Postglacial species arrival and diversity buildup of northern ecosystems took millennia

Abstract

What drives ecosystem buildup, diversity, and stability? We assess species arrival and ecosystem changes across 16 millennia by combining regional-scale plant sedimentary ancient DNA from Fennoscandia with near-complete DNA and trait databases. We show that postglacial arrival time varies within and between plant growth forms. Further, arrival times were mainly predicted by adaptation to temperature, disturbance, and light. Major break points in ecological trait diversity were seen between 13.9 and 10.8 calibrated thousand years before the present (cal ka BP), as well as break point in functional diversity at 12.0 cal ka BP, shifting from a state of ecosystem buildup to a state where most habitat types and biotic ecosystem components were in place. Trait and functional diversity stabilized around 8 cal ka BP, after which both remained stable, although changes in climate took place and species inflow continued. Our ecosystem reconstruction indicates a millennial-scale time phase of formation to reach stable and resilient levels of diversity and functioning.

Fig. 1.. Appearance of 315 taxa in northern Fennoscandia in relation to deglaciation and climate.

(A) The 10 study sites (black dots) in relation to the reconstructed ice sheet (white) and mean daily 2-m air temperature for July (other colors) (106). Dark gray lines indicate reconstructed surface elevation at 1000-m steps above sea level (a.s.l.) starting at 0 m a.s.l. (106) with modification for the Baltic region (107). For additional time slices, see fig. S1. (B) Accumulated richness of all growth forms aggregated into 500-year time slices. The gray lines represent numbers of species appearing in nine samples per time slice (based on 100 resampling of all lakes within 500-year time slices; lines are not visible when they overlap with the full data), and dotted lines represent estimated first arrival date (taking into account variation in detectability; see Materials and Methods), whereas solid lines are the observed first appearance. (C) Box plot of observed first appearances (Ob, filled, vertical line = median) and estimated first arrival (Es, unfilled, vertical line = median) for each growth form. The estimated first arrival takes into account uncertainties in detection of taxa. (D) The relative abundance of each growth form is estimated as the proportion of all PCR repeats for each 500-year time slice across all lakes.

Fig. 2.. Impact of trait variables on the estimated arrival time.

(A) The relative importance of different trait variables in explaining the estimated arrival time of 227 vascular plant taxa to northern Fennoscandia based on BRT analysis. (B to D) Relationship between the trait value and establishment time predicted by the BRT analysis for the three variables with the highest relative importance: temperature optimum (B), soil disturbance (C), and light requirement (D).

Fig. 3.. Northern ecosystem changes through 16 millennia based on abundance of plant taxa with different trait values.

(A to I) Dashed vertical lines indicate the break point in diversity of each trait (see fig. S8). The break points in cal ka BP are given in brackets. Note that the break point in pollinator dependence is negative, indicating low diversity. The plots include both terrestrial and aquatic taxa except for moisture, which is for terrestrial taxa only. Taxa are quantified as detection in proportion of PCR repeats. Gray areas represent undefined traits for some taxa. For full description of traits and categories, see the Supplementary Materials. AM, arbuscular mycorrhiza; EM, ectomycorrhiza; ErM, ericoid mycorrhiza.

Fig. 4.. Breakpoints estimated by segmented regression on Shannon diversity of traits based on the proportions of PCR repeats.

Estimated break points follow facet names, and are indicated by dashed line. The blue lines represent fitted values of break point analysis, the solid circles represents the Shannon diversity, and pink shadings represent approximate 95% confidence intervals (Estimates ± 2 SE) of the break points.

Fig. 5.. New appearance of taxa in relation to climate, niche saturation, and ecosystem diversity.

(A) Mean July temperature (°C) for the 10 sites (lines) and observed number of new taxa per 500 years according to growth form (bars). (B) Accumulated trait values for each of the nine traits. (C) Shannon diversity of trait values. (D) Functional diversity measured as functional dispersion (pink fill for 95% confidence intervals of break point).