Comment on Colin D. Wren, Ariane Burke, Habitat suitability and the genetic structure of human populations during the Last Glacial Maximum (LGM) in Western Europe. Research Article | published 19 Jun 2019 PLOS ONE. https://doi.org/10.1371/j...
Figures from this comment are available at https://figshare.com/arti...

I will not discuss the method of Habitat suitability construction.
Still I have some problems related to the data used in that contribution.
1. The time duration considered
The authors write “…human populations living in Western Europe during the Last Glacial Maximum 19–23 kyrs BP).”
“The Last Glacial period (encompassing Marine isotope stages 4–2) is characterised by a global trend towards cooler and drier conditions, culminating with the Last Glacial Maximum (LGM), which correlates with Greenland Stadial 2.1.”
The authors explore a time duration of 19-23 kyrs BP.

There is a problem related to the time duration of GS 2.1 the authors have used. Rasmussen e.a. 2014 consider another time for GS 2.1, being 14700 -22900 kyrs BP, which is quite different from the duration the authors consider.

I do not understand why the authors use another time. Their time stops in the middle of GS-2.1b (Fig. 1), where there seems not to be a special climatic event observed in de GSP2 16O/18O graphic (software CalPal). Why cut it there? The authors don’t discuss their choice.

Fig. 1 – Calibrated radiocarbon dates (used in Colin D. Wren, Ariane Burke 2019.

When I use the Radiocarbon Palaeolithic Europe Database v.26 (RPED v26 ) list of radiocarbon dates (see 2) it appears that an increase of population (as reflected in het number of 14C data) is just starting where the authors cut their list.

Fig. 2 – Calibrated radiocarbon dates (Weninger, Jöris and Danzeglocke, 2013) from RPED v26 using the same time duration as Colin D. Wren, Ariane Burke 2019.

When I make the same graph using all data from GS-2.1 (being the time period 22900 -14700 yrs BP corresponding with uncalibrated 14C dates 19090-12450 BP available in RPED v26, I obtain following graph, quite different from the author’s results. The graph is based on 1905 reliable dates. 273 dates from the list were considered as unreliable.


Fig. 3 – Calibrated radiocarbon dates from RPED v26 for the time duration of the GS-2.1.



The use of a strange time duration by Colin D. Wren, Ariane Burke 2019 is disturbing and certainly is not representative for Greenland Stadial 2.1.

 

2. The sites considered

Another problem is related to the choice the authors made for the sites of Fig. 1. The authors do not specify the database they have used. I presume that they use the sample used by Burke e.a. 2017 (Comparing the map of Fig 1 of the authors with that of Burke, I observe that only 1 site has been added in the map of Fig. 1). There I read that “A sample of Western European archaeological sites was created by cross-referencing three existing databases: PACEA (d'Errico et al., 2011), S2Ages (Pettitt et al., 2003) and Climex (Margottini and Vai, 2004). We extracted archaeological occupations with 14C dates falling between 19 and 23,000 cal years B.P., calibrated using OxCal 4.2, (Ramsey, 2009), from the existing databases.” That list is quite old and does not represent the present state of the art.

When I transpose the time period of 19 and 23,000 cal years B.P into uncalibrated 14C ages, it corresponds (following the list of Burke e.a.) to the time period of 19970 (Lyon-1675 ) - 15173 (BM-1452) uncalibrated BP, comprising 242 dates. That list is very restricted and seems not to comply with the filters (indeed, the Burke list includes 7 dates with a standard deviation larger than 500 yrs and 3 dates are attributed to the Aurignacian, which is not fitting with the filter used).



Fig 4 – Map from Colin D. Wren, Ariane Burke 2019, fig. 1.

A query (for the same time duration as used by the authors) in the RPED 26 results in a list of 884 reliable dates, none of them attributed to the Aurignacian and all with a standard deviation less than 501 yrs. Importing those dates on Google Earth produces a map (Fig. 5), which is quite different from Colin D. Wren, Ariane Burke 2019 map.


Fig. 5 – Map with the RPED v2 dates for the same time duration as Colin D. Wren, Ariane Burke 2019

For the time duration (19090-12450 uncal BP) of Greenland Interstadial 2.1, the RPED v26 query produces 1905 reliable dates and 140 unreliable dates. Importing the reliable dates on a Google Earth produces the map of Fig. 6. That map is quite different from the Greenland Stadial 2.1 map by Colin D. Wren, Ariane Burke 2019,

Fig. 6 – Map with the RPED v2 dates for the Greenland Stadial 2.1 period.

3.→Conclusion
The database the authors are using is not representative for the area and the time duration on which they apply the results of their Habitat Suitability project. As such, the contribution discussing HS during the Greenland Stadial 2.1 is misleading.
4. Reference
Rasmussen SO, Bigler M, Blockley SP, Blunier T, Buchardt SL, Clausen HB, et al. A stratigraphic framework for abrupt climatic changes during the Last Glacial period based on three synchronized Greenland ice-core records: refining and extending the INTIMATE event stratigraphy. Quaternary Science Reviews. 2014; 106: 14–28. https://doi.org/10.1016/j...

Weninger B, Jöris O, Danzeglocke U, 2013. CalPal-2007. Cologne Radiocarbon Calibration & Palaeoclimate Research Package. Available from: http://www.calpal-online....

Competing interests statement
I would like to clarify that the database referenced in the above comment is entirely constructed and continuously updated by myself based on available published data in journals and books and is openly available on the internet for use by researchers.
Pierre M. Vermeersch, Department of Earth and Environmental Sciences, KULeuven, Celestijnenlaan 201, 3001 Leuven, Belgium