Tectonic drivers for vegetation patterning and landscape evolution in the Albany River region of the Hudson Bay Lowlands


Paul H. Glaser (e-mail glase001@umn.edu).


  • 1Groundwater–peatland interactions were assessed by a regional survey in the Hudson Bay Lowlands, where the rapid rate of isostatic uplift has perturbed hydrological flow systems across a 6000-year chronosequence.
  • 2A 24 000 km2 study area along the Albany River consists of 55% fen, 35% bog and 10% mineral soil. The peatland vegetation may be further subdivided into 11 noda, which are closely related to different water levels, ranges in water chemistry, and peat landform type. Species richness generally declines with increasing water level and acidity, whereas the gradient from bog to extremely rich fen is marked by the changing abundance and occurrence of fen-indicator species.
  • 3Bog landforms are restricted to physiographic settings where surface waters flow downwards and the bog vegetation is therefore isotated from the influence of geogenous waters. In contrast, fens are located in areas where mineral solutes are transported to the peat surface either by upwelling groundwater or by advective/dispersion along lateral flow paths.
  • 4Peatlands spread across the study area between 6000 and 3000 bp, coinciding with the emergence of new land from the sea. The release of organic acids from the nearly continuous peat cover acidified this calcareous landscape, leading to the convergence of the surface-water chemistry into four discrete groupings of pH vs. calcium.
  • 5Isostatic uplift, however, continues to alter the topography, fluvial geometry and groundwater flow systems of the lowlands, maintaining diverse peatland types on land surfaces of similar age. The formation of water-table mounds under the interfluvial divides and rising moraine system spurred the development of raised bogs, whereas the formation of regional seepage faces for goundwater on the margins of the moraine and rivers of the till plain maintains large areas of fen.
  • 6Although peatland succession seems to follow predictable pathways within a given hydrogeological setting, these pathways are locally altered by tectonic drivers that continually modify surface and groundwater flow systems. In this large peat basin the pace and pathway of peatland succession seems to be driven by tectonic rather than climatic forcing.


Four physiographic regions enclose the largest peat basins of the northern hemisphere. Here low relief, thick glacial deposits, and a cool moist climate combined to form an interlocking web of peatlands that blanket the landscape. The largest of these peatland regions are the West Siberian (> 500 000 km2) and Pripyat (100 000 km2) basins of Eurasia (Neustadt 1984) and the Hudson Bay Lowlands (320 000 km2) and Glacial Lake Agassiz basins of North America (Zoltai & Pollet 1983). Despite their global significance as major carbon reservoirs, the dynamics of these wetland ecosystems remains to be established.

Recently, groundwater was identified as a major determinant for peatland development in the Glacial Lake Agassiz peatlands of northern Minnesota (Siegel 1983; Romanowicz et al. 1993; Glaser et al. 1997). Regional surveys determined that the bogs, fens and spring fens of this region are closely adjusted to climate-driven flow systems and the local hydrogeological setting (Siegel et al. 1995; Glaser et al. 1997). This unexpected relationship was clarified by observations during the regional drought cycle, which simulates an experimental manipulation of groundwater flow systems. The prominent role of groundwater drivers in the Lake Agassiz basin raises questions about the general applicability of these findings to other large peat basins. The groundwater hypothesis was therefore tested in another large peat basin, where an unusual environmental setting should elucidate different aspects of this relationship.

The Hudson Bay Lowlands have been rising isostatically since deglaciation at a rate matched only by that of northern Fennoscandia (Andrews & Peltier 1989). Isostatic uplift during the past 8000 years created a regional chronosequence as new land emerged from the sea and altered flow systems that continually adjust to the changes in base level, topography and drainage networks. The response of peatlands to this dynamically changing landscape is manifested by the highly developed vegetation/landform patterns that are visible on remote sensing imagery at an array of scales. We therefore selected a 24 000 km2 study area along the lower Albany River (51–52° N and 82–84° W) to investigate the role of these peat landforms as indicators for the vegetation, hydrogeology and developmental pathways, This area spans a mature section of the chronosequence with an array of peatland patterns typical of those described from northern, southern and maritime areas of boreal America by Glaser & Janssens (1986). This area lacks permafrost but still remains close to the centre for isostatic uplift near the mouth of James Bay.

