Walrus and polar bear population size changes in the N. Atlantic over the last 20k years

Spread the love

From polarbearscience

This is a lesson in how to assess the potential worth of scientific papers. One of two similar Arctic evolution studies got media attention, at least in Canada — about the polar bears, of course — but in my opinion the walrus research conclusions are much better supported, less biased by climate change rhetoric, and lack the hubris present in the polar bear paper.

Both studies use similar sample sizes for the regions they had in common (North Atlantic) and used computer models to determine genetic diversity and population size changes since the LGM. However, the tone of the walrus paper was less emotionally-charged and the caveats of the work were appropriately stated. In my opinion, papers like the polar bear example contribute to eroding the public’s trust in science.

The last Ice Age peaked between about 27,000 and 19,000 years ago. At this time the Arctic was buried under kilometers of glacial ice sheets, and so marine mammals were pushed southwards to areas of ice floes and more open water. Walrus survived in some areas of the Atlantic located further to the south, and as soon as climates warmed again, the ice edge retreated and walrus populations pushed quickly northwards again. This combination of warming and climate-driven dispersal led to local walrus populations becoming more genetically differentiated. Walrus study, Lund University press release 27 September 2023

The walrus paper

By Emily Ruiz-Puerta and 17 other authors, called “Holocence deglaciation drove rapid genetic diversification of Atlantic walrus.”

  • In the introduction, cites the IPCC and several other papers to support a statement that the Arctic is warming faster than other regions and is predicted to lead to future ecological changes without hyperbole
  • Shows sea ice changes over the last 30k years, with illustrative figures and references based on a variety of proxy datasets
  • Uses both historic-era and fossil samples up to about 8.5k years old to discern past genetic profiles
  • State outright that historic reductions in population numbers by humans may have impacted their interpretation of past population size changes due to climate change (e.g. Crockford 2023a)
  • State outright that their proposed “Last Glacial Maximum (LGM) refugia” in NW Greenland around Kane Basin may not be well supported (i.e., that it may not have been ice-free and available habitat 20k years ago)
  • Conclude that walrus populations have been divided into several fairly isolated populations over the last 8-10k years, which may make them vulnerable to future localized threats; while this seems to be a fairly well-supported conclusion, the authors admit further research is required

Determining the full effect of historic and recent exploitation and climatic warm periods on walrus genetic diversity may require larger sample sizes and the addition of nuclear genome data from both ancient, historic and contemporary populations” [Ruiz-Puerta et al. 2023: 8]

The polar bear paper

By Michael Westbury and 19 other authors, called “Impact of Holocene environmental change on the evolutionary ecology of an Arctic top predator.” [they mean polar bears: why don’t they just say so?]

  • In the introduction, cites several articles promoting an unverified ‘tipping point’ concept to support statements that a warming Arctic is headed for “irreversible ecosystem change” and that the polar bear is a “sentinel species for detecting ecosystem tipping points” (Crockford 2022a)
  • Uses only modern and late historic-era samples (1936-2019) to characterize genetic data over 11k years
  • Uses only sea surface temperature data to infer past sea ice extent over 20k years of polar bear evolutionary history; no other proxy data cited and no illustration is provided (cf. Crockford 2023b; Ruiz-Puerta et al. 2023)
  • Uses inferences about “preferred habitat” of polar bears (e.g. citing Amstrup et al 2008, and Durner et al. 2009, which were used to support the original listing of polar bears on the US Endangered Species List) that have since been exposed as a failed concept (Crockford 2017, 2019)
  • Makes dubious inference that sea ice extent at the LGM (ca. 20k years ago) represented optimal polar bear habitat (i.e., that ‘more is better),’ apparently failing to account for the fact that virtually all current polar bear habitat was covered in perennial sea ice too thick to support mammalian life (cf. Crockford 2022b, 2023b; Ruiz-Puerta et al. 2023)
  • Cites only Miller et al. 2012 and Cahill et al. 2013 as previous genetic studies on polar bear evolution, leaving out other recent papers that are potentially relevant (summarized in Crockford 2023b)
  • There is no mention of the well-documented but seldom mentioned slaughter of polar bears in the late 1800s-early 1900s and the follow-up overhunting from 1945-1970 that led to the species requiring international protection (1973): therefore, there is no discussion of what impact these events would have had on long term genetic signals of increasing/decreasing population sizes the authors are trying to interpret, especially given they used no prehistoric-era samples (Crockford 2019, 2023b; Honderich 1991; cf. Ruiz-Puerta et al. 2023)
  • Conclude that ancient polar bears populations around Greenland declined in step with sea ice declines over the last 11k years (despite the above caveats) and that this strengthens predictions future sea ice declines pose a serious threat to long-term polar bear survival. However, I am unconvinced their evidence warrants such a confident conclusion.

Although our paleoclimate data and habitat suitability reconstruction extend across the Holocene only, the initial rapid decline in Ne [i.e., effective population size] observed in west Greenland bears of ~19 ka may signal the end of the Last Glacial Maximum in the region, a period of massive sea ice loss and increasing temperatures. An association between Ne and changes in environment is further evidenced when comparing our genomic demographic reconstructions with paleoclimate and suitable habitat over the past 11,000 years (Fig. 4). Overall, we observe a clear pattern of Ne decline when suitable habitat decreased, which was associated with periods of warmer SST and reduced sea ice cover.” [Westbury et al. 2023: 5]

Model-generated population size decline from ~20k years ago that seems to assume, without evidence, that LGM sea ice habitats in the North Atlantic were optimal for polar bears (blue is East Greenland, orange is West Greenland). Westbury et al. 2023, Fig. 1D.


Crockford, S.J. 2019The Polar Bear Catastrophe That Never Happened. Global Warming Policy Foundation, London. Available in paperback and ebook formats.

Crockford, S.J. 2022a. Sir David Attenborough and the Walrus DeceptionAmazon Digital Services, Victoria.Crockford, S.J. 2022b. Polar bear fossil and archaeological records from the Pleistocene and Holocene in relation to sea ice extent and open water polynyas. Open Quaternary 8(1):7.

Crockford, S.J. 2023a. The Polar Wildlife Report. Global Warming Policy Foundation Briefing 63, London. pdf here.

Crockford, S.J. 2023b. Polar Bear Evolution: A Model for How New Species Arise. Amazon Digital Services, Victoria.  Available in hardcover, paperback and ebook formats https://www.amazon.com/dp/1778038328 and https://www.amazon.co.uk/dp/1778038328 [On Sale! 8 November to 6 December 2023 only]

Honderich, J.E. 1991. Wildlife as a hazardous resource: an analysis of the historical interaction of humans and polar bears in the Canadian arctic. MA thesis, University of Waterloo, Ontario.

Ruiz-Puerta, E.J., Keighley, X., Desjardins, S.P.A., and 15 other authors. 2023. Holocence deglaciation drove rapid genetic diversification of Atlantic walrus. Proceedings of the Royal Society B 290: 20231349. https://doi.org/10.1098/rspb.2023.1349

Westbury, M.V., Brown, S.C., Lorenzen, J. and 17 other authors. 2023. Impact of Holocene environmental change on the evolutionary ecology of an Arctic top predator. Science Advances 9(45): eadf3326. https://www.science.org/doi/10.1126/sciadv.adf3326