Using 19th century land use maps to study forest vegetation of former rotational slash and burn sites

Historical land use have designed numerous valuable semi-natural habitats in Europe. Among other traditional agricultural practises the slash and burn cultivation (SB) has received little attention, though the fire cultivation lasted up to the beginning of 20th century in Northern Europe.
The aim of current study is analyse the differences in vegetation of former slash and burn cultivation (SB) sites and continuous forest land and the impacts of fire cultivation to the soil and vegetation using historical maps.
In 19th century cadastral maps are designated the rotational SB sites and permanent forest land. The map analyse showed, that most of these sites are covered by forests now (Tomson et al, 2015). The most common forest type in former slash and burn sites are Oxalis type. (Tomson et al, 2016). Fieldworks were carried out in 80 forest stands, where 45 stands were chosen on former SB land and 35 in forest stands identified on 19th century maps. Study areas were strategically selected inside 5 protected areas in Southern Estonia. During the fieldwork ploughing terraces were found also in some places which were mapped as forest in 19th century. Therefore the statistical analyses were performed in three groups: recent SB sites, older SB sites (mapped as forest but cultivation signs present) and permanent forest land. The univariate analyses as permutation analysis of variance and Fisher exact test were applied in the first step. The principal component analysis was used to discover basic patterns in vegetation coverage the multi-response permutation procedure and permutational multivariate ANOVA were used to test the difference in vegetation coverage between land use groups. Between groups principal component analysis was used to discover the patterns in vegetation coverage distinguishing areas with different land use history.

Results of statistical analyses reflect the different landscape characteristics of SB sites and forest land, the first tends to situate in hillocks and nearby borders of forest areas. Analyses did not revealed differences in soil characteristics, species richness and species affinity to human activities. Ancient forest indicator species were present in both sites. The less number of myrmechohores in SB sites reflect the dispersal delay. The results did not show, that the SB had changed the soil characteristics for certain direction permanently, impact of fire to the soil fertility were not identified. The environmental factors affect the vegetation more than previous land use. 3.9-5.2% of vegetation variability could be explained by historical SB. The land use status in 20th century explained more vegetation variability than SB and therefore the forests in former SB sites could be considered as well restored post-agricultural forests