Equal year weights reduce the estimated May–October advance in mean beetle catch date from 5.1 to 0.8 days at seven UK ECN sites, 1994–2003 to 2004–2013

We tested seasonal timing in ground-beetle catches at seven UK Environmental Change Network sites. We used 21 fixed transects and 347 complete May–October transect-years in 1994–2003 and 2004–2013. We corrected monthly catches for trap effort and split the change in mean date into changes within species and changes in species shares. The planned estimate gave years weights set by their catch indices. It showed a 5.09-day advance: 4.70 days within species and 0.40 day from species shares. An equal-year check, added after the planned analysis, reduced the total advance to 0.79 day. Its within-species part was −3.12 days and its species-share part was +2.33 days. The 4.30-day difference between the total estimates comes exactly from the covariance of yearly catch size and yearly mean date. High-catch years had later dates in the early period; this relation was weaker in the later period. The weighting effect remained under stricter effort coverage and a common-species filter. Intervals that resample both sites and year blocks include zero for both total changes and their difference. The supported result is an effect of year weights on the estimate in this fixed panel. The sites are not a random UK sample. Pitfall catches measure activity and capture, not breeding dates. The package includes source data, code, provenance, figures and reproduction checks.

ground beetles · carabids · phenology · Environmental Change Network · species composition · temporal weighting · pitfall traps · reproducible analysis