Materials and methods


The entire field area was first surveyed from aerial photographs and Landsat imagery to determine the relationship of the vegetation patterns to the physiographic setting. Peat landforms were distinguished by the spectral signatures produced by the vegetation assemblages and water levels, and also by their surface patterns, which indicate landform margins and topography. The classification scheme was checked by the various methods described below. The relative abundance of bog landforms, fen landforms and mineral soil/standing water was later determined from the Landsat image (Fig. 1) by merging the 15 spectral classes identified from an unsupervised classification of field-control sites into a supervised classification of data from the entire study area. A digital elevation model (DEM) was constructed for the study area using SRTM (shuttle radar topography mission) data and visualization software. Local peat slopes were determined by standard surveying techniques following Almendinger et al. (1986). Peat cores were collected from 14 representative landform types with a piston corer 10 cm in diameter, equipped with a serrated cutting edge (Wright et al. 1984). Peat samples were dated by radiocarbon by Beta-Analytic Inc (Miami, Florida, USA).

Figure 1.

Vegetation/landform units of the Albany River study area (51°18′–51°51′N, 82°10′–83 55′ W). The different types of peat landforms are distinguished by their spectral properties on this false-colour composite Landsat image (scene ID L4TM8447 4006215332; 9/16/82). The long white arrows mark regional seepage faces for groundwater. Drainage is toward the northeast. The image covers an area 100 km across.
Fig. 2 Digital elevation model of the study area. Elevation (30-120 m) has been exaggerated to emphasize the fluted moraine (blue) and till plain (green). Other features labelled are the gorge of the Albany River (white arrows) and former channels of the Albany and Kenogomi Rivers (black arrows). This image covers the same area shown in Figure 1, but the orientation is reversed to reveal the edge of the moraine.


A helicopter provided comprehensive access to all the different peat-landform sites within a 160-km radius of a base camp along the Albany River in July of 1985 and 1992. The vegetation associated with each landform type was determined by the relevé method using the methods previously described by Glaser et al. (1981). The relevés were evenly distributed across the study area on representative landform types to capture a major portion of the variability observed in the field. A plot of 10 × 10 m was used for non-forested stands, whereas a larger 20 × 20 m plot was laid out in forested stands (Westhoff & van der Maarel 1973). The relevés were arranged in tables partly following Shimwell (1971) and also by using detrended correspondence analysis (DCA) ordinations (Hill l979; Jongman et al. 1987; ter Braak 1987). Separate DCA ordinations were made for the bog and fen relevés to avoid clumping of the bog samples, which are species-poor. All samples were included in the ordination, and the individual species were weighted according to their cover value. The vegetation tables were then condensed into summary tables for publication. Nomenclature follows Fernald (1970) for vascular plants, Anderson et al. (1990) for mosses, Anderson (1990) for Sphagnumand Stotler & Crandell-Stotler (1977) for liverworts.


A water sample was collected from each relevé plot in an acid-washed 250-mL plastic bottle. The samples were analysed for pH with a field pH meter at the end of each field day. The samples were then filtered and acidified for later analysis of the major cations by direct-current plasma-emission spectrometry. Piezometer nests were installed on a representative set of peat landforms across the study area using methods described by Siegel & Glaser (1987). The piezometers were generally inserted to depths of 1, 2 and 3 m to measure hydraulic head and to extract pore-water samples. The difference in the water levels within the piezometers at each site indicates whether a site is in a groundwater recharge or discharge zone. In recharge areas, the deeper a piezometer is placed below the water table, the lower the water level will be below the water-table elevation. Conversely, in upwelling areas, the deeper the piezometer, the higher the water level above the water table. The absence of head gradients with depth indicates lateral flow.

The pore waters collected from piezometers were analysed for alkalinity in open containers from filtered (0.2 µm) untreated samples by strong acid titration using the Gran method (Stumm & Morgan 1981). Raw samples for analysis of acid anions were collected and stored in a cooler, whereas those for cations were filtered and acidified in the field. Metal cations were determined by direct-current emission spectroscopy, acid anions with an ion chromatograph, and DIC (dissolved inorganic carbon) and DOC (dissolved organic carbon) with a carbon analyser. DIC was measured after acidification and a helium purge, whereas DOC was measured after filtration through a 0.45-µm glass filter, acidification and oxygenation promoted by ultraviolet persulphate. Carbonate species were calculated from alkalinity and pH. Metal complexation was not considered in the calculation, as samples were typically dilute and had low pH. The Oliver et al. (1983) model was used to estimate the charge-balance error if a significant organic anion was present (CBE-OM) and the amount of organic acid required to balance charge (OM-A). Charge-balance errors were calculated from the measured mineral ions, assuming that the contribution of other ions is not significant. This mineral-ion balance error (CBE-NoA)) was calculated from the following equation:

%IBE = (100) (Σ Cations − Σ Anions)/ (Σ Cations + Σ Anions)


physiographic setting

The Albany River study area spans two distinct physiographic units: a till plain to the north and a fluted moraine to the south-east (Figs 1–9). The till plain has a broad nearly level surface that gently dips to the north-east with a gradient of only 4 mm km−1. A dense network of parallel flowing streams divides the surface of the plain into a series of elongated interfluves that are very subtly mounded. The plain is also cut by a series of narrow U-shaped river valleys with underfit streams and by the Albany River, which flows in a gorge 30 m deep. The till plain has only a few lakes and an extensive cover of clayey silt under the peat.

