Holocene fen–bog transitions,current status in Finland and future perspectives

Abstract

Minerotrophic fens and ombrotrophic bogs differ in their nutrient status, hydrology, vegetation and carbon dynamics, and their geographical distribution is linked to various climate parameters. Currently, bogs dominate the northern temperate and southern boreal zones but climate warming may cause a northwards shift in the distribution of the bog zone. To more profoundly understand the sensitivity of peatlands to changes in climate, we first used the plant macrofossil method to identify plant communities that are characteristic of past fen–bog transitions. These transitions were radiocarbon dated, to be linked to Holocene climate phases. Subsequently, palaeoecological data were combined with an extensive vegetation survey dataset collected along the current fen–bog ecotone in Finland where we studied how the distribution of the key plant species identified from peat records is currently related to the most important environmental variables. The fossil plant records revealed clear successional phases: an initial Carex-dominated fen phase, an Eriophorum vaginatum–dominated oligotrophic fen phase followed by an early bog phase with wet bog Sphagna. This was occasionally followed by a dry ombrotrophic bog phase. Timing of initiation and phase transitions, and duration of succession phases varied between three sites studied. However, the final ombrotrophication occurred during 2000–3000 cal. BP corresponding to the neoglacial cooling phase. Dry mid-Holocene seems to have facilitated initiation of Eriophorum fens. The peatlands surveyed in the fen–bog ecotone were classified into succession phases based on the key species distribution. In 33% of the studied peatlands, Sphagnum had taken over and we interpret they are going through a final transition from fen to bog. In addition to autogenic processes and direct climate impact, our results showed that ecosystem shifts are also driven by allogenic disturbances, such as fires, suggesting that climate change can indirectly assist the ombrotrophication process in the southern border of the fen–bog ecotone.

Keywords

Boosted Regression Trees analysis fen–bog ecotone fen–bog transition Holocene peatland plant macrofossil analysis

Introduction

Peatlands are essential ecosystems in the northern hemisphere because of their large areal extent and their large carbon (C) stock. Recently, this stock was estimated at 436 Gt C for peatlands above the 45°N latitude (Loisel et al., 2014), a significant amount when compared with the atmospheric C reservoir of about 750 Gt C. Therefore, the fate of peatlands currently under threat, particularly from climate change, is of vital importance (Dise, 2009; Moore, 2002). Climate warming is predicted to have the strongest influence on high-latitude ecosystems (Dise, 2009; IPCC, 2013), where warming will result in a northwards shift of vegetation zones (Walther et al., 2002). The ecotone between minerotrophic fens and raised bog zones is likely to be among the most vulnerable (Allen and Breshears, 1998; Hansen et al., 1992; Neilson, 1993). In Fennoscandia, climate change is expected to result in raised mean annual temperatures, increased winter precipitation, and increased summer evapotranspiration (IPCC, 2013), which are important climatic parameters in regulating peatland development, that is, vertical growth and related changes in trophic status (Andersen et al., 2011; Breeuwer et al., 2009; Hughes and Dumayne-Peaty, 2002; Murphy and Moore, 2010). In particular, the predicted extension of the growing season, along with increased evapotranspiration rates that change the effective moisture conditions may accelerate the ombrotrophication process where the peatland-type changes from groundwater-fed fen to atmospheric-fed bog (Roulet et al., 1992; Turetsky et al., 2014). In brief, fens are more nutrient rich, dominated by minerotrophic Sphagna and brown mosses and sedge vegetation, and have higher decomposition rates, while bogs are dominated by ombrotrophic Sphagnum mosses and dwarf shrubs, which produce highly resilient litter (Straková et al., 2011). Generally, fens support larger methane emissions and are weaker C sinks than bogs (Saarnio et al., 2009; Turetsky et al., 2014; Turunen et al., 2002).

The border line between southern raised bogs and northern fen massifs zones in Finland roughly follows the 1100 temperature sum (degree days (d.d.)) isocline (Ruuhijärvi, 1960), although isolated raised bogs also occur inside the fen zone in the north and some large peatlands have remained as fens in the south because of terrain and local hydrology. In autogenic-driven ombrotrophication, there is a gradual increase in peat thickness, and this eventually decreases the impact of groundwater, which in turn reduces the amount of available nutrients, lowers the surface pH and slows down the decomposition process, and leads to further acceleration of the rate of peat accumulation (e.g. Pastor et al., 2002; Soudzilovskaia et al., 2010). Palaeoecological studies have shown that boreal and sub-arctic peatlands that generally initiated as minerotrophic fens (Supplementary Table 1, available online and references therein) south of the 1110 d.d. isocline evolved to ombrotrophic raised bogs over the course of time. It is not yet fully resolved as to how climate conditions determine peat accumulation processes, although it is speculated that under a warmer climate peatland plant species productivity might increase, thus leading to accelerated peat accumulation (Charman et al., 2013). Consequently, if warmer climate could trigger ombrotrophication processes in fens it may also lead to a northward movement of the raised bog zone, thereby diminishing the spatial coverage of the fen zone, especially if this zone does not shift northward at the same rate (Breeuwer et al., 2009; Moore, 2002; Parviainen and Luoto, 2007; Vitt, 2006). However, these processes are not straightforwardly linked only to climate per se but are driven by several interlinked autogenic and allogenic processes (Foster and Wright, 1990; Gignac and Vitt, 1990; Payette, 1988; Tahvanainen, 2011; Tuittila et al., 2007).

Knowledge of past peatland-type changes, that is, fen–bog transitions (FBTs), when linked to climate reconstructions, may provide valuable information on the sensitivity of peatlands to future changes in climate. As such, past FBT dynamics have been relatively widely studied, especially in Europe (Supplementary Table 1, available online and references therein). However, inconsistencies occur in terminology and in classification, which obscures the interpretations and site comparisons. In many previous studies, the bog phase is assigned to start with dominance of Eriophorum vaginatum, while in other studies the presence of certain Sphagnum species is taken as an indicator of the FBT (Supplementary Table 1, available online and references therein). Moreover, in many cases transitional phases have been determined by visual inspection of the peat content in the field or laboratory (Almquist-Jacobson and Foster, 1995; Korhola, 1992; Pajula, 2000) rather than by detailed analysis of species composition under a microscope.

This study aims to contribute to the practice of how to interpret past mire plant community shifts, most importantly the FBTs, and consequently we aim to discuss how these data can be used to recognize possible ongoing change in the fen–bog ecotone. The FBT question is approached here with two independent and complementary research strategies. First, we identify key peat plant species from the peat records, which were associated with major peatland-type shifts in the past. Then, by using an extensive vegetation survey dataset collected along the fen–bog ecotone in Finland, we study how the distribution of the key species is currently related to the environmental variables that are expected to be the major drivers in the future. Major regime shifts are radiocarbon dated to link the transitions to past regional climate phases and previous studies. We hypothesize that (1) a common indicator species of FBT can be distinguished, and (2) on the basis of the FBT indicators, the trajectories of current fens near their southern border can be predicted and linked to the regional climate. On the basis of the results, we discuss possible future FBT dynamics in boreal peatlands and frame our findings in a broader context.

