In 2,202 Neotoma pollen records, denser late-Holocene sampling causes half of the European and most of the North American late-Holocene rise in rate of change when smoothing uses five levels

Question: Does the higher density of pollen levels near the top of cores inflate the late-Holocene acceleration of vegetation rate of change (RoC) that Mottl et al. (2021, Science) reported? Approach: We built 2,202 dated fossil pollen records (1,088 Europe, 677 North America) from the Neotoma snapshot of 2026-10-05. We computed RoC with an R-Ratepol-like pipeline (500-yr moving-window bins, one random level per bin, 150 grains, chi-squared dissimilarity). We varied the smoothing and thinned each record to equal level spacing. The test statistic is the late-Holocene excess (LHE): mean RoC at 0-2 ka over mean RoC at 4-8 ka, minus 1. We also linked the published per-record RoC of Mottl et al. to local level density. Results: Level density at 0-0.5 ka is about twice that at 4-8 ka. With 5-level age-weighted smoothing, thinning to >=200 yr spacing lowered the 95%-quantile LHE from 50% to 24% in Europe and from 31% to 5% in North America (paired differences 25 and 27 points, 95% CI 19-32 and 19-32). Without smoothing, thinning did not change the LHE. A fixed 250-yr time-window smoother gave values near the thinned values, so the bias comes from smoothers that use a fixed number of levels. In the published RoC, the 95%-quantile LHE fell from 36% to 12% in Europe and from 23% to 18% in North America when we used only windows with at most 4 levels per 1000 yr. The published RoC also increased with local density (elasticity 0.17 and 0.22). A smaller European late-Holocene rise (median +28 to +46%) is robust in all density-equal pipelines. The North American median shows no excess. With equal density, late-Holocene 95%-quantile RoC is below deglacial RoC in both continents. Limitations: We could not confirm the published smoothing setting or reproduce its Bayesian age ensembles. The published-data test is observational. Thinning cannot equalise the shorter time span per level near core tops.

fossil pollen · rate of change · Neotoma · Holocene · vegetation change · sampling resolution · smoothing artefact · R-Ratepol · palaeoecology · replication