Abstract
Ericaceous dwarf shrubs comprise a key component of the vegetation in several types of northern peatlands. Widespread draining of the peatlands is known to favour forest species (such as Vaccinium myrtillus and Vaccinium vitis-idaea) over mire species (such as Andromeda polifolia and Vaccinium oxycoccos), but it is unclear to what extent such assemblage shifts should form a target for ecological restoration. In this paper, we analyse the performance of eight co-occurring dwarf shrub species in a large-scale ecological restoration experiment in Scots pine-dominated wetlands that had been drained over 40 years ago in Estonia. We address two related questions: (1) how major ecosystem-change-related factors and within-assemblage interactions affect the 1-m(2) scale presence of each species in the drained landscape, and (2) to what extent their cover responses to ditch blocking and partial harvest over 6 years reveal a reversal of the drainage-caused succession. We explored those factors and the treatment effects using general linear mixed modelling of the species' presence and cover. At least four species were responding negatively to drainage, predominantly along with the stand successional stage. However, the results infer that most species were probably enhanced in the early post-drainage phase. The presence of each species was independently enhanced by the presence of other species; the only antagonistic relationship found was between V. myrtillus and V. uliginosum. Ditch blocking had a clear effect only on V. oxycoccos, which increased along with Sphagnum moss cover. In several species, we found a temporary decline in some treatments. Overall, the ericaceous shrub cover appeared rather resistant to the fundamental environmental changes investigated and it may serve as a relatively stable functional component both in natural and anthropogenic peatland transitions. In addition to clone longevity, the facilitative mechanisms suggested by co-occurrence patterns may play a role in this and deserve further study.