Mean within-Sweden abundance centroids of 154 bird species did not move north with the isotherms in 1998–2025 fixed-route counts, with route and observer effects fixed

Question: from 1998 to 2025, did breeding bird abundance in Sweden move north as fast as the temperature isotherms? Data: Swedish Bird Survey fixed routes (716 routes on a 25 km grid, 11,871 line-transect surveys, 798 observers; GBIF doi:10.15468/hd6w0r) and GHCN_CAMS gridded temperature. Method: for each of 154 species, a Poisson model with route x observer and year fixed effects estimated the annual change in the latitude slope of log abundance. We converted this change to a centroid shift (km/yr) over the fixed route set. We compared the shifts with isotherm speed, with the shift that each species would show if its latitude profile had moved with the isotherms, and with simpler estimators. A route-cluster bootstrap (200 replicates) gave the intervals. Results: isotherms moved north at 2.6 km/yr (April-June) and 3.7 km/yr (annual). The mean species shift was -0.77 km/yr (95% CI -1.86 to 0.33), and 41% of species moved north. Under perfect tracking, the expected mean was +0.75 km/yr. The observed mean was 1.52 km/yr lower (0.42 to 2.61). The fitted tracking fraction was 0.01 (-0.34 to 0.38). Species differ (15 significant northward, 23 southward shifts), but the rankings reproduced weakly: the split-half Spearman correlation was 0.24 across independent route halves and 0.04 between 1998-2011 and 2012-2025. Raw-count centroids gave +1.30 km/yr because the mean latitude of the surveyed routes increased by 2.0 km/yr. Route effects without observer effects gave +0.38 km/yr. Limitations: the method measures latitude tilt inside Sweden only. It cannot see a uniform change of a log-linear profile, range change outside Sweden, or elevation shifts. The expected shifts assume quadratic profiles and isotherms that move north only.

breeding birds · range shift · climate change · isotherm velocity · climatic debt · Swedish Bird Survey · fixed routes · observer effects · Poisson fixed effects · abundance centroid · reproducibility · GBIF · GHCN_CAMS