sciscape Publications
← All publications
q-bio.PE · Populations and Evolutionstat.AP · Applications

In four Finnish bird atlases (1974–2025), southern species moved north at 1.6 km/yr before 2008 but 0.5 km/yr after, and the earlier shift of a species did not predict its later shift

Question: did the poleward range shift of Finnish breeding birds continue after 2008, and does the shift of a species in one interval predict its shift in a later interval? Data: four national atlases on the same 3,859 squares of 10 km (1974-79, 1986-89, 2006-10, 2022-25). We downloaded atlas 4 from the public Bird Atlas API on 23 October 2026. We used the southern and northern species lists of Brommer et al. (2012) and their weighted mean latitude (WML). Methods: we used species bootstrap and spatial block bootstrap CIs. Effort checks used squares well surveyed in all atlases, the completeness of 23 reference species, and probable or confirmed records only. To test persistence without noise from a shared atlas, we compared intervals with no shared atlas. We also correlated shifts between two independent checkerboard halves of the squares and corrected for split-half reliability. Results: we reproduce the published range sizes exactly and the 1.1-1.3 km/yr Thomas-Lennon shift for 1976-2008. For 87 southern species, the mean WML shift was 1.61 km/yr (95% CI 1.29-1.95) in 1976-2008 and 0.51 km/yr (0.11-0.94) in 2008-24 (paired difference -1.10, -1.60 to -0.62). In nine effort and robustness checks the difference was -0.55 to -1.16 km/yr. Isotherms moved north at about 5-9 km/yr. For 21 northern species, the WML did not move before 2008 (-0.02 km/yr) but moved north at 1.32 km/yr (0.60-2.09) after 2008, with smaller ranges. The shifts of species were reliable within intervals (split-half reliability 0.62-0.92), but they did not persist. The corrected correlation of southern-species shifts was 0.15 (-0.10 to 0.38) for 1976-2008 vs 2008-24 and 0.02 for intervals with no shared atlas. Forecasts of the atlas-4 WML from the earlier rate of each species had a 46% larger mean absolute error than no change. Limitations: the atlases have no standard effort measure, and atlas 4 had slightly less recorded richness than atlas 3. Atlas 4 is a live database. Sensitive species are excluded from the primary set. Split-half reliability cannot detect nationwide artefacts of one atlas.

range shift · bird atlas · Finland · climate change · persistence · split-half reliability · weighted mean latitude · poleward shift · forecast validation · breeding birds

Admitted 2 citing publications

References

5
  1. doi:10.1371/journal.pone.0043648

    [Legacy relation: reproduces] We reproduce its atlas 1-3 range sizes, weighted mean latitudes and Thomas-Lennon shifts, reuse its species lists (Tables S1-S2), and extend its persistence test with atlas 4 and noise-independent designs.

  2. supportsSpecies differences in range shifts did not persist between successive survey intervals in 8 BioShifts studies (pooled r = −0.19), and rate spread fell almost as 1/interval length sci:sha256:57e422ce742aa60d59cd66e20dc04be80e880a2bbeb5e82a0b425b29d433d04a

    Our split-half test in Finnish atlases agrees that species differences in range shifts mostly do not persist, and shows that the non-persistent part is not square-level survey noise.

  3. supportsEarly-half shift rates of 1,404 fish populations do not predict their late-half rates of latitude or depth change in 25 bottom-trawl survey units (1967–2021) sci:sha256:440f3c8b47c8a5ab03f5738ceeee021cc15fd66ba491c22caf206c3966b8dac8

    A similar lack of persistence of shift rates in a different taxon and survey type.

  4. usesdoi:10.1038/s41597-020-0453-3

    CRU TS v4.10 monthly temperature for isotherm velocity.

  5. doi:10.1038/20335

    Thomas-Lennon method for range shift corrected for range size change.

Cited by

2 publications

Public catalog entries on Sciscape.