Urban population density is a hot topic – some people complain that it’s getting too high in Auckland, while others worry that it’s too low to get the urban outcomes we want. Either way, density matters – it can have a big impact on:

  • The efficiency of infrastructure provision and public transport services
  • Urban productivity and levels of competition in industries like retail
  • Amenity for residents – higher density can support cultural institutions and local vibrancy, but some people may prefer more open space
  • Preservation of open space and agricultural land on the urban fringes
  • Cities’ energy efficiency and use of resources.

However, we seldom ask: What is population density and are we measuring it correctly?

With the support of my employer, MRCagney, I’ve just written a short working paper that looks at this issue: Population-weighted densities in New Zealand and Australian cities: A new comparative dataset. It can be read in full here (pdf) and comes with an interactive spreadsheet that allows easy comparisons between cities.

This paper presents a new, more robust measure of density: population-weighted density. In contrast to simple average density measures, which basically measure the number of people living in the average hectare of land in the city (which is generally found in a low-density fringe suburb), this measure estimates the density of the neighbourhood in which the city’s average resident lives. As a result, this measure is much more representative of the lived experience of a city’s residents. The US Census Bureau has used a similar approach in their recent population density profiles.

While a wider range of results are available in the full paper, some of the most interesting findings actually relate to Auckland. In short, there have been some massive changes to the city over the last decade:

  • Auckland’s population density has increased 33% since 2001 – the city’s population-weighted density is now around 43 people per hectare after a decade of infill development and intensification
  • As a result, Auckland is now the third-densest city in Australasia – behind Sydney (76 people/hectare) and Melbourne (45 people/hectare) but significantly ahead of Wellington (38 people/hectare), Perth (30 people/hectare), and Brisbane (34 people/hectare).
  • There is no geographic reason why good public transport will not work in Auckland. Cities that are less dense than Auckland have successful, high-patronage PT systems.

And now, some visualisations!

Here are maps of Auckland’s population density in 2001 and 2013. As you can see, the city’s gotten incrementally denser throughout its entire area – look at the slightly darker shades of blue appearing all over the isthmus, on the North Shore, in the West and in Manukau.

Auckland density 01-13 v2

If we zoom into centres of New Zealand’s three largest cities, we can see that there have been big changes in city centre density in both Auckland and Wellington. Christchurch, on the other hand, as suffered from the effects of the 2011 Canterbury Earthquake, which demolished much of the city centre and reversed the city’s tentative moves towards apartment living. In spite of rapid demographic transition in Auckland’s city centre, rising density hasn’t really spilled over to surrounding suburbs.

Akl Wgtn Chch city centre comparison v1

Comparisons between cities also provide some interesting insights. The following chart, from the interactive spreadsheet, compares population density profiles in Perth and Auckland. Perth has frequently been cited as an example for Auckland due to the success of its electrified rail network, which now carries over 60 million annual boardings. However, Perth’s less dense, and significantly more sprawling, than Auckland. Clearly, low density isn’t holding back Auckland’s PT system.

Auckland Perth density profiles

The following chart makes that point even more clearly. It shows population-weighted densities and annual public transport ridership in eight large Australasian cities. Although Auckland is the third-densest city, it’s got the lowest PT boardings per capita.

Density and PT ridership

In short, Auckland is in an especially good position to benefit from a virtuous cycle in its transport system. Recent increases in density throughout the urbanised area have contributed to rising ridership on public transport, strengthening the case for further investment in projects like the City Rail Link, the AMETI busway, and the city’s New Network. Delivering a better public transport system will, in turn, encourage further land use change.

If you’re interested in this topic, take a look at the working paper and the interactive spreadsheet. What do you think about Auckland’s changing urban form?

Share this

27 comments

  1. Good to see posts on density, and the up to date density maps are much better than the older versions recently in use.

    I do think citywide average stats can be misleading, necessary though some people think they are. Wellington must be the classic: it’s two city patterns really, the CBD and inner burbs distinct from the further flung valleys and coast. Averages disguise the strength or weakness of centres, a key factor in transport patterns, and diminish the importance of the dynamics of change in these spatial patterns. Not looked at the paper yet, but did your study map these patterns of change?

    Having just returned from visiting friends in the outer burbs of Perth, I would observe a large part of their success in PT planning is having land development delivery aligned with PT investments. Auckland has no alignment in this space: visions, planning tools and development delivery are totally unconnected by comparison.

    New Lynn is the closest we have, a credit to years of WCC work, but is painfully slow and still challenging.

    1. Sorry, dumb question on mapping change. What I specifically meant was: did you produce maps of proportionate changes rather than just before and after?

      1. Just the before/after maps I’m afraid… although if you take a look at the spreadsheet you should be able to graph the changes between Census years.

  2. Excellent work Peter, and I’ll have to have a read of the paper too. This post shows yet another reason why Auckland’s PT network has massive potential to grow and provide a better and more efficient service for the city.

  3. I love this measure. Something about area density never sat right with me, and this explains it so simply in a real “but of course!” way.

    What I would love to see is a infographic type chart that plots Aucklands population weighted density on one side and *number* of people who live at that density on the other, perhaps with a line showing the average area density measure too.

    I imagine this would show the spread drags the average very low, despite the fact most live much higher.

  4. This has long been my favorite image for explaining the difference between average density and population weighted density. Both cities have the same average density but would have a very different feel to them

  5. Excellent work, thank you Peter. Really interesting read. Nice to see something other than by Demographia, which always seems a bit suss. One thing I’d be interested to know – how does this relate to other measures of density such as used by the people at AT&T publishers, where they define density as number of habitable dwellings per hectare (from memory). I feel that this is part of the slightly confusing aspect regarding density, whereby people are using different measures / scales. Have you done any work on that aspect?

