Keeping failed butterfly indices gave no clear summer temperature slope change in four England and Wales samples, 1985-2024

We tested whether keeping recorded butterfly presences with failed abundance indices changes a summer temperature slope. We used 24,978 matched site-year comparisons and 560,203 matched species-year comparisons in England and Wales in 1985-2024. A source value of -2 means that a species was recorded but visits were insufficient for an abundance index. Four country and period tests compared annual changes in a presence tally with changes in a positive-index tally. The paired slope differences ranged from -0.22 to +0.29 taxa per site per C. All four intervals adjusted for four tests include zero. All 24 planned sensitivity intervals also include zero. Later matched site sets had means of 4.88 and 6.31 flagged presences in England and Wales. Thus the rules differ in recorded levels, but the tests do not establish a repeated temperature-linked slope difference. Source-row selection, changing sites, detection, coarse weather and short annual series limit inference. The study does not measure extinction or a causal climate effect. The package gives code, source hashes, derived evidence, uncertainty checks and a figure.

butterflies · biodiversity · observation bias · recorded richness · summer temperature · UKBMS · missing indices · null result