Introduction

Peatlands are wetlands formed through the accumulation of partially decayed organic material (peat); they represent 50–70 % of global wetland resources (Chapman et al. 2003). These ecosystems are important for biodiversity conservation, climate regulation and human welfare. For instance, owing to their peat accumulation processes, peatlands play an important role in the global carbon (C) cycle, having accumulated approximately 600 GtC during the Holocene (Yu et al. 2010). They also store fresh water and help regulate regional hydrologic fluxes (e.g., Levison et al. 2014). Furthermore, they support specialized flora adapted to harsh prevailing conditions, notably acidic and water logged soils. In temperate regions, their flora contrasts sharply with surrounding environments and contributes to increased regional diversity (Ingerpuu et al. 2001). The maintenance of the above functions and services is nevertheless threatened when peatland plant communities are altered. With intensification of the human footprint on these ecosystems, it becomes crucial to better understand the factors driving their compositional and structural changes through time, in order to better protect them.

Increases in woody plant cover and density on historically treeless or sparsely treed sites is one of the most drastic changes recently reported in ombrotrophic peatlands (bogs) of northern Europe (Freléchoux et al. 2000; Linderholm and Leine 2004; Kapfer et al. 2011) and North America (Pellerin and Lavoie 2003; Berg et al. 2009; Talbot et al. 2010; Ireland and Booth 2012; Pasquet et al. 2015). Although the progression of bog communities from Sphagnum to tree dominated has been considered as a natural stage of wetland succession (e.g., Mitsch and Gosselink 1986), the results of recent studies indicate that direct and indirect anthropogenic disturbances favour the current phenomenon. Indeed, human-induced drainage, eutrophication, fire and contemporary climate change (warmer and/or drier climatic conditions) are the main factors usually pinpointed to explain the present-day woody encroachment in bogs, through their influence on water and nutrient availability (e.g., Frankl and Schmeidl 2000; Linderholm and Leine 2004; Pellerin and Lavoie 2003; Berg et al. 2009; Pasquet et al. 2015). For example, using paleoecological techniques, Ireland and Booth (2012) have shown that the shift from Sphagnum to white pine (Pinus strobus) dominance on a floating kettle peatland in north-western Pennsylvania (USA) coincided with enhanced nutrient availability due to an increase of mineral dust deposition that concurred with upland deforestation by European settlers (ca. 1900). Using similar techniques, Pellerin and Lavoie (2003) attributed the recent (20th century) tree expansion (Pinus banksiana and Picea mariana) in bogs of southeastern Québec (Canada) to the impact of combined fire and drainage disturbances associated with short dry climatic periods. In Europe, the recent pinewood development on bogs in the Jura Mountains has also been linked to artificial drainage and peat cutting (Freléchoux et al. 2000).

The aims of this study were 1) to reconstruct the recent dynamics of two bog zones located in a large wetland complex in southern Québec (Canada), and 2) to determine which factors influenced tree expansion on those sites. The vegetation and hydrological dynamics of the bogs were reconstructed using a combination of paleoecological, paleohydrological, and historical approaches. As the study sites are located in protected areas, and in accordance with Ireland and Booth (2012), it is hypothesized that upland deforestation indirectly induced peat sediment enrichment through an increase of atmospheric mineral dust deposition rich in nitrogen and phosphorous, triggering forest expansion on the sites.

