Episode 2

Energy Ratings, Roof Colour and the Urban Heat Island Debate

4 March 2026 · 31 minutes · Hosted by Brian Haines, Matthew Graham and Andrew Hooper

Summary

A widely reported study claimed that households with dark roofs pay close to seven hundred dollars a year more to keep their homes cool. This episode examines what that headline does and does not tell you, and why the answer changes completely depending on where in Australia the house is built.

The team works through how roof colour actually behaves in the modelling: strongly beneficial in Queensland, marginal in Melbourne, and complicated everywhere by roof material, solar absorptance values that are darker than they look, and ceiling insulation levels. Matthew brings modelled results for the same project across all sixty-nine NatHERS climate zones in light, medium and dark, and the spread is smaller than the public debate suggests.

The conversation then widens to the urban heat island effect itself, which the team argue is largely a planning question — black roads, minimal canopy, shrinking lots and rising site coverage — rather than something an individual homeowner solves by changing a roof colour. It closes on the uncomfortable question underneath the whole debate: an assessor works for one client and one dwelling, so who is looking after the street?

In this episode

• 0:35The study, the headline, and the numbers behind it

• 3:03Why Sydney and Melbourne should not appear in the same sentence

• 4:16Concrete tiles, solar absorptance, and roofs that only look light

• 5:24Performance-based rating: the whole point is that you can trade

• 6:13Queensland versus Melbourne: how much a roof colour is actually worth

• 7:56The disconnect between the media headline and the modelling engine

• 8:46Painting roads white: the Sydney and United States trials

• 10:56Why councils cannot afford the fix, and pass it to the homeowner

• 12:52Section J bans dark roofs, and JV3 goes around it

• 14:17Victorian planning, town planning permits, and who could enforce colour

• 15:35Queensland practice, and the thing that hurts ratings more than roofs

• 16:49Is the heat island already inside the climate files?

• 22:24The same project modelled in all 69 climate zones, light to dark

• 24:00Optimising a St Kilda house to nine stars without touching the windows

• 25:36Colour as the cost-neutral lever assessors forget

• 20:07The New South Wales policy that developers stopped

• 28:29Working for the client, or working for the street?

Key points

The headline is a Sydney answer applied nationally

The reported figures — an average New South Wales electricity bill of $1,827, with lighter roofs paying up to $694 less, and a dark roof said to raise Sydney bills by thirty-eight per cent — describe a specific climate and a specific kind of development: new estates with dark roofs, black roads and almost no tree cover. The team do not dispute the Sydney finding. Their objection is to the generalisation, because Brisbane, Perth and Melbourne each behave differently, and Perth's typical brick cavity construction changes the picture again.

In Melbourne, a dark roof can be the better outcome

Melbourne is heating dominated. There are more hours in the year when the heater is running than when the cooler is, so a dark roof saves more over winter than it costs over summer. The modelled difference between light and dark in Melbourne is small — around two per cent — and much of the retained heat is in the concrete tiles regardless of colour, because a tile that appears light often still has a medium to dark solar absorptance.

The value of roof colour scales with latitude

A useful rule of thumb from the episode: in Queensland, changing a dark roof to a light one might move a dwelling from 6.5 stars to 6.8 or 6.9. In Melbourne, changing a light roof to dark might gain one or two tenths. The further north you go, the more the choice matters; the further south, the more it washes out.

Sixty-nine climate zones, and the spread is small

Matthew modelled the same project in every NatHERS climate zone in light, medium and dark, changing only the roof. In many zones there is no difference at all. Where there is one it is typically two to three tenths of a star at most. In climate zone 9 the result was 4.5 stars dark against 4.7 light and medium; in zone 10, 5.3 against 5.4. His point is not that colour is irrelevant — two tenths of a star for a cost-neutral change is worth having — but that it is not the lever the public debate imagines. Window frame colour, he notes, moves the dial considerably harder.

