A warmth link to insect-sample coverage in later Madagascar did not persist with 7-day or 28-day controls

We tested whether processed insect samples covered warmer days more often within the same trap and season. We used daily weather and measured collection windows at 191 Swedish traps in 2019 and 50 Madagascar traps in 2019-2020. We split each network year into two periods. Coverage within observed collection spans was 94-98 percent in Sweden and 73-74 percent in Madagascar. Only later Madagascar gave a clear primary temperature slope with trap, latitude-season and 14-day calendar controls: 27.01 percentage points per 5 C, with a four-test interval of 0.18 to 53.83. The slope was -1.89 with 7-day controls and -0.20 with 28-day controls; both intervals included zero. Thus, the positive result depends on the calendar model. The other three primary intervals included zero. This is an observation-process diagnostic, not evidence of insect abundance change. One year per network, observed rather than known deployment spans, few time clusters and coarse weather limit inference. We test no pollution or long-term climate effect. The package gives code, source hashes, aggregate results, figures and an independent numerical check.

insect biomass · observation bias · sample coverage · temperature · seasonality · Madagascar · Sweden