Search

Pollen data represent one of the most widely available and spatially resolved sources of information about the past land cover and climate of the Last Glacial Maximum (LGM; 21 000 yr BP). Previous pollen data compilations for Europe, the Mediterranean, and the Middle East, however, have been limited by small numbers of sites and poor dating control. Here we present a new compilation of pollen data from the region that improves on both the number of sites (63) and the quality of the chronological control. Data were sourced from both public data archives and published (digitized) diagrams. The analysis is presented based on a standardized pollen taxonomy and sum, with maps shown for the major pollen taxa and biomes and the total arboreal pollen (AP), and on quantitative reconstructions of forest cover and of winter, summer, and annual temperatures and precipitation. The reconstructions are based on the modern analogue technique (MAT) adapted using plant functional type (PFT) scores and with a modern pollen dataset taken from the latest Eurasian Modern Pollen Database (EMPD) (∼8000 samples). A site-by-site comparison of the MAT and the inverse modelling method shows little or no significant difference between the methods for the LGM, indicating that the presence of low-CO 2 conditions and no modern analogue during the LGM does not appear to have had a major effect on MAT transfer function performance. Previous pollen-based climate reconstructions using modern pollen datasets show a much colder and drier climate for the LGM than both inverse modelling and climate model simulations do, but our new results suggest much greater agreement. Differences between our latest MAT reconstruction and those in earlier studies can largely be attributed to bias in the small modern dataset previously used and to differences in the method itself (Brewer et al., 2008; Salonen et al., 2019). We also find that quantitative forest cover reconstructions show more forest than previously suggested by biome reconstructions but less forest than suggested by simply the percentage of arboreal pollen, although uncertainties remain large. Overall, we find that LGM climatic cooling and drying were significantly greater in winter than in summer but with large site-to-site variance that emphasizes the importance of topography and other local factors in controlling the climate and vegetation of the LGM. [ABSTRACT FROM AUTHOR]

Pollen data represents one of the most widely available and spatially-resolved sources of information about the past land cover and climate of the Last Glacial Maximum (21,000 years BP). Previous pollen data compilations for Europe, the Mediterranean and the Middle East however have been limited by small numbers of sites and poor dating control. Here we present a new compilation of pollen data from the region that improves on both the number of sites (63) and the quality of the chronological control. Data has been sourced from both public data archives and published (digitized) diagrams. Analysis is presented based on a standardized pollen taxonomy and sum, with maps shown for the major pollen taxa, biomes and total arboreal pollen, as well as quantitative reconstructions of forest cover and winter, summer and annual temperatures and precipitation. The reconstructions are based on the modern analogue technique (MAT) with a modern calibration dataset taken from the latest Eurasian Modern Pollen Database (~8000 samples). A site-by-site comparison of MAT and Inverse Modelling methods shows little or no significant difference between the methods for the LGM, indicating that no-modern analogue and low CO2 conditions during the LGM do not appear to have had a major effect on MAT transfer function performance. Previous pollen-based climate reconstructions based on MAT show a much colder and drier climate for the LGM than both Inverse Modelling and climate model simulations, but our new results suggest much greater agreement. Differences between our latest MAT reconstruction and those in earlier studies can be largely attributed to bias in the small modern calibration dataset previously used. We also find that quantitative forest cover reconstructions show more forest than that previously suggested by biome reconstructions, but less forest than that suggested by simple percentage arboreal pollen, although uncertainties remain large. Overall, we find that LGM climatic cooling/drying was significantly greater in winter than in summer, but with large site to site variance that emphasizes the importance of topography and other local factors in controlling the climate and vegetation of the LGM. [ABSTRACT FROM AUTHOR]