Figure 3.

Aerial photograph of forested and non-forested bogs near Blackbear Island (51°46′ N, 83°8′ W). Note the streamlined margins of the bog landforms (lighter tones) that have either a forested crest (1) or non-forested plateau (2) with intricate pool patterns. The fen water tracks (darker tones) are marked by open stands of Larix laricina (3) or systems of pools and peat ridges (4). Arrows show the direction of drainage towards the tributary streams (5). The entire area is covered by peat except for mineral outcrops along the streams. The photographs measure 6.9 km across.

Figure 4.

Aerial photograph of bogs and water tracks near Belec Lakes (51°37′ N, 82°20′ W). The large bogs (lighter tones) contain concentric pool systems (1) or are fragmented into ovoid lobes by internal water tracks (2). The outer margins of the bog landforms are trimmed by fen water tracks (3, darker tones) or tributary streams (4). Reticulate fens occur downslope from some of the bogs (5). The arrows indicate directions of drainage. The photograph measures 12.8 km across.

Figure 5.

Spring-fen channels and minor moraines near Cheepay Island. The spring-fen channels (1) originate along the flanks of a moraine and drain downslope through a series of minor moraines (2) into a featureless swamp forest (3). The summit of the moraine is covered by non-forested bogs (4). Note the outcrops of mineral soil (5) along the banks of the Albany River and tributary streams. Direction of flow is indicated by a white arrow. The photograph covers an area 5.9 km across.

Figure 6.

A large bog complex near the headwaters of the North Wabassie River. The raised bog (light tones) has a central forested crest (1) from which water drains centripetally (white arrows) towards the nearest stream (2). Fen water tracks (dark tones, black arrows) arise on the lower bog flanks fragmenting the bog into large ovoid lobes (1a) or small streamlined islands (3). The aerial photograph covers an area 12 km across that is almost entirely covered by peat.

Figure 7.

Origin of bog islands in water tracks. The raised bogs (light tones) are largely non-forested (1a) with forested strips (1b) at their edges and upslope margins. The bogs are trimmed by fen water tracks (dark tones) that contain either open stands of Larix laricina (2a) or networks of pools and peat ridges orientated perpendicular to the slope (2b). Local flooding occurs near flow obstructions producing a scalloped bog margin. Note wet (black) indundations that fragment the bog margin into small tree islands (black arrow). Surface runoff drains towards the lower end of the photograph. The aerial photograph measures 2.8 km across.

Figure 8.

Bog formation related to fluvial erosion. Raised bogs (1, lighter tones) are located within the interfluvial divides between parallel draining streams (2). The streams are eroding into a featureless forested fen (3). Note the bog crests with radiating lines of forest (black arrow) and the many small bogs (4) forming between the smallest tributaries. Runoff drains towards the upper portion of the photograph. The aerial photograph measures 6.8 km across.

Figure 9.

Bogs and kettle-hole lakes east of Jaab Lake. The summit of the moraine separating the Albany and Moose River drainages is marked by numerous small non-forested bogs (1), narrow fen water tracks (2), and sinuous outcrops of glacial till (3). The kettle-hole lakes contain lake sediments. The image covers an area 12 km across.

The southern edge of the till plain is defined by a slightly mounded moraine, which rises to a summit east of Jaab Lake. The moraine has a fluted appearance on the SRTM DEM, with tapered fingers grading out onto the till plain. The moraine is marked by large numbers of kettle lakes (Fig. 9) filled with lake sediments and also by flights of minor moraines (Figs 1, 2 and 5) along its northern flanks. The moraine has a combination of parallel and dendritic drainage, with numerous exposures of fine sand, gravel and silty till with numerous stones. Both the moraine and the till plain have a subtle fluted topography, which probably served as the template for the dense parallel drainage systems (cf. Hack 1965).

Approximately 90% of the study area is covered by thick deposits of peat 1–4.5 m thick. The peat cover is deepest under the raised bogs and shallowest under the spring fens (Table 1). The basal peat dates generally are older with increasing distance from the coast and/or increasing elevation on the moraine. These dates range in age from 3900 bp at Belec Lake, which is about 80 km from the coast, to 5920 bp at Oldman Bog, which is 250 km from the coast (Table 1).