Methods

Study Area

The Lanoraie Wetland Complex is located 40 km northeast of Montréal, on the north shore of the St. Lawrence River, southern Québec (Canada), and covers an area of 76 km2 (Fig. 1). Inherited from the St. Lawrence River paleochannels, the complex lies over low-permeability clayey silts of the Champlain Sea and is bordered by deltaic terrace sands (Rosa and Larocque 2008). Organic accumulation began around 7400 cal BP (Comtois 1982), and peat thickness ranges between 0.3 and 8.4 m (Rosa et al. 2009; Tousignant et al. 2010). The complex is mainly composed of forested fens dominated by Acer rubrum, Alnus incana subsp. rugosa and Ilex verticillata. Scattered areas of bogs are also present, and dominated by Picea mariana, Larix laricina, ericaceous shrubs (mostly Chamaedaphne calyculata, Kalmia angustifolia, Rhododendron groenlandicum and Vaccinium angustifolium) and Sphagnum mosses (mostly S. capillifolium, S. fallax and S. magellanicum). Swamps and riparian marshes occur along the four rivers that run through the complex (Fig. 1). The complex is part of a vast hydrosystem composed of five catchment areas where surface water and groundwater are intensively used for drinking water purposes and irrigation (Bourgault et al. 2014).

Fig. 1
figure 1

Location of the Lanoraie Wetland Complex (southern Québec, Canada) and of the two coring sites: Ecological Reserve (ER) and St. Jean (SJ)

Full size image

European settlement in the region began around 1630 (Morneau et al. 2012). However, the region remained sparsely populated until 1850, when agricultural activities began to increase. Land clearance peaked around 1940–50 with the mechanization of drainage and agricultural practices, and during this period, summer aeolian erosion events were frequent (Morneau et al. 2012). Today, less than 10 % of the lands surrounding the complex are forested. The region has a strong agricultural vocation with approximately 85 % of the territory used for agriculture, mainly fruit and vegetable production.

The lands surrounding the wetland complex have been transformed by human activities, and the complex itself has also been highly disturbed, mainly in the forested swamps at its margins. Since the beginning of European settlement, the complex has been used for wood gathering, especially Thuja occidentalis trees, which were used for posts and fences (Tousignant et al. 2010). Over the last 100 years, approximately 28 % of the complex’s area has been converted to agricultural use or housing (Tardy and Pellerin 2006). Moreover, it is now partitioned by several roads, highways, railroads and power line rights-of-way, while about 200 irrigation reservoirs and drainage ditches have been built in and around it. According to a recent study of its present-day vegetation patterns (Tousignant et al. 2010), anthropogenic disturbances (agriculture, drainage, logging, transportation infrastructures) had detrimental effects on the richness of bryophyte and peatland species (i.e., species that occur always, preferably or frequently in peatlands), but favoured grasses as well as non-peatland and exotic species diversity. In 1994, approximately 4 km2 were set aside to ensure long-term protection, with the creation of the Tourbières-de-Lanoraie Ecological Reserve by the Government of Québec (Fig. 1). Additional areas have been protected since then by a local non-governmental conservation organization (Fig. 1).

The regional climate is temperate and continental. The mean annual air temperature (1915–2010) at the nearest meteorological station (Joliette; 15 km) is 5.5 °C, and varies from a minimum of 3.2 °C in 1926 to a maximum of 7.9 in 2010 (see Online resource 1 for complete climatic data). Mean annual precipitation is about 910 mm (20 % of which falls as snow), and varies from a minimum of 567 mm (1953) to a maximum of 1346 mm (1981). The mean annual potential evapotranspiration is 588 mm and varies from 528 mm (1926) to 642 mm (2010). Mean temperature, total precipitation and potential evapotranspiration time series show a statistically significant increase trend (Mann-Kendall test; p < 0.001).

Coring Sites and Peat Extraction

In 2010, two 1-m long cores were collected using a Box corer (100 × 8 × 8 cm; Jeglum et al. 1992) in two protected ombrotrophic sectors of the complex (Fig. 1). Both cores were extracted from a small hollow in a Sphagnum-dominated environment surrounded by a dense cover of ericaceous shrubs and a semi-open forest canopy of Larix laricina and Picea mariana. Following extraction, the sediments were wrapped in cellophane and aluminium foil for transport to the laboratory, where they were stored at 4 °C until analysis.

