Pollen anthropogenic indicators revisited using large-scale pollen and archaeological datasets: 12,000 years of human-vegetation interactions in central Europe
Abstract
Tracing human-vegetation interactions that occurred in the past has always been one of the key topics of paleoecology. Here we use the pollen and archaeological databases available for the Czech Republic to determine links between individual pollen taxa and archaeological data and search for the spatial scales of comparability.
The datasets include 1,500 pollen samples and 65,000 archaeological components covering the period from 12,000 to 700 cal. BP, divided into time windows of 250 years.
Spearman's rank correlation was used to measure the link between pollen and archaeological data at different sites. Using generalized additive models for the whole dataset, we explained the variance of pollen by archaeologically registered human activities and by two environmental variables.
The first was the overall trend for each taxon in the Holocene representing the long-term dynamics of the species, the second was the elevation of pollen sites. Both factors affect species representation over the whole period studied or/and the area and cannot be statistically separated from human-induced changes.
Both decrease the indicative strength of anthropogenic pollen; however, elevation did so more than the Holocene trend, since past human activities and elevation are strongly correlated and account for the first main gradient. The pollen taxa with a positive correlation with the level of past human activity, indicated by all methods, are: Plantago lanceolata, Artemisia and Amaranthaceae, resprouting edible trees that tolerate fire and pruning (Quercus) and pioneer trees (Pinus).
Probability indicating the presence or absence of archaeological evidence when pollen of these species is present or absent is high (0.56-0.76). However, explained variability by the full model is low (0.01-0.09).
Fagus, Carpinus and Abies expand during the late-successional stages after human disturbance, therefore their relationships to past human activity are negative when con-sidering a 250-year time window. Secale does not correlate at the level of individual sites due to its late appearance during the Holocene.
We ascribe the weak relationship between archaeolog-ical data and pollen of Cerealia to inconsistent determinations. The radius of comparability of pollen and archaeological evidence is around tens of kilometres due to the spatial resolution of archaeology is the area of a parish, but lower for herbaceous plants (15-20 km) than for trees (30-40 km).
This critical comparison delimits overlaps and gaps between widely-used assump-tions and data-based evidence.