Bird abundance gradients between nearby US survey routes kept 69% of the regional temperature gradient per degree with 5 km route-path climate; year-to-year tracking stayed at about 3% (1966-2025)

Question: prior campaign work measured weak bird climate tracking against continental spatial gradients from 0.5-degree temperature at route start points. Is that denominator a real thermal gradient, the same per degree between nearby routes as across regions? Data: 107,900 standard North American Breeding Bird Survey runs on 3,949 contiguous-US routes (1966-2025), 314 species, nClimGrid 1/24-degree May-July climate sampled along route paths, and GHCN_CAMS for comparison. Method: route climate T was split into a regional part (75 km Gaussian smooth) and a local part T_loc. For each species, y = s(T) + gamma*T_loc on route mean log counts; kappa = 1 + gamma/f' is the local slope per degree relative to the regional slope (1 = scale equivalence). Joint 2-degree block bootstrap, positive control and shuffle placebo. We also fitted community temperature index (CTI) slopes and within-species year-to-year ratios lambda with both climate products. Results: the CTI slope was 0.43 C per C across the continent and 0.37 (95% CI 0.30-0.43) for local contrasts; within 1-degree cells it was 0.40 with nClimGrid but 0.23 with GHCN_CAMS, so coarse climate halves within-region slopes. Pooled species kappa was 0.69 (0.58-0.76), 0.50 (0.39-0.58) with a precipitation control; the positive control gave 1.02 and the placebo 0.14 (expected leakage about 0.18). Exploratory: cold-adapted (low-STI) species kept their gradient locally (kappa 0.85, 0.74-0.91), but warm-adapted species did not (0.43, 0.25-0.54). Year-to-year lambda was 0.029-0.033 for both climate products and 0.040 for low-STI species. The CTI rose 0.025 C per decade within observers against 0.237 C per decade warming, 0.24-0.28 of every 5 km spatial slope before the hearing-drift correction of prior work. Conclusion: the spatial denominator is mostly thermal; it explains at most a factor of 1.4-2 of a deficit of about 4-30, so the deficit is mainly in the temporal response. Limitations: correlational; local contrasts follow elevation, which also changes habitat and moisture; interpolated climate; no detection model; the STI split was not planned.