The Ecological Reserve coring site (ER; 45°59′26.5″N, 73°18′00″W) is located in the Point-du-Jour River catchment near the junction with the St. Joseph River catchment, and about 240 m south of a railway constructed between 1847 and 1850, and abandoned in 1894. A large (~1.5 m), deep (~2 m) artificial drainage ditch was dug in the southwestern part of the sector (about 1000 m from the coring site) at the end of the 1960s (Fig. 1). The St. Jean coring site (SJ) is located in the headwaters of the St. Jean River catchment (45°59′46.6″N, 73°13′22.8″W), and about 30 m from an unpaved road and a large (~1.5 m), deep (~2 m) artificial drainage ditch dug in the 1980s (Fig. 1).

Chronology

For each core, one sample (consisting of a few mg dry weight of bulk sediments) from 46 cm (ER site) and 50 cm (SJ site) was submitted for accelerator mass spectrometry (AMS) radiocarbon dating (Table 1). One sample was pre-treated and prepared at Laval University’s 14C Laboratory and dated at the Keck-Carbon Cycle AMS facility (University of California, Irvine). The other sample was dated at the Beta Analytic laboratory (Florida). The two radiocarbon dates (14C yr. BP) were calibrated (cal yr. BC/AD) using the CALIB 6.0.1 program (Stuiver and Reimer 1993) and the INTCAL09 dataset (Reimer et al. 2009). Calibrated dates were rounded to the nearest 10 years using a 2-sigma cal age range. A 210Pb chronology was also obtained for each core at 4 cm-intervals from the surface to 50 cm downcore. Measurements were performed on an alpha-spectrometer at GEOTOP Research Center (Université du Québec à Montréal, Montréal). For each level, a dry 0.5 g subsample was analyzed for 210Pb activity after spiking with a 209Po yield tracer. The constant rate of supply model was applied to calculate age (Appleby and Oldfield 1978; Appleby 2004). Results are presented in calibrated years (years AD; Table 2).

Table 1 Radiocarbon dates of the cores collected from the Lanoraie Wetland Complex
Full size table
Table 2 210Pb concentrations and dates from the upper part of the peat cores collected from the Lanoraie Wetland Complex
Full size table

Macrofossil Analysis

First, both cores were cleaned and cut into contiguous 1-cm thick slices. Since the aim of our study was to reconstruct vegetation changes following the onset of anthropogenic activities in the vicinity of the wetland complex, analyses and results described in this paper are restricted to the upper 50 cm of each core (data and results from 50 to 100 cm are available in Boucheny 2014). Intervals of 1 cm were used for macrofossil analysis. For each level, a 20-cm3 subsample was prepared according to Bhiry and Filion (2001). Macroremains were separated from the organic matrix by boiling the material for about 3 min in a weak 5 % KOH solution to induce deflocculation. The material was then wet-screened through a series of sieves of 0.850, 0.425 and 0.180 mm mesh. The percentage of volume occupied by each of the different botanical groups (wood remains, roots and rootlets, herbaceous plants, brown mosses, Sphagnum, ericaceous plants) was first estimated. Then, remains of vascular plants were picked out, identified and counted using a stereomicroscope at 4 to 40× magnification (results were expressed in number of macrofossils per 20 cm3 of sediments).

Testate Amoebae Analysis and Water Table Level Reconstruction

Subsamples of 2 cm3 were taken for testate amoebae analysis at 2 cm-intervals. Standard methods were used to isolate testate amoebae from the organic matter (Hendon and Charman 1997; Charman et al. 2000; Booth et al. 2010). Subsamples were boiled in distilled water (10 min) and then screened through a series of sieves with 0.350 and 0.150 mm mesh. Residual material was stained, mounted on glass slides and analyzed under a microscope (400× magnification). For each level, at least 150 tests (shells) were identified and counted (Payne and Mitchell 2009). Test identification was performed using the Charman et al. (2000) identification key with the modifications suggested by Booth and Sullivan (2007). The relative abundance of each taxon was calculated as a percentage of the total count. Past water table depths (cm) were inferred using a transfer function developed by Lamarre et al. (2013). Diagrams were built with the most abundant observed species throughout both cores.

