Cloudy survey ends were linked to higher relative bird richness in US records from 2005-2024

We tested whether cloudy or overcast end sky was linked to a lower last-minus-first ten-stop bird taxon count. We used 58,792 accepted Breeding Bird Survey visits in Canada and the US in 1985-2024. Four country-period models controlled for route-observer and state-year effects, sky, wind, temperature and survey timing. All four point estimates were positive. Only the US 2005-2024 estimate excluded zero after correction for four tests: +0.565 recorded taxa, with an interval of +0.235 to +0.895. Route trends and a low-wind limit gave similar late US estimates. The other three adjusted intervals included zero. The late US direction also held for an offset log richness ratio. Full-survey richness gave no clear main-model association. These results do not establish a common effect across countries and periods. They can reflect bird activity, detection or local community differences. Endpoint weather, selected roadside visits, fixed route order and approximate cluster intervals limit inference. The study does not test true richness, pollution, weather extremes or climate-change damage. The package includes a derived panel, raw source hashes and retrieval code, all estimates, a figure, numerical checks and a successful raw-data rebuild check.

bird richness · sky cover · observation bias · Breeding Bird Survey · weather · regional replication