The importance of statistical geography - what's in a boundary?

It's not exactly headline news, but the release of a new Australian Statistical Geography Standard (ASGS) matters more than it sounds, especially in the lead-up to the Census. Since 2011, the geographic areas the ABS uses have been designed to stay stable for the five years between Censuses. Every five years a new version of the standard is released, about a year ahead of the Census results and even a month ahead of the Census itself. That gives users of Census data time to prepare, and an early hint at how our population has shifted before the actual numbers land.

It depends how you draw the boundary

One of my favourite lines about geographic data is 'it all depends on how you draw the boundaries'. I use it most often answering questions like 'is Melbourne bigger than Sydney?', because it genuinely does depend on where the line goes. Urban and rural areas, government areas, electorates: wherever we draw a boundary, we're deciding how we think about a place, who belongs to it and where people feel they belong.

Take me. I'd say I live in Melbourne, but I actually live in the suburb of Ringwood, in the City of Maroondah, in the federal electorate of Deakin. If Australia defined cities the way the United States does, by 'city limits', which are government boundaries, I almost certainly wouldn't call myself a Melburnian. In the US you're only 'in' Los Angeles if you're inside the City of Los Angeles local government boundary, which excludes most of the built-up area around it.

Try it yourself: search 'Adelaide' in Google Maps and it shows the urban extent of Greater Adelaide. Search 'Los Angeles' and it shows the City of Los Angeles boundary, excluding most of the surrounding urban area. Australians and Americans are, quite literally, looking at different shapes when they picture a city.

In Australia we lean towards urban areas rather than government boundaries, to the point where we use 'suburb' and 'locality' almost interchangeably. So when the question is whether Melbourne or Sydney is bigger, it comes down to calls like whether the Central Coast counts as Sydney, or Melton as Melbourne. It's a judgement call, which is exactly why we rely on the ABS to set an official standard, one that weighs urban extent, labour markets and community identity. It isn't perfect, but it gets revisited every five years, which is what they have just done in preparation for the 2026 Census.

The ASGS levels: from Mesh Blocks to SA4

The ASGS has been in place since 2011, when it replaced the Australian Standard Geographical Classification (ASGC), one letter different in the abbreviation, a very different structure underneath. One change that matters most for Informed Decisions' clients: Local Government Areas (LGAs) are represented in the ASGS, but they no longer combine neatly from smaller areas. SA1s and SA2s can cross LGA boundaries, as anyone who has used Informed Decisions' atlases will recognise. To fully cover an LGA, we need every SA1 that falls within it, even partly.

The ASGS 'Main Structure' is built from Statistical Areas at four levels, SA1 through to SA4, sitting on top of the smaller Mesh Block.

Mesh Blocks and SA1s

Mesh Blocks are the base building block of the whole standard, typically 30 to 60 dwellings. They're small enough that the ABS releases very little data at this level to protect confidentiality, though you can aggregate them up to a custom area to get the full range of Census data.

SA1s sit above Mesh Blocks and are the smallest area for which the full range of Census data is released. They carry a code rather than a name, and are designed to hold between 200 and 800 people, think a small neighbourhood or a handful of street blocks, larger in rural areas. They're the basis for Informed Decisions' Social Atlas and the best unit we have for fine-grained mapping, though mostly only Census data is available at this level. The SA1 replaced the old 'Census Collection District' (used up to the 2006 Census), the area covered by one Census collector at the time, which doubled as the smallest output unit.

SA2s

SA2s are the geography most people actually use, often described as interchangeable with 'suburb' or 'town', though they're neither. An SA2 can cover several suburbs, even where its name only mentions one, and larger suburbs are often split across more than one SA2. The target population is 10,000 to 20,000 people, though SA2s range from around 3,000 to 25,000, and a handful sit close to zero where they isolate an industrial area with no residents.

Regional towns are often split into a 'town' SA2 and a 'surrounds' SA2, while the built-up area might stretch across both plus neighbouring SA2s. Smaller towns keep their surrounding area within a single SA2.

SA2s are designed to be a flexible unit: a place people identify with, but with enough resolution to release small-area statistics and stay stable over five years. They're a Census building block, but their real value is that they also underpin most non-Census data, building approvals, JobSeeker figures, and health and crime data from outside the ABS. Census goes down to SA1 level; almost everything else starts at SA2.

SA3s

Between SA2 and SA4 sits the SA3, built by grouping whole SA2s that share a common area, typically 30,000 to 130,000 people. SA3s are designed for regional-scale output, though they aren't widely used and are a little hard to describe simply. In many of our biggest cities they roughly approximate a Local Government Area, or two to three smaller LGAs combined, but they're an ABS-maintained part of the classification, not a separate structure like LGAs are.