Organic Matter, Nitrogen and Phosphorus Analyses

Subsamples (2 cm3) were taken from both cores at regular intervals (2 cm) and analyzed for percent-weight organic matter by loss-on-ignition (550 °C; Heiri et al. 2001). To measure nitrogen (N) and phosphorus (P) contents, supplemental subsamples of 1- cm3 were collected. Subsamples were subjected to a wet digestion procedure based on the method developed by Parkinson and Allen (1975). Total N concentrations were quantified using flow injection analysis (FIA QuikChem4000, Lachat Instruments Division, Milwaukee, USA), and total P with an inductively coupled plasma optimal emission spectrometer (Optima 4300 DV ICP, PerkinElmer Instruments, Norwalk, USA). Analyses were performed at Laval University’s Laboratoire de pédologie forestière.

Aerial Photograph Analyses

To assess recent tree encroachment history on the two studied bogs, we used grey-scale aerial photographs from1933, 1964, 1979, 1997 and 2012. Aerial photographs were selected based on cloud-free conditions and absence of distortion. All photos were digitized and georeferenced using QGIS (QGIS, Open Source Geospatial Foundation Project, Beaverton, OR, USA). Sectors surrounding the coring point that showed no obvious signs of extensive tree cutting since 1933 were first delineated (ER: 67.5 ha; SJ: 250 ha). Forested areas in these sectors were then manually delineated on each aerial photo based on colour, brightness, contrast and texture. Forested areas roughly correspond to those with more than 35 % coverage of tall trees (>2 m; S. Pellerin, unpublished data). Automatic methods, such as thresholding, were not suitable due to high variability in the background colour of the photos (especially with the oldest photos). Visual interpretation of vegetation structure was confirmed by stereoscopic viewing of all aerial photographs.

The perimeters of disturbed surfaces (agricultural lands, tree cutting areas, roads, etc.) and drainage ditches located within or bordering the peatlands were then digitized on each georeferenced photograph. The percentage of disturbed surfaces and ditch density (m/ha) within a radius of 500 m from each coring site were then calculated.

Results

Vegetation History and Human Disturbances

For the ER site, a radiocarbon date of 1860 AD (mid-point age) was obtained at a depth of 46 cm while 210Pb results indicated a date of 1907 AD at 24 cm. For the SJ site, a radiocarbon date of 1660 AD (mid-point age) was obtained at a depth of 50 cm while 210Pb results indicated a date of 1880 AD at 32 cm (Tables 1 and 2; Fig. 2).

Fig. 2
figure 2

Macrofossil diagrams (black bars = botanical groups expressed as percentage of organic matter volume; grey bars = number of vascular plant macrofossils/20 cm3, selected taxa only) for the Ecological Reserve (ER) and St. Jean (SJ) cores. Grey bands begin at 1850 and end at 1950, corresponding respectively to the onset of agricultural activities and maximum deforestation in the area (Morneau et al. 2012). Evolution of the tree cover on the sites according to aerial photograph analyses is also shown, along with percentage of disturbed areas and ditch density in a radius of 500 m of the coring point. The asterisks indicate the approximate date when the main ditches were dug. Chronological data are shown on the left axis (black squares =210Pb; black circles =14C)

Full size image

In both cores, Sphagnum remains were abundant throughout the record, suggesting ombrotrophic conditions (Fig. 2). Remains of ericaceous shrubs (mainly Chamaedaphne calyculata, Kalmia angustifolia and Rhododendron groenlandicum; data not shown) increased in abundance towards the top of both cores, with a similar pattern for Larix laricina needles in the SJ core and brown mosses (mainly Polytrichum strictum) in the ER core. The percentage of volume occupied by roots and rootlets (mainly monocots roots) was quite high at the bottom of the ER core. This section of the core corresponds to the fen/bog transition (see Boucheny 2014 for complete macrofossil results). In the ER, the percentage of volume occupied by roots and rootlets (mainly dycots roots) peaked between 24 and 16 cm and between 10 and 8 cm. These peaks likely correspond to the rooting zones of present-day shrubs and trees and to those that established at the beginning of the 1990s.

