Moving toward quantitative landscape reconstruction: The influence of landscape context on relative pollen productivity and relevant source area, a study from Patagonia

Quantitative pollen-vegetation models offer a route to reconstruct past land cover, but the absence of regionally derived estimates of pollen fall speeds and relative pollen productivity (RPP) has constrained their application in Patagonia. Here, we provide the first estimates of pollen fall speeds and relative pollen productivity (RPP) for major taxa in west-central Patagonia, using modern pollen assemblages collected from moss polsters and vegetation data expressed as distance-weighted plant abundance (DWPA). Our study was conducted in two forest-steppe transition landscapes: Monumento Natural Dos Lagunas and Parque Nacional Patagonia-Jeinimeni to test the consistency of parameters across geographically close but environmentally distinct ecotonal settings. Fall speeds were estimated for 17 pollen types, with Poaceae showing notably higher values than those commonly used in Northern Hemisphere studies. Extended R-value (ERV) analyses revealed a clear difference in the spatial scale of pollen representation: the dataset from MN2L has a relevant source area of pollen (RSAP) between 150 and 450 m depending on sub-model and reference taxon, whereas JEI has a more sharply defined RSAP of 750–850 m regardless of reference choice. The productivity ranking of the two taxa that dominate these assemblages, Nothofagus and Poaceae, is reversed between the two landscapes irrespective of sub-model, reference taxon, and the removal of individual samples in a leave-one-out analysis; estimates for the less abundant taxa proved sensitive to sample removal. Our results indicate that, within the Patagonian forest-steppe ecotone, both RSAP and relative productivity structure are strongly context-dependent, implying that quantitative reconstructions should be calibrated to landscape configuration rather than assuming uniform parameter transferability across sites.