Ask for the embodied carbon of a window and you'll usually get back one number: a kilograms-of-CO₂e-per-square-metre figure, lifted from a catalogue or a generic dataset, quoted with a confidence it hasn't earned. Drop it into your Green Star or NABERS tally and move on. The trouble is that number quietly assumes a window is a single, uniform material smeared evenly across its area. It isn't — and the maths gives it away.
Frame is lineal. Glass is areal.
A window is really two things that scale by two different laws. The frame is a run of profile around the edge — its mass tracks the perimeter, in lineal metres. The glass fills the opening — its mass tracks the area, in square metres. Those don't move together. Perimeter grows with the sum of the sides; area grows with their product. So as a window changes size, the balance of frame-to-glass shifts, and with it the carbon.
Perimeter and area don't scale together — so neither does the carbon per square metre.
Picture the extremes. A small window is almost all edge: a lot of frame wrapped around a little glass, so per square metre it's frame-heavy and its embodied carbon runs high. Make the same window bigger and the glass area races ahead of the perimeter — the frame becomes a thinner and thinner slice of the whole, and the per-m² figure drops and shifts toward the glass make-up. Now add a mullion, or a transom, or split it into more lites, and you've added frame back in without adding area — the number moves again. One catalogue kg/m² is an average of all of that. It fits almost no real window on your project.
Why this is about to matter more
For a long time nobody checked, so a rough figure was fine. That's changing. Embodied carbon is moving from a nice-to-report line into something Green Star and NABERS increasingly want substantiated — and if your window carbon comes from a static per-m² number, you're either over-stating it (and handicapping your own rating) or under-stating it (and carrying a number you can't defend). Either way it's the wrong figure for the windows you're actually installing.
The fix is to compute it, not quote it
The honest version is dynamic: build the actual window and work the carbon out from its real geometry. That's what BTW does, off the same model that gives you the thermal numbers — frame mass from the modelled cross-section material areas times the real member lengths, glass mass from pane area times the make-up times density, each multiplied by published carbon factors. Change the size, the split or the system and the number changes with it, because it's derived from the thing in front of you rather than borrowed from a thing that looks like it. It's a screening estimate, not a certified EPD — but it's your window's estimate, and it moves when your window does.
There isn't a lot of rigorous, geometry-based work on this in Australia yet, which is exactly why the shortcut persists. A single number is easy to pass around. It's just not true for more than one window at a time.
Want the real number for one of your windows?
Send us a window — a size, a system, a glass make-up. We'll model it and return the embodied-carbon estimate computed from its actual geometry, so you can see how far it sits from the catalogue figure you've been using.
Model one window → Try BTW yourself →About these figures. Embodied-carbon results from BTW are a screening estimate computed from modelled cross-section material areas, member lengths and glazing make-up against published factors (NABERS Australia) — they are not a certified EPD or a full life-cycle assessment. Modelled results are indicative to the systems, sizes and make-ups provided. BTW models glazed façade products (windows, doors, glazing). AFRC Lab (Advanced Fenestration Research & Certification) is an independent consultancy; it is not the AFRC and is not affiliated with WERS or the Australian Fenestration Rating Council.