Based on analysis of time series of abundances of insect species in the United States, Crossley et al. reported in Nature Ecology & Evolution (August 2020) that there is no evidence of a global decline in insect abundance or diversity in the United States, this for both natural and anthropogenic sites. Their study described the apparent robustness of insect populations in the U.S. as reassuring, in contrast to recent studies reporting dramatic declines in their abundance around the world.
However, a multidisciplinary consortium including researchers from INRAE, the University of Rennes and CNRS identified major problems in the Crossley et al. paper concerning: 1) statistical analysis and 2) inconsistencies in data selection. The consortium demonstrates, in a commentary published in Nature Ecology & Evolution Desquilbet et al. , that these biases call into question the conclusions of Crossley et al.
This is the second time (see here) that a high-ranking publication downplaying insect declines has been subject to methodological criticisms. These studies raise the question of a "biodiversity-skepticism" within the scientific community. In order to implement an appropriate protection of biodiversity, public decision-makers need an informed diagnosis that is not clouded by biased studies which slow down decision-making.
Desquilbet M, Cornillon PA, Gaume L & Bonmatin JM (2021)
Matters arising: Adequate statistical modelling and data selection are essential when analysing abundance and diversity trends
Nature Ecology & Evolution doi : https://www.nature.com/articles/s41559-021-01427-x