Critique of Sala et al. 2021 Published by Nature

Last year, Sala et al. 2021 made waves in both the scientific community and mainstream press with its publication in Nature. The paper claimed that increasing MPAs to stop fishing would lead to more seafood harvest, more biodiversity, and a reduced carbon footprint—a true win-win-win for the ocean. The press release that accompanied the paper highlighted an eye-popping statistic that bottom trawling released more carbon than all airline travel; stories covering Sala et al. 2021 appeared in hundreds of press outlets worldwide.However, the three computer models used to make each of the “win-win-win” claims have been under increased scrutiny and many scientists doubt their conclusions.It started with the food provisioning model initially published in the Proceedings of the National Academy of Sciences (PNAS) in 2020. Inexplicable assumptions in the model and several data errors were missed by an inadequate peer review—likely due to a conflict of interest by the PNAS editor. The journal retracted it in October 2021.You can read the whole breakdown of the retraction and the science of the food model here.Retractions are rare in science and generally only used in cases of misconduct. Poor science is hardly ever retracted for its flaws—instead, it gets officially criticized and/or updated in the literature.That process has now started for Sala et al. 2021, with the first official critique (and response) published today in Nature (though several critiques have been available on preprint servers).The comment, by Ray Hilborn (founder of this site) and Michel Kaiser, points out inconsistent parameters and assumptions and criticizes the overall approach to global MPA science and advocacy.

Inconsistent assumptions in Sala et al. 2021

According to Hilborn and Kaiser, the most severe flaw in Sala et al. 2021 is the inconsistent accounting of fishing effort in the author’s MPA scenarios.In the carbon and biodiversity model, Sala et al. assumes that when an area is placed into an MPA, the fishing effort that was previously there disappears. But, in the food provisioning model, the paper assumes fishing effort moves to other areas open to fishing. This upwardly biases their claims that MPAs could simultaneously reduce carbon footprint, improve biodiversity, and increase catch:

In their calculations of biodiversity conserved and CO2 emissions reduced, the authors assume that fishing effort disappears, which would decrease total harvest at the point when the MPAs are established. Yet in the base case for the fisheries harvest section, the authors assume that fishing effort moves to areas open to fishing, keeping fishing harvests high.

MPAs certainly reduce fishing effort inside a protected area, but in the real world, fishing effort does not simply disappear—it moves outside the MPA to places where fishing is still allowed. In this scenario, the benefits to carbon emissions and biodiversity presented in Sala et al. would significantly decrease, perhaps even show a net negative response because:

Fishing effort generally goes to places with high catch rates, and if forced to fish elsewhere, more effort is required to achieve the same catch.

In their response to Hilborn and Kaiser, the original authors acknowledge that the attention-grabbing statistic in the press release that bottom trawling releases more carbon into the ocean than all airline emissions would only be true if fishing effort disappears.

For carbon, we acknowledge that relocating bottom trawling effort would reduce potential benefits, particularly if relocated to areas with little or no previous trawling.

Nature has yet to publish a comment pointing out other issues in the carbon model that show Sala et al. overestimated their projections by up to 100x.Regardless, picking and choosing fishing effort assumptions for different computer models is a clear bias—one that was especially deceiving as the assumptions were not evident in the text of the paper, only deep into the supplementary materials.

For each topic in the base case the authors have made assumptions about vessel movement that maximizes the benefits of MPAs, but they cannot have it both ways. To fully support their analysis, they must use the same assumption about effort displacement; either the effort disappears, or it does not.

A comment from a peer reviewer of the Hilborn and Kaiser comment sums it up nicely:

Thanks to the discussion and communication, it now becomes clear that Sala et al. considered the effect of effort displacement only for food provisioning and NOT for biodiversity and carbon. (This was hidden deep in the supplementary information). Effort displacement likely diminishes the effect on biodiversity and almost certainly for carbon… The overall message of Sala et al. that MPA are win-win-win for food, biodiversity and carbon is thus undermined and findings may be inconclusive for this reason.

(comment provided thanks to peer reviewer)

Marine protected areas as a panacea

Hilborn and Kaiser also have issues with how the paper presented itself as a global solution to fisheries management. The critique points out that nearly a third of global fisheries are missing from the data, particularly in Asia and Southeast Asia. Further, fisheries are highly local and regional—prescribing a global solution to local and regional fisheries (while leaving out a third of the world) makes for bold claims and notable headlines, but misses the reality of the work needed on the ground to improve fisheries around the world.

Certainly, protection of the oceans is needed, but the paper by Sala et al. suggests that protection can be achieved primarily by using no-take MPAs, and does not include a suite of strategies and tools that have proved to be effective. Almost all the large-scale successes in rebuilding fish stocks and protecting biodiversity have resulted from fisheries management measures such as limits on how many fish can be caught, restrictions of when and where fisheries can operate, and gear limitations, not from no-take MPAs. Put simply, sustainable fisheries are managed, informed by science and have enforcement. The same cannot be said for most of the world’s MPAs.

Though it garnered big headlines and more attention than any other ocean science paper of the last few years, Sala et al. 2021 does a disservice to marine conservation with its analysis based on incomplete data and erroneous assumptions. Policy that follows its recommendations would potentially waste conservation effort and money on strategies that would not deliver on goals, e.g., proposing a network of MPAs where fisheries are already well managed.

Read the whole criticism and response here.Read the breakdown of Cabral et al. 2020, the precursor to Sala et al. 2021

Max Mossler

Max is the managing editor at Sustainable Fisheries UW.