End wind codes gave no clear change in bird richness temperature slopes in Canada and the US in 1985-2024

We tested whether end wind codes changed the link between survey end temperature and recorded bird richness in the last ten stops relative to the first ten. We reused a verified panel of 64,682 accepted Breeding Bird Survey runs. Four Canada and US models covered 1985-2004 and 2005-2024. Models controlled for route-observer and state-year effects, start temperature, wind, sky, timing and assistant status. The high-minus-low wind slope estimates ranged from -0.185 to +0.072 taxa per 5 C. All four family-adjusted intervals included zero. High wind means end Beaufort codes at least 3; low wind means codes 0-2. The tests do not show a clear shared change in temperature slopes. They do not prove that the slopes are equal. Selection, weather measurement, habitat change and detection can affect the estimates. Routes overlap between periods. We did not test extreme heat, climate trends or air pollution. The package gives the input panel, source hashes, code, all estimates, a figure and numerical checks. The original raw files were not rebuilt in this study.

biodiversity · birds · weather · temperature · wind · observation bias · interaction · null result · Breeding Bird Survey