North American breeding-bird community temperature index fell in 1970–1996 and rose in 1997–2024 in all 30 tested BBS regions, at 25% of full tracking, but not more where June warming was larger

Question: the community temperature index (CTI) of North American Breeding Bird Survey (BBS) communities has a near-zero 1966–2024 trend. Is this because two opposite periods cancel, does the change of direction occur in each region, and which species cause it? Data: BBS 2025 release (127,646 standard runs, 5,306 routes, 10,229 observers, 534 species, 1966–2024); June temperature from GHCN_CAMS, checked with GISTEMP, CRU TS 4.09, Berkeley Earth and May–July values; habitat classes from the PIF ACAD. Method: count-weighted species temperature index from route June climate; log-weighted CTI; annual indices with route × observer and year fixed effects and a first-year-observer term; period trends, a one-break fit, and a 2° spatial-block bootstrap. Full tracking is the spatial CTI–temperature slope (0.52) times the warming trend. An exact additive decomposition gives species and habitat contributions. Results: the CTI decreased by 0.034 °C per decade in 1970–1996 (95% CI −0.044 to −0.024) and increased by 0.062 in 1997–2024 (0.054 to 0.069). The break year was 2000.5 (1998.3 to 2002.0). June warming increased from 0.17 to 0.48 °C per decade at the same time. The 1997–2024 tracking ratio was 0.25 (0.21 to 0.29; 0.20 to 0.32 in 17 variants), and the 1970–2024 ratio was 0.04 (−0.005 to 0.10). The 1997–2024 increase occurred in all 5 regions, both countries, and all 30 Bird Conservation Regions with at least 40 routes (lower limit above zero in 25). The CTI trend did not increase with warming across 98 cells of 4° (Spearman 0.02; slope −0.03 vs 0.52 expected) or across regions. Forest species gave 77% of the 1997–2024 increase; Eurasian Collared-Dove gave 13%. The 1997–2024 increase was stable when species were resampled (0.014 to 0.116; STI-shuffle p = 0.056), but the 1970–1996 decrease was not (−0.092 to 0.019). The increase also occurred within observers in their first 8 years. Limitations: roadside counts without detection models; observer drift is only partly controlled; a large same-observer CTI step from 2019 to 2021 (+0.049 °C) is not explained by temperature, but the trend to 2019 was still 0.054; STI is truncated at the BBS area; the break model is descriptive and does not identify a cause.

community temperature index · climatic debt · North American Breeding Bird Survey · thermophilization · climate change · bird communities · observer effects · breakpoint · species decomposition · Bird Conservation Regions