Corrected 17 August 2026, two hours after publication. The first version said the 1 mg/m³ welding fume figure “takes legal effect on 1 December 2026” and that a control set against an older number was about to stop counting. That was wrong in the direction that matters: it could be read as meaning a workplace still working to the old 5 mg/m³ has until December to move. It does not. WHS ministers reduced the standard from 5 to 1 mg/m³ on 18 January 2024, mandatory as each jurisdiction implemented it, and 1 December 2026 simply carries the same figure onto the new limit list. The standfirst, the table note and the closing practical point are rewritten; Safe Work Australia’s announcement of the 2024 change is added as a source. No figure in the table changed.
Substantially revised 17 August 2026, later the same day. We found and read the independent review behind the 1 mg/m³ figure, which we had recorded as unreachable. It changes the piece in two ways and both are corrections rather than additions. First, our framing was wrong. The original said the blanket limit was “the weakest constraint in the table” and that using it “when you know the consumable is the error”. The review recommends keeping a total-fume standard precisely because the constituents are not fully known, and says so directly; a new section sets out its reasoning. The compliance point stands, that a respirable manganese limit is fifty times tighter and must be met on its own terms, but the two limits are parallel duties and we had implied one supersedes the other. Second, aluminium (welding fumes) is not in the same position as welding fumes (NOC) and the table now says so. No figure in the table changed.
The blanket number, and what it is for
Safe Work Australia’s Hazardous Chemical Information System carries an approved record for welding fumes (not otherwise classified): a time-weighted average of 1 mg/m³ measured on the inhalable fraction. It has no CAS number, because it is not a chemical. It has no short-term limit, no peak limitation and no IDLH value. There is a second, near-identical record for aluminium (welding fumes) (as Al), also 1 mg/m³ inhalable, under CAS 7429-90-5.
“Not otherwise classified” is the whole design. The blanket limit is the
catch-all for fume whose composition is not covered by a more specific entry. It is a floor
under the general duty in regulation 49, which requires a PCBU to ensure
no person at the workplace is exposed to a substance or mixture in an airborne
concentration that exceeds the exposure standard for the substance or mixture
. If you
know what is in your fume, the specific entry is the one that binds.
What is actually in the fume, and what each part is allowed
Welding fume is condensed metal vapour from the consumable, the parent metal and any coating. What that contains depends entirely on what is being welded and with what. Every figure below is the approved record in the national register, read from it directly.
Approved workplace exposure limits, time-weighted average, from the national register on 17 August 2026. These are the figures on the new limit list, which replaces the exposure standards on 1 December 2026; for welding fumes the same 1 mg/m³ has applied since the ministers’ January 2024 reduction. Aluminium (welding fumes) is the exception on this list and the one to check: the January 2024 reduction did not cover it, a consultation on cutting it from 5 to 1 mg/m³ ran from 5 December 2024 to 28 February 2025, and the new limit list carries 1 mg/m³. If you work to the current exposure standard rather than the new list, confirm which figure your jurisdiction has in force for aluminium before relying on either. The size fraction is part of the limit, not a footnote: an inhalable and a respirable figure are not interchangeable.
| Register entry | Size fraction | TWA |
|---|---|---|
| Welding fumes (not otherwise classified) | Inhalable | 1 mg/m³ |
| Aluminium (welding fumes) (as Al) (see note) | Inhalable (as Al) | 1 mg/m³ |
| Iron oxide fume | Inhalable (as Fe) | 5 mg/m³ |
| Chromium (metal), and chromium (II) and (III) compounds | Inhalable (as Cr) | 0.5 mg/m³ |
| Nickel, metal and insoluble compounds (as Ni) | Inhalable (as Ni) | 0.1 mg/m³ |
| Manganese fume, dust and compounds | Inhalable (as Mn) | 0.1 mg/m³ |
| Manganese fume, dust and compounds | Respirable (as Mn) | 0.02 mg/m³ |
| Chromium (VI) compounds | none recorded | No number |
The gap the table shows
Read the first row and the second-last row together. The blanket welding-fume limit is 1 mg/m³. Respirable manganese is 0.02 mg/m³. That is a factor of fifty, and manganese is present in most steel welding consumables. Air monitoring that samples total inhalable fume and reports a comfortable result against the blanket figure has not measured the thing with the tightest limit, and the respirable fraction is a different sample, not a different calculation on the same one.
The same logic applies to nickel and chromium in stainless work. Both duties apply at once: the blanket figure and the specific one are separate limits, not alternatives, and meeting the looser one is not a defence on the tighter one.
Why Australia keeps a total-fume number at all
It is tempting to conclude from the table that the blanket figure is the weak link and the constituent limits are what really matter. The independent review that recommended the 1 mg/m³ figure argues the opposite, and it is worth reading before drawing that conclusion.
