Bumblebee southern limits moved north from 1901–1974 to 2011–2025 in effort-matched GBIF records, but cool limits did not follow warming in Europe and North America

Question: Kerr et al. (2015) reported that bumblebee southern limits moved north from 1901–1974 to 1999–2010, but northern limits did not. Did this pattern continue in 2011–2025, after their data ended? Approach: We used 2,466,999 cleaned species-level Bombus records from the GBIF snapshot of 1 October 2026 (Canada, USA, Europe). Records in 2011–2025 are 9 to 24 times more numerous than in 1901–1974, and their locations are different. We compared the Kerr-style metric (mean latitude of the 5 most extreme records) with equal record numbers and with cell-matched rarefaction. In the cell-matched method, we used only 100-km cells with records in both periods, and we thinned each cell to equal Bombus record numbers. A 300-km injected shift was the positive control. Odd against even years in 2011–2025 was the null. Results: With cell matching, southern limits moved north by 112 km [95% CI 38, 195] in Europe (44 taxa) and 328 km [223, 455] in North America (42 taxa). Outer northern limits did not move north by a detectable amount (+11 km [-65, 76]; -72 km [-192, 21]). The coolest occupied cells warmed by as much as local warming (tracking 0.05 [-0.11, 0.21] and -0.12 [-0.53, 0.22]). European range bodies moved north by about 100 km (90th percentile, centroid and 10th percentile). The null gave values near zero. With all records, the naive metric showed a false northern expansion of 189 km in Europe, because of changes in effort. The southern shift was northward in the main analysis and in all seven sensitivity analyses. Limitations: In North America, most northern limits are at the edge of the sampled cells, so the positive control shows that the data cannot test them. With specimens only, the North American southern shift decreased to 83 km [15, 152] and the European northern limit moved 74 km [10, 137]. Thus, records from photographs change the magnitudes. The effort measure is Bombus records, so changes in the composition of reports can bias the results. We used no phylogenetic model and no elevation analysis.

bumblebees · Bombus · range shift · range limits · climate change · GBIF · sampling effort · rarefaction · community science · out-of-sample replication · thermal tracking · Europe · North America