The increase of shrub (both sites) and tree (SJ) remains concurred with the increase of tree cover (according to aerial photograph analyses) on both sites (Fig. 2). Widespread forest expansion has indeed occurred in both sectors since 1933 (Fig. 3). In the ER sector, the percentage of the area occupied by forest increased from 4 to 45 % between 1933 (year of the older aerial photograph) and 2012 (the most recent photograph), which represents an overall gain of 27 ha of forest habitat. The tree cover increase accelerated after 1979. During the same time period, disturbed areas have increased from 10 to 30 %, and artificial drainage density from 4.3 to 6.3 m/ha (Fig. 2). In the SJ sector, the percentage of the area occupied by forest increased from 44 to 95 % between 1933 and 2012, representing an overall gain of 143 ha of forest habitat (Fig. 3). Disturbed areas have increased from 32 to 71 %, and artificial drainage density from 7.8 to 9.0 m/ha (Fig 2).

Fig 3
figure 3

Spatiotemporal evolution of the forest cover (grey areas) of the Ecological Reserve (ER) and St. Jean (SJ) sectors of the Lanoraie Wetland Complex (sites are not to scale), reconstructed from aerial photographs. Percentage of the bog area covered by forest is indicated for each year

Full size image

Inferred Water Table Depths

Testate amoebae were identified throughout both cores (Fig. 4). As living tests were found in the uppermost 10 cm, their assemblages have not been interpreted as a record of past hydrological changes.

Fig. 4
figure 4

Testate amoebae diagrams (selected taxa only) and inferred water table depth (WTD ± standard error [SE]) for the Ecological Reserve (ER) and St. Jean (SJ) cores. Living tests were found in the uppermost 10 cm of both cores; their assemblages have thus not been interpreted as a record of past hydrological changes. Grey bands begin at 1850 and end at 1950, corresponding respectively to the onset of agricultural activities and maximum deforestation in the area (Morneau et al. 2012). The asterisks indicate the approximate date when the main ditches were dug. Chronological data are shown on the left axis (black squares =210Pb; black circles =14C)

Full size image

At ER, a gradual rise of the water table from 29 cm (ca. 1840 at 50 cm depth) to 13 cm (ca. 1900 at 28 cm depth) was inferred with the decline of Hyalosphenia subflava and Cyclopyxis arcelloides. This was followed by a relatively stable water table (mean 20 ± 0.8 cm) recorded between 28 and 14 cm depths (ca. 1900 to 1970) and characterized by the co-dominance of Difflugia pristis, D. pulex, Hyalosphenia subflava and Nebela militaris. Finally, a decline of the water table from 21 to 35 cm was registered between14 cm (1970) and 10 cm (1985) depths, associated with the decline of D. pristis, D. pulex and N. militaris, and the increase of Trinema-Corythion, Assulina muscorum and C. arcelloides. The mean inferred WTD was 32 ± 3.7 cm during this last period.

At SJ, the inferred water table was relatively stable (15 ± 2.8 cm) between 50 and 20 cm depths (ca. 1660 to 1946). This period is characterized by a high abundance of Difflugia pulex, a species with a wide tolerance to surface wetness level (Lamarre et al. 2013), and other species associated to similar conditions (e.g., Archerella flavum, Cyclopyxis arcelloides, Pseudodifflugia fulva). This period was followed by a decrease of the water table from 20 to 10 cm depths (1946 to1988) inferred by the abundance of Hyalosphenia subflava and the co-occurrence of Nebela militaris, Heleopera sylvatica, Trigonopysis arcula and D. pristis.

