Origin and temporal development of macro‐scale vegetation patterns in the cultural landscape of Denmark

1 Since the introduction of agriculture, 4000 bc, the Danish region has witnessed a change from almost complete woodland cover to the landscapes present at ad 1800 that were dominated by heathland, cultivated areas, grassland or patchy woodland. We used a combination of pollen records and landscape data to identify the timing and major pathways of this land‐cover transformation. 2 Pollen assemblages dated to about ad 1800 were collected from 20 Danish lakes with a median size of 26 ha. Fossil pollen records for the interval 4500 bc − ad 1800 were compiled from 11 of these 20 sites and divided into 11 time units corresponding to major archaeological periods. 3 Land cover around the 20 sites was digitized from historical maps, dating from the same period as the pollen assemblages. Percentage cover of woodland, heathland, wetland and agricultural land within 2‐ and 5‐km radii were calculated (calculations were also made within 10‐ and 20‐km radii for woodland). Environmental data within 5 km of each site were collected from soil, geological and topographical maps. 4 The ad 1800 pollen assemblages formed three clusters characterized by either a dominance (> 20%) of Calluna (group A), a dominance (> 10%) of Fagus (group B) and a dominance of Poaceae with Fagus < 10% and Calluna < 20% (group C). Respectively, these clusters correspond well to sites where heathland is the dominant land cover, where it has a strong woodland element and where agricultural land is dominant with little, if any, woodland or heathland. 5 Direct and indirect ordinations suggest that the main variation in the pollen assemblages from the 20 sites is related to gradients in soil clay content and in the degree of landscape openness. Land‐cover data from within a 2‐km radius account for almost as much of the variation in pollen data as do land‐cover data from within a 5‐km radius. 6 Indirect ordinations of ad 1800 samples with the fossil records as passive samples reveal that long‐term records from cluster B sites (the ‘1800 woodland type’) show little landscape variation through time. Samples from cluster C sites (the ‘1800 agricultural area type’) show the same starting point as cluster B but their development diverge to more open landscapes. Cluster A sites (the ‘1800 heathland type’) either develop directly from rather wooded landscapes to heathland or by heathland expansion after initial development towards landscapes producing cluster C type pollen assemblages. 7 Chord distances between the ad 1800 sample and the (pre)historic samples at 10 long‐term sites strongly suggest that the between‐site variation in land‐cover characteristic of the early 19th century was already in existence by 1000 bc. Direct ordination of the fossil samples using woodland cover at ad 1800 as an environmental variable and soil clay content as a covariable also indicates that the macro‐scale land‐cover patterns of the historical cultural landscape in Denmark has been present for more than 3000 years.