourne. This is what engineer CL Fouvy envisaged in 2004, for completion in 2025.
CL Fouvy worked for the Tramways as an engineer for many years. He did much of the modelling for the 1969 Melbourne Transportation Plan. More about that later but let's first look at the scheme he proposed in 2004.
The Fouvy scheme
In 2004 he presented a paper to the Railway Technical Society of Australia proposing an underground Metro system for Melbourne. This was not just a loop under the CBD (like the City Loop) nor a single line under the suburbs (like the Suburban Rail Loop) but a grid network of lines spaced every 2 kilometres under most inner and middle suburbs. There would be hundreds of kilometres of new tracks in a General Scheme for Rapid Transit Development.
Fouvy's proposed concept is below:
A close-up of the CBD area is below:
Of note is a north-south line not dissimilar to the now operating Metro Tunnel under Swanston St. This fanned out to two lines at Parkville and Southbank. Stations were typically closer together than the spacing chosen for the Metro Tunnel.
Fouvy's proposal was published on the RTSA website. Unfortunately RTSA deleted it in a website revamp. But not before it was captured by the Web Archive at:
https://web.archive.org/web/20080721025337/http://rtsa.com.au/assets/2008/03/rapid-transit-by-l-fouvy.pdf
I recommend you read the whole thing. But the main points were:
* Large urban areas require rapid transit of 40km/h or more to cater for trips longer than 6 to 8 km. Current services are just too slow, especially when compared to car travel.
* From 1854 to 1929 Melbourne's rapid transit rose in line with the population. Since then (writing in 2004) there were few developments, resulting in a 75 year backlog.
* Fixing this backlog requires a 2km grid of radial and ring urban rapid transit routes across Melbourne's suburbs. Construction would cost $1.5b per year for the next 20 to 25 years. Fouvy considered that affordable, comparing it with the $2.8b we apparently spend on remedying road traffic crash data. Funding would come from fuel and congestion taxes on motorists.
* A new rapid transit grid rail network could double the speed of many if not most public transport trips, particularly those between non-CBD points. It could attract serious mode share from driving.
* Lines could be at ground level, tunnelled or elevated. Some would be electrified existing lines while others (the bulk of the network) would be new.
* There would be 72 radial routes and 21 proposed ring rapid routes. Each route would run every 10 minutes during peak times and (apparently) 15 minutes off-peak. Tracks would be designed for 2 minute headways with up to 5 routes per track.
* Fouvy does not address the reliability risks that a combination of relatively infrequent lines sharing the same track would pose. Also that 10 and 15 minute frequencies impose significant transfer penalties and platform crowding risks for what is envisaged to be a high volume metro network.
* These electric rapid transit lines would keep Melbourne going as world oil supplies decline over the next 50 years.
In other words, whereas universal automobility was predicated on cheap oil and continued high road construction, this transit-oriented alternative was justified by declining oil supplies. Rapid transit was necessary for cities with urbanised radii larger than 5 or 6 km. Melbourne had passed that long ago.
CL Fouvy quotes the 1929 plan that found that Melbourne could accommodate 6 million people within a 21km radius. He lists various plans (1929, 1954, 1969) and their extension proposals, most of which were not implemented. The vision was for a compact, transit-oriented city (as opposed to a more car-oriented sprawly city).
Fouvy says that he acknowledged the need for ring rapid transit routes (ie effectively SRL predecessors) as far back as 1975. This was driven by the incapability of the current radial rail network to serve many trips and the depletion of oil reserves by 2050 (with no substitute).
His scheme envisaged about 8 times more route kilometres than operated at the time to provide a useful grid based on nearly everyone in the urbanised area being within 2km of a station. Track kilometres would rise to about 3.6 times current. Thus many tracks would have several routes, which he describes as being more efficient. Hold that thought as I'll come back to his network philosophy later.