Material and methods

Palaeoecological study sites

To define the key species indicative of ombrotrophication, we studied vegetation succession from peat cores sampled from three open bog sites located in different parts of Finland within the boreal zone (Figure 1). Peat samples were cored with a Russian peat corer in the summer of 2011. The mires were selected on the basis of their location on the north–south fen–bog gradient and their differential development phase in relation to ombrotrophication. The most southern site Siikaneva (61°50′N, 24°10′E) is situated in the southern boreal vegetation zone. It is a large (1300 ha) oligotrophic fen (OF) that contains a number of isolated ombrotrophic cupola and therefore presents an interesting setting to explore the factors that trigger FBTs. The majority of Siikaneva is in a natural state. The peat core (450 cm) was taken from a Sphagnum magellanicum lawn. The other two sites, Honkaneva and Puohtiinsuo, are located at the fen–bog ecotone in the middle-boreal vegetation zone. Honkaneva (64°34.39′N, 25°20.35′E) is situated in western Finland (Figure 1). Despite the presence of ditches and peat mining in an adjacent area, our study area in Honkaneva has remained in its natural state (Ronkainen et al., 2013). The peat core (280 cm) was taken from a Sphagnum balticum dominated lawn. The Puohtiinsuo complex (62°43′N, 31°06′E) in eastern Finland contains ombrotrophic areas in the upslope parts where the sample core was taken. The coring point was an S. magellanicum–dominated lawn and the peat core was 480 cm long.

Figure 1.

Locations of the study sites. Three palaeoecological sites are marked by triangles, 1 = Honkaneva, 2 = Siikaneva, 3 = Puohtiinsuo, and 21 vegetation survey sites by circles. The dashed line indicates roughly the current border line that separates the southern bog zone from the northern fen zone. The solid lines indicate the approximate historical position of the Baltic Sea ca. 10,000 and 7000 years ago, respectively, redrawn from Eronen et al. (2001). The current land uplift rate near the study site 1 is 7 mm yr⁻¹ (see also Tuittila et al., 2013), while in the southern coast of Finland the uplift rate is <2 mm yr^–1.

Plant macrofossil analysis

The peat cores were subsampled into 2-cm slices and the final analysis resolution was 6 cm. A 5-cm³ subsample was taken for macrofossil analysis. The subsample was cleaned under running water using a 140-µm sieve. A stereomicroscope was used for identification and the percentage cover of different plant remains was determined on a petri dish with the aid of scale paper. We applied a modified Quadrat and Leaf Count protocol (QLC) (Mauquoy and van Geel, 2007; Väliranta et al., 2007). Seeds, leaves, charcoal particles and E. vaginatum spindles were counted as absolute numbers. Species-level identification was performed with a high power light microscope when the proportion of bryophytes was more than 15%.

Radiocarbon analysis

To define the timing of peatland initiation and the major plant community shifts in the studied peat profiles, a total of 10 peat samples were sent for AMS radiocarbon dating to the Finnish Museum of Natural History (LUOMUS). Dated levels were chosen based on palaeobotanical variations in the plant macrofossil record (Figure 2). Radiocarbon ages Before Present (BP) were calibrated using Intcal09-age calibration curve (Reimer et al., 2009) (Table 1) and 2-sigma range median values are used for discussion.

Figure 2.

Selected plant macrofossil taxa and chronological data for the three study sites: Honkaneva, Siikaneva and Puohtiinsuo, Finland. Macroscopic charcoal finds are counted as numbers and indicated by #; other remains are expressed as percentages proportions.

Table 1.

Radiocarbon analyses results of peat profiles from the three study sites: Honkaneva, Siikaneva and Puohtiinsuo, Finland.

Lab. code	Sample	¹⁴C BP age	Calibrated 2-sigma range and the median
Hela-2857	Honkaneva 118–120 cm	1115 ± 30	1120–935, 1015
Hela-2858	Honkaneva 142–144 cm	3144 ± 31	3445–3270, 3375
Hela-2859	Honkaneva 262–264 cm	4099 ± 36	4815–4445, 4615
Hela-2860	Honkaneva 276–278	4133 ± 36	4825–4530, 4675
Hela-2868	Siikaneva 272–274 cm	3197 ± 32	3470–3360, 3415
Hela-2869	Siikaneva 406–408 cm	6858 ± 46	7795–7590, 7690
Hela-2870	Siikaneva 446–448 cm	8481 ± 54	9550–9420, 9495
Hela-2865	Puohtiinsuo 122–123 cm	2072 ± 32	2130–1945, 2040
Hela-2866	Puohtiinsuo 284–286 cm	5966 ± 43	6900–6675, 6800
Hela-2867	Puohtiinsuo 476–478	9005 ± 58	10,255–9915, 10,180

Palaeoecological data analysis

We applied detrended correspondence analysis (DCA) to study the main variations in historical vegetation. In order to analyse species replacement along the major gradient, we used Generalized Additive Models (GAM) (Guisan et al., 2002) to define species response to the first DCA axes.

In the GAM model, a logarithmic link function was applied assuming a Poisson distribution of errors.

We applied variation partitioning (Borcard et al., 1992; Vandvik and Birks, 2002) of macrofossil data to examine the relative importance of autogenic development related to growth in peat thickness and to the disturbance succession related to fire events for the composition of peatland vegetation during its history. Explanatory variables were split into three groups: (a) peat thickness, (b) fire history (including time since fire event and occurrence of charcoal particles of >1 mm and charcoal particles of <1 mm) and (c) peatland site. The latter was included to address site-specific variations in autogenic development and fire succession. We used canonical correspondence analysis (CCA) that is suitable for data with high compositional variation and excluded species with less than six occurrences. Data were log-transformed to stabilize the variance and to obtain a more symmetric distribution. Significance of variables (p < 0.05) was assessed using the Monte–Carlo permutation test before they were included. Analyses were conducted with CANOCO 5.0. (Ter Braak and Šmilauer, 2002).