Table 1.  Basal radiocarbon dates and hydrogeology from the Albany River area. The symbols indicate bog crest (BC), bog plateau (BP), internal water track (INT), fen swamp forest (SWF), spring-fen channel (SFC), reticulale fen (RF), slight recharge (S), underpressure (UP) or overpressure (OP) in the middle of the profile, and probable lateral flow (P)
IDSite nameLatitudeLongitudeLandscape unitSample depth (cm)Peat depth (cm)Kilometres to coastMineral substratumLaboratory number14C age (BP)Hydrology
8501Albany River51°15′83°35′Till plain259–264266170Clayey siltBeta-445374810 ± 70
9201Albany River BP51°26′83°40′Till plain239170Clayey siltRecharge
8502Oldman51°04′84°30′Till plain441–444445250Clayey siltBeta-423815920 ± 90
8502Oldman51°01′84°34′Till plain460250Silty gravelRecharge
8507ABelec Lake51°37′82°17′Till plain231–235236 80Clayey siltBeta-545983960 ± 60Recharge
8508ABlackbear51°43′83°14′Till plain218–223223127Clayey siltBeta-459053730 ± 50
9204Sesi-bog BP51°15′83°20′Till plain200–205220172Clayey siltBeta-649095220 ± 80Recharge
9207Sesi-moraine − BP51°14′83°07′Moraine280156Silty/gravelRecharge (S)
9214Wabassie bog51°43′83°38′Till plain200–205209145Clayey siltBeta-649254500 ± 90
9214Wabassie bog − BC51°43′83°38′Till plain245145Clayey siltRecharge (S, UP)
9212Cheepay bog51°23′83°22′Moraine100–105127165Fine sand/gravelBeta-667363700 ± 70Recharge (S, UP)
9208Jaab Lake − BP51°12′83°53′Moraine240140Silty/gravelOP
9210BBelec Lake − INT51°37′82°17′Till plain103–108109 80Clayey siltBeta-667353840 ± 70
9205Sesi-bog − INT51°16′83°18′Till plain210–215234172Clayey siltBeta-667325260 ± 70
9205Sesi-bog (INT)51°16′83°18′Till plain287172Clayey siltRecharge (S)
8506ASesi-moraine − SWF51°14′83°02′Moraine136–138149156Clayey siltBeta-545945200 ± 60
9210ABelec Lake − INT51°37′82°17′Till plain 92–97 98 80Silty fine sandBeta-667334010 ± 80OP
9215Wabassie (INT)51°43′83°35′Till plain229140Clayey siltUP
9206Sesi-water track51°15′83°23′Till plain347177Clayey siltLateral (P)
9202Albany River (INT)51°26′83°37′Till plain257170Clayey siltLateral (P)
9203Albany River (INT)51°29′83°37′Till plain228170Clayey siltOP
9216Oldman bog (INT)51°01′84°34′Till plain338250Silty gravelRecharge
Spring fens
9211Cheepay SFC-151°23′83°22′Moraine100–105127161Fine sand/gravelBeta-649204550 ± 70Discharge
9213Cheepay SFC-251°25′83°22′Moraine 98–103104160Fine sand/gravelBeta-649213910 ± 80Discharge
9209Belec Lake RF51°38′82°25′Till plain187 82Sand/gravelDischarge
8505ASesi-moraine − SFC51°37′82°17′Moraine162–166166148Clayey silt/gravelBeta-530645370 ± 80

peat landforms

Field surveys determined that the surface patterns visible on remote sensing imagery are closely related to changes in water level, water chemistry and species assemblages. In addition these surface patterns are consistently aligned to local slopes as determined by the DEM or topographic surveys, indicating that these patterns also indicate landform morphology. These close relationships are reflected in the various peat landform classifications (e.g. Sjörs 1948, 1963; Heinselman 1970; Glaser 1987a, 1992) most applicable to the Albany River region. A classification of the Landsat imagery (Plate 1) determined that the study area is composed of 55% fen landforms, 35% bog landforms and 10% mineral soil.

Raised-bog landforms

Raised-bog landforms have distinctive spectral signatures on remote-sensing imagery and surface patterns that indicate centripetal drainage from an inner crest, dome or plateau to the lower bog margins (Figs 1–9). Bog crests are marked by forested fingers that radiate downslope onto more gently sloping flanks (sensuHeinselman 1970; Glaser 1987a), whereas bog plateaux are often fringed by stunted tree cover. In both cases the forest patterns are associated with steeper bog slopes (Fig. 10), whereas gentler slopes are marked by lawns (sensuGlaser et al. 1981) or networks of hummocks, hollows and shallow pools aligned perpendicular to the slope (sensuSjörs 1948; Sjörs 1950). All bogs larger than 20 km2 are dissected by narrow strips of fen vegetation that widen downslope and divide the lower bog flanks into streamlined lobes or ovoid-shaped islands (sensuGlaser et al. 1981; Glaser 1987a,b). These islands have rounded heads facing upslope and tapered tails downslope. The largest raised bogs (> 200 km2) are found on the broader interfluves of the till plain where former channels of the Albany and Kenogomi rivers intersect (Figs 1 and 2). Many of these bogs have sharp streamlined margins trimmed by adjacent fens. Smaller bogs (1–20 km2) occur on the moraine or as fields of streamlined bog islands in fen water tracks on the till plain.

