Surface pollen-vegetation relationships in the forest-steppe, taiga and tundra landscapes of the Russian Altai Mountains

Detailed knowledge of the relationships between modern pollen deposition and actual vegetation on both regional and local scales is valuable for correct interpretation of fossil pollen samples. We collected surface pollen samples and recorded all vascular plant species and their cover in 55 plots of 100 m ², located along five transects in the steppe, forest and tundra of the valleys of the Russian Altai Mountains. We compared proportions of plant species in actual vegetation and their pollen types in surface pollen spectra along each transect. Of the total of 144 pollen taxa, 54 were represented sufficiently in both pollen and vegetation to allow correlation analysis, and 47 showed a significant positive correlation (p < 0.05) between their proportion in the pollen and the proportion of corresponding plant species in the vegetation. These included trees and shrubs (Betula nana type, Larix, Lonicera, Picea obovata, Salix and Spiraea), many herbs or dwarf shrubs, and strong (e.g. Artemisia, Chenopodiaceae) as well as weak pollen producers (e.g. Aconitum, Astragalus/ Oxytropis type, Bupleurum, Dryas oxyodonta, Hedysarum/ Onobrychis type, Polemonium, Swertia obtusa, and Trollius). Pollen of Larix was a good indicator of Larix forests if they occurred in an open steppe or forest-steppe landscape, but became a poor indicator in more forested landscapes where pollen of other trees confounded the Larix pollen signal. Artemisia pollen, although generally considered a poor indicator that is overrepresented in pollen spectra, appeared to be a fairly good indicator on a larger spatial scale, discriminating between large steppe areas on the one hand and forest-steppe and forest-tundra landscapes on the other. We conclude that considering the whole pollen spectra, including weak pollen producers, is important for pollen-based reconstruction of past vegetation diversity. Our results also suggest that pollen spectra can change remarkably over short distances, especially in landscapes with a high diversity of vegetation types, which should be taken into account when we interpret vegetation diversity on the basis of fossil pollen spectra.