Vegetation survey along the current FBT zone

To quantify the ecological niche of the key species indicative of ombrotrophication, and to estimate the current successional phases of peatlands in the fen–bog ecotone, we collected plant species and environmental data from 21 peatlands north of the geographical margin of bog and fen complex types in Finland (Figure 1). Quantification of the ecological niche was based on the assumption that species have their highest abundance at their optimal conditions. Estimation of successional phase was based on the assumption that historical phases identified with palaeoecological data analysis from peat cores could also be seen in current vegetation. From aerial photographs, we selected basins or sub-basins that appeared as typical fens. The vegetation data were collected along transects that crossed the peatlands with a study point interval of 20 m. At each study point, pH, peat thickness and water table (WT) levels were measured and vegetation coverage (%) surveyed from three replicates using circular relevés of 30 cm in diameter. A total of 272 transect points with 673 vegetation relevés were surveyed in August 2011.

The western part of the ecotone is characterized by more enhanced post-glacial rebound that continuously exposes new land and enables primary paludification (Ekman and Mäkinen, 1996). Topographically, western Finland is lowland and the topography in the eastern part of the ecotone is more variable with some supra-aquatic high areas that were never inundated by any of the post-glacial Baltic Sea phases (Figure 1; Tikkanen and Oksanen, 2002). Accordingly, eastern peatlands are on average older than the western peatlands at the same latitude.

Mean annual precipitation (P) in the western and eastern parts of the study area is on average 563 and 650 mm, respectively. The mean annual temperature is 1.6°C throughout the studied transition areas. The January mean temperature is slightly lower in the east (−11.1°C) than in the west (−10.7°C), although the July mean temperatures are similar (15.5 and 15.6°C). Temperature sum in d.d. is 1000 to 1200 and the annual potential evapotranspiration (PET) is about 287 mm. In the west, the difference between annual precipitation and annual potential evapotranspiration (P-PET) is smaller (286 mm) than in the east (362 mm) (Finnish Meteorological Institute, 1961–1990 measurement period). In the west, the climate is slightly more oceanic and the growing season longer (Karlsen et al., 2006).

Analysis of the relationship between living plant species and environmental parameters

The key plant species indicative of ombrotrophication based on palaeobotanical data were selected for environmental response analyses. We used Boosted Regression Trees (BRTs) to model the species response to WT, peat thickness, pH, and the 30-year means of climate parameters (T and P) (Finnish Meteorological Institute, 1961–1990 measurement period). BRTs have been proven to perform well in studies on species responses to environmental variables (De’ath, 2007; Elith et al., 2006; Leathwick et al., 2006; Moisen et al., 2006). BRTs rely on numerical optimization that minimizes residuals and consist of two algorithms: regression trees based on binary splits, and boosting that connect the regression trees into a final model (Elith et al., 2008). As a result, BRT assigns relative importance of each model parameter to the response variable based on how many times each of the parameters was allocated to the regression trees. The possible interactions are automatically counted (Elith et al., 2008) and each tree is made from a random subset. Generalized Boosted Regression Models were made with libraries gbm (Ridgeway, 2012) and dismo (Elith et al., 2008) in R software (R Development Core Team, 2010). Interaction depth was set to 3, learning rate to 0.001, and Gaussian distribution was used. Multicollinearity was taken into account by choosing only parameters that did not correlate strongly |r| < 0.70. In BRTs, a number of simple regression-tree models are combined for each of the key species to produce a final predictive model. The final models were selected by cross-validating the deviance explained (Elith and Leathwick, 2015).

The current succession phase of each studied peatland was defined by relating the current vegetation composition to the vegetation composition in successional sequences defined from the peat cores. We calculated the proportion of key species and plant groups for the succession phases, identified from palaeobotanical data, from the total coverage of the current vegetation. The determination was based only on the vegetation data collected from the centres of the study basins (defined as central 50% of the transect length, n = 371 vegetation plots) because the edge vegetation in fen massifs is typically formed by bog species.

Results

Variation in palaeoecological records

The palaeobotanical records from the three peatlands revealed distinct succession phases (Figure 2), which were also clearly seen in the groupings of species response curves along the main gradient in the historical vegetation (Figure 3). These were (1) an initial Carex-dominated fen (CF) phase that was characterized by key minerotrophic species such as Carex spp., Scheuchzeria palustris, Equisetum spp. and Menyanthes trifoliata (right end of axis 1 in Figure 3). The CF phase was followed by (2) an OF phase, where E. vaginatum clearly dominated the plant assemblages (middle of axis 1 in Figure 3), and (3) a Sphagnum-dominated bog (SB) phase that typically initiated as a wetter Sphagnum phase with species such as S. balticum, S. angustifolium and S. magellanicum (SB I). In some cases, SB I was replaced by hummock communities that included for instance S. fuscum and S. rubellum (SB II) (left end of axis 1 in Figure 3). The dry phase (SB II) was not always clearly visible in the palaeo-communities as the coring took place in a lawn surface. Our records suggest that the FBT is characterized by replacement of Eriophorum by ombrotrophic Sphagna: S. balticum, S. angustifolium and S. magellanicum. Accordingly, these four species are identified as key FBT species. In Honkaneva, the transition from CF to OF occurred at 4615 cal. BP. The change from OF to SB was gradual but can be assigned to around 3300 cal. BP, when the amount of S. angustifolium started to increase. The minerotrophic and Eriophorum phases lasted 60 and 1240 years, respectively. In Siikaneva, the transition from CF to OF occurred at 7690 cal. BP and the change from OF to SB, when the amount of E. vaginatum clearly decreases, occurred at 3415 cal. BP (Figure 2). The CF and OF phases lasted 1800 and 4280 years, respectively (Table 2). In Puohtiinsuo, the transition from CF to OF occurred at 6800 cal. BP, and the transition from OF to SB occurred at 2040 cal. BP. The CF and OF phases lasted 3335 years and 4760 years, respectively.

Figure 3.

Generalized Additive Models (GAM) response patterns of the identified key species illustrate the different succession stages. In axis 1, early succession fen species occur on the right while bog species become more abundant on the left side of the axis. Eriophorum vaginatum occurs in the middle of the succession where it acts as an ecosystem engineer.

Table 2.

Duration of different peatland succession phases and the timing of the fen–bog transition in the three study sites: Honkaneva, Puohtiinsuo and Siikaneva, Finland.

Site	Mire initiation cal. BP	Duration of minerotrophic fen phase	Timing of minerotrophic fen–Eriophorum fen transition	Duration of Eriophorum fen phase	Timing of fen–bog transition
Honkaneva	4675	60	4615	1240	3375
Siikaneva	9495	1800	7690	4280	3415
Puohtiinsuo	10,180	3335	6800	4760	2040

Variation partitioning indicated that peat thickness, time since fire event and peatland initiation together explained 43.4% of the variation in the palaeobotanical records of the three Finnish mires (Supplementary Figure 1, available online). Autogenic succession driven by peat thickness was site-specific while peat thickness alone explained only 2.6% of the variation in palaeovegetation composition, when combined with site-specific succession it explained 11.4% of the variation. Species succession following fire was more common between the sites; time since fire alone explained 3.6% of the variation and when combined with site-specific fire succession it explained 6.3%. Differences in the palaeobotanical records between the three Finnish peatlands were responsible for 15.3% of the variation.