Many new stations on existing lines were proposed. These range from sensible (eg Campbellfield) to batty (eg Ormond North which would make the Frankston line excruciatingly slower than it currently is). The current Premier would however note Keilor East's inclusion.
* Large urban areas require rapid transit of 40km/h or more to cater for trips longer than 6 to 8 km. Current services are just too slow, especially when compared to car travel.
* From 1854 to 1929 Melbourne's rapid transit rose in line with the population. Since then (writing in 2004) there were few developments, resulting in a 75 year backlog.
* Fixing this backlog requires a 2km grid of radial and ring urban rapid transit routes across Melbourne's suburbs. Construction would cost $1.5b per year for the next 20 to 25 years. Fouvy considered that affordable, comparing it with the $2.8b we apparently spend on remedying road traffic crash data. Funding would come from fuel and congestion taxes on motorists.
* A new rapid transit grid rail network could double the speed of many if not most public transport trips, particularly those between non-CBD points. It could attract serious mode share from driving.
* Lines could be at ground level, tunnelled or elevated. Some would be electrified existing lines while others (the bulk of the network) would be new.
* There would be 72 radial routes and 21 proposed ring rapid routes. Each route would run every 10 minutes during peak times and (apparently) 15 minutes off-peak. Tracks would be designed for 2 minute headways with up to 5 routes per track.
* Fouvy does not address the reliability risks that a combination of relatively infrequent lines sharing the same track would pose. Also that 10 and 15 minute frequencies impose significant transfer penalties and platform crowding risks for what is envisaged to be a high volume metro network.
* These electric rapid transit lines would keep Melbourne going as world oil supplies decline over the next 50 years.
In other words, whereas universal automobility was predicated on cheap oil and continued high road construction, this transit-oriented alternative was justified by declining oil supplies. Rapid transit was necessary for cities with urbanised radii larger than 5 or 6 km. Melbourne had passed that long ago.
CL Fouvy quotes the 1929 plan that found that Melbourne could accommodate 6 million people within a 21km radius. He lists various plans (1929, 1954, 1969) and their extension proposals, most of which were not implemented. The vision was for a compact, transit-oriented city (as opposed to a more car-oriented sprawly city).
Fouvy says that he acknowledged the need for ring rapid transit routes (ie effectively SRL predecessors) as far back as 1975. This was driven by the incapability of the current radial rail network to serve many trips and the depletion of oil reserves by 2050 (with no substitute).
His scheme envisaged about 8 times more route kilometres than operated at the time to provide a useful grid based on nearly everyone in the urbanised area being within 2km of a station. Track kilometres would rise to about 3.6 times current. Thus many tracks would have several routes, which he describes as being more efficient. Hold that thought as I'll come back to his network philosophy later.
Many new stations on existing lines were proposed. These range from sensible (eg Campbellfield) to batty (eg Ormond North which would make the Frankston line excruciatingly slower than it currently is). The current Premier would however note Keilor East's inclusion.
Despite the closer spaced stations, trains would still be faster than buses. To answer objections against extending rail in favour of adding feeder services, the latter was considered to add too much time, and thus be ineffective. Overall trips would be vastly faster than now due to rail's much extended coverage. A few examples are given later.
At $4.2m per kilometre of electrified double track, construction costs appear very low by modern standards. Even after accounting for CPI. Something that would have made construction cheaper is acceptance of new level crossings - something that would not be tolerated today.
What difference would the Fouvy scheme make to travel speed? Like the Suburban Rail Loop project, Fouvy published before and after travel times for cross suburban trips. Here's a sample:
Caulfield - Camberwell 35 > 14 min
Coburg - Melbourne Airport 51 > 25 min
Doncaster Shoppingtown - Ringwood 37 > 12 min
Hampton - Dandenong 65 > 25 min
Knox City - Dandenong 40 > 16 min
Monash - Doncaster Shoppingtown 61 > 24 min
Sydenham - Werribee 77 > 51 min
These are basically car competitive travel times, if not better, especially with higher congestion. Their reduction would not just contribute to modal shift but also likely generate new trips.
