Get reliable science

Questions
(a) What are directions of spontaneous succession; in particular, do target stages (identified as shrubby grassland and semi-natural deciduous woodland) develop, and if so, which species are involved? (b) Are the target stages predictable? (c) How do species richness and environmental characteristics change during succession? (d) What are the consequences for restoration and landscape management?

Location
The Bohemian Karst Protected Landscape Area, SW of Prague, Czech Republic (49° 52?–50° 00? N, 14° 03?–14° 21? E, 251–488 m a.s.l.).

Methods
In a repeated analysis, phytosociological relevés recorded in 4 m x 4 m plots in 58 old-fields initially surveyed in 1975 were compared to those from 28 still existing fields in 2008–2009. Average Ellenberg indicator values were calculated for each relevé. Aspect and slope were measured and potential radiation calculated. pH was measured from soil samples. Species were grouped according to their affiliation to the phytosociological classes Querco-Fagetea, Festuco-Brometea, Trifolio-Geranietea, Molinio-Arrhenatheretea, and weedy and ruderal vegetation. Those belonging to the first three classes were considered target species. The data were analysed using multivariate (ordination methods) and univariate statistics.

Results
The spontaneous succession in old-fields proceeded towards target stages, either deciduous woodlands or shrubby grasslands. Their establishment can be tentatively predicted by soil pH and early occurrence of grassland species. Except pH, all Ellenberg indicator values changed during succession. pH values (Ellenberg and measured) were higher in shrubby grasslands than in woodlands. The total number of species decreased during succession, the number of target woodland species increased, and target grassland species remained the same in the shrubby grassland stages but decreased in the woodland stages during the past 33 yr of succession.

Conclusions
Target shrubby grasslands, resembling natural steppe-like communities typical of the region and valuable from a conservation point of view, can be restored by spontaneous succession within a few decades in about one-third of the studied old-fields. Other fields developed into deciduous woodland. Restoration of well-developed target woodland will take longer, but the trend is already obvious, although less desirable nitrophilous woodland might also alternatively develop. Repeating earlier chronosequence studies may provide valuable information useful in restoration ecology and landscape management.