Figure 10.

Topographic survey of the Albany River bog (8501). The vegetation and surface features are closely related to peat slopes but vary independently from the underlying mineral topography. The slope at the bog margin descends 30 m to the Albany River gorge.

Water-track landforms

Fen landforms in contrast are distinguished by darker tones on remote-sensing imagery and by surface patterns indicative of a slightly concave topography (Figs 1–9). These landforms function as conduits for runoff across large peatlands and have therefore been called water tracks by Sjörs (1948, 1963) and Heinselman (1970). The largest water tracks in the Albany River region form featureless plains covered by open swamp forests, although Landsat imagery shows striking vegetation banding aligned parallel to flow paths. These featureless water tracks (sensuGlaser 1987a, 1992) grade into more open areas where vegetation lines are conspicuously orientated parallel to the local slope or into highly patterned areas of sinuous pools (i.e. flarks) and peat ridges (i.e. strings) orientated perpendicular to the slope. These pool networks can be quite large and deep, with robust peat ridges supporting trees. Small narrow fen water-tracks with conspicuous pool networks originate within the interiors of all raised bogs larger than 20 km2. These features are called seeps by Sjörs (1963) or internal water tracks by (Glaser 1987a,b, 1992).

Spring-fen and reticulate-fen landforms

Spring-fen channels (sensuGlaser 1987a) are much smaller and narrower than fen water tracks (Figs 4, 5). They are distinguished as series of narrow non-forested channels that contain fields of very small forested peat islands (< 1 km2 in area), which appear as fine stippling on 1 : 56 000 aerial photographs. Some spring-fen channels grade downslope into water tracks with the pool and peat-ridge networks characteristic of a patterned fen. Reticulate fens, in contrast, have reticulate networks of pools separated by much wider peat ridges.

vegetation assemblages

The vegetation assemblages of the Albany River peatlands were divided into 11 discrete noda and one transitional grouping (bog-lawn transition) that are closely related to the different landform types (Fig. 11a, b). These vegetation noda were distinguished on the basis of their (i) floristic composition, (ii) cover values for dominant species, (iii) number of plant taxa, and (iv) the results of the DCA ordinations (see Table 2, Fig. 11a,b and Appendices S1 and S2 in Supplementary Material). Bog and fen assemblages are clearly distinguished by the presence/absence of fen-indicator species (sensuSjörs 1963) and total number of species occurring in a standard plot. Bogs consistently have less than 15 species of vascular plants and 15 species of bryophytes per relevé plot, whereas all fens with the exception of a few deep fen pools have more species. The most important fen-indicator species within the study area are Utricularia intermedia, Carex lasiocarpa, C. livida, C. chordorrhiza, Equisetum fluviatile, Betula pumila var. glandulifera and Juncus stygius. Although no species are solely restricted to bogs, different species of Sphagnum (e.g. S. fuscum, S. angustifolium, S. capillifolium, S. magellanicum and S. majus) are characteristic bog dominants.

Figure 11.

(a) Detrended correspondence analysis of the raised bog relevés. The ordination includes both bryophyte and vascular plant species with the species weighted according to their cover values. Poor fens are labelled PF. (b) Detrended correspondence analysis of the fen relevés.

