Lagging adaptation to warming climate in Arabidopsis thaliana

Abstract

If climate change outpaces the rate of adaptive evolution within a site, populations previously well adapted to local conditions may decline or disappear, and banked seeds from those populations will be unsuitable for restoring them. However, if such adaptational lag has occurred, immigrants from historically warmer climates will outperform natives and may provide genetic potential for evolutionary rescue. We tested for lagging adaptation to warming climate using banked seeds of the annual weed Arabidopsis thaliana in common garden experiments in four sites across the species' native European range: Valencia, Spain; Norwich, United Kingdom; Halle, Germany; and Oulu, Finland. Genotypes originating from geographic regions near the planting site had high relative fitness in each site, direct evidence for broad-scale geographic adaptation in this model species. However, genotypes originating in sites historically warmer than the planting site had higher average relative fitness than local genotypes in every site, especially at the northern range limit in Finland. This result suggests that local adaptive optima have shifted rapidly with recent warming across the species' native range. Climatic optima also differed among seasonal germination cohorts within the Norwich site, suggesting that populations occurring where summer germination is common may have greater evolutionary potential to persist under future warming. If adaptational lag has occurred over just a few decades in banked seeds of an annual species, it may be an important consideration for managing longer-lived species, as well as for attempts to conserve threatened populations through ex situ preservation.

Keywords: adaptation lag; cliimate adaptation; local adaptation; provenance test.

Fig. 1.

Map of common garden sites and sites of origin of the 241 native A. thaliana accessions represented in our experiments.

Fig. 2.

Local adaptation by region in accessions of A. thaliana. In autumn cohorts at three of our four field sites—(A) Norwich, United Kingdom; (B) Halle, Germany; and (C) Valencia, Spain—accessions from the same region as the garden site had the highest fitness. In D, Oulu, Finland, accessions from Germany had higher fitness than more local Nordic accessions. Box-and-whisker plots of relative fitness for ecotypes from each region within each autumn planting are displayed. For each experimental planting, Kruskal–Wallis tests demonstrated that accessions from these four regions differed in relative fitness (α = 0.05).

Fig. 3.

Lagging adaptation to April temperature in a large sample of accessions from throughout the native range of A. thaliana. In autumn cohorts in (A) Valencia, Spain; (B) Oulu, Finland; (C) Norwich, United Kingdom; and (D) Halle, Germany, historic April temperatures were cooler than experienced April temperatures, which were closer to the predicted optimum. Selection also favored accessions from historically warmer climates in (E) summer and (F) spring cohorts planted into the Norwich, UK, common garden site. Lines of linear (solid) or quadratic (dashed) best fit, significant at P ≤ 0.006 are shown (Table S4). Warmer-colored symbols denote accessions from more southerly locations, with each color and shape combination characterizing a single country of origin (legend details in Fig. 1). Vertical gray lines show the recent historic April temperature, whereas vertical black lines show the April temperature during the experimental planting at each site.