There's a loophole in the Montreal Protocol: will ozone monitoring change?

For years, the Montreal Protocol’s feedstock exemption looked like a manageable compromise. 

Certain ozone-depleting substances could still be produced and used as industrial feedstocks, because policy-makers believed only a tiny fraction would leak into the atmosphere. 

A new study suggests that assumption no longer holds. 

Find your next air quality monitoring in our international directory.

Researchers say feedstock-related emissions are now high enough to delay mid-latitude ozone recovery by about seven years if no additional action is taken.

How will this affect the Montreal Protocol?

The Montreal Protocol has always depended on a combination of national reporting, technical assessment and atmospheric science. But this latest finding points to a weakness in the balance between those three things. 

The exemption was built around an expectation of minimal leakage. What scientists are now showing is that real-world emissions may be materially higher than the system was designed to assume.

The new paper, published in Nature Communications, examines emissions from ozone-depleting substances used as feedstocks in the manufacture of other chemicals, including polymers and fluorochemicals.

 The authors compare the old working assumption of roughly 0.5% leakage with present-day estimates closer to 3.6% for many feedstock uses. In their baseline scenario, these elevated emissions push ozone recovery from the mid-2060s into the early 2070s. 

In other words, one of the main remaining threats to ozone recovery may now sit not in banned legacy uses, but in permitted industrial processes that were supposed to be tightly contained.

That distinction matters for anyone involved in environmental measurement. Article 7 of the Montreal Protocol requires parties to provide statistical data on production, imports and exports of controlled substances. 

It is a robust framework for tracking regulated chemicals through the formal economy, and it has been central to the treaty’s success. But production and trade data are not the same as verified emissions data. A system can be excellent at recording what entered a process while still being weak at proving what escaped from it.

Are we monitoring incorrectly?

If feedstock emissions are higher than previously assumed, then the question facing regulators is no longer simply whether countries are reporting feedstock use correctly. 

It is whether they can show, with enough confidence, that industrial losses are being minimised in practice. That is a more demanding task. 

It means greater attention to leak pathways, intermediates, by-products, destruction efficiency and abatement performance. It also means placing more weight on the relationship between plant-level estimates and the independent atmospheric observations that can confirm or challenge them.

New standards in development

There are signs that the policy community already sees the issue in those terms. At the 47th meeting of the Montreal Protocol’s Open-ended Working Group in July 2025, feedstock uses and associated emissions were discussed as an active issue. 

Delegates debated the level of monitoring required, considered what emissions data parties could provide from processes involving feedstock production or use, and discussed the role of atmospheric monitoring in identifying and validating emissions patterns. 

The same process also included work on regional atmospheric monitoring capacity, including suitable station locations, calibration, validation, data-sharing and long-term support.

That does not automatically translate into continuous perimeter monitoring at every chemical site. In many cases the first changes will probably be more incremental. 

National ozone units and their technical partners may ask for better mass-balance accounting, more detailed emissions factors, stronger evidence for claimed containment rates, and clearer data on where losses occur across the process chain.

But once atmospheric networks indicate that sector-wide assumptions are too optimistic, the direction of travel becomes harder to ignore. Inventory-based compliance starts to look insufficient on its own.

The end of averaging?

For the monitoring community, that is the most important implication of the loophole. Environmental regulation often works on the basis of categories and averages until measurement technology becomes good enough to expose what those categories hide. That is effectively what has happened here. 

Long-term atmospheric observations have helped reveal that a legally permitted use may still be creating a significant environmental burden. Once that happens, the burden of proof changes. Operators, regulators and treaty institutions are pushed away from assumed low leakage and towards demonstrated low leakage.

The Montreal Protocol has often been praised because it adapts when science changes. If feedstock emissions are now one of the largest remaining obstacles to full ozone recovery, adaptation will not just mean rewriting treaty language. It will mean changing what counts as acceptable evidence. For those responsible for monitoring, the likely shift is from tracking quantities handled to proving quantities emitted. That is a subtle change in wording, but a major change in approach.

Read the full paper here.