Current vegetation: Key species’ relationship to environmental factors

According to BRT analysis, WT, pH and peat thickness were clearly the most powerful factors (relative contribution of the variables typically over 10%) that determine species distribution in the studied fens (Figure 4a and b). Grouping of individual climate variables was found to have a less powerful impact on WT than when the variables were inspected separately. The key species had different WT optima, pH and peat thickness values, and the level of significance of the environmental variables also varied (Table 3). E. vaginatum and S. balticum were more abundant on sites with lower pH and a thicker peat layer, whereas S. angustifolium and S. magellanicum thrived on sites with higher pH (Figure 4a and b). Less important environmental factors were PET for E. vaginatum, P for S. balticum and the water balance for S. magellanicum. The Spearman rank correlations between the observed and predicted values were good for the E. vaginatum and S. balticum models; however, the S. magellanicum model performed very poorly (Table 3).

Figure 4.

Response of living key plant species; Eriophorum vaginatum and S. balticum (4a), and Sphagnum angustifolium and S. magellanicum (4b), to the various environmental factors based on Boosted Regression Tree model (BRT). Columns present the relative influences of each variable. The most important factors are peat thickness, pH and water table level. Partial dependence plots for each parameter in the BRT models. Predictors are centered and are on a standard deviation scale. Partial dependency functions (Friedman 2001) visualize the dependency between fitted response (vertical axis) and each predictor. The vertical axes in partial dependency plots are not comparable with the range of the response variable. Nevertheless, the shape of the partial dependency function reflects the effect shape and the range is proportional to the relative contribution of the predictor (Friedman 2001). Note the different vertical axes ranges in plots. Note that the response curves to parameters with low relative influence (< 5%) are unlikely to contain much relevant information.

Table 3.

Correlations between pH, water table level and peat thickness and key species coverage (%). Values in bold are statistically significant (p < 0.01).

	E. vaginatum	S. angustifolium	S. balticum	S. magellanicum
pH	−0.31	0.14	−0.48	0.00
Peat thickness	0.11	−0.17	0.38	−0.02
Water table level	−0.13	−0.27	0.10	−0.12

Current succession phases along the fen–bog gradient and their environmental characteristics

The key species for the different succession phases were defined from the palaeobotanical data, namely E. vaginatum, S. angustifolium, S. balticum and S. magellanicum (Figures 2 and 3). The current succession phase of each peatland along the ecotone was defined based on the proportional abundance of those species. All defined succession phases based on palaeobotanical data were represented in our plant community inventory. From the study sites located in the recent FBT zone, 57% were categorized as a minerotrophic fen phase based on the dominance of Carex spp. in the field layer, five peatlands in the western study area and six peatlands in the eastern study area (Figure 5). Only 10% of the surveyed sites were currently in the OF phase that precedes the FBT, while 33% were currently in the middle of a transition from fen to bog. This phase is characterized by a proportional increase in Sphagnum. Median pH for CF, OF and FBT sites was 4.17, 3.92 and 3.89, respectively. Mean peat thickness was 2.2, 2.1 and 2.8 m for the CF, OF and FBT sites, respectively.

Figure 5.

Inferred current succession stages of the studied peatlands along the fen–bog ecotone based on the vegetation survey. White dots indicate minerotrophic fens, the grey dots represent sites that are in an Eriophorum-phase preceding the final fen–bog transition, and black dots indicate peatlands where the amount of bog Sphagna was notable, that is, bogs.

Discussion

Past succession phases, timing of transitions and climate implications

OF phase

In all our palaeobotanical study sites, an OF phase dominated by E. vaginatum preceded the FBT, that is, emergence of bog Sphagnum species in all peat cores. This corresponds to what has been reported by previous studies (e.g. Dudová et al., 2013; Hughes, 2000; Hughes and Barber, 2004; Hughes and Dumayne-Peaty, 2002) (Supplementary Table 1, available online). E. vaginatum has been identified as an ‘ecosystem engineer’ that changes the habitat conditions so that they are more suitable for the establishment of oligotrophic–ombrotrophic Sphagnum mosses (Hughes and Dumayne-Peaty, 2002). The long and aerenchymatous roots that are a characteristic feature of E. vaginatum allow large WT fluctuations to be tolerated. Moreover, the highly resistant remains of E. vaginatum promote peat accumulation and consequent ombrotrophication (Coulson and Butterfield, 1978; Wein, 1973).

In Siikaneva and Puohtiinsuo, the Eriophorum-phase corresponds to the dry and warm mid-Holocene period between ca. 8500 and 5000 cal. BP (Seppä et al., 2005; Snowball et al., 2004). Low effective moisture likely resulted in lowered WT levels (Korhola et al., 2005; Mathijssen et al., 2014; Väliranta et al., 2005), and this would have benefited tolerant plant species, such as E. vaginatum. In contrast to the other sites, the initiation of the Eriophorum-dominated OF phase at the youngest site Honkaneva only seems to start long after the large-scale climate shift towards cool and moist conditions around 5000 cal. BP. This suggests that post-glacial rebound has been an important driving factor in changing peatland hydrology. Previous data, for instance from Ireland and England, show that the duration of the Eriophorum-dominated OF phase varied considerably (380 to 1800 years) between sites (Supplementary Table 1, available online).

Fen–bog transition

The key transitional Sphagnum species in our study, S. angustifolium, S. balticum and S. magellanicum, have also been identified by earlier studies as important early-stage bog species (e.g. Bohlin, 1993; Hughes, 2000; Loisel and Garneau, 2010; Robichaud and Bégin, 2009; Sillasoo et al., 2007; Tuittila et al., 2007). Hummock communities with S. fuscum, S. rubellum and dwarf shrubs are seen as climax bog communities (Klinger, 1996). These species are commonly found in hummock microforms in our study sites, and more generally in the region of this study. Our results taken in conjunction with others would suggest that bog development can proceed either through a wet or a dry Sphagnum phase (Almquist-Jacobson and Foster, 1995; Hughes, 2000; Pajula, 2000).

