Changes in available ozone sites altered annual exposure estimates at US bird routes in 1995-2024

We tested whether changes in available ozone sites altered annual exposure contrasts at repeated bird surveys. We used 28,955 consecutive-year survey pairs with the same route and observer in the lower 48 US states in 1995-2024. Exposure was previous May-September ozone. We compared inverse-distance means from all eligible sites within 50 km with means from sites available in both years. Site sets changed in 12.1-23.5% of pairs across East and West and two periods. Root mean square differences in annual change were 0.45-1.02 ppb. The common-site estimate had support in 99.55% of pairs. Differences were at least 2 ppb in 617 supported pairs. Sixty-nine pairs changed sign when both annual changes had magnitude at least 1 ppb. Stricter coverage and shorter distance did not remove the discrepancy. These are exposure checks, not effects on birds. Shared weather, site placement, missing days, repeated routes, and selected survey coverage limit inference. Neither measure is true route exposure. The package includes verified parent inputs, code, site and pair evidence, source hashes, uncertainty and independent arithmetic checks.

ozone · birds · exposure measurement · monitor availability · observation bias · Breeding Bird Survey · regional comparison