Interpreting modern and fossil pollen data along a steep environmental gradient in northern Patagonia

Vegetation reconstructions rest on modern vegetation–pollen rain relationships and deductive reasoning. Establishing this relationship is a nontrivial task because differences among pollen assemblages are not necessarily proportional to differences in vegetation. This task is particularly challenging in Patagonia, where some tree taxa have indistinguishable pollen, and pollen grains can be transported long distances. In this study, we describe the modern pollen of 48 lake and wetland samples from northern Patagonia (40.5–44°S) to better discriminate the major vegetation zones of the region through pollen analysis. Specifically, we focus on the performance of three methodological approaches, namely, pollen indicators, classification trees, and optimal thresholds of dissimilarity. As a proof of concept, we use the modern pollen–vegetation relationships to reconstruct the vegetation history at Laguna el Trébol (41.07°S; 71.5°W). Our results revealed that (1) pollen sums exceeding 260 grains ensured replicable vegetation reconstructions, (2) modern vegetation zones could not be separated solely by visual inspection of their pollen spectra, (3) the classification tree and optimal thresholds of dissimilarity permitted discrimination of most vegetation zones, (4) detection of nonanalog communities required use of pollen indicators or optimal thresholds of dissimilarity, and (5) vegetation at L. el Trébol was likely dominated by late glacial shrubland with no modern analogs in the study area (15,000–12,180 cal. yr BP), modern shrubland (12,180–6500 cal. yr BP) and mixed forest (6500 cal. yr BP–present). This study allows a more realistic understanding of the pollen–vegetation relationship and provides new tools for interpreting past vegetation in northern Patagonia.