Study Sites and Data
In Estonia, located in North-Eastern Europe, mires cover 5.5% (245,000 ha) of territory, peat is actively extracted on more than 20,000 ha (Paal & Leibak 2011), and the total area of extracted abandoned peatlands is 9,371 ha (3.8% of the area of raised bogs). In the coming decades, the area of abandoned peatfields will double because of the depletion of excavation areas in use. The annual precipitation in Estonia varies from 600 to 700 mm/yr−1. Temperature ranges from 17°C in July to −6°C in February.
The study consists of an inventory of almost all abandoned extracted peatlands in Estonia (64 peatlands, Fig. 1) carried out by the Geological Survey of Estonia by the request of Ministry of the Environment. Studied peatlands were abandoned 5–50 years before the survey, mostly 20–30 years ago, that is in the period when there was no statutory obligation to restore the areas after the cessation of peat excavation. Numerous extracted peatlands were sparsely re-vegetated with a mean projective cover of 10–20% (Orru 2010). The small-area peatlands were inventoried as one study unit. In the larger-area peatlands, where different peatfields were abandoned at different time or had a contrasting outlook because of treatments, peatfields were inventoried as separate study areas. We did not include the data of those peatlands, which were without established plant species and/or areas where peat excavation history was unknown. In total, we had 114 peatfields as survey data.
Figure 1. Location of studied abandoned extracted peatlands in Estonia (centroid of study areas 58°41′42′′, 25°25′59′′, radius approximately 150 km).
Download figure to PowerPoint
In all inventory areas the presence or absence list of the plant species was compiled separately in three types of microtopographic habitats: flats (central parts of peatfields, maximum width of the peatfield was 16 m), ditch margins (0–2 m from the ditch), and ditches. In analysis of microhabitats, we used only data, where at least two species were recorded (112 flats and ditch margins, and 49 ditches). Approximately in the center of each inventory area the water table, the depth of slightly (decomposition degree <25%) and well-decomposed peat (decomposition degree >25%; according to the Records of Mineral Resources of the Environmental Register), and the total depth of residual peat layer were measured. The water level of ditches was obtained from neighboring flats. From the uppermost 50 cm residual peat layer the peat sample core was taken for the laboratory analyses. In laboratory the following parameters were estimated: (1) peat botanical composition (described with microscope), (2) ash content (measured with the loss of weight in burning at 450°C), (3) pH (measured in the pHKCl), and (4) degree of decomposition (estimated by centrifugal method). According to the botanical composition, that is by the fragments of mosses and vascular plants and peat decomposition degree, the trophicity level of upper peat layer (0–0.5 m) was ascertained as: (1) oligotrophic, (2) mesotrophic, or (3) eutrophic, which refers to the gradient from the rain-fed type to groundwater-fed type peatland.
Environmental factors such as (1) time since extracted peatland abandonment, (2) the total area of abandoned extracted peatfields in the bog, (3) visual signs of burning, and (4) special management operations were considered in the analysis. Special management operations used to promote the re-establishment of bog vegetation on some peatfields were cutting of young birch trees, sowing of seeds of some species (Oxycoccus palustris, Vaccinium angustifolium, Betula sp., Pinus sylvestris, Picea abies), or fertilization with P2O5 mixed with sawdust. Supplementary characteristics of the landscape around the excavation areas and the distance from the sea were obtained from maps and aerial photos, available from the WMS-service of the Estonian Land Board Web site (www.maaamet.ee). The dominant habitat adjacent to the extracted peatlands was categorized as (1) an active peat mining area, (2) mire in natural state, (3) forest, or (4) area of intermixed habitats.
Occurrence of certain species groups, such as (1) trees and bushes—Pinus sylvestris, Betula spp., Picea abies, Salix spp., (2) dwarf shrubs—Andromeda polifolia, Calluna vulgaris, Empetrum nigrum, Ledum palustre, Vaccinium uliginosum, V. myrtillus, V. vitis-idaea, (3) fen species—Carex spp., Juncus spp., Eriophorum angustifolium, Typha latifolia, Phragmites australis, Potentilla palustris, (4) bog bryophytes—Sphagnum magellanicum, Polytrichum strictum, Aulacomnium palustre, and (5) lichens were considered as binary categorical factors in the evaluation of composition.
The nomenclature of vascular plant species follows Tutin et al. (1964–1980), and that of bryophytes Hill et al. (2006), and that of lichens Randlane & Saag (1999).
The generalized pattern in floristic composition was analyzed by the detrended correspondence analysis (DCA; McCune & Mefford 1999). Cluster analysis was applied to define the species community (resp. assemblage) types for microtopographic forms. The chord distance was used as the dissimilarity measure, and the flexible beta method (β = −0.8) as the grouping algorithm (McCune & Mefford 1999). The distinctness of community types was tested by the multi-response permutation procedure (MRPP; McCune & Mefford 1999), taking into account the Bonferroni correction for multiple comparisons in pair-wise tests. In the multivariate analyses (DCA and clustering), the relevé data were filtered by selecting species occurring in at least three micro-habitats. For each revealed community type, the characteristic species were ascertained according to the indicator species analysis (Dufrêne & Legendre 1997) implemented in PC-Ord ver. 5.2 (McCune & Mefford 1999).
To obtain the normal distribution of residuals, the environmental variables such as the total area of abandoned extracted peatfields in bog, the distance from the sea, the depth of residual peat layer, the depth of slightly decomposed peat, the depth of well-decomposed peat, and ash content in residual peat were log-transformed.
The difference in the mean values of environmental variables in established community types was tested by means of ANOVA and Tukey HSD test (StatSoft Inc. 2004).