Put in another way, on Melbourne's current PT network you struggle to have much less than 60 minutes travel time for all but direct rail trips to the CBD from most train stations. Whereas Fouvy's network roughly doubles travel speed to make many trips possible within 30 minutes, including those between suburbs. Thus it repositions public transport as being practical for a huge number of cross-suburban trips where it currently has a modal share of maybe 5% or less.
Put in another way, on Melbourne's current PT network you struggle to have much less than 60 minutes travel time for all but direct rail trips to the CBD from most train stations. Whereas Fouvy's network roughly doubles travel speed to make many trips possible within 30 minutes, including those between suburbs. Thus it repositions public transport as being practical for a huge number of cross-suburban trips where it currently has a modal share of maybe 5% or less.
Background on CL Fouvy and his network philosophy
Other material on Fouvy is published in past Australian Timetable Association magazines. These are hosted on its website at https://timetable.org.au under the Back Copies section at https://www.timetable.org.au/publications/back-copies .
Sadly at the time of writing its links are broken. The lesson is always save copies of historical important documents on your own storage - you never know when they might become unavailable.
Fortunately I was able to retrieve the April 2020 issue of The Times which says something about CL Fouvy's approach. Not from the current link which is at https://www.timetable.org.au/wp-content/uploads/2023/06/thetimes202004issue.pdf but inoperative when checked yesterday. Nor at a previously used link at http://cdn.timetable.org.au/thetimes202004issue.pdf . But an archive of that previous link at https://web.archive.org/web/20221027140521/http://cdn.timetable.org.au/thetimes202004issue.pdf was working.
The relevant article, on page 3, is by Ian Manning. This gives some background to transport modelling, as used by (often) American consultancies to do transportation (sic) studies for Australian cities in the 1960s. These studies assumed continued cheap oil and road building that could keep pace with traffic resulting in a shift to driving for most trips other than (in the larger cities) CBD commuting.
Fouvy's work on Melbourne's 1969 plan included measures of public transport service quality. Here's an interesting quote:
Fouvy's work on Melbourne's 1969 plan included measures of public transport service quality. Here's an interesting quote:
‘Convenience of service is assumed greater for two routes [or two rail stations] each with two vehicles an hour, than for one route with four vehicles an hour, since wider areas are served. The use of the square root of the service frequency places more emphasis on the number of routes serving the area than on the service frequency of an individual route...'
This is a different approach to younger planners such as the late Paul Mees and Jarrett Walker who have favoured simpler but frequent grids featuring turn-up-and-go frequencies in suburban areas. It's as if the "network effect", which needs high frequency to work, wasn't then a thing in his mind. Also that perfect coordination with multiple routes in multiple directions was possible (which it isn't). And that people are more willing to wait up to half an hour for a line that takes them exactly where they want to go rather than a lesser time to a route that may or may not, not much waiting is involved to connect to the route to their destination.
In the case of rail multiple routes on the one corridor can lead to knock-on effects where a delay on one line also disrupts other lines. Hence the philosophy of 2010's era rail planning to operationally separate lines but run each at high frequency with the best possible interchange between them. At the time Fouvy presented his Melbourne plan in 2004, metropolitan rail punctuality had just started a long-term reliability fall with this not to seriously recover until nearly a decade later. Part of that recovery was driven by operational segregation, notably the formation of the cross-city group that took lines out of the City Loop.
Based on the quote about his assumptions, it would seem that if Fouvy was allowed to plan a network it would be more like Melbourne's (often) complex and infrequent buses than the train or tram network (with a smaller number of more frequent routes). Or, in capital city terms, preferring Adelaide or Brisbane (both of which have been weak on bus network reform and have inferior frequencies) over Perth (with its simpler frequent routes, including on some non-CBD corridors).
