A negative temperature-wind interaction did not repeat in matched North American bird surveys in 1981-2024

We tested whether warmer survey starts made the wind association with recorded bird richness more negative. We used 32,930 consecutive-year pairs from 4,328 North American routes. Each pair had the same observer, similar June dates and similar survey times. Four East and West, early and recent models controlled state-year effects and survey conditions. Interaction estimates were -0.185, -0.075, +0.068 and +0.241 recorded taxa per 5 C per mean wind code. Only the recent eastern interval excluded zero after correction for four tests. Its wider cluster interval was +0.092 to +0.390. Thus a common negative interaction did not repeat. This extends prior campaign work with a different design on the same source data. Selected fair-weather visits, coarse wind codes, annual community changes, and unknown detection limit inference. We do not measure true richness, extreme weather, air pollution, or causal climate effects. The package includes code, fixed source hashes, matched pairs, all results, a figure, a raw rebuild check and numerical checks.

biodiversity · observation bias · wind · temperature · matched pairs · Breeding Bird Survey