A simple climate-driven semi-mechanistic vegetation model can explain the pan-Asian extent of the glacial mammoth steppe

Abstract

Graminoid and forb tundra vegetation has been hypothesized to have covered a vast area in northern Eurasia during the Late Pleistocene cold phases, providing habitat for many now-extinct megaherbivores. The formation and persistence of this now nearly-lost vegetation have been either attributed to the cold and arid climate of the glacials (climate hypothesis) or the action of its large herbivores (keystone herbivore hypothesis). Here, we test whether the extent of this plant community can be modelled mechanistically using a few bioclimatic variables with clear physiological effects on plants. We fitted the bioclimatic limits based on the distribution of the closest modern analogues of the graminoid-forb-tundra. Next, we applied this parameterization to conditions of the Last Glacial Maximum, given by the gridded climate-model-based CHELSA-TraCE21k paleoclimate dataset. The projected glacial distribution of the graminoid and forb tundra mainly expands in central Siberia. The model failed to reproduce fossil-inferred occurrences of graminoid and forb tundra in glacial Europe and northern Eurasia, however, possibly because of inaccuracies in CHELSA-TraCE21k, with temperatures being too high in Europe and too low in northern Siberia. This initial study suggests that climate might have played a substantial role in forming the glacial graminoid and forb tundra, but analyses with more climate models will be necessary to corroborate this finding. Our parsimonious, transparent, and process-based model holds the potential for testing the climate hypothesis more rigorously with additional and improved palaeoclimate data.