The heat island is a planning problem wearing a roofing costume

Andrew's argument is that the responsibility is being placed on homeowners while the larger levers sit with councils and government. Roads are black. Lots are shrinking, coverage is rising, nature strips are narrowing, and canopy and parkland are not keeping pace. Trials in Sydney and the United States that painted road surfaces with reflective coatings measured air temperature reductions of one to two degrees — a whole-of-street effect no individual roof colour can match — but repainting and repaving at that scale costs money councils do not have.

Nobody has a clean way to enforce colour

Section J prohibits dark roofs in commercial work, and a JV3 performance solution routinely goes around it. The Victorian planning scheme's recent amendments encourage light to medium colours, but planning controls only bite where a town planning permit is required. A New South Wales policy to mandate lighter colours was abandoned after developer objection. The team's conclusion is that in a performance-based code, a homeowner who wants a particular colour can offset it elsewhere — and that an assessor's job is to find that path, not to refuse the colour.

The awkward question at the end

Getting a single dwelling to comply is single-factor analysis. It does not account for what thirty of those dwellings do to the air temperature of the street they share. As Andrew puts it, assessors work for their client, not for the grey suburb — and that is precisely why the team keep returning to the idea that this belongs in planning policy rather than in an individual homeowner's colour selection.

From the comments

Viewers raised several points worth recording alongside the episode. One noted that ceiling insulation levels moderate the impact of roof colour, so hot climates where lower ceiling insulation is common — parts of Queensland where a roof blanket alone is typical — will show a larger colour effect than climates where higher insulation is standard.

Another pointed out that New South Wales places western Sydney in climate zone 6 alongside Melbourne, which permits a dark roof there, and argued the better fix is to reclassify western Sydney to climate zone 5. The same commenter noted that under NCC 2022 the Sydney elemental pathway does not permit a roof with solar absorptance above 0.64, and questioned whether a dark roof could add six hundred dollars a year in cooling costs at all where R5.0 ceiling insulation is present — while asking for a link to the underlying report.

A further comment challenged the premise more broadly, arguing that the urban heat island consequences of built environment decisions reach further than household energy cost comparisons account for, and that a fractional star benefit is not sufficient justification for dark roofs in any climate.

Documents referenced

• National Construction CodeAustralian Building Codes Board — the code itself, and the Housing Provisions referenced throughout.

• NatHERS publicationsTechnical Note, assessor handbook and scheme guidance material.

Transcript

Lightly edited for readability. Hesitations and repetition have been removed and technical terms corrected; the substance is unchanged.

0:04Welcome to The Rating Room, the podcast where we unpack everything NatHERS. I am Brian Haines, and alongside me are my co-hosts Matthew Graham and Andrew Hooper. Together we dive into the technical, the practical and sometimes the controversial aspects of the NatHERS scheme and everything in between. Whether you are an accredited assessor, a building designer, an energy efficiency specialist, or just trying to navigate the ever-evolving world of residential energy performance in Australia, this is the room where the ratings get real.

0:35This is a current study that has just been released talking about dark roofs. I am not sure why it is such a controversial subject — light roofs, dark roofs, different climate zones. According to this one, people with dark roofs are spending almost seven hundred dollars extra a year to keep their house cool. Have we run that through the other CSIRO calculation engine to see whether it is accurate? It certainly could be done.

1:07I often wonder where this information comes from, and whether it is nuanced enough to talk about different climate zones. It specifies that a dark roof means you pay considerably more to keep your house cool in summer, and that last year the average household in New South Wales paid $1,827 for electricity, while those with lighter coloured roofs can pay up to $694 less due to lower cooling needs. Put another way, it says a dark roof in Sydney drives up your power bills by thirty-eight per cent.

1:53In some climate zones in Sydney, we can probably agree with that. But it also talks about dark coloured roofs in climates like Melbourne — new housing estates in the hottest areas around Sydney and Melbourne being dominated by dark rooftops, black roads and minimal tree cover, and dark colours trapping and holding heat rather than reflecting it. We have all done NatHERS assessments in colder climates like Melbourne. What happens when you change a house with a dark roof to a light roof and calculate, then back again?

3:03If we are talking Melbourne, that is a very different climate from Sydney. They should not be in the same sentence. The study is written by a professor at the university who is very well regarded and knows his subject. I have read a number of his studies and they are very good. But he is very Sydney oriented, and rightly so — he is from Sydney. There are a lot of reports about Parramatta, and photographs of completely treeless subdivisions which are all black roofs and black roads with no shading.

