Frozen bird thermal indices had spatial slopes of 0.33-0.44 C per climate C in four East and West samples in 1970-2024

We tested whether a frozen bird community temperature index had a spatial slope of one degree per degree of route climate. Four East and West samples in 1970-1996 and 1997-2024 used 120,034 surveys on 4,703 routes from the BBS 2025 release. Equal-route slopes were 0.408, 0.438, 0.331 and 0.373 index C per climate C. All four corrected upper limits were below 0.48. Survey-condition adjustments changed little. Within ecological regions, slopes were lower, at 0.201-0.310. We prove a matched-weight variance identity and give an exact example in which all species track their climate niches but the frozen index changes by only 0.20 C per 1 C of warming. Thus a raw index-change / temperature-change ratio alone does not measure the fraction of climate tracking. This challenges the direct percentage interpretation of a prior campaign ratio. It does not change the prior temporal trend estimates or identify a causal response. Shared species values, spatial habitat differences, observation bias, coarse weather data and selected routes limit inference. The package includes verified inputs, code, proof, all tables, figures, source hashes and a successful retained-input rebuild.

community temperature index · spatial calibration · climate tracking · Breeding Bird Survey · observation bias · thermal niche · variance identity · reproducible methods