Safe Work Australia commissioned Professor Tim Driscoll, an occupational physician and epidemiologist, to review the standards for welding fumes (NOC) and aluminium (welding fumes). His report notes that several countries have gone the constituent-only way. The United Kingdom withdrew its 5 mg/m³ total welding fume standard in 2005; in the United States the American Conference of Governmental Industrial Hygienists withdrew its equivalent 5 mg/m³ value in 2003, in favour of values for individual constituents only.
He recommended Australia not follow them, in terms:
This approach is not recommended for Australia because the harmful constituents of welding
fumes are not known in their entirety
. The reasoning is that the International Agency for
Research on Cancer classified welding fumes as a human carcinogen
based on total welding fumes, not on any specific component of welding fumes
, and that
Focussing only on the individual constituents runs the risk of not measuring the presence
of a relevant constituent or combination of constituents and thus underestimating the
relevant exposure of workers
.
The other thing the report is candid about is what the number is not. It states there is
insufficient published information to confidently propose either a
health-based or
risk-based Workplace Exposure Standard for welding fumes (NOC)
, and that
Dose-response data for lung cancer are insufficient to allow lung cancer to be used as the
critical health effect
. 1 mg/m³ is therefore not a threshold below which welding fume is
established as safe. It is a figure chosen for consistency with comparable countries,
which range from 1 to 5 mg/m³, and deliberately at the protective end of that range.
Treat it as a control trigger, not a clean bill of health.
The last row is a different kind of gap and we have covered it in full separately: hexavalent chromium has no exposure number at all, because no safe airborne concentration has been identified. It is flagged in the register as a non-threshold genotoxic carcinogen rather than given a figure, and stainless steel welding is one of the ways it is generated. If that is your work, the blanket welding-fume limit is not the relevant control and neither is any other number. See our piece on the fifty-four chemicals with no exposure number.
Health monitoring: the regulator says the guide does not exist
Safe Work Australia publishes health monitoring guidance for specific hazardous chemicals,
currently 34 chemical-specific guides and 4 general ones. In its August 2026 consultation
paper on health and air monitoring it records that an expert review of that guidance is under
way, and that The project will also include the development of additional health monitoring
guides, such as welding fumes, where gaps are identified.
Welding fumes is the example the regulator reaches for when naming a gap. That is worth knowing on both sides of the duty. For a PCBU it means there is no chemical-specific guide to hand a doctor for this exposure today. For a worker it means the absence of a guide is not evidence the exposure is minor; it is evidence the guidance has not caught up, and the regulator has said so in writing.
The same consultation is reviewing when health monitoring is required at all, and we have covered that separately: the threshold turns on a term the Regulations never define. See our piece on the review. Submissions close 21 September 2026.
What this means in practice
Three things follow, and none of them requires waiting for the guidance.
Know your consumable before you pick a number. The safety data sheet for the electrode or wire tells you which metals are in play. That determines which row of the table binds, and the answer is often not the blanket one.
Match the sample to the limit. A respirable limit needs respirable sampling. Reporting an inhalable result against a respirable figure is not conservative, it is a different measurement.
Do not wait for 1 December 2026 on this one. That is the date the new workplace exposure limit list replaces the old exposure standards, and it matters for chemicals whose number is changing. Welding fume is not one of them. Its figure was cut from 5 to 1 mg/m³ by WHS ministers on 18 January 2024, becoming mandatory as each jurisdiction implemented it in its own WHS or OHS laws, and 1 December 2026 carries the same 1 mg/m³ onto the new list. A control still set against 5 is not early. It is two and a half years late.
Methodology
Every exposure figure here was read directly from Safe Work Australia’s Hazardous Chemical Information System on 17 August 2026, through the public API the site’s own search uses, filtered to approved records only. The register holds draft and superseded records alongside approved ones and they are not distinguished on screen, so reading it any other way returns figures that are not the operative ones.
What the table is and is not. It is the register’s approved TWA for each named entry. It is not a claim about what any particular welding process produces, and we have not measured anything. Which rows apply to a given job depends on the consumable, the parent metal and any coating, which is the point of the piece rather than a caveat to it. Where an entry carries both an inhalable and a respirable figure, both are shown, because the size fraction is part of the limit.
Hexavalent chromium. The blank is in the register: the approved records for chromium (VI) compounds carry no TWA, no short-term limit and no peak limitation, and are annotated with the register’s note k. We reported that class in full on 6 August and have not re-derived it here.
What we have not done. We have not contacted Safe Work Australia and would publish a response. We have not read the 34 existing health monitoring guides for this story; the statement that welding fumes is a named gap is the consultation paper’s own words, quoted above.