1. Conclusions reached in meta-analyses of changes in insect communities may be influenced by method-specific sampling biases, which may lead to inappropriate conservation measures.

2. We argue that the contradictory conclusions regarding terrestrial insect biomass, abundance and richness patterns are, at least partly, due to methodological limitations that reflect taxon-specific responses to environmental changes.

3. In this study, light and Malaise traps were simultaneously deployed to sample insects at 52 plots in a temperate forest in Germany along gradients of elevation (> 1000 m) and canopy openness (3 - 100 %). These gradients were used as predictors in models of total arthropod biomass according to the two trapping methods, and in models of abundance and richness of three commonly targeted groups: nocturnal moths, sampled using light traps, and hoverflies and bees, collected with Malaise traps.

4. A comparison of the total arthropod biomass obtained with the two methods revealed contrary results along the canopy openness gradient. Biomass in light traps showed a decreasing trend with increasing canopy openness while biomass in Malaise traps increased. The same opposing pattern was found for the abundance of selected taxa.

5. The different patterns describing spatial variation of arthropod communities obtained using light and Malaise traps can be explained by differences in the taxa predominantly collected. Regarding the ongoing debate on insect decline, our results demonstrate that comparing different taxa from different taxon-specific traps is inappropriate. Thus, we recommend that future meta-analyses take into account the sampling methods and taxon-specific responses to environmental changes.

Comparison of two insect trapping method with regard to biomass, abundance and species richness.