Shared missing dates ruled out date-sensitive ozone labels in 17 selected US bird survey windows in 2005-2019

We tested date-sensitive ozone labels near selected US bird surveys in four states in 2005-2019. A new exact method gives one state to each missing date within overlapping windows. It ruled out date-sensitive labels in 17 of 1,592 nearest-site records where separate label bounds allowed them. Checking all 15 integer survey date shifts from -7 to +7 days found no added changes in 951 complete primary records. Other thresholds and run lengths did show added changes. A witness proof and exhaustive checks validate the method. Bounds are sharp per record. Their averages are sample outer bounds because records can share dates. Selected routes, repeated monitors, rare episodes and uncertain observed states limit use. This is an exposure support result, not an estimate of bird harm or climate effects.

ozone · bird surveys · missing data · exposure support · observation bias · partial identification · ecological timing