In our study sites, FBT occurred during the period 3375–2040 cal. BP (Table 2) when the cool moist climate had already prevailed for millennia. Many, but not all, previous studies have reported that initiation of the ombrotrophication process occurred during the late-Holocene period (Bohlin, 1993; Hughes and Dumayne-Peaty, 2002; Janssen, 1992; Korhola, 1992) (Supplementary Table 1, available online). In some southern Finnish peatlands, the FBT already occurred in the early Holocene (Tolonen, 1966). As a general rule, in Europe, towards north the ombrotrophic peat layer is typically younger than in the south (Tolonen, 1987). Moreover, Eriophorum- and bog Sphagna–dominated peat layers alternate and co-occur throughout peatland history often as a result of fire events and no clear FBT may be visible (Mathijssen et al., 2016; Tuittila et al., 2007). Such discrepancies highlight the fact that peatland succession patterns are also regulated by factors other than climate or autogenic succession.

Initiation of bog development in moist conditions is in slight contradiction to the view that low moisture availability, for instance as a result of autogenic vertical peat growth or catchment processes, such as drainage (Tahvanainen, 2011), is an essential factor triggering the ombrotrophication process (Hughes, 1997, 2000; Leah et al., 1998). An alternative hypothesis proposes that increased precipitation may wash the nutrients from the already slightly convex peatland surface thereby creating suitable ombrotrophic habitats for bog Sphagnum establishment (Hughes, 2000). A cool and humid climate can also slow down decomposition rates thus accelerating peat accumulation (Andersson and Schoning, 2010), and consequently promote ombrotrophication. This fundamental process is also included in the peatland development model of Frolking et al. (2010).

Even though the late-Holocene has been characterized as a moist climate phase, macroscopic charcoal records from boreal bogs, which indicate repeated intensive fire frequency periods, suggest that prolonged summer droughts have synchronously occurred during the last 5000 years (Morris et al., 2015; Sillasoo et al., 2011; see also Ronkainen et al., 2013). The importance of fire events was also observed in the variation partitioning where fire succession explained 6% of the compositional variation during the history of the peatlands. The presence of macroscopic charcoal particles in the Siikaneva record during the Eriophorum-phase suggests local fires. Severe fires may actually have a twofold influence: the burning may result in temporary wet conditions when burning lowers the surface level close to the WT (Sillasoo et al., 2011), and/or the reduction in vegetation may decrease evapotranspiration rates (Tuittila et al., 2007). In any case, E. vaginatum has been found to benefit from fires (Tuittila et al., 2007) and other disturbances: it is an early colonizer where resources become rapidly available (Komulainen et al., 1999; Tuittila et al., 2000). Its appearance may further assist the ombrotrophication process.

Future trajectory of succession at the Finnish fen–bog ecotone

The succession phases that were defined from the palaeoecological data were also detectable in the plant survey data. Peatland succession towards the bog phase seems to have proceeded further in the east than in the west of the country (Figure 5). This may be partly linked to landscape evolution and land uplift so that the peatlands are younger in western Finland (Figure 1; Huttunen and Tolonen, 2006; Tikkanen and Oksanen, 2002). It should be noted, however, that age is not a very strong succession phase determinant; peatlands in Lapland had already initiated during the early Holocene period, but have remained as fens since mainly because of climate (e.g. Mäkilä and Moisanen, 2007; Väliranta et al., 2014). Climate envelope models suggest that temperature is the most important individual climate parameter to define the current distribution of bogs and fens (Parviainen and Luoto, 2007). Moreover, strong spring floods caused by snow melt, are considered to be essential for the maintenance of fens as they restrict Sphagnum growth (Granath et al., 2010).

Our results also reveal the importance of pH and WT level for peatland vegetation and are in agreement with several earlier studies (e.g. Bubier, 1995; Gignac, 1992; Hughes and Barber, 2004; Jeglum, 1971; Lemly and Cooper, 2011; Økland et al., 2001; Vitt and Slack, 1984) (Figure 4a and b). Peat thickness, pH and WT are local factors that have a direct effect on plants. However, these factors are strongly linked to temperature, precipitation and evapotranspiration, and this complicates the interpretation.

According to recent climate scenarios, the climate is unlikely to change unanimously all over Fennoscandia for instance (Gong et al., 2012), and consequently, assessment of the implications for ecosystem functioning is a challenge. WT levels in peatlands correlate strongly with the precipitation–evapotranspiration ratio, in particularly with the summer moisture deficit (Charman et al., 2006). While evapotranspiration models differ from region to region, higher temperatures generally lead to increased evapotranspiration rates and a lowering of WT levels (Heijmans et al., 2013). A recent model prediction for Finland based on ACCLIM climate change scenarios (Gong et al., 2012) showed that average water levels over the period 2000–2099 would decrease by 2–4 cm, and by up to 11 cm in the south-western parts of the fen region. Water level draw-down is predicted to be most pronounced in late spring (Gong et al., 2012) thereby diminishing surface water flow and the spring flood effect, and this may stimulate Sphagnum production and thus bog development (Bauer et al., 2003). Lowered WT levels can also promote tree establishment on peatlands (Päivänen, 1999; Päivänen and Hånell, 2012). This may increase evapotranspiration and lower the WT even further (Sarkkola et al., 2010), which again benefits the establishment of dry bog species, such as S. rubellum and S. fuscum. Under suitable conditions, the FBT can happen quickly, even within years if bog Sphagna invade the peatland (Granath et al., 2010; Kuhry et al., 1993; Loisel and Yu, 2013; Tahvanainen, 2011; Tuittila et al., 2013). Once established, Sphagnum mosses have the ability to change the ecosystem to their advantage by acidifying the habitat and creating after their death a water logged, anoxic, nutrient poor environment that impedes other species.

There is evidence from Finnish Lapland to show that the dry mid-Holocene climate, which can be used as an approximate analogue for a future climate, and associated lowering of the fen WT did not trigger bog development, but instead produced a dramatic slow-down in peat accumulation (Mathijssen et al., 2014). It may be that spring flood patterns did not change in this region and thus pH levels remained high, which prevented bog Sphagnum establishment (Granath et al., 2010).

According to the peatland habitat theory, the concave shape of fens should mean that the centres of fens support mainly minerotrophic lawn or flark species, vascular plants and bryophytes. Thus, in this study those fens where the key FBT species, E. vaginatum, S. angustifolium, S. balticum and S. magellanicum, were abundant in the centre, were classified as FBT sites. It can be speculated that the abundant presence of various bog Sphagnum species may result in increased peat accumulation rates in the near future (Breeuwer et al., 2009; Gorham, 1991).

Conclusion

The palaeobotanical analyses clearly revealed some key species that have been involved in past FBTs after the initial fen phase (hypothesis 1). These were E. vaginatum, S. angustifolium, S. balticum and S. magellanicum. The following succession phases were detected from the peat records: a CF phase, an E. vaginatum–dominated OF phase and a true bog phase, where wet bog Sphagna replaced Eriophorum. Bog succession may continue and eventually wet bog communities could be replaced by dry bog species. The duration of these phases varied but the FBT mainly occurred during the late-Holocene. Based on our knowledge of how nutrient levels and hydrological functioning differ between oligotrophic fens and raised bogs, we suggest that in the future only the transition that separates the Eriophorum-dominated phase from the phase dominated by ombrotrophic Sphagna, wet or dry, be called an FBT, even though we acknowledge that this transition is not always clearly definable.