3:54Talking about Melbourne, we have similar outcomes come summer: a dark roof is going to be hot. There is no getting around that, especially with concrete roofs as opposed to steel roofs, which might have insulation or a blanket under them.

4:16We can end up with about a 2.3 per cent difference between a light coloured roof and a dark coloured roof. But a lot of that heat is still retained in the concrete tiles even with a light coloured roof, which appears light — yet if you look at the actual solar absorptance it is still medium to dark in some cases. So the concrete is still absorbing even for a light roof. Material should come into play as well, which means you can change your insulation properties and perhaps get a better effect that way.

4:55There are a lot of generalisations in that article. They are absolutely applicable to Sydney, but not necessarily everywhere else. Brisbane will be different again. Perth will be different, and Perth has different construction with brick cavity walls.

5:24The great thing about NatHERS is that it is not a prescriptive energy rating. It is performance based. We can model every house for what it is actually being built as, and we can change the materials, look at the outcomes, and say that in Melbourne a dark roof is going to save more over winter than you will pay in summer. There are more hours in the year when your heater is going in Melbourne than your cooler is in summer. So it makes sense that darker colours in Melbourne reduce reliance on mechanical heating during winter — more days of use, more carbon, more cost.

6:13As you get further up the country to Queensland, a light roof makes a dramatically positive difference to the rating. As you come further down, that difference reduces. If you had a dark roof on a house in Queensland rating at 6.5 stars and changed it to a light roof, there is every chance it goes to 6.8 or 6.9. If you go down to Melbourne and change a light roof to a dark roof, it will more than likely improve, but only by a tenth or two.

7:06So the generalisations are probably not great. There need to be nuances, and studies done across all or most of the climate zones. The other thing I find interesting is the urban heat island effect itself, and what that does to the cost of mechanical heating and cooling in an individual house. It would be very hard to do a proper study, because you would effectively have to build it one way, study it, build it the other way, and compare.

7:56What I want to know is where the disconnect is. You have CSIRO doing this study, and you have the CSIRO Chenath engine. We know the Chenath engine puts a more positive emphasis on dark colours in a colder climate zone. When these studies reach the media and people latch onto them without understanding those nuances, where is the disconnect? Why do we not go deeper into the individual climate zones, get proper data, and produce a graph of which colours are better in which zones? Because it is not all the same.

8:46I am going to put my tinfoil hat on at this point. Look at any given suburb: our roads are black. That has a really big effect on the heat island effect, and the heat island effect exists — I will not argue that. There is a lot of good research out of Berkeley Labs. One of the trials in three suburbs of Sydney painted the roads white, using the same technology the American Air Force uses on runways. They painted sections of road and measured the differences in surface temperature.

9:44America has run these trials as well, and they determined that the Australian studies were too small to draw a strong conclusion, but the American ones found air temperature decreased by one to two degrees. So that is your measurement of what a cool surface does. It reduces the temperature around the houses, which brings the temperature of the houses themselves down, which means less heating and cooling in that environment. The biggest problem with the white roads was the hoons who wanted to leave their signature — and one council was very heavily lobbied to paint them black again.

10:56So the studies are there. And as I said, tinfoil hat on: it comes down to money. Councils cannot afford to do this. They cannot afford white roads, and they cannot afford to repave with a newer product — a lighter grey material that helps break down chemicals and gases and brings temperatures down. It is money the councils and governments do not have. So they put it on the homeowners. They say this is your responsibility, we want you to have a white roof, while we are not going to provide more canopy trees, more parks or bigger nature strips. Everything is getting smaller: more concrete, more road, smaller lots, which means more housing with roof area.

11:55I look at this and say councils and governments need to step up and take more responsibility, because the studies show these things make a difference. Separate the houses a little further. Insist the blocks be a little bigger. You can still keep your coverage percentages, so that you have an amount of green — nature strips, tree canopies, parkland, waterways, stormwater management. There is more that could be done without putting it all on the homeowner and saying, well, you have a dark roof.

