North American breeding-bird community temperature index did not follow 1966–2024 May–June warming at BBS routes on continental, route or yearly scales; residents followed more than migrants

Question: Did the community temperature index (CTI) of North American breeding-bird communities follow the warming at the survey routes? We test the continental trend, the trend at each route, and the response to yearly temperature anomalies. Data: North American Breeding Bird Survey 2025 release (1966–2024; 127,473 standard runs, 5,290 roadside routes, 339 landbird species) and NOAA GHCN_CAMS 0.5° monthly temperature. Method: The species temperature index is the count-weighted mean 1981–2010 May–June temperature of the routes where a species occurs. The CTI is the abundance-weighted mean index of a run. Year effects come from route × observer and year fixed effects. Full tracking is the spatial CTI–temperature slope (0.495) times the temperature trend. Intervals come from a route-cluster bootstrap. Results: May–June temperature increased by 0.260 ± 0.031 °C per decade. Full tracking predicts +0.129 °C per decade of CTI. The observed CTI trend was +0.003 °C per decade: tracking ratio 0.025 (95% CI −0.018 to 0.070). Other CTI definitions gave −0.12 to 0.17. The CTI decreased until about 1996 and then increased; for 1990–2024 the ratio was 0.36. At 2,682 routes (1980–2024), the local warming rate did not predict the local CTI trend (slope 0.02 ± 0.05; full tracking 0.495). Yearly CTI increased by 0.0070 ± 0.0015 °C per °C of anomaly (1.4% of full tracking); a next-year placebo gave −0.0018 ± 0.0014. Survey-morning temperature had an effect of the same size per standard deviation. Year-round residents (eBird class, 81 species) had a tracking ratio of 0.26 (0.15 to 0.34), and migratory species (222) had −0.06 (−0.12 to 0.00). Migratory species caused 90% of the 1966–1996 CTI decrease. Exclusion of open-habitat species did not increase tracking. The 2024 release (1966–2023) gave the same results. Limitations: Observational roadside counts with no detection model. The index uses BBS ranges, which are truncated at the survey limits. Only May–June temperature was tested, so the effect of winter temperature on residents is not known. Guild classes come from external species-level data.

community temperature index · climatic debt · Breeding Bird Survey · North America · climate change · species temperature index · thermophilization · bird communities · migratory birds · fixed effects · GHCN_CAMS · detection