All succession phases identified from our palaeobotanical data were also detected in the current fen–bog ecotone. Our analysis confirmed our second hypothesis whereby on the basis of the FBT indicators the trajectories of current fens located near the 1100 d.d. isocline border can be predicted and linked to regional climate. Our results also showed that 43% of the ecotonal peatlands surveyed in this study are currently undergoing a change towards more ombrotrophic conditions. Climate change accompanied by autogenic succession may drastically alter the spatial proportions of fens and bogs by increasing the amount of bog area. Our study provides the basis for future studies to investigate if and how peatland succession proceeds.

Footnotes

Funding

Funding from the Academy of Finland (codes 140868, 287039) is gratefully acknowledged.

References

Allen

Breshears

(1998) Drought-induced shift of a forest–woodland ecotone: Rapid landscape response to climate variation. Proceedings of the National Academy of Sciences 95: 14839–14842.

Almquist-Jacobson

Foster

(1995) Towards an integrated model for raised-bog development: Theory and field evidence. Ecology 76: 161–174.

Andersen

Poulin

Borcard

. (2011) Environmental control and spatial structures in peatland vegetation. Journal of Vegetation Science 22: 878–890.

Andersson

Schoning

(2010) Surface wetness and mire development during the late-Holocene in central Sweden. Boreas 39: 749–760.

Bauer

Gignac

Vitt

(2003) Development of a peatland complex in boreal western Canada: Lateral site expansion and local variability in vegetation succession and long-term peat accumulation. Canadian Journal of Botany 81: 833–847.

Bohlin

(1993) Botanical composition of peat. PhD Thesis, Swedish University of Agricultural Sciences, 396 pp.

Borcard

Legendre

Drapeau

(1992) Partialling out the spatial component of ecological variation. Ecology 73: 1045–1055.

Breeuwer

Robroek

Limpens

. (2009) Decreased summer water table depth affects peatland vegetation. Basic and Applied Ecology 10: 330–339.

Bubier

(1995) The relationship of vegetation to methane emission and hydrochemical gradients in northern peatlands. Journal of Ecology 83: 403–420.

10.

Charman

Beilman

Blaauw

. (2013) Climate-related changes in peatland carbon accumulation during the last millennium. Biogeosciences 10: 929–944.

11.

Charman

Blundell

Chiverrell

. (2006) Compilation of non-annually resolved Holocene proxy climate records: Stacked Holocene peatland palaeo-water table reconstructions from northern Britain. Quaternary Science Reviews 25: 336–350.

12.

Coulson

Butterfield

(1978) An investigation of the biotic factors determining the rates of plant decomposition on blanket bog. Journal of Ecology 66: 631–650.

13.

De’ath

(2007) Boosted trees for ecological modeling and prediction. Ecology 88: 243–251.

14.

Dise

(2009) Peatland response to global change. Science 326: 810–811.

15.

Dudová

Hájková

Buchtová

. (2013) Formation, succession and landscape history of Central-European summit raised bogs: A multiproxy study from the Hrubý Jeseník Mountains. Holocene 23: 230–242.

16.

Ekman

Mäkinen

(1996) Recent postglacial rebound, gravity change and mantle flow in Fennoscandia. Geophysical Journal International 126: 229–234.

17.

Elith

Leathwick

(2015) Boosted Regression Trees for ecological modeling. Available at: http://cran.r-project.org/web/packages/dismo/vignettes/brt.pdf.

18.

Elith

Leathwick

Hastie

(2008) A working guide to boosted regression trees. Journal of Animal Ecology 77: 802–813.

19.

Elith

Graham

Anderson

. (2006) Novel methods improve prediction of species distributions from occurrence data. Ecography 29: 129–151.

20.

Eronen

Glückert

Hatakka

. (2001) Rates of Holocene isostatic uplift and relative sea-level lowering of the Baltic in SW Finland based on studies of isolation contacts. Boreas 30: 17–30.

21.

Foster

Wright

(1990) Role of ecosystem development and climate change in bog formation in central Sweden. Ecology 71: 450–463.

22.

Friedman

(2001) Greedy function approximation: A gradient boosting machine. The Annals of Statistics 29: 1189–1232.

23.

Frolking

Roulet

Tuittila

E-S

. (2010) A new model of Holocene peatland net primary production, decomposition, water balance, and peat accumulation. Earth System Dynamics 1: 1–21.

24.

Gignac

(1992) Niche structure, resource partitioning, and species interactions of mire bryophytes relative to climatic and ecological gradients in western Canada. The Bryologist 95: 406–418.

25.

Gignac

Vitt

(1990) Habitat limitations of Sphagnum along climatic, chemical, and physical gradients in mires of western Canada. The Bryologist 93: 7–22.

26.

Gong

Wang

Kellomäki

. (2012) Modeling water table changes in boreal peatlands of Finland under changing climate conditions. Ecological Modelling 244: 65–78.

27.

Gorham

(1991) Northern peatlands: Role in the carbon cycle and probable responses to climatic warming. Ecological Applications 1: 182–195.

28.

Granath

Strengbom

Rydin

(2010) Rapid ecosystem shifts in peatlands: Linking plant physiology and succession. Ecology 91: 3047–3056.

29.

Guisan

Edwards

Hastie

(2002) Generalized linear and generalized additive models in studies of species distributions: Setting the scene. Ecological Modelling 157: 89–100.

30.

Hansen

Risser

di Castri

(1992) Epilogue: Biodiversity and ecological flow across ecotones. In: Hansen

Di Castri

(eds) Landscape Boundaries. New York: Springer, pp. 423–438.

31.

Heijmans

MMPD

van

der

Knaap

YAM

Holmgren

. (2013) Persistent versus transient tree encroachment of temperate peat bogs: Effects of climate warming and drought events. Global Change Biology 19: 2240–2250.

32.

Hughes

(1997) The palaeoecology of the fen-bog transition during the early- to mid-Holocene in Britain. PhD Thesis, University of Southampton.

33.

Hughes

(2000) A reappraisal of the mechanisms leading to ombrotrophy in British mires. Ecology Letters 3: 7–9.

34.

Hughes

Barber

(2004) Contrasting pathways to ombrotrophy in three raised bogs from Ireland and Cumbria, England. The Holocene 14: 65–77.

35.

Hughes

Dumayne-Peaty

(2002) Testing theories of mire development using multiple successions at Crymlyn Bog, West Glamorgan, South Wales, UK. Journal of Ecology 90: 456–471.