Fouvy's bias against frequency is confirmed when you look at his General Scheme profiled above. 10 and especially 15 minute frequencies might be fine for suburban buses but impose significant transfer penalties for those who need to change on a high volume Metro-style train station. His acceptance of up to five lines sharing tracks also imposes significant reliability challenges that modern Metro rail planning (including the Metro Tunnel) has tried to untangle.
In the case of rail multiple routes on the one corridor can lead to knock-on effects where a delay on one line also disrupts other lines. Hence the philosophy of 2010's era rail planning to operationally separate lines but run each at high frequency with the best possible interchange between them. At the time Fouvy presented his Melbourne plan in 2004, metropolitan rail punctuality had just started a long-term reliability fall with this not to seriously recover until nearly a decade later. Part of that recovery was driven by operational segregation, notably the formation of the cross-city group that took lines out of the City Loop.
Based on the quote about his assumptions, it would seem that if Fouvy was allowed to plan a network it would be more like Melbourne's (often) complex and infrequent buses than the train or tram network (with a smaller number of more frequent routes). Or, in capital city terms, preferring Adelaide or Brisbane (both of which have been weak on bus network reform and have inferior frequencies) over Perth (with its simpler frequent routes, including on some non-CBD corridors).
Fouvy's bias against frequency is confirmed when you look at his General Scheme profiled above. 10 and especially 15 minute frequencies might be fine for suburban buses but impose significant transfer penalties for those who need to change on a high volume Metro-style train station. His acceptance of up to five lines sharing tracks also imposes significant reliability challenges that modern Metro rail planning (including the Metro Tunnel) has tried to untangle.
Puzzlingly Fouvy's preference for an extensive infrequent one-seat network over a frequent one that may require changes is despite him working for the Tramways, which (more so than the Victorian Railways, especially post WWII) had a culture of high frequency all day service. Although one could argue that the St Kilda Rd operations where each route had its own line to the CBD with no changing required is somewhat Fouvy-like.
In a way one could consider Fouvy in the spirit of Daniel Andrews and Jacinta Allan. As premier and transport minister they staked their political life on building new rapid transit routes (eg the SRL) but in government were reluctant to champion rail frequency upgrades and bus network reform. Though to be fair Fouvy did envisage reform of feeder routes and good interchange points.
In a way one could consider Fouvy in the spirit of Daniel Andrews and Jacinta Allan. As premier and transport minister they staked their political life on building new rapid transit routes (eg the SRL) but in government were reluctant to champion rail frequency upgrades and bus network reform. Though to be fair Fouvy did envisage reform of feeder routes and good interchange points.
Other big rail schemes
https://handwiki.org/wiki/Engineering:Proposed_Melbourne_rail_extensions
https://www.wikiwand.com/en/Proposed_Melbourne_rail_extensions
https://www.wikiwand.com/en/Proposed_Melbourne_rail_extensions
Conclusion
Based on low construction costs and ignoring the electrification of road transport, some might consider Fouvy's ambitious rail scheme as fanciful. It nevertheless still deserves study and reflection. Especially when one looks at the big expansion of road freeways in the last 60 years whereas electrified metropolitan passenger rail per capita is has fallen by maybe two-thirds in that period.
In contrast to the dense grid of the Fouvy plan, the SRL is the sole current orbital rail project with gap-filling concepts like extending the Alamein line or even tram routes to nearby stations seemingly not on the agenda. Electrification to growth areas like Melton, Donnybrook and Clyde has also lagged in the last decade, and these are just basic radial routes.
The Fouvy scheme's audacity can inspire us all. But also be aware of its assumptions and limitations with regards to the broader transport context and network design, including frequency and reliability.
Read it here: https://web.archive.org/web/20080721025337/http://rtsa.com.au/assets/2008/03/rapid-transit-by-l-fouvy.pdf
No comments:
Post a Comment