LAUSR

Abstract Climate change is likely to alter wildfire regimes, but the significance of climate-driven factors in regional fire regimes over extended temporal scales is poorly understood. Comparison of the reconstructed fire dynamics from charcoals in sediments with modern active fires may provide clues about the drivers of wildfire activities, and help us validate models of fire activity for both the past and the future. Microscopic charcoals from Miocene sediments in Wenshan Basin (Yunnan, southwestern China), were used to reconstruct fire dynamics in deep time. Palaeoclimatic data were obtained from the previous quantitative reconstruction using pollen samples from a sedimentary sequence. The relationship between palaeoclimatic parameters and the ratio of charcoals during the Miocene was explored, and compared with maximum entropy (Maxent) model results throughout different time periods. Our results indicate that the temperature in the dry season was the main factor controlling the frequency of fire in Wenshan during the Miocene. Maxent modeling results based on the modern active fire dataset of Yunnan Province from NASA Earth observations are consistent with the results from fossil sediment analysis, that dry season temperature is the main driver of fire activity. Our findings suggest a significant fire-temperature relationship under the monsoonal climate since the Neogene. Furthermore, models suggest that the last-glacial maximum (LGM) had lower levels of fire activity than present, which increased during the Holocene prior to reaching present levels. Models also predict increases in fire activity across most of Yunnan in the future.