Our Chief Scientist, Professor Ed Mitchard, is lead author on a new paper “Serious errors impair an assessment of forest carbon projects: A rebuttal of West et al. (2023)”
Professor Ed Mitchard, alongside Riccardo Cosenza, Professor Sassan Saatchi, and others have published a paper stating that the methodology used to discredit carbon credits generated by preventing deforestation contains serious errors.
They are calling for the West paper to be retracted or heavily revised, as its conclusions are highly uncertain and based on data with serious flaws.
Mitchard, a world-leading carbon specialist and Chief Scientist of nature data company Space Intelligence, which counts Apple and Equinor among its clients, has been joined by a similarly respected group of co-authors. They have affiliations with NASA, Conservation International, UCLA and the University of Edinburgh’s School of Geosciences.
The group analysed a paper published by West et al, which claimed deforestation rates were hardly affected by forest carbon projects and that too many carbon credits had been awarded.
Key Flaws and Findings Identified
The global dataset used by West et al underestimated avoided deforestation by 90% or more. The global deforestation dataset used was found to be inappropriate as it inevitably contains random errors and its sensitivity changed through time as available satellites changed. This meant that projects that successfully reduced deforestation were less likely to be detected as such. The authors refer to a large study in sub-Saharan Africa that assessed the deforestation dataset used and found using it would result in a project that was 100% effective only being credited with being 10% effective.
West and colleagues made numerical errors when calculating the carbon benefits of projects their analysis found were effective at stopping deforestation. There were two different calculation errors that together meant the proportion of credits they found that delivered real carbon benefits should be increased by 62%.
The comparison sites West et al used to estimate what would have happened in the REDD project sites if no intervention was made to prevent deforestation were found to be completely inappropriate. For example, Peru and Colombia project areas were compared to sites on the other side of the Andes mountain range. They were therefore incomparable in universally recognised key factors that influence deforestation such as the biome, crop species grown, and whether there was access to international markets.
Impacts of Flawed Analysis
Stopping deforestation will not on its own fix climate change: to do that we have to reduce greatly our use of fossil fuels (the other 80%). But we cannot keep to 1.5 degrees without also stopping deforestation, and quickly. Deforestation is responsible for about 20% of carbon emissions. There is no other funding mechanism in place to get large amounts of money to the places where deforestation is happening apart from the voluntary carbon markets.
This flawed analysis of 24 projects unfairly condemned all 100+ REDD projects, and risks cutting off finance for protecting vulnerable tropical forests from destruction at a time when funding needs to grow rapidly. Unwarranted criticism of the carbon markets right now greatly increases the risks of us missing the target of keeping global warming below 1.5 degrees centigrade.
What is the Alternative?
Critique of NBS projects is absolutely essential to ensure high quality projects, but the approach and method taken needs to be valid and accurate.
We believe that any assessment of a REDD+ project must:
Use either locally tuned forest change data with known accuracies or point-based sampling approaches to quantify deforestation. This is what has been used by all Verra REDD projects to date, and the requirements to pass the audit procedures required by the carbon standards. Global forest area and change datasets are not sufficiently accurate to make these comparisons.
Furthermore, matching approaches must always lead to meaningful comparisons between forests of the same ecological type and legal status, and should consistently pass rigorous validation checks before conclusions are drawn from them.
We insist that broad conclusions about avoided deforestation projects as a whole should not only satisfy these requirements, but also require a formal hypothesis testing framework and feature representative sample size.