SA4s, capital cities and urban areas

SA4s are broad regions within a state or city, usually named for their location, think 'North East Melbourne', 'South West Sydney', 'Sunshine Coast' or the 'Mandurah' region. Greater Capital City Statistical Areas (GCCSAs) are larger again: built from SA4s, they define the functional area of each state and territory capital, and they're the main benchmark geography on Informed Decisions' community profile sites.

There's a second way of defining a city: the Significant Urban Area (SUA), built from SA2s rather than SA4s. An SUA covers any urban centre of 10,000 people or more, so it captures cities well beyond the state capitals, and a GCCSA can contain more than one SUA, as Greater Sydney does with Sydney and the Central Coast.

Other geographies on Informed Decisions' tools

On Informed Decisions' community profile and forecast tools, we can also build suburb and locality boundaries, planning areas, wards, structure plans and other areas that matter to a council or community (SA2s are often labelled 'suburbs', but they aren't the same thing). This only works for Census data, and only by aggregating smaller areas like SA1s up to approximate the larger boundary.

LGAs themselves sit outside the ABS structure altogether. They're a 'non-ABS structure', built from Mesh Blocks, which the ABS represents but doesn't maintain, and, as above, they can cross SA1 and SA2 boundaries. The same applies to electoral divisions, maintained by the AEC or state equivalents, and to postal areas, which is ABS terminology for the best-fit approximation of postcodes. Postcodes are one of the most common ways survey data gets grouped, and usually the first thing people are asked when identifying where they live. But they're built for mail delivery, not statistics, and can include non-spatial things like delivery routes and PO boxes. That's why they're not a standard ABS geography, and why we rarely use them in Informed Decisions' output.

What's changing in the 2026 release

A new ASGS release matters well beyond the Census, it underpins most of the ABS's statistical collections. The ABS has released the new Main Structure and Greater Capital City Statistical Areas on 22 July 2026. So what's actually changing?

What has been included in this release is only the 'Main Structure' shown above, and the Greater Capital City Statistical Areas. None of the GCCSAs have changed between 2021 and 2026, we are using the same boundaries for another five years. This is because they are quite wide boundaries which fully encapsulate the urban area. You still have the slightly strange boundary for Greater Brisbane, which surrounds the Gold Coast (part of Regional Qld) without including it. So places like Beaudesert are considered to be Greater Brisbane, but Surfers Paradise is not.

It also means we can't say whether Melbourne or Sydney is going to be bigger based on the Significant Urban Area geography for a few months more yet, that's out in October.

At the lower geographic levels, there are more SA2s. Growth areas have been split, usually without boundary adjustment, so the new areas should still add up to match the old ones for comparison. Some growth suburbs have been split into several SA2s; others have simply been tidied up. We've added 98 SA2s nationwide, for a total of 2,571. The largest changes to these are in our growth areas on the urban fringes, particularly in Melbourne. The former SA2 of Fraser Rise in Melbourne's west has been split into five different SA2s due to population growth.

The biggest changes are at SA1 level. There are too many to list individually, but the ABS publishes a concordance showing how each SA1 has changed and what it corresponds to previously. That's what lets us show two Census periods on comparable SA1 geography in the Social Atlas, though only two; anything more than that becomes almost impossible to represent visually. In some high-growth fringe areas, a single 2021 SA1 has been broken into 10 or more smaller areas this time round. But for most established areas they remain the same for data comparability. Overall the number of SA1s nationally has risen from 61,845 to 63,485, an increase of 1,640 areas. Informed Decisions tries to get ahead of the changes in growth areas with our 'forecast small areas' geography in Placemaker, which pre-splits large SA1s in growth areas, though we don't know exactly where the ABS will draw the new lines until they're released, so we still need to reconcile our areas against the final SA1s, which we will receive data from the ABS based on.

Got a question about the 2026 Census?

This is one of many questions Informed Decisions can help you make sense of before the 2026 Census results land. If there's something about the Census, or how to use its data with confidence, that you're not sure about, ask us anything on our Census 2026 hub and we'll put it to our experts and publish the answer.

Sources

Australian Bureau of Statistics, Australian Statistical Geography Standard (ASGS) Edition 3, July 2021 - June 2026, abs.gov.au (accessed July 2026).

Australian Bureau of Statistics, ASGS Edition 4 consultation and release schedule, consult.abs.gov.au (accessed July 2026).

https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs/latest-release