Where does Melbourne end (and sprawl begin)?

Drive out towards Warburton and it seems easy to see where Melbourne ends and rural life begins. One minute you’re driving through houses, shops and businesses, when all of a sudden you’ve arrived in country. Except you’re actually still in Melbourne because the official boundary of the metropolitan area lies on the other (eastern) side of Warburton!

People seem to like a hard edge – a clear and unambiguous boundary – between city and country. But it only works if the non-developed land is “pure” bush or bucolic farming land, without service stations, hobby farms or other urban detritus. Head out of Melbourne in most other directions and development – almost all of it tacky and ugly – tracks you like a mangy dog.

The continuous built-up area of Melbourne (the pink bit in the middle of the map) occupies less than 2,000 km2. This is much less than is commonly assumed by the media and is just a little more than a quarter of the area covered by the official or administrative boundary, which is 7,672 km2. There are a number of “islands” of development within the boundary (also shown in pink), like the townships of Melton and Sunbury, that are officially part of the metropolitan area but separated from “mainland Melbourne” by green wedges. It makes sense to count a place like Melton township as part of Melbourne because 65% of workers living there travel across 9 km of green wedge to work in mainland Melbourne.

These islands make discussions about sprawl particularly fraught. Is it just the central core of continuous urbanised development that sprawls or should all the islands within the boundary also be included? If they are, then that not only includes towns such as Melton, Sunbury and Pakenham, but also towns like Warburton, Healesville and Gembrook that appear to the first-time visitor to be country towns. And given that island townships like Garfield and Bunyip in the outer south-east corridor are officially part of Melbourne, it’s reasonable to wonder why towns that lie just outside the boundary, like Drouin and Warragul, aren’t also seen as part of Melbourne’s sprawl.

This story from a 2003 issue of The Age shows how closely linked many country towns located outside the boundary are to Melbourne:

Census 2001 figures cited by a Monash University Centre for Population and Urban Research report for the Southern Catchments Forum show that, remarkably, more than half of the working residents of the Macedon Ranges area are employed in Melbourne. Similarly, about 40 per cent of the working residents of the Moorabool region (which includes Bacchus Marsh) and the Melbourne side of the Greater Geelong area commute to Melbourne for work. It’s clear, the report says, that these areas are “largely dormitory towns servicing the metropolis.

According to another report in The Age published in November last year, “Melbourne has sprawled 50 per cent further than its official urban growth boundary and is overrunning small country towns ……… Developers are building large suburban-style estates as close as three kilometres to the boundary, marketing to metropolitan commuters while avoiding the infrastructure levy”. The “50 per cent further” claim is a ludicrous exaggeration but the general picture isn’t. The paper quotes the secretary of the Macedon Ranges Residents Association, Christine Pruneau, who says such estates make a mockery of the urban growth boundary:

Every small town near the boundary is really being hammered ……… These are little towns getting development that looks like it belongs in Essendon and it’s changing the character of the places into suburbs of Melbourne”.

This sort of development could be interpreted favourably as “decentralisation” rather than negatively as “over-spill”. Perhaps this is why cities like London, where development has jumped beyond the greenbelt and into tightly defined satellites cities, are often assumed to be less sprawled than Melbourne despite having much larger populations. I think a lot of the widespread dislike of sprawl comes down to the absence of a hard edge. If Melbourne had reserved a real greenbelt in the 70s or 80s and prevented tacky roadside development, it’s not that fanciful to think today’s outer suburbs might be seen as satellite cities – as decentralisation rather than sprawl. But in terms of the downsides of fringe development, I’m not sure we’d be any better off. We might in fact be worse off because of all that extra travelling over the greenbelt.

I’ve previously argued that a sensible way of defining where a city “ends” is the extent of contiguous urbanisation combined with that area’s labour catchment. The latter metric reflects the fact cities exist because of agglomeration economies, especially in labour. Satellite cities that are closely tied to the continuously urbanised area by commuting should be considered part of Melbourne. Work travel isn’t the only possible measure – there could be analogous social linkages – but the significance of commuting, I would argue, is the amount of time spent physically within Melbourne. Commuting typically involves spending in the order of eight hours within Melbourne’s boundary on around 270 days per year. The really hard bit is figuring out just what proportion of workers would have to commute to urbanised Melbourne to qualify the township as part of Melbourne. It’s probably far-fetched to think that this kind of thinking would lead to the official boundary being extended to include places like Bacchus March, but it would provide a more sophisticated way of assessing where sprawl begins.

11 Comments on “Where does Melbourne end (and sprawl begin)?”

  1. rohan says:

    Great map – you can see that the green wedges have virtually disappeared already – there’s only one major one left really. And the townlets are along rail or roads I presume, very linear. If only the green wedges had actually been firmly established in the 70s, then we would have a radial city, with green space not too far on either side, and of course the ‘spokes’ would be much thicker by now. But I do think that would have been better than city as blob.

