Technical Comments

Comment on “Glacial Survival of Boreal Trees in Northern Scandinavia”

Science  09 Nov 2012:
Vol. 338, Issue 6108, pp. 742
DOI: 10.1126/science.1225345


Parducci et al. (Reports, 2 March 2012, p. 1083) fail to present convincing evidence for glacial survival of Pinus and Picea in northern Scandinavia. Their methodology does not exclude contamination. Additionally, they should consider the lack of suitable habitats, the apparent extinction of both taxa after deglacial warming, and alternative hypotheses for the distribution of the Picea genetic marker haplotype A.

Parducci et al. (1) propose that Picea (spruce) and Pinus (pine) survived the Last Glacial Maximum (LGM) in northern Scandinavia, based on the amplification of ancient chloroplast DNA (cpDNA) in lake sediments from Endletvatn on the Andøya strandflat, northwest Norway (69°N) and genetic evidence from extant spruce populations. We comment first on their methodology and then on their interpretations.

(i) How reliable is ancient DNA (aDNA) from lake sediments? Although knowledge of aDNA, including sedimentary aDNA (sedaDNA), and environmental DNA is advancing rapidly, Hofreiter et al. (2) warn that the reliability of all aDNA sequences must be critically assessed, major concerns being contamination and the variable quality of aDNA from waterlogged sites. They urge “a critical attitude towards whether the results obtained do make sense.” This new technique was apparently not validated by Parducci et al., who ignored the criteria and advice of Gilbert et al. (3) to minimize contamination and replicate results.

The Geonor stationary-piston sediment corer, like most piston corers, is not a contamination-free corer. A contamination-free corer (4) should have been used.

The core studied by Parducci et al. was opened in 2002, sampled, and stored at 4°C for 7 years until sampled for DNA. How certain is it that no contamination occurred during potentially unsterile sampling in 2002 from aerial pollen, dust, wood, or paper? The reported contamination by an Urtica plant fragment is worrying.

(ii) Was glacial survival on Andøya possible? The LGM ice sheet advanced over Andøya, including Endletvatn, reaching a 2- to 300-m altitude. Mountain glaciers left only the highest summits and plateaus as ice-free nunataks (5). These are mostly covered by block fields (5), which are not suitable habitats today for Picea or Pinus. As ice retreated at ~22 thousand years ago (ka), the extensive lowlands, including Endletvatn, were below sea level until a marine regression ~18 ka (5). Glacial and marine conditions prevailed at Endletvatn at the times of the cpDNA records.

(iii) Where were Pinus and Picea growing? They could not survive on exposed nunataks in the inferred polar-desert conditions (1, 6). Picea, but not Pinus, may persist vegetatively for centuries in sheltered locations. To deliver DNA to the lake, they should have grown nearby, but the strandflat was glaciated or submerged (driftwood could provide aDNA) at the critical periods. Therefore, where did the cpDNA come from? Willerslev et al. (7) proposed long-distance dispersal of conifer DNA to Greenland ice. Likewise, extremely low Pinus values in Andøya pollen diagrams could derive from long-distance dispersal. The rather surprising absence of conifer pollen in their samples led Parducci et al. to exclude this source.

(iv) Does haplotype A (HapA) reflect Picea spreading directions? Early Holocene Picea mitochondrial DNA (mtDNA) from lake sediment and younger pollen in Trøndelag shows HapA (1), probably a single 21–base pair deletion. Western glacial survival is inferred from its modern distribution. Old cpDNA from Andøya can provide no support for this hypothesis. The distribution of other mtDNA and nuclear DNA markers (8)—combined with palaeoecological evidence of small scattered, perhaps palynologically silent, early Picea populations in Scandinavia (9), including a possible Late Glacial (but not LGM) occurrence on a Swedish nunatak (10)—suggest several immigrations along both northern and southern routes after deglaciation. This scenario would explain the observed genetic diversities from Russia to Scandinavia (8). Furthermore, LGM Picea populations south of the ice with HapA could have spread north during deglaciation, meanwhile becoming extinct in their southern refugia. Populations only expanded after ~3000 years before the present (BP) (9). HapA could have arisen in a pioneer southern population, increased in small populations, and spread toward the west and north, mixing with lineages sharing HapB. Multiple immigrations from western and southern directions would tend to block the spread of the northern ones (11). Even if HapA survived on Andøya, it would be unlikely to spread southward. Although Parducci et al. argue that a distinct genotype cluster in central Scandinavian populations (8) indicates a western origin, they should also consider the several different genotype clusters in Scandinavia that may be explained by leading-edge and surfing phenomena. Interestingly, these clusters map poorly with HapA, but the spatial resolution is coarse. To clarify colonization processes, more sampling and genetic analyses are needed to map the distribution of other individual mutations and markers. Parducci et al. demonstrate that Picea’s history in Scandinavia was probably more complex than previously thought, but they cannot claim that the lineage of the widespread HapA survived in northern Scandinavia and spread south and east.

(v) Why are Pinus and Picea absent on Andøya after 17,700 years BP? Late-Pleniglacial refugial tree populations with low palynological presence typically respond to deglacial warming with increases in pollen percentages. For example, low Pleniglacial Picea pollen percentages at Lake Galich, in northern Russia, increased to more than 20% during the Late Glacial interstadial (12). No Late Glacial expansion of Picea and Pinus is recorded on Andøya (6), and neither grew on Andøya during the early Holocene (13). Picea’s present northwestern limit is ~500 km to the south (14). Andøya’s cool oceanic climate today does not support Picea, and Pinus is rare (14). It must have been much more inhospitable during the glacial period, with a climate resembling that of Svalbard today (6).

We conclude that Parducci et al. have not justified in sufficient detail “how their data were obtained and why they should be believed to be authentic” (3). Their methodology is questionable, due to the unsuitable corer, the age and storage conditions of the sediment core, and the potentially unsterile sampling. The taphonomy of DNA in lake sediment needs rigorous investigation. One cpDNA record of Picea and two of Pinus are insufficient to postulate glacial survival. No account was taken of the lack of suitable habitats on Andøya during the glacial period. Other parts of northern Scandinavia were ice-covered. The absence of Pinus and Picea during the deglacial and Holocene is contrary to all ecological expectations if small LGM populations survived near Endletvatn. The distribution of Picea mtDNA haplotypes can be more readily explained by postglacial colonization than by a northern LGM refuge. We consider the evidence presented by Parducci et al. to be unconvincing, so we reject their hypothesis that Picea and Pinus survived the last glacial period in northern Scandinavia.


View Abstract


Navigate This Article