Ice-age endurance: DNA evidence of a white spruce refugium in Alaska

Abstract

Paleorecords offer key information for evaluating model simulations of species migration in response to forecast climatic change. However, their utility can be greatly compromised by the existence of glacial refugia that are undetectable in fossil records (cryptic refugia). Despite several decades of investigation, it remains controversial whether Beringia, the largely unglaciated area extending from northeastern Siberia to the Yukon Territory, harbored small populations of certain boreal tree species during the last glaciation. Here, we present genetic evidence for the existence of a glacial refuge in Alaska that helps to resolve this long-standing controversy. We sequenced chloroplast DNA (cpDNA) of white spruce (Picea glauca), a dominant boreal tree species, in 24 forest stands across northwestern North America. The majority of cpDNA haplotypes are unique, and haplotype diversity is relatively high in Alaska, arguing against the possibility that this species migrated into the region from areas south of the Laurentide Ice Sheet after the end of the last glaciation. Thus, white spruce apparently survived long glacial episodes under climatic extremes in a heterogeneous landscape matrix. These results suggest that estimated rates of tree migration from fossil records may be too high and that the ability of trees to track anthropogenic warming may be more limited than previously thought.

Fig. 1.

Sampling locations and haplotype diversity of white spruce. (a) Locations of Alaskan (red) and non-Alaskan (blue) populations sampled for this study (all 12 non-Alaskan populations are outside the boundary of Beringia). The modern white spruce range is indicated in green. (b) Means and 95% confidence intervals of white spruce cpDNA haplotype diversity in relation to the number of sequenced individuals based on rarefaction resampling (using the statistical program EstimateS). (c) Means and 95% confidence intervals of white spruce cpDNA haplotype diversity in relation to the number of sequenced populations based on rarefaction resampling.

Fig. 2.

Haplotype network assembled by using the method of Templeton et al. (13). Each haplotype is represented by a number–letter combination; red haplotypes occur only in Alaskan populations, blue haplotypes occur only in non-Alaskan populations, and white haplotypes occur in both regions. Circle sizes are proportional to haplotype frequencies in all 24 populations, with the largest circle representing the most abundant haplotype.

Fig. 3.

Geographic patterns of white spruce cpDNA haplotypes in northwestern North America. (a) Modern white spruce range (green), haplotype diversity (circles), and locations of unique haplotypes (labeled). Intrapopulation haplotype diversity was calculated by using the computer program Rarefac to account for differences in population sample sizes. Orange and black circles indicate diversity values above and below the mean (1.97) of all 24 populations, respectively. Circle sizes are proportional to deviations from the mean, with the largest orange circle representing the highest diversity, the largest black circle representing the lowest diversity, and the smallest circles having values closest to the mean. Haplotypes unique to each region are indicated by the number–letter combinations. (b) Frequency patterns of haplotypes Ia (white circles) and Ih (blue circles) suggest bidirectional mixing of white spruce cpDNA. Circle sizes are proportional to intrapopulation haplotype frequencies, with the largest circle representing the highest frequency. Light-blue shading indicates the extent of the Laurentide and Cordilleran Ice Sheets during the LGM (17).