12:41We have NatHERS tools that show us what it is doing and we can make informed decisions house by house, but bring in the bigger scheme and you might start making a bigger difference to the heat island effect.

12:52In Section J commercial you are not allowed to have dark roofs, and then everybody does a JV3 and gets around it anyway. And who would enforce something like that? It would have to be a council matter, not necessarily an Australian Building Codes Board item, because if it were in Volume Two of the NCC you should be able to do a performance solution and get around it.

13:46That is the point of having a performance-based building code, where we have Deemed-to-Satisfy solutions or we can have a performance-based solution. We are allowed to vary it. Most people do not, because it is more complicated, but there is nothing saying you cannot.

14:17Under the Victorian planning scheme, the latest amendments are more encouraging of light to medium colours for exactly the same reason. But that is town planning, which generally means units, and it supersedes a building permit. So it could be done that way — although the requirements for town planning have become a lot less restrictive in the last few months with regard to environmentally sustainable design requirements. Council can say things like put shading on these orientations, or it has to be ten per cent above NatHERS requirements, or ten per cent above Section J, or 6.5 stars back in NCC 2019. Council can do whatever they want, but it is only applicable where there is a town planning permit. So I find it hard to see how they would enforce something like this.

15:35Do you do much work in Queensland? Predominantly what sort of roof colour do you see there? The majority are light coloured roofs. And we mostly have issues with dark coloured roofs — it is highly detrimental in those warmer climate zones. But the thing that kills those energy ratings the most, believe it or not, is dark coloured window frames. That one scales up badly.

16:06I do see a lot of dark coloured roofs there, and people are very reluctant to change the colour of their houses. It feels like an aesthetic thing — wanting to look modern, or wanting to look the same as your neighbour. But as Andrew says, it is a performance-based code. If we can offset it another way, I do not see it as an issue. Why would we be so restrictive about somebody's colour palette when they can offset it elsewhere?

16:49The argument in this article is about the heat island effect, so it is not looking at the individual home — it is looking at thirty individual homes within an area, and that heat island increasing the air temperature, which then affects each individual home. Some of that is modelled into the climate data files, I believe. That effect is already happening in parts of Melbourne and Sydney. So is the modelling already taking it into account? Are we doubling up?

17:35We have spoken before about the future climate files. Did you run those with a dark coloured roof? I cannot remember. They are the sort of thing that would be interesting to run. That is a forecast, and you mentioned that in many of those the star rating actually increased, because the heating requirement decreased. A dark coloured roof is definitely going to reduce the heating requirement, but it will exacerbate the cooling requirement. Proportionally you might still get away with it — but would we not also expect Melbourne to become more like Sydney under those files, where the roof colour does have a real effect?

18:35I think it is partly terminology as well. The article says lighter. It does not mean light, it means lighter. Dark and lighter. For me, perception of light in New South Wales would be a medium. What is a galvanised steel roof? Most people think that is light, and it is not — I think the NCC calls it up at around 0.49, which is medium.

19:05With the climate data we are potentially extending our summers as well. We talked before about Melbourne being heating dominated, so the dark roofs are still helping. But if we are extending summer through climate change, that can have a greater effect on the heat island, increasing the air temperature around the dwelling. So we have this see-sawing effect between the individual home and the greater community. We can see the performance of the individual home, but it is harder to take the same data across 2050, 2070 and 2090.

20:07On the question of who is going to police it, there was a policy in New South Wales that came out of these studies — I am not sure of the date — where New South Wales was going to enforce lighter colours. The developers got up in arms very badly and said no, and it was shut down before it became live. So there must have been some method they were going to use to police it.

20:43Why do you think the developers objected? Developers have a lot of land, they build a lot of houses, and they choose the colours. It would not be individual people complaining about the colour of their own roof, because they are choosing it themselves. Presumably the developers are trying to build a house that presents in a way they think will sell fastest. I see projects going through planning with light and medium colours called up and they look fantastic, so I am not convinced aesthetics is the reason. Maybe there are fewer options in lighter colours, or maybe they want the full gamut rather than a subset.