36.

Huttunen

Tolonen

(2006) Mire development history in Finland. In: Lindholm

Heikkilä

(eds) Finland – Land of Mires (The Finnish Environment, 23). Vammala: Vammalan kirjapaino Oy, pp. 79–88.

37.

IPCC (2013) Annex I: Atlas of global and regional climate projections. In: Stocker

Qin

Plattner

G-K

. (eds) Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press, pp. 1311–1393.

38.

Janssen

(1992) The Myrtle Lake peatland. In: Wright

Jr Coffin

Aaseng

(eds) The Patterned Peatlands of Minnesota. Minneapolis, MN: University of Minnesota Press, pp. 223–235.

39.

Jeglum

(1971) Plant indicators of pH and water level in peatlands at Candle lake, Saskatchewan. Canadian Journal of Botany 49: 1661–1676.

40.

Karlsen

Elvebakk

Høgda

. (2006) Satellite-based mapping of the growing season and bioclimatic zones in Fennoscandia. Global Ecology and Biogeography 15: 416–430.

41.

Klinger

(1996) The myth of the classic hydrosere model of bog succession. Arctic and Alpine Research 28: 1–9.

42.

Komulainen

Tuittila

E-S

Vasander

. (1999) Restoration of drained peatlands in southern Finland: Initial effects on vegetation change and CO₂ balance. Journal of Applied Ecology 36: 634–648.

43.

Korhola

(1992) Mire induction, ecosystem dynamics and lateral extension on raised bogs in the southern coastal area of Finland. Fennia: International Journal of Geography 170: 25–94.

44.

Korhola

Tikkanen

Weckström

(2005) Quantification of Holocene lake-level changes in Finnish Lapland using a cladocera – Lake depth transfer model. Journal of Paleolimnology 34: 175–190.

45.

Kuhry

Halsey

Bayley

. (1992) Peatland development in relation to Holocene climatic change in Manitoba and Saskatchewan (Canada). Canadian Journal of Earth Sciences 29: 1070–1090.

46.

Kuhry

Nicholson

Gignac

. (1993) Development of Sphagnum–dominated peatlands in boreal continental Canada. Canadian Journal of Botany 71: 10–22.

47.

Leah

Wells

Stamper

. (1998) The Wetlands of Shropshire and Staffordshire, North West Wetlands Survey. Lancaster: University of Lancaster, Archaeological Unit, 252 pp.

48.

Leathwick

Elith

Francis

. (2006) Variation in demersal fish species richness in the oceans surrounding New Zealand: An analysis using boosted regression trees. Marine Ecology Progress Series 321: 267–281.

49.

Lemly

Cooper

(2011) Multiscale factors control community and species distribution in mountain peatlands. Botany 89: 689–713.

50.

Loisel

Garneau

(2010) Late-Holocene paleoecohydrology and carbon accumulation estimates from two boreal peat bogs in eastern Canada: Potential and limits of multi-proxy archives. Palaeogeography, Palaeoclimatology, Palaeoecology 291: 493–533.

51.

Loisel

(2013) Recent acceleration of carbon accumulation in a boreal peatland, south central Alaska. Journal of Geophysical Research: Biogeosciences 118: 41–53.

52.

Loisel

Beilman

. (2014) A database and synthesis of northern peatland soil properties and Holocene carbon and nitrogen accumulation. The Holocene 24: 1028–1042.

53.

Mäkilä

Moisanen

(2007) Holocene lateral expansion and carbon accumulation of Luovuoma, a northern fen in Finnish Lapland. Boreas 36: 198–210.

54.

Mathijssen

Tuovinen

J-P

Lohila

. (2014) Development, carbon accumulation and radiative forcing of a subarctic fen over the Holocene. The Holocene 24: 1156–1166.

55.

Mathijssen

PJH

Väliranta

Korrensalo

. (2016) Holocene carbon dynamics reconstruction from a large boreal peatland complex, southern Finland. Quaternary Science Reviews 142: 1–15.

56.

Mauquoy

Van Geel

(2007) Mire and peat macros. In: Elias

(ed.) Encyclopedia of Quaternary Science, vol. 3. Amsterdam: Elsevier Science, pp. 2315–2336.

57.

Middleton

Wells

Huckerby

(1995) The Wetlands of North Lancashire (North West Wetlands Survey, vol. 3). Lancaster: Lancaster Imprints, 279 pp.

58.

Moisen

Freeman

Blackard

. (2006) Predicting tree species presence and basal area in Utah: A comparison of stochastic gradient boosting, generalized additive models, and tree-based methods. Ecological Modelling 199: 176–187.

59.

Moore

(2002) The future of cool temperate bogs. Environmental Conservation 29: 3–20.

60.

Morris

Väliranta

Sillasoo

. (2015) Re-evaluation of fire histories of three boreal bogs suggest a link between bog fire and climate. Boreas 44: 60–67.

61.

Murphy

Moore

(2010) Linking root production to above-ground plant characteristics and water table in a temperate bog. Plant and Soil 336: 219–231.

62.

Neilson

(1993) Transient ecotone response to climate change: Some conceptual and modelling approaches. Ecological Applications 3: 385–395.

63.

Økland

Rydgren

(2001) A Scandinavian perspective on ecological gradients in north-west European mires: Reply to Wheeler and Proctor. Journal of Ecology 89: 481–486.

64.

Päivänen

(1999) Tree stand structure on pristine peatlands and its change after forest drainage. International Peat Journal 9: 66–72.

65.

Päivänen

Hånell

(2012) Peatland Ecology and Forestry – A Sound Approach (University of Helsinki, Department of Forest Sciences Publications, vol. 3). Sastamala: Vammalan Kirjapaino Oy, 267 pp.

66.

Pajula

(2000) Spatio-temporal development of the Soomaa mire system in SW Estonia. Proceedings of the Estonian Academy of Sciences 49: 194–208.

67.

Parviainen

Luoto

(2007) Climate envelopes of mire complex types in Fennoscandia. Geografiska Annaler: Series A, Physical Geography 89: 137–151.

68.

Pastor

Peckham

Bridgham

. (2002) Plant community dynamics, nutrient cycling, and alternative stable equilibria in peatlands. The American Naturalist 160: 553–568.

69.

Payette

(1988) Late-Holocene development of subarctic ombrotrophic peatlands: Allogenic and autogenic succession. Ecology 69: 516–531.

70.

R Development Core Team (2010) R: A Language and Environment for Statistical Computing. Vienna: R Foundation for Statistical Computing, 409 pp.

71.

Reimer

Baillie

Bard

. (2009) IntCal09 and Marine09 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 51: 1111–1150.

72.