Table 2.  (a) Summary phytosociological table for raised bogs in the Albany River peatlands. For each vegetation type the following information is listed: (1) number of relevés included in the sample, (2) ranges in water chemistry, (3) number of vascular plants or bryophytes in a relevé plot, and (4) species composition. Each species is assigned numbers for its frequency and (2) average cover value. The cover symbols are: 0.5, sparsely present; 1, plentiful but small cover value; 2, very numerous, or covering 1/20 of area; 3, any number of individuals covering 1/41/2 area; any number of individuals covering 1/23/4 area; 5, covering more than 3/4 area. (b) Summary phytosociological table for fens in the Albany River peatlands. Fen indicator species marked by an asterisk (*). The symbols are the same as in (a)
Releve numbern = 6n = 13n = 5n = 5n = 4
Vascular plant taxa/plot8 to 107 to 1313 to 177 to 125 to 7
Bryophyte taxa/plot9 to 136 to 156 to 154 to 73 to 5
Gaultheria hispidula0.660.5
Ledum groenlandicum1.
Kalmia angustifolia1.
Picea mariana1.
Rubus chamaemorus1.
Eriophorum spissum0.801.
Chamaedaphne calyculata0.662.
Kalmia polifolia0.500.
Vaccinium oxycoccos1.
Carex pauciflora0.500.
Smilacina trifolia0.800.
Drosera rotundifolia0.660.
Sarracenia purpurea0.
Larix laricina0.
Carex paupercula0.
Andromeda glaucophylla0.20.3110.90.60.4
Carex oligosperma0.10.04130.40.2
Scheuchzeria palustris0.
Carex limosa0.
Rhynchospora alba0.20.412.90.751.125
Scirpus cespitosus0.
Drosera anglica0.
Menyanthes trifoliata0.20.111.8
Utricularia cornuta0.
Vaccinium vitis-idaea
Ledum palustre0.080.04
Lophozia sp.0.330.3
Cephaloziella sp.0.830.80.20.2
Ptilidium ciliare0.660.70.20.4
Ptilidium pulcherrimum0.660.
Cephalozia connivens0.330.
Pohlia sp.0.330.
Pohlia nutans0.
Sphagnum lindbergii0.
Dicranum undulatum0.830.
Pleurozium schreberi1.
Mylia anomala1.
Sphagnum fuscum1.
Sphagnum angustifolium0.500.
Sphagnum capillifolium1.
Sphagnum magellanicum0.330.
Cladopodiella fluitans0.
Sphagnum tenerum0.
Warnstorfia fluitans0.
Odontoschisma denudatum0.20.2
Dicranum groenlandicum0.10.1
Cephalozia sp.0.20.2
Sphagnum balticum0.
Warnstorfia exannulatus0.
Sphagnum majus0.
Cephalozia connivens0.10.1
Sphagnum pulchrum0.
Sphagnum rubellum0.
Sphagnum papillosum0.
Sphagnum platyphyllum0.30.5
Number of relevesn = 9n = 3n = 3n = 3n = 5n = 2n = 4
Vascular plant taxa/plot13 to 2618 to 2724 to 3417 to 3514 to 3333 to 4813 to 30
Bryophyte taxa/plot2 to 105 to 69 to 169 to 2111 to 2122 to 2621 to 33
 FrequencyCoverFrequencyCoverFrequencyCoverFrequency CoverFrequency CoverFrequencyCoverFrequencyCover
  1. Additional Species: Vaccinium angustifolium, Carex rostrata (8549: +; 8557: 1), Sparganium minimum (8549: +; 8559: +), Juncus cf. canadensis (8532: +; 8533: +), Drosera sp. (8545: +), Typha latifolia (8557: +), Vaccinium vitis-idaea (8562: +), Calla palustris (8559: +; 8537: +), Juniperus communis (8559: +; 8555: +), Xyris montana (8543: 1), Rhynchospora fusca (8543: 2), Eriocaulon septangulare (8543: +), Eriophorum virginicum (8543: +), Drosera linearis (8543: +), Viola sp. (8543: +; 8537: +; 8563: +; 8541: +), Eriophorum tenellum (8533: +), Utricularia cornuta (8533: +), Tofieldia pusilla (8533: +; 8535: +), Rubus sp. (8555: +; 8546: +), Selaginella selaginoides (8528: +), Campanula aparinoides (8555: +), Graminaea (8555: +; 8541: +), Eriophorum sp. (8555: +; 8563: +; 8541: +), Eriophorum viridi-carinatum (8528: +; 8548: +), Carex prairea (8528: +; 8541: +; 8563: +), Lysimachia thyrsiflora (8563: +; 8537: +), Carex pauciflora (8563: +; 8561: +), Eriophorum chamissonis (8561: +; 8531: +), Carex cephalantha (8561: +), Salix sp. (8537: 1); 8537: +; 8548: +), Salix candida (8537: +; 8536: +), Carex diandra (8537: +; 8548: +). Bryum pseudotriquetrum (8545: 1; 8528: 1; 8537: 1; 8534: 1), Moerckia hibernica (8562: 1; 8528:1; 8537: 1), Sphagnum papillosum (8519: 5), Loeskynum badium (8543: 1; 8533: 1; 8550: 1), Sphagnum teres (8543: 1; 8534: 1), Marchantiaceae (8533: 1; 8540: 1), Fissidens osmundoides (8540: 1), 8528: 1; 8537: 1), Lophozia rutheana (8540: 1; 8528: 1), Sphagnum squarrosum (8533: 1), Bryum sp. (8540: 1), Cephaloziella sp. (8555: 1; 8534: 1; 8536: 1), Riccardia multifida (8528: 1; 8534: 1), Aneura pinguis (8528: 1), Hypnum pratense (8528: 1), Marchantia polymorpha (8528: 1), Sphagnum subnitens (8528: 1), Brachythecium sp. (8563: 1; 8536: 1), Dicarnum groenlandicum (8550: 1; 8561: 1; Sphagnum jughunianum var. pseudomolle (8561: 1), Campylium hispidulum (8561: 1), Sphagnum section Acutifolia (8518: 4), Pohlia nutans (8518: 1; 8531: 1), Meesia triqetra (8537: 1; 8548: 1), Plagiochila asplenoides (8537: 1), Calliergon giganteum (8537: 1), Dicranum sp. (8548: 1), Calliergon richardsonii (8548: 1; 8534: 1), Sphagnum wulfianum (8548: 1), Calliergon aftonianum (8548: 1), Onocophora wahlenbergii (8552: 1).

