Thinking

Are we designing homes for the climate they will actually live in?

Are we designing homes for the climate they will actually live in?

A home designed today may still be occupied in 2070. Australian research is exposing a simple mismatch: we build for decades, but much of our regulatory thermal assessment still describes climate conditions drawn from the past.

A home designed today may still be occupied in 2070. Australian research is exposing a simple mismatch: we build for decades, but much of our regulatory thermal assessment still describes climate conditions drawn from the past.

A house completed today may outlive its first owners.

Over that time its occupants will change. Materials will age. Trees will grow. Parts of the building will be repaired, altered and replaced.

The climate around it will change too.

Yet much of the thermal information used to assess Australian buildings remains tied to historical weather data. That is increasingly difficult to reconcile with the lifespan of the thing being designed.

Recent Australian research makes the consequence tangible.

A 2026 study modelled NCC-compliant apartments in Redfern and Penrith against projected future climates. Under the study’s upper-bound 2050s scenario, the frequency of overheating increased from 1 to 12 per cent in Redfern and from 10 to 23 per cent in Penrith. The researchers found that apartments compliant with today’s standards could experience conditions outside acceptable comfort ranges for the equivalent of about four weeks each year in Redfern and more than seven weeks in Penrith.

This is not a prediction that every Sydney apartment will experience those exact conditions. The study tested particular apartment configurations against a range of projected climates, and the most severe numbers belong to an upper-bound model.

But the question it exposes is difficult to avoid.

If a building is expected to last for fifty or more years, should we test it only against the climate at the beginning of its life?

Compliance and safety are different questions

An energy rating is useful. It gives a consistent way to estimate the heating and cooling energy associated with a design under defined assumptions.

But energy demand and thermal safety are not the same thing.

A building can require relatively little annual cooling and still perform poorly during a short period of extreme heat. Average performance can conceal the few days when indoor conditions matter most.

Those days also bring another vulnerability. Extreme heat places pressure on electricity networks. Cooling equipment can fail. Power can be interrupted. Some occupants will limit air-conditioning because of cost.

That is why researchers increasingly use the idea of thermal autonomy or passive survivability: how long a building can maintain tolerable conditions when mechanical systems cannot compensate for the weather outside.

This changes the design question.

It is no longer simply, How efficiently can the house maintain comfort?

It becomes, What can the building itself do when conditions are difficult?

The useful decisions are often architectural

The Sydney study is particularly interesting because more insulation was not a complete answer.

External shading reduced modelled overheating. Secure night ventilation helped release accumulated heat. Roof colour mattered, especially for upper-level apartments. Orientation and floor position changed exposure. Combinations of envelope improvements, ventilation and external shading reduced weighted overheating exceedance hours by as much as 94 per cent in the modelled cases.

The lesson is not that one of these measures is universally correct.

It is that future performance is produced by relationships between decisions.

A better window cannot undo uncontrolled western sun. More insulation can slow heat moving in, but it can also slow heat leaving. Natural ventilation is only useful when outside conditions are favourable and the opening can actually be used. Mechanical cooling can provide safety during extreme conditions, but only while energy, equipment and affordability allow it.

Architecture is where these decisions are brought together.

We can already test future conditions

This is not a speculative capability waiting for future technology.

CSIRO has already produced projected hourly weather files for 83 Australian locations covering 2030, 2050, 2070 and 2090 under three climate scenarios. They are available for building simulation software and for NatHERS tools operating in non-regulatory mode.

In 2026, the national Building for the Future research programme is going further, developing prototype future typical and extreme-weather files intended to test thermal comfort, energy use and heatwave performance.

So the question is no longer whether future-climate testing is possible.

It is when it is worth doing.

Not every decision needs another model. But some decisions are difficult or expensive to reverse: orientation, window size, external shading, roof form, ventilation paths and the basic environmental strategy of the house.

Testing those decisions against more than one climate can reveal vulnerabilities while they are still cheap to change.

Enough information for a long-lived building

Future climate models are uncertain. They describe ranges of plausible conditions, not the weather on a particular Tuesday in 2050.

That is not a reason to ignore them.

Architecture already makes decisions under uncertainty. We design for wind, flood, bushfire, rainfall, structural loads and material deterioration without knowing exactly when any particular event will occur.

The useful question is whether a design remains reasonably capable across a range of conditions.

Perhaps a house does not need to be “future-proof”. Nothing is.

It does, however, deserve to be tested against more than the weather of the past.

A regulatory rating remains an important baseline. But for a building intended to last generations, the more useful question may be:

Will the decisions we make today still protect the people living here when the climate around the building is different?

That is not asking a model to predict the future.

It is asking architecture to take the life of the building seriously.


Sources