Duchenne, F.; Fontaine, C.

In a recent publication, Muller et al. (2023) re-analysed, in light of new data, the dataset of the highly cited paper of Hallmann et al. (2017) showing a strong decline in insect biomass in Germany between 1989 and 2016. In their re-analysis, Muller et al. completed Hallmann et al\'s model with a focus on modelling the effects of weather conditions on insect biomass. They also included temporal changes in habitat as additional predictors, using the same variables as Hallmann et al., which although not entirely satisfactory due to the scarcity of historical habitat data, represent the best available data. While they trained their model on the Hallmann et al.\'s dataset, Muller et al. validated it with an independent dataset. These upgraded analyses are a nice demonstration of the strong impact of climatic conditions on annual insect biomass. However, Muller et al. conclusion that \"temporal variation in weather conditions explained most of the temporal changes in insect biomass whereas temporal changes in habitat conditions played only a minor role\" was overstated. Here we argue that their methodological approach was unsuitable to draw such conclusion, because of omitted variable bias. We show that more appropriate analyses produce a pattern opposite to the main conclusion of Muller et al.: there is a significant temporal decline in insect biomass not explained by weather conditions.