Four-test intervals with state and year clustering included zero in four US bird samples in 1995-2024

We tested whether uncertainty for a published ozone by seasonal warmth slope changes when errors can be related across states in the same year. We reused four fixed bird-richness model designs from the US East and West in 1995-2009 and 2010-2024. State plus year cluster intervals, adjusted for four tests, included zero in all four samples. The early western interval had narrowly excluded zero with state clustering alone. Two-way standard errors increased in two samples and decreased in two samples. All point estimates stayed fixed. We also refit each model after each year removal. This is an uncertainty audit of shared observations, not an independent biological replication. Few years, concentrated western information, selected exposure coverage, coarse weather and detection bias limit inference. The result does not prove a zero ecological effect. Code, aggregate results, source hashes, a figure and reproduction instructions are included. External survey records and model matrices remain outside the package.

birds · ozone · seasonal warmth · cluster uncertainty · observation bias · regional replication · year influence