Reconstructing past arboreal cover based on modern and fossil pollen data: A statistical approach for the Gredos Range (Central Spain).

Quantifying and reconstructing past vegetation changes is an important step to fully understand human-environment interactions in the past. In this study we present a reconstruction of arboreal cover of six studysites in the Gredos Range (central Spain) over the last 3000 yr based on multivariate statistical analysis (clusteranalysis and non-metric multidimensional scaling, NMDS) of 186 modern pollen samples, modern vegetationdata and six detailed fossil pollen records. The integrated approach allows distinguishing eight modern vegeta-tion communities and linking the fossil pollen records with these vegetation communities as modern analogues.The information of the arboreal cover of the modern pollen sites was used to estimate the past arboreal cover ofthe fossil pollen sites in the Gredos Range. This estimated arboreal cover shows a higher level of landscape open-ness than suggested by the original pollen percentages data. Our results show that the evolution of the arborealcover through time differs along an altitudinal gradient, with a decrease in arboreal cover during the Roman andVisigoth periods (2000–1240 cal BP) at low altitudes and only during the Christian/Feudal period (850–500 calBP) at high altitudes. The applied methodology, integrating fossil pollen data with modern pollen and vegetationdata in one NMDS, allows reconstructing past arboreal cover in a quantitative way.