Organic Matter, Nitrogen and Phosphorus Contents

Organic matter content was high (> 94 %) and quite stable along both cores (Fig. 5). Concentrations of peat N ranged from 0.99 to 2.03 % (mean: 1.34 %) in the ER core and from 0.89 to 2.07 % (mean: 1.27 %) in the SJ core. In the ER core, concentrations of N were lower than the mean, from the top of the core to 37 cm depth (ca. 1880) with the exception of a small increase at 9 cm depth. In the SJ core, concentrations of N were lower than the mean from the top of the core to 17 cm depth (ca. 1970) as well as from 31 to 36 cm depth (ca. 1886 to 1835). In the ER core, P concentrations remained relatively stable from the bottom up to 20 cm (ca. 1943) ranging from 0.06 to 0.09 % (mean: 0.08 %), after which they increased up to the surface (ranging from 0.10 to 0.57 %; mean: 0.18 %). Concentrations of peat P ranged from 0.04 to 0.08 % (mean: 0.06 %) in the SJ core and were stable throughout.

Fig. 5
figure 5

Changes in organic matter, nitrogen and phosphorus contents (dashed lines = mean level) through time for the Ecological Reserve (ER) and St. Jean (SJ) cores. Grey bands begin at 1850 and end at 1950, corresponding respectively to the onset of agricultural activities and maximum deforestation in the area (Morneau et al. 2012). Chronological data are shown on the left axis (black squares =210Pb; black circles =14C)

Full size image

Discussion

Disentangling drivers of vegetation changes in peatlands during recent centuries is a challenge, as natural drivers can be confounded with in situ and regional anthropogenic factors. A multi-proxy approach, as used in this study, is expected to facilitate the task by making it possible to explore several drivers acting at different temporal and spatial scales (Pellerin and Lavoie 2003; van der Linden and Van Geel 2006; Lamentowicz et al. 2008; Ireland and Booth 2012). Overall, the paleoecological and historical analyses showed that tree encroachment on both sites began recently (last 100–50 years) but was not synchronous. This suggests that the factors driving the phenomenon are different at ER and SJ, even though they are situated in proximity to each other (5.7 km). This general result concurs with a previous study conducted in 16 bogs of southeastern Québec, that concluded that local historical factors of each site (e.g., fire, drainage) override regional factors (e.g. climate anomalies, upland deforestation) in explaining rapid and abrupt changes (Pellerin and Lavoie 2003).

In a recent Pennsylvania study, Ireland and Booth (2012) showed that upland deforestation by European settlers indirectly induced enrichment of organic deposits due to increased deposition of atmospheric mineral dust rich in nitrogen, phosphorus and potassium and likely caused tree encroachment in a kettle peatland. It was hypothesized that a similar phenomenon occurred on the studied sites, as they were both affected by a very low level of direct human disturbances. However, there was no evidence to support this in the present study. Nitrogen content was lower in both sites following the onset of upland deforestation (1850), while phosphorus content remained relatively stable at the SJ site. At the ER site, a sharp increase in phosphorus content was only registered at the top of the core (ca. 1995–2000), i.e., much later than the peak of upland deforestation. This increase is likely related with the conversion of agricultural practices from tobacco to potato in the vicinity of the ER site (Morneau et al. 2012). Phosphorus loss (through runoff or soil aeolian erosion) in areas of potato production is indeed a widespread concern, because potatoes are a high P-demanding crop that uses applied P inefficiently, leading to frequent over-application of P to crop fields (Ruark et al. 2014).