  2. Sphagnum balticum (8552: 1), Sphagnum fallax (8552: 1), Sphagnum rubellum (8552: 1), Calypogia sp. (8534: 1), Bazzania trilobata (8534: 1).

Utricularia cornuta0.
Drosera intermedia0.
Drosera anglica0.
Carex livida*
Rhynchospora alba1.
Utricularia intermedia*
Menyanthes trifoliata1.
Carex limosa1.
Sarracenia purpurea0.
Carex lasiocarpa*
Scirpus cespitosus0.
Scirpus hudsonianus*
Carex chordorrhiza*
Andromeda glaucophylla0.
Equisetum fluviatile*
Malaxis uniflora0.
Pogonia ophioglosoides0.
Myrica gale0.
Scheuchzeria palustris0.
Juncus stygius*
Muhlenbergia glomerata*
Galium labradoricum0.
Eleocharis compressa0.
Agrostis scabra*
Solidago uliginosa0.
Carex exilis0.
Tofieldia glutinosa*
Potentilla fruticosa*
Triglochin maritima*
Lobelia kalmii*
Calamagrostis inexpansa0.
Lycopodium innundatum0.
Eriophorum angustifolium0.
Thuja occidentalis*
Viola conspersa1.00.5
Rubus pubescens0.
Epilobium palustre0.30.110.5
Carex interior0.
Aster juniformis0.
Rhamnus alnifolia0.
Habenaria hyperborea/dilatata0.
Lonicera villosa0.
Alnus rugosa0.
Drosera rotundifolia0.
Rubus acaulis0.
Vaccinium oxycoccos0.
Kalmia polifolia0.
Larix laricina0.
Picea mariana0.
Ledum groenlandicum0.
Chamaedaphne calyculata0.
Betula pumila glandulifera*
Smilacina trifolia0.
Salix pedicellaris*
Potentilla palustris0.
Carex tenuiflora*
Gaultheria hispidula0.
Carex paupercula0.
Kalmia angustifolia0.
Linnea borealis10.50.30.1
Carex disperma0.
Carex gynocrates10.750.30.1
Rubus chamaemorus0.80.9
Sphagnum jensenii0.20.2
Sphagnum majus0.10.1
Calliergon trifarium*
Sphagnum contortum0.
Sphagnum subsecundum s.s.*
Cinclidium stygium0. 
Scorpidium scorpioides*
Campylium stellatum*
Warnstorfia exannulatus0.
Sphagnum warnstorfii*
Sphagnum centrale0.
Tomenthypnum nitens0.
Limprichtia revolvens0.
Aulacomnium palustre0.
Sphagnum angustifolium0.
Sphagnum magellanicum0.
Dicranum undulatum0.
Cladopodiella fluitans0.
Calliergon stramineum*
Sphagnum fuscum0.
Mylia anomala0.
Lophozia sp.
Polytrichum strictum0.
Pleurozium schreberi0.
Pohlia sp.
Scapania sp.
Hypnum lindbergii0.30.3110.30.3
Ptilidium pulcherrimum0.
Sphagnum capillifolium0.
Tomenthypnum falcifolium0.
Ptilidium ciliare0.
Cephalozia sp.
Sanionia uncinata0.
Rhizomnium pseudopunctatum0.
Hylocomium splendens110.81.0
Blepharostoma trichophyllum0.
Ptilium crista-castrensis0.
Lepidozia reptans0.50.5
Tetraplodon angustatus0.50.5
Dicranum flagellare0.50.5
Dicranum ontariense0.50.5
Ptilidium sp.0.30.3
Rhizomnium gracile0.30.3
Dicranum polysetum0.30.3
Dicranum sp.0.30.3
Kurzia setacea0.30.3
Lophocolea heterophylla0.30.3
Plagiothecium denticulatum0.30.3
Tetraphis pellucida0.30.3
Drepanocladus brevifolius0.

