Negative residual changes had no clear excess pair overlap over positive changes in nine British butterfly indices, 1981-2024

We tested whether large negative annual residual changes occurred together across butterfly species more often than large positive residual changes. Nine species had complete country indices and at least five sites per year in England, Scotland and Wales in 1980-2024. Six country and period tests used annual changes in 1981-2002 and 2003-2024. Models removed time trends and changes in contributing site count. Each species supplied five lower-tail and five upper-tail years per period. Lower-minus-upper pair overlap ranged from -0.0177 to +0.0114. All six 95 percent year-block bootstrap intervals included zero. Signs differed between periods in England. Same-year summer weather adjustment did not establish a common explanation. This result does not show equal ecological risk. Short samples, index error, site changes and coarse weather limit inference. The package includes frozen inputs, exact hashes, code, all tests and a figure.

butterflies · weather · tail dependence · synchrony · observation bias · null result · regional comparison