The study focused on a controversial type of carbon credit, which is based on protecting forests against any future deforestation.

Under this scheme, a business buys one credit for protecting a forest in such a way that one tonne of carbon has been prevented from being emitted into the atmosphere.

The peer-reviewed study, published in the journal Global Environmental Change, looked at how these credits are measured.

It found that the baseline for the calculations -- one credit for one tonne of carbon dioxide saved -- lacked detail and accuracy, and the parameters were too flexible.

As a result, this has led to "over-crediting" that benefits Reducing Emissions from Deforestation and Forest Degradation (REDD+) projects.

These projects have generated 436 million carbon credits, about one fifth of the two billion credits issued worldwide since the early 2000s.

Two of the study's authors, Thales West and Barbara Haya, were involved in other recent studies which tarnished the credibility of carbon credits, notably those generated by Verra, a US organisation that has certified close to two-thirds of existing credits.

The researchers went further in the latest study by claiming that four of Verra's methodologies use criteria far too vague to estimate the deforestation that would have happened without the implementation of a protection project funded by carbon credits.

This is key data to calculate the amount of CO2 not released into the atmosphere due to the project.

According to the study, the methodologies rely on "simplistic forecasts" that are based on "historical deforestation trends or averages observed over a 10-year period, disregarding changes in political or economic contexts known to influence deforestation" over the course of a project.

Instead, researchers recommended that the protected zone be compared with a similar area that is not funded to determine if the project actually prevented deforestation.

- Contributions versus offsets -

Faced with criticism, Verra announced in recent months that it would replace its four methods with a new one, which has so far been applied to five projects in development, without generating credits at this stage.

Meanwhile, more than half of credits generated through old methodologies were already used by companies to sell "carbon neutral" flights, shampoo and coffee.

The rest is still circulating in the market and can be used despite being shown that the benefits are largely exaggerated.

The Integrity Council for the Voluntary Carbon Market (ICVCM), a private group established in 2021 after several scandals to evaluate the methodologies of certifiers certification, is currently studying Verra's new protocol to decide whether to award it as a quality carbon credit.

Verra asked to exclude its former methods in the evaluation, an ICVCM spokesperson told AFP.

The decision by the ICVCM committee, composed of experts, market stakeholders and funded mainly by foundations, is expected in the coming months.

But one author of Tuesday's study, Haya, said the new method remains based on a flawed approach.

"The improvements likely will only reduce the amount of over-crediting while still leaving significant room for over-crediting," she told AFP.

"The methodologies should be treated as guilty until proven innocent," she said, adding that companies should stop using carbon credits and consider financing forest protection projects as a "contribution" toward mitigating global warming rather than as an offset for their emissions.