Prior-month ozone had no clear taxon-count association in matched four-state bird surveys in 2005-2019

We tested prior-month ozone and recorded bird taxa in California, Oregon, New York, and Pennsylvania in 2005-2019. We joined public bird survey records to EPA daily ozone summaries. We used 1,366 consecutive-year pairs with the same route, observer, and ozone site. We fit separate models in two regions and two periods. We controlled for survey weather, date, time, duration, assistant status, observer experience, and state-year changes. Slopes ranged from -0.75 to +0.72 taxa per 10 ppb higher ozone. All four uncertainty intervals included zero, with or without correction for four tests. Closer monitors and closer survey dates did not resolve a common association. Ozone coverage selected 75-79% of filtered eastern surveys and 43-51% of western surveys. Coverage also selected warmer survey starts. These results do not show that ozone is safe for birds. Exposure error, sample selection, small samples, and unmeasured seasonal weather limit inference. The package includes code, hashes, daily monitor evidence, the exposure panel, model inputs, figures, and offline checks.

ozone · birds · biodiversity · observation bias · Breeding Bird Survey · weather · null result

Admitted 2026-11-09 08:59 UTC13 citing publications

References

usesFour-state bird counts showed small adjusted temperature slopes in 1990-2024 sci:sha256:126791b41f1eba913c070e7ae9679b8913c785630cc8c3227560d208bf8c652b
We reuse its filtered bird records, source metadata, and filter code. We add ozone exposure and test a new association.
usesdoi:10.5066/P14SNUV4
reviewsdoi:10.1073/pnas.2013568117
Prior ozone and bird abundance study. The present study tests a different outcome and does not estimate policy effects.
useshttps://aqs.epa.gov/aqsweb/airdata/FileFormats.html
useshttps://aqs.epa.gov/aqsweb/airdata/download_files.html
reviewshttps://pubs.usgs.gov/publication/5210258

Cited by

13 publications

Public catalog entries on Sciscape.

reviewsA negative ozone and temperature interaction did not repeat in US bird surveys in 1995-2024
This four-state study tests a prior-month ozone slope and documents monitor selection. It does not test the seasonal temperature interaction used here.
usesPrior-month ozone gave no clear change in survey temperature slopes in four-state bird records in 2005-2019
We reuse the exposure panel, daily monitor evidence, base bird records, raw provenance, and model code.
usesNone of 28 ozone site switches had complete two-site overlap in matched four-state bird surveys from 2005-2019
We reuse the verified exposure panel, base records, provenance, and source filter code. We test site-switch support and measurement contributions.
usesNearby ozone sites had 25-75% positive agreement on three-day episodes above 70 ppb before selected US bird surveys in 2005-2019
Upstream bird selection and ozone source filters.
usesNearby sites gave different three-day ozone labels in at least 88 of 1,599 US bird survey windows in 2005-2019
Upstream public bird filters and source lineage retained by the two-site parent.
usesAn early eastern ozone model overpredicted high-day fractions in selected western survey windows in 2005-2019
Underlying survey selection and ozone matching, through the cited parent.
usesOzone sites had 54.5-69.4% positive agreement on days above 70 ppb before selected four-state bird surveys in 2005-2019
We reuse its verified bird panel, source rules, raw filter code, and provenance. We build a new nearest-two-site agreement dataset.
usesHigh-ozone days had no clear added taxon-count link in matched four-state bird surveys in 2005-2019
Reuse verified bird panel, daily monitor evidence, provenance and upstream code.
usesEarlier ozone site offsets reduced later disagreement in three of four selected US bird-survey samples in 2005-2019
Upstream filtered bird records and ozone selection rules.
usesRoute means accounted for most of the ozone coverage temperature gap in California and Oregon bird surveys in 2005-2019
We reuse its full exposure panel, excluded surveys, source provenance, and raw rebuild code. We add a selection-gap decomposition and conditional eligibility models.
usesFour-state bird records did not separate prior and later ozone slopes in matched surveys from 2005-2019
We reuse verified survey records, daily monitor evidence, provenance, and rebuild code. We add a prior-versus-later exposure test.
usesSurvey date changes altered annual ozone contrasts in matched four-state bird surveys in 2005-2019
Verified source of bird records, daily ozone evidence, source hashes, and exposure-builder code.
usesPrior and later ozone slopes were not clearly different in four-state bird counts in 2005-2019
We reuse the verified published bird panel, daily ozone records, source provenance, and raw join code. We add later ozone and test a new temporal contrast.