21:57Have you looked at how many colours there are below a solar absorptance of 0.3? Not that many. Even white is around 0.2. Surfmist is 0.33.

22:24Since you mentioned this earlier, I fired up a project that simulates very quickly, and using one of my favourite tools I have run it in all sixty-nine climate zones in light, medium and dark. Unfortunately it does not map each NatHERS climate zone to a familiar name, so this is a test of how well you know your zone numbers. But in many of these zones it does not actually make much difference between light, medium and dark. Some are simply the same. I am changing the roof only — I figured that was a good test, because adding window frame colour would really move the dial.

23:35It is a little dependent on the climate zone. There is no massive swing. For many, none. For some, a tenth of a star. At the most you are looking at two or three tenths. I still think two tenths is important — if I could get that on colour alone, for sure.

24:00When I go through an optimisation session with my students it is really interesting where they start, because I do a lot of the driving — I make the changes to a particular project, but I want them to suggest what they think should be done to improve the design. We take it from slightly below six stars, about 5.8. The first target is to hit seven, and the next is to take it as far as it will go. Each cohort is different; some top out around 9.3 stars, some get all the way to ten. What is fascinating is that this project — a single-storey brick veneer in St Kilda in Melbourne — can get all the way to nine stars with single glazing.

25:04They start off a little rough and they are not thinking about cost, so invariably a student suggests jumping straight to the windows. I tell them I am not touching the windows until they can get the project to nine stars, because I know it can. Then they think a bit harder. It comes through design changes and external wall construction — they can go to town, reverse brick veneer and all of that. Out of all the sessions I run it is the one I enjoy most.

25:36What is interesting is that they do not start with colour. Colour, for me, is cost neutral. It does not really cost anything — there may be a small difference depending on the product, but changing floor, wall, roof and especially window frame colour is fairly cost neutral. There are also cost savings students quite often miss: that project has a fireplace, and there is a saving that can disappear. A few other things fall into the same category, but colour is an important one. It is something I tested more when I presented at a CSIRO conference many years ago.

26:44The roof area on this one is not extensive — it is a smaller project. The results exaggerate as the roof area increases. We could take a bigger project and run it again. But you can see there is just not a lot in it when you look at these.

27:11What climate zone do you do a lot of work in? Nine and ten in Queensland. In zone 9: 4.5 for a dark roof, 4.7 for a light, 4.7 for a medium. So a slight penalty of two tenths for the dark. Zone 10: 5.3, 5.4 and 5.4.

27:48This reinforces what Andrew was saying: there is a trade-off, and providing there is a solution, you lose a tenth of a star and do something else somewhere else. You can have your light coloured roof if you really want it, and our role as assessors is to make that possible. I do not tell clients they cannot have the colour. If there is no alternative and they ask for my advice, I tell them that in a cool climate darker roof colours will give them a little bit of a lift. But at the end of the day, if they want Surfmist, fine.

28:29That is single-factor analysis though — getting the NatHERS assessment to comply without thinking about the urban heat island effect. We are not trained for that. We are working for our client, not for the grey suburb. It is one of those questions of where your responsibility lies. We are working for an individual. Do we talk to that individual and say, have you thought about everybody else? I think most people just want the colour they want. It is my house, and I like whatever colour I like. Precisely.

29:17Which is why the question of whether this should be mandated comes into play — so that the responsibility is taken away from the individual, and the government thinks about the greater good. But it still gets stuck on the homeowners.

29:36There would be other reasons to choose darker colours depending on where the dwelling is located. Personally, on a more rural project, I think darker colours work better. They sit more comfortably within the Australian landscape and do not pop or stand out. It depends on what you are trying to do. And I have had clients who want lighter colours on large acreage.

30:22That wraps up this episode of The Rating Room. Thanks for joining myself, Matthew and Andrew as we uncovered another corner of the NatHERS world. Your thoughts, questions and insights are incredibly important to us, and this podcast is as much about the broader industry conversation as it is about our own views. Jump onto the NatHERS Assessor Network Facebook page, share your perspective, or leave a comment below the video. We would love to hear what you think and what topics you would like us to explore next. Until next time, keep rating.