Abstract

The development of quantitative temperature reconstructions in regions of paleoclimate interest is an important step for providing reliable temperature estimates in that region. Fossil chironomid assemblages have been studied in Patagonia showing great promise for reconstructing paleotemperatures; however there is still a lack of robust temperature inference models in that area. To contribute to the understanding of climate change, a transfer function using chironomids preserved in 46 lakes in Chile and Argentina was developed. The best performing model to infer the mean air temperature of the warmest month was a 3-component WA-PLS model with a coefficient of correlation (r(jack)(2)) of 0.56, a root mean square error of prediction (RMSEP) of 1.69 degrees C and a maximum bias of 2.07 degrees C. This model was applied to the chironomids preserved in the sediment of the Huelmo mire (41 degrees 31'S, 73 degrees 00'W), in the lake district of northwestern Patagonia. The reconstruction showed several cold spells (one at 13,200 to 13,000 cal yr BP and a cooling trend between 12,600 and 11,500 cal yr BP) associated with the Younger Dryas and/or Huelmo-Mascardi Cold Reversal (HMCR). Our findings support climate models proposing fast acting inter-hemispheric coupling mechanisms including the recently proposed bipolar atmospheric and/or bipolar ocean teleconnections rather than a bipolar see-saw model. (C) 2014 Elsevier B.V. All rights reserved.