Bird-count adjustment did not establish a repeated ozone and richness link in US surveys in 1995-2024

We tested whether total bird count explains the link between previous-season ozone and recorded species richness. We used a published panel of US Breeding Bird Survey, EPA ozone, and NOAA weather data. Paired models used the same 25,608 surveys on 1,993 routes, with route-observer and state-year effects and survey and weather controls. A cubic bird-count adjustment left the pooled slope near its initial value: -0.367 species per +10 ppb ozone, with a 95% interval of -0.689 to -0.045. The paired change was +0.008, with an interval of -0.187 to +0.204. Four separate region and time samples gave no clear conditional link after correction for four tests. The pooled interval included zero with route trends. Thus, count adjustment did not explain the pooled slope, but a repeated negative link was not established. This is a descriptive test. Count can be a mediator or reflect detection. Selected routes, missing controls, coarse exposures, and few western state clusters limit inference. The package has the exact input, source recipes and hashes, code, all model results, a numerical check, and a figure.

ozone · recorded richness · bird count · observation bias · regional replication · null result