Ridgeway

(2012) Package ‘GBM’. Available at: http://cran.r-project.org/web/packages/gbm/gbm.pdf (accessed 15 November 2012).

73.

Robichaud

Bégin

(2009) Development of a raised bog over 9000 years in Atlantic Canada. Mires & Peat 5: 1–19.

74.

Ronkainen

Väliranta

Tuittila

E-S

(2013) Fire pattern in a drainage-affected boreal bog. Boreal Environment Research 18: 309–316.

75.

Roulet

Moore

Bubier

. (1992) Northern fens: Methane flux and climatic change. Tellus B 44: 100–105.

76.

Ruuhijärvi

(1960) Uber die regionale Einteilung der nordfi nischen Moore. Annales Botanici Societatis Zoologicae Botanicae Fennica ‘Vanamo’ 3: 1–360.

77.

Saarnio

Winiwarter

Leitão

(2009) Methane release from wetlands and watercourses in Europe. Atmospheric Environment 43: 1421–1429.

78.

Sarkkola

Hökkä

Koivusalo

. (2010) Role of tree stand evapotranspiration in maintaining satisfactory drainage conditions in drained peatlands. Canadian Journal of Forest Research 40: 1485–1496.

79.

Seppä

Hammarlund

Antonsson

(2005) Low-frequency and high-frequency changes in temperature and effective humidity during the Holocene in south-central Sweden: Implications for atmospheric and oceanic forcings of climate. Climate Dynamics 25: 285–297.

80.

Sillasoo

Väliranta

Tuittila

E-S

(2011) Fire history and vegetation recovery in two raised bogs at the Baltic Sea. Journal of Vegetation Science 22: 1084–1093.

81.

Sillasoo

Mauquoy

Blundell

. (2007) Peat multi-proxy data from Männikjärve bog as indicators of Late-Holocene climate changes in Estonia. Boreas 36: 20–37.

82.

Smith

Goddard

(1991) A 12 500 year record of vegetational history at Sluggan Bog, Co. Antrim, N. Ireland (incorporating a pollen zone scheme for the nonspecialist). New Phytologist 118: 167–187.

83.

Smith

Morgan

(1989) A succession to ombrotrophic bog in the Gwent Levels, and its demise: A Welsh parallel to the peats of the Somerset Levels. New Phytologist 107: 145–167.

84.

Snowball

Korhola

Briffa

. (2004) Holocene climate dynamics in high latitude Europe and the North Atlantic. In: Battarbee

Gasse

Stickling

(eds) Past Climate Variability through Europe and Africa. Dordrecht: Springer, pp. 465–494.

85.

Soudzilovskaia

Cornelissen

During

. (2010) Similar cation exchange capacities among bryophyte species refute a presumed mechanism of peatland acidification. Ecology 91: 2716–2726.

86.

Straková

Niemi

Freeman

. (2011) Litter type affects the activity of aerobic decomposers in a boreal peatland more than site nutrient and water table regimes. Biogeosciences 8: 2741–2755.

87.

Svensson

(1988) Bog development and environmental conditions as shown by the stratigraphy of Store Mosse mire in Southern Sweden. Boreas 17(1): 89–111.

88.

Tahvanainen

(2011) Abrupt ombrotrophication of a boreal aapa mire triggered by hydrological disturbance in the catchment. Journal of Ecology 99: 404–415.

89.

Ter Braak

Šmilauer

(2002) CANOCO Reference Manual and CanoDraw for Windows User’s Guide: Software for Canonical Community Ordination (version 4.5). Ithaca, NY: Microcomputer Power, 500 pp.

90.

Tikkanen

Oksanen

(2002) Late Weichselian and Holocene shore displacement history of the Baltic Sea in Finland. Fennia: International Journal of Geography 180: 9–20.

91.

Tolonen

(1966) Stratigraphic and rhizopod analyses on an old raised bog, Varrassuo, in Hollola, South Finland. Annales Botanici Fennici 3: 147–166.

92.

Tolonen

(1987) Natural History of Raised Bogs and Forest Vegetation in the Lammi Area, Southern Finland Studied by Stratigraphical Methods (Annales Academiae Scientarium Fennicae: Series A, III Geologica-Geographica). Helsinki: Suomalainen Tiedeakatemia, 144 pp.

93.

Tuittila

E-S

Juutinen

Frolking

. (2013) Wetland chronosequence as a model of peatland development: Vegetation succession, peat and carbon accumulation. The Holocene 23: 25–35.

94.

Tuittila

E-S

Rita

Vasander

. (2000) Vegetation pattern around tussocks of Eriophorum vaginatum L. in a cut-away peatland in Southern Finland. Canadian Journal of Botany 78: 47–58.

95.

Tuittila

E-S

Väliranta

Laine

. (2007) Quantifying patterns and controls of mire vegetation succession in a southern boreal bog in Finland using partial ordinations. Journal of Vegetation Science 18: 891–902.

96.

Turetsky

Kotowska

Bubier

. (2014) A synthesis of methane emissions from 71 northern, temperate, and subtropical wetlands. Global Change Biology 20: 2183–2197.

97.

Turunen

Tomppo

Tolonen

. (2002) Estimating carbon accumulation rates of undrained mires in Finland – Application to boreal and subarctic regions. The Holocene 12: 69–80.

98.

Väliranta

Korhola

Seppä

. (2007) High-resolution reconstruction of wetness dynamics in a southern boreal raised bog, Finland, during the late-Holocene: A quantitative approach. The Holocene 17: 1093–1107.

99.

Väliranta

Kultti

Nyman

. (2005) Holocene development of aquatic vegetation in a shallow Lake Njargajavri, Finnish Lapland with evidence of water level fluctuations and drying. Journal of Paleolimnology 34: 203–215.

100.

Väliranta

Oinonen

Seppä

. (2014) Unexpected problems in AMS ¹⁴C dating of fen peat. Radiocarbon 56: 95–108.

101.

Vitt

(2006) Functional characteristics and indicators of boreal peatlands. In: Wieder

Vitt

(eds) Boreal Peatland Ecosystems. Berlin, Heidelberg: Springer-Verlag, 435pp.

102.

Vitt

Slack

(1984) Niche diversification of Sphagnum relative to environmental factors in northern Minnesota peatlands. Canadian Journal of Botany 62: 1409–1430.

103.

Walther

G-R

Post

Convey

. (2002) Ecological responses to recent climate change. Nature 416: 389–395.

104.

Wein

(1973) Biological flora of the British Isles: Eriophorum vaginatum. Journal of Ecology 61: 601–615.

105.

Wells

Huckerby

Hall

(1997) Mid- and late-Holocene vegetation history and tephra studies at Fenton Cottage, Lancashire, U.K. Vegetation History and Archaeobotany 6: 153–166.