The disappearance of insects is a widely documented fact, more his high speed was questioned by two studies published in 2020 in prestigious scientific journals Science and Nature Ecology & Evolution . However, they have been the subject of numerous criticisms, including two published in the same journals by a multidisciplinary consortium including researchers from INRAE, CNRS and Rennes 2 University.
Good long-term (sub)tropical data on insects are rare and contradict the idea that freshwater insects are increasing at a global level as per Van Klink etal 2020.#insectapocalypse
Super cool to be involved with this paper! Nice job @gqromero, @mormul_rp, Pavel, Pablo and Dieison https://t.co/iK3ycovybg— Liam N. Nash (@LNMNash) June 14, 2021
The authors of these comments, the last of which is on April 5, 2021, point to numerous methodological errors in these studies and are concerned about the decisions that could result biased analyzes of their colleagues. They wonder about the manifestation, even within the scientific community, of a “biodiversity-skepticism”, which fuels an unfounded doubt about the decline of biodiversity.
–
The timarch or maritime blood-spitter, a small beetle incapable of flight. © archives Thierry Creux, Ouest-France
“By wrongly minimizing the decline of insect populations, these publications can only slow down public decisions, thinks Laurence Gaume-Vial, CNRS researcher in Montpellier and co-author of the commentary. The title of the first meta-analysis, published by Van Klink in early 2020, contrasts the presumed increase in the number of freshwater insects with the decrease in terrestrial insects. It is shocking to put it this way because it misleads the reader, suggesting that the increase in some outweighs the decrease in others. However, fresh water accounts for only 2.5% of the earth’s surface. It’s not much ! This may be a desire to hook the reader, but on such subjects, scientists must ask themselves the question of the use of their results. “
Thirty pages of errors
Meta-analyzes, these studies bringing together data from hundreds of other scientific publications, are indeed important milestones in their disciplines because they present the state of play. “You have to be extremely careful when publishing the conclusions of meta-analyzes, because they serve as a reference, continues the researcher. The golden rule is to focus on a specific goal. However, the meta-analyzes that we have commented on are not robust: their objectives are too broad and they end up mixing leeks and carrots. “
–
The golden cetonia is a metallic green beetle with golden reflections © archives Thierry Creux, Ouest-France
In the end, the authors of the commentary identified thirty pages of errors of all kinds on the meta-analysis of Van Klink and his co-authors. As is customary in these cases, the authors of the commented study answer and present a erratum of their work. But Laurence Gaume-Vial and her colleagues are far from being convinced: “They only took into account 5% of our comments and corrected only the most flagrant errors, such as that of a code 101, corresponding to an absence of measurement in a data set, and which was transformed into 101 counted specimens. “ Another error, not rectified, is the use of the results of studies observing the repopulation of insects in a disaster area, for example after a fire. By integrating the data as is in their work, the authors reap growing numbers when it comes to specific situations that cannot be extrapolated.
–
However, current research is unambiguous on the disappearance of insects. © Ouest-France infographic
But how could a study with such a weak methodology have climbed into the pages of a journal as prestigious as Science ? “A meta-analysis contains thousands of data and requires a very important proofreading effort, enlightens Laurence Gaume-Vial. It also requires having a very good level in statistics. Faced with the amount of data to be analyzed, it is undoubtedly necessary to introduce new rules in the proofreading system of the journals that publish them. “
–
A ladybug. © archives Thierry Creux, Ouest-France
Currently, any scientific publication must be reread by other researchers working in the same specialty. The latter may or may not validate the working method, but this proofreading time is voluntary and is constantly shortening in the face of the increasing quantity of scientific production. The authors of the commentary took several months to analyze all the data used and identify any errors in interpretation.
– .
