Vegetation cover of Brazil in the last 21 ka: New insights into the Amazonian refugia and Pleistocenic arc hypotheses

Aim: The two main hypotheses about the Neotropical palaeovegetation, namely that of Amazonian refugia by Haffer and of the Pleistocene arc by Prado and Gibbs, are still constantly debated. We offer new insights on this debate using ecological niche modelling with combined climate-soil predictors to test both hypotheses, reconstruct the palaeovegetation of the Last Glacial Maximum (LGM; 21 ka) and Mid-Holocene (Mid-H; 6 ka) and indicate the configuration of refugia areas. Location: Brazil. Time period: Last 21 ka. Major taxa studied: Biomes. Methods: We modelled the environmental space of the 10 most representative biomes with the RandomForest classifier, using climate predictors from three atmospheric general circulation models (CCSM4, MPI-ESM-P and MIROC-ESM) and soil predictors, the same for the different situations. Based on the consensus among the models, we reconstructed the palaeovegetation cover for LGM and Mid-H and used fossil pollen sites to validate the reconstructions in a direct comparison. Results: The climate in the past was cooler and wetter throughout most of the territory. The Amazon basin region was the most affected by climate change in the last 21 ka, with equatorial rain forest retracting to refugia areas, while the tropical rain forest (with climatic preferences similar to the Atlantic forest) expanded in the basin. In southern Brazil, the mixed forest (Araucaria forest) shifted to lower latitudes, while the grasslands expanded. In most biomes, the greatest changes occurred in the ecotonal zones, supported by pollen fossils. Main conclusions: With regard to Haffer's hypothesis, the forests of the Amazonian lowlands retreated to refugia areas, while the colder and wetter climate of the basin created a favourable niche for another type of forest, instead of savanna. The advance of dry vegetation was restricted to ecotonal conditions, preventing the formation of a continuous Pleistocene arc, predicted by Prado and Gibbs's hypothesis.