    1. Hi Guy – Good question! I considered using dwelling density rather than population density and did some analysis of these figures. Jane Jacobs is an advocate of this measure, as it gives a better perspective on the scale of the area. (Also, people per bedroom can vary over time and between areas in the same city. In gentrifying areas, it’s common to see dwelling density increase, due to new construction, while population density stays constant or decreases, due to reduced overcrowding.)

      I ultimately decided not to publish figures on dwelling density due to data availability problems. Statistics NZ and the Australian Bureau of Statistics publish detailed information on _occupied_ private dwellings, but don’t release detailed data on unoccupied dwellings. As a result it wasn’t possible to estimate the total number of dwellings in an area. Which is a bit of a shame!

  6. Fascinating! Wish I could see the results for Wellington TA – or zones where smaller individual investments are planned in Auckland (rather than broad networks)

    Either way great work. Brisbane, Perth and Auckland look to be remarkable comparators.

  7. The 2001 v 2013 maps are telling. Look at that leap in density along the CRL route!
    And how persistently dispersed ChCh has been reinforced by the quakes.

    Auckland and Christchurch are clearly heading in diametrically opposed directions. This is an interesting experiment, all those dispersal advocates can move to Chch or at least watch to see if that is a successful urban form for our times. Of course we won’t really see the full economic consequences till the artificial stimulus of the rebuild winds down…. Although we can already see the traffic outcomes…

  8. Auckland certainly has the density for PT and seems to underperform on PT for density. It would be interesting also develop an index of the “average topographical challenge” for a city somehow. Crow-flying distances in Auckland (and Wellington) can be OK, but water and/or hills get in the way (especially for rail, which can’t handle steeper inclines). I suspect Auckland is the outlier for PT re: density partly because of topography, but it could be being over-used as an excuse. I can’t tell.

    1. I’d actually expect Auckland’s geography to favour efficient public transport services. There are a number of natural “choke points” on the approach to the isthmus – the Harbour Bridge, the SH16 causeway, the narrow bit of land between the Tamaki estuary and Manukau Harbour, etc.

      These choke points make it hard to provide enough road capacity (or alternative road links). Conversely, they advantage PT by ensuring that there will be a lot of trips on the single corridor through the choke points – i.e. higher cost recovery.

    1. Because there city is built for people. Also if you go to the outer suburbs of Melbourne they are probably worse than Auckland. Tourists to Melbourne only ever see the tram suburbs.

    2. I would say Sailor Boy is correct. The prewar city on the rail fingers and tram network is lovely and vibrant, walkable and dense. The other 2/3 of the city is complete sprawlsville. Tourists usually spend all their time in the former, but as someone who used to train and bus to work in an outer business park out in the burbs it is worse than anything Auckland has to offer.

    3. Maps confirm the on-the-ground view that Nick and Sailor Boy provide! If you take a look at the maps of the Aussie cities in the paper, you see these vast low-density yellow areas radiating out in all directions. Auckland, by contrast, is a consistently medium-density city.

      1. No worries Patrick… I knew what you meant!

        Perhaps you’re on to something. It probably is beneficial for New Zealand to have a few different types of cities – ensuring better matching of preferences to places. I’d suggest that Auckland, Wellington, and (to a lesser extent) Dunedin have got momentum when it comes to getting denser and more urban. Let’s let that happen. Tauranga and (perhaps) Hamilton seem to be headed for flatland instead. Fair enough.

        Christchurch is a bit of a harder case, to my mind. There were signs of rising urbanisation prior to the earthquake, and, as the Crampton article points out, its de-densification since 2011 has been more a result of planning and political failures than market demand. I’d put them in the “wait, see, and hope” category honestly.

        Finally, it’s essential that people understand THEY HAVE OPTIONS. Don’t like the fact that Auckland’s about to build a proper rail system? Fine – move to a city without one!

        1. Yes the Crampton is very good and therefore very very concerning. Even a brief visit there will show that, for anyone like me who considers hyper-dispersal unsustainable, enormous damage has been done to our second city in the name of the rebuild; mandated low quality sprawl. And when I say unsustainable I mean it in the broadest sense, economically as much as ecologically.

          The fact that this is ideologically driven by vested interests and the local Demographica fan club and reinforced by poor regulation and suspect governance is even more depressing.

          1. Patrick we get it that you don’t like Demographia and Hugh Pavletich and you will therefore take every opportunity to denigrate them. But What is happening in Christchurch has nothing to do with the Demographia crowd. Hugh has been ignored by Brownlee, Parker and now Dalziel. Crampton makes that clear as did I here http://www.interest.co.nz/opinion/70493/fridays-top-10-brendon-harr%C3%A9-national-vs-labour-housing-affordability-uk-councils-spy- (no.7). There is more than one way to look at Hugh’s work. There are ways to balance housing affordability with different transport options including PT and a range of housing density choices. You just need to take your blinkers off. A polarised debate is not going to get us anywhere. I acknowledge that Hugh himself takes an extreme position but not every one in the Demographia crowd does.

            P.S I agree with all your other concerns Patrick. I am worried.

    1. Thanks Peter for standing up for Christchurch and thanks for the link. Most balanced read about the Christchurch rebuild that I have encountered.

  9. Figure 1: Comparing density measures in a hypothetical city
    Easily to see this in larger cities, but looking at medium size cities none of them actually show this, especially when give flat land. They tend to be empty near the centre. Maybe the reason Obiter bus route has been successful in Hamilton.
    If anyone is interested http://hamiltonurbanblog.co.nz/ has post on pop density of Electorates.

Leave a Reply