Trees or shrubs encroachment in peatlands has often been ascribed to warmer and/or drier climatic periods (e.g., Birks 1975; Pilcher et al. 1995; Berg et al. 2009). Such conditions are known to favor the lowering of the water table and facilitate the survival and growth of trees in waterlogged ecosystems. In SJ, the beginning of the water level drawdown and of the forest expansion (ca 1946) occurred during a period of low annual precipitation and of high potential evapotranspiration (Online resource 1). In ER, the beginning of the water table drawdown (ca. 1970) and of forest expansion (ca. 1979) occurred rather during a period of high annual precipitation and of average level of potential evapotranspiration. Although we cannot rule out that the climate have contributed to enhance tree expansion, the fact that the tree encroachment and water level drawdown were not synchronous on both studied sites despite close proximity, suggest that climate was likely not the main factor initiating the phenomenon. A more detailed climatic analysis combined with tree rings analyses would be needed to evaluate adequately the influence of climate on the observed changes. An alternative explanation for tree encroachment could be related to gradual drying of the peat deposits associated with in situ drainage or by the development of agriculture and associated drainage network on mineral soils in the surrounding catchment. Several paleoecological and ecological studies have indeed shown increased tree or shrub density following the establishment of drier conditions induced by artificial drainage (e.g., Pellerin and Lavoie 2003; Linderholm and Leine 2004; Talbot et al. 2010; Pasquet et al. 2015). The ditch density measured on both study sites of the Lanoraie Wetland Complex is quite low (<10 m/ha), and much lower than the level of 200 m/ha recommended to spur tree growth in bogs (Roy et al. 2000). However, on both study sites, a decrease of the water table depth (inferred from testate amoebae) coincides with the beginning of forest encroachment. At the SJ site, the water table began to decrease around 1946, i.e. three decades before the drainage ditches in the bog were dug (1980s). However, this water level drawdown matches the peak of agricultural activities and ditch digging mechanization in the area (Morneau et al. 2012). This suggests that the main driver of the tree encroachment in this site was the gradual drying of the peatland surface due to a change in the regional hydrology associated with an intensification of modern agricultural activities on adjacent mineral soils. In situ ditches and climate warming/drying may have facilitated further tree expansion and eventually impeded the return to previous open Sphagnum-dominated state. At the ER site, forest cover expansion (ca. 1979) occurred much later than the peak of agriculture (ca. 1950), but only ca. 15 years after a large, deep artificial drainage ditch was dug in the southwestern part of the sector (Fig. 2). This lag is likely associated to the time needed for a tree to be distinguishable on an aerial photograph following its establishment. Consequently, human-induced drainage likely triggered tree encroachment in the sector, an assertion supported in part by the fact that forest expansion began near the ditch (see Fig 1).

Conclusion

This multi-proxy approach reveals that the recent tree encroachment in the two studied bogs was triggered by a drying of the bog surface likely caused by a combination of both in situ and regional anthropogenic drainage. Although climate warming/drying could have contributed to the tree encroachment, its effect does not appear to be dominating. Contrary to the prediction, upland deforestation seems to have had no or little effect on woody encroachment, as no evidence was found of nutrient enrichment concurrent with this transformation of the landscape. Whatever the underlying causes of the forest expansion in the two studied bogs, the phenomenon is expected to cause deleterious effects on the diversity and functioning of these protected sites. For instance, a recent study on the flora of these sites showed that forested bog habitats were characterized by fewer true ombrotrophic species, such as Sphagnum magellanicum, S. angustifolium and Kalmia angustifolia and more minerotrophic, non-peatland or exotic species, such as Calamagrostis canadensis, Osmundastrum cinnamomeum, Oxalis stricta and Solidago rugosa, than open bog sectors (Tousignant et al. 2010). Moreover, the observed plant succession toward a globally more forested system may create conditions that will hamper peat accumulation, notably through the decrease in Sphagnum growth under forest cover (Eppinga et al. 2009). Finally, this study shows that the vegetation dynamics of even preserved bogs located in humanized landscapes are sensitive to activities in the surrounding catchments, but that the local history of disturbances seems more influential than the regional setting.