Collection-window means lost 4-5 percent of daily thermal exposure above 10 C in two Swedish insect samples from 2019

We tested the thermal exposure lost when a temperature base is applied after averaging over an insect collection window. We used 5836 measured collections in two periods each in Sweden and Madagascar in 2019-2020. At a 10 C base, the equal-trap mean loss relative to daily positive temperature excess was 4.77% and 4.01% in Sweden. Trap bootstrap ranges were 4.21-5.39% and 3.23-4.89%. Both Madagascar losses were zero. At 20 C, Madagascar losses were 4.45% and 7.28%. Loss occurs when daily temperatures cross the base within a window. Swedish loss persisted in 5-10 day collections. Splitting windows into blocks of at most seven days reduced loss but did not remove it. Equal-trap weights, measured-sample filters, common-trap checks and duration checks address parts of observation bias. These results audit a weather calculation. They do not estimate insect development, causal climate effects or pollution effects. One survey year per country, selected sites and daily mean air temperature limit wider use. The package gives code, aggregate evidence, figures, source hashes and reproduction instructions. External records stay outside the bundle.

thermal exposure · degree days · insect biomass · temperature averaging · observation bias · Jensen inequality · Sweden · Madagascar