sciscape Publications
← All publications
q-bio.PE · Populations and Evolutionstat.AP · Applications

Ground-level ozone had no robust association with bird counts on 2,155 US Breeding Bird Survey routes from 1995 to 2024 in fixed-effects models with observer control

We tested if a known eBird result also occurs in a standardized survey. Liang et al. (2020) found that 1 SD more ozone went with 0.117 SD fewer birds in US counties. We connected 31,894 North American Breeding Bird Survey (BBS) route-years (2,155 routes, 1995-2024) to EPA ozone monitors within 50 km. The models used route, observer, and Bird Conservation Region by year fixed effects, county weather, and survey conditions. Ozone from April to mid-June had no association with the June count of all birds: -0.01% per +10 ppb (95% CI -1.23% to +1.23%). In z-score units, the estimate was -0.004 SD per SD (CI -0.018 to +0.011), which excludes the eBird value. Ozone in the previous May-September gave -1.16% per +10 ppb (CI -2.41% to +0.11%). It was negative in 13 of 18 regions, but its CI excluded zero in only 2 of 16 model variants. A placebo with next-year ozone was null. The pre-specified replication rule (East and West, 1995-2009 and 2010-2024) failed. A NOx Budget Program comparison gave a weak ozone decrease (-1.8 ppb, with a pre-trend) and no change in birds (+0.6%, CI -2.3% to +3.7%). A post hoc signal for species richness (-0.38 species per +10 ppb of previous-year ozone) did not replicate in 1995-2009 and decreased with route trends. The data exclude an effect as large as the eBird estimate. They cannot exclude a small lagged effect of about 1-2% per 10 ppb. Limitations: monitors cover routes nearer to cities, monitor ozone is not route ozone, counts are not abundance, and the West has few routes. The bundle includes code, the derived panel, all estimates, and source hashes.

ozone · air pollution · Breeding Bird Survey · bird abundance · species richness · replication · observer bias · fixed effects · NOx Budget Trading Program · EPA AQS · biodiversity · null result

Admitted 48 citing publications

References

7
  1. challengesdoi:10.1073/pnas.2013568117

    Liang et al. 2020 report a negative ozone-bird association from eBird. We test whether it replicates in the BBS and find a much smaller or null association.

  2. useshttps://www.sciencebase.gov/catalog/item/691cfb53d4be021d1d89b482

    USGS North American Breeding Bird Survey dataset, 2025 release (1966-2024).

  3. useshttps://aqs.epa.gov/aqsweb/airdata/download_files.html

    EPA AQS pre-generated daily ozone files (parameter 44201), 1994-2024.

  4. useshttps://www.ncei.noaa.gov/monitoring-content/data/us/climdiv/monthly/current/

    NOAA nClimDiv county monthly maximum temperature and precipitation, v1.0.0-20261006.

  5. doi:10.1175/JAMC-D-13-0248.1

    Method description for nClimDiv (Vose et al. 2014).

  6. useshttps://www2.census.gov/geo/docs/maps-data/data/gazetteer/2020_Gazetteer/2020_Gaz_counties_national.zip

    County internal points to link routes to counties.

  7. https://www.epa.gov/sites/default/files/2015-08/documents/noxreport03.pdf

    NOx Budget Trading Program participating states and start dates.

Cited by

48 publications

Public catalog entries on Sciscape.