Eurasia has served as refugia and an intersection of important migration routes throughout the Last Glacial Maximum (LGM). Therefore, predicting the LGM climate impacts on Eurasian vegetation is crucial for an accurate representation and understanding of past migration events. Accordingly, the Quaternary vegetation of Eurasia has been studied both through fossil pollen records, and simulations with vegetation models. A mismatch between the reconstructed vegetation from pollen data (less tree cover) and simulated vegetation by models (more tree cover) persists for the LGM. It has been suggested that anthropogenic fire impact can explain this discrepancy (Kaplan et al., 2016). However, the fire regimes determined from the charcoal records do not show a particular increase for this period. Instead, we suggest that the lower atmospheric CO2 values combined with LGM's cold and dry climate to be the main reason why tree cover was reduced. To test these hypotheses, we run a dynamic vegetation model, LPJ-GUESS, through Predictive Ecosystem Analyzer (PEcAn) with various CO2 concentrations and fire recurrence interval scenarios and the climate information provided by PMIP3 snapshot simulations with the MIROC model (Watanabe et al., 2011). For this study, we conducted these simulations for a core site Lake Iznik, (Turkey) where a pollen record of the last ca. 31 ka cal BP is available (Miebach et al., 2016), covering both the mid-Holocene (MH) and the LGM. Lake Iznik is geographically located at the intersection of Europe and Anatolia, its pollen record is well interpreted, and it has been an old human settlement.
Results/Conclusions
Present-day and MH simulations agree well with the present vegetation composition of Iznik region as reflected in core data. All our simulations showed a substantial shift from temperate to boreal vegetation for LGM climate conditions. As expected, we also simulated increasing tree cover with a hypothetical increase in CO2 values and decreasing fire frequencies in our scenarios. Our simulations with reduced atmospheric CO2 values to the 190 ppm (actual LGM values) and low fire frequency, and with 280 ppm and relatively high fire frequency showed cold-type grasses increasing at the expanse of trees during the LGM and the best agreement with the pollen data. However, the high fire frequency scenario is not supported by the proxy data from this region. Therefore, our results indicate that rather than the anthropogenic effect in terms of fire manipulation, low CO2 values explain the reduced tree cover and steppe herb abundance during the LGM.