  2. Andrew says:

    Rohan has nailed it well. Naively, I used to think that the green wedges were firmly in place. But as you suggest, a hard edge with satellites would have been good too.

  3. wilful says:

    What ‘official’or administrative boundary are you talking about? I tell you, I know a bit about property and admin law, and I have no idea what boundary you mean. That blue line means nothing to me.

    The taxi boundary, or the CFA boundary, or the UGB boundary, these could be some form of surrogate.

    Warragul and Drouin (where I live) have a healthy commuter population, but that’s just people going to Pakenham and the far SE suburbs. Some people go from those suburbs out into the country every day. Nobody would say that they’re a melburnian, their sense of identity is certainly not as part of melbourne.

    • Alan Davies says:

      The blue line is the Melbourne Statistical Division (excluding the bit sticking out in the north east) — the Stat Div is the customary definition of cities in Australia e.g it’s used by Melbourne 2030. The people of Drouin and Warragul might not feel they’re Melburnians (and formally they’re not) but the question is whether or not their growth is part of Melbourne’s sprawl.

  4. Chris Curtis says:


    Even when I expand the map, I cannot make out where it comes from. I’d like to be able to so do, because it just doesn’t look right. More specifically, the pink blobs in the North-East don’t look right. I cannot tell which townships they are, but they do not look like three townships of Hurstbridge, St Andrews and Panton Hill that are in that area outside the main urban growth boundary.

    The initial green wedges and development corridors strategy for Melbourne was to set the permanent shape of the city, but the green wedges have been nibbled way at for decades. Mr J. A. Hepburn, the chief planner for the Board of Works, said, “All future development will be in corridors. There will be no urban development in the green wedges – that is the basis of the whole policy.” (“The plan”, The Sun, 30/11/1971) If only! That is why the urban areas do not bear the distinct shape set out 40 years ago.

    In this post-modern deconstructionist age, everything means whatever you want it to mean, so the discussion of the meanings of the end and beginning of Melbourne’s sprawl could go on forever. We could also argue about what sprawl itself is. Is it anything better than 10 houses per hectare? Or is that not enough? Does it have to be anything better than 30 houses per hectare? Or is it anything where people actually dare to use cars? Is it any area above 1,000 square kilometres? Or does it have to exceed 2,000 square kilometres?

    The Board of Works got it right 40 years ago by determining that urbanisation would be confined and the rural environment would be reasonably close to all city-dwellers. If that means that some people have to drive through green wedges to get from one urban area to another, that it not a problem. That is the intention.

    I have given lots of relevant facts (including links to maps) and figures in my posts on What should we do about Melbourne?).

    • Alan Davies says:

      They are indeed Hurstbridge, St Andrews and Panton Hill. The pink layer is a government map of Victorian “urban centres” and as I noted above is a bit over ten years old (its age is immaterial to the points I’m making in the post). I’ve superimposed it on a more recent Melway layer. So, it’s not a map of the area within the UGB but rather a map of urbanised areas as they stood in the second half of the 90s.

      Re post-modernism etc: Every time anyone mentions “sprawl” they are implicitly making an assumption about what it is. Likewise, when The Age yet again says Melbourne spreads more than any other city in the world, it is implicitly making an assumption about where Melbourne “ends”. These assumptions need to be explicit — good policy can’t be made on the basis that we assume we know what everyone else means.

  5. Oz says:

    Defining our urban boundary for development of policy and planning as the ABS Melbourne Statistical Division (MSD) totally defined by the 31 jurisdictions of the Local Government Authorities (LGAs) is always a meaningful start. However Lewis Carroll’s writing also applies to defining urban sprawl.
    Lewis Carroll wrote…””When I use a word,” Humpty Dumpty said in rather a scornful tone. “It means just what I choose it to mean – neither more or less.”
    “The question is,” said Alice, “whether you can make words mean so many different things.”
    “The question is,” said Humpty Dumpty, “which is to be master – that’s all.”

    • Alexander says:

      Oz, point of clarification, the ABS MSD excludes about half of the Yarra Ranges Shire, which is grouped in with Gippsland for the time being. The area excluded is somewhere near 1000 km ², but contains only a few hundred people. It’s that area outlined by blue that goes out the end of the Alan’s map.

      Furthermore, on the 1 July the ABS is going to stop using statistical divisions and use Greater Capital City Areas (I think that’s the name). Melbourne’s GCCA will include parts of frex Moorabool, Macedon Ranges, and Murrindindi Shires, but continues to exclude the Yarra Ranges beyond East Warburton (which now goes to the Hume “SA4” region).

  6. […] issue – I’ve discussed how Melbourne’s boundary might be defined before (here and here). -37.781700 145.039432 LD_AddCustomAttr("AdOpt", "0"); LD_AddCustomAttr("Origin", "other"); […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s