Raised bogs

Each of the five discrete bog noda has a narrow range of variability despite the wide distribution of each across the study area. Most of the variation in these assemblages is expressed along the first DCA axis, which corresponds to microtopographic changes in the elevation of the water table. In general, woody vascular plants decline in cover with rising water levels as non-woody vascular plants become more dominant. The five major groupings identified by the floristic tables and DCA ordination correspond to the microtopographic gradient of hummock, lawn, mud bottom and pool previously defined by Sjörs (1948, 1950), plus a forested type on steeper peat slopes.

The bog-forest nodum dominated by Picea mariana occurs on steeper, better-drained slopes along the outer margins and interior crests of bogs. The open stands of trees are up to 10 m high, with a continuous understorey of bog ericads and ground layers dominated by Sphagnum. On more level slopes the bog-forest nodum gives way to networks of hummocks, lawns, mud bottoms and pools (Fig. 10). The hummocks are low undulating ridges of peat with a firm surface covered by prostrate clumps of Picea, a higher cover of shrubs such as Chamaedaphne calyculata, and a field layer of Sphagnum and lichens. The water table is generally 10–20 cm below the hummock surface. In contrast, Carex oligosperma forms extensive lawns with a nearly continuous cover of Sphagnum sp. on softer peat surfaces, where the water table is near the peat surface. In sinuous depressions (i.e. mud bottoms) where the water table is at the peat surface, the sparse vegetation is dominated by Rhynchospora alba, Carex limosa and Cladopodiella fluitans. The Albany River bogs also have shallow bog pools generally less than 3 m deep with sparse stands of Carex limosa and Scheuchzeria palustris, although patches of Menyanthes trifoliate also occur in the deepest water.


The fen vegetation in the study area has a much larger number of plant taxa, but only a few of these species attain high cover values. Most of the variation in the fen assemblages is again expressed along the first DCA axis, which corresponds to water-table elevations, whereas the second DCA axis corresponds to water-chemical gradients. In general, the least variable assemblages are associated with fen pools (i.e. flarks), whereas the most variable are on peat ridges (i.e. strings) or forested stands. Dominance shifts from trees to shrubs to sedges with increasing elevation of the water level.

The pools (cf. flarks of Sjörs 1963) of the water tracks and spring fens are dominated by sedges (e.g. Carex limosa and Rhynchospora alba), including many fen indicators (e.g. C. lasiocarpa and C. livida), with submerged carpets of Scorpidium scorpiodes. These pools typically have 10–20 cm of standing water, but with increasing water depth the vegetation becomes sparser and ultimately disappears. A few water tracks have reticulate networks of pools with a different assemblage of sedges (e.g. Carex exilis, Scirpus hudsonianus and S. cespitosus) and a large group of rare species (e.g. Eriocaulon septangulare, Xyris montana and Rhynchospora fusca). This reticulate pool nodum resembles that of the spring fen channels, which have shallower water depths and a large group of species, which are indicative of extremely rich fen (e.g. Thuja occidentalis, Tofieldia glutinosa and T. pusilla).

The most variable vegetation noda are associated with the strings and swamp forests. Strings are dominated by Carex exilis, Potentilla fruticosa and Myrica gale when these landforms occur downslope from spring-fen channels. However, in other water tracks strings are consistently dominated by Betula pumila, but otherwise they resemble the assemblages in fen pools or swamp forests, depending on string height and width. Swamp forests in contrast are consistently dominated by Larix laricina and Carex chordorrhiza but otherwise have a large assemblage of shrubs and herbaceous plants, depending on the microtopographic relief of hummocks and depressions. Picea maraiana with Chamaedaphne calyculata and Ledum groenlandicum, however, dominate the small tree islands in spring fens, which have less microtopographic relief and fewer species, particularly bryophytes.

species richness

The flora of the Albany River peatlands consists of 111 taxa of vascular plants and 102 bryophytes. The bog flora contains only 26 taxa of vascular plants and 49 bryophytes, whereas the fen flora is much larger, with 107 vascular plants taxa and 84 bryophytes. The species-richness patterns are closely related to landform type, water chemistry and water levels. Most of the main landform types are distinguished by the ratio of vascular plants to bryophyte taxa (Fig. 12a,b). In addition the ratio of vascular-plant to bryophyte taxa declines at a nearly linear rate with increasing water depth for most bog (r2 = 0.694) and fen (r2 = 0.664) landforms (Appendices S3 and S4). The only exceptions to this relationship are the noda of the bog hummocks/forests, water-track strings and spring-fen forests. The number of species in a relevé plot is also significantly related to pH (r2 = 0.399) and calcium concentrations (r2 = 0.485) in the surface water.

Figure 12.

(a) The ratio of vascular plant and bryophyte taxa on bog landforms within the Albany River region. The circles with crosses are bog lawn types whereas the shaded circles are lawns that are transitional to hummocks (i.e. mixed stands). (b) The ratio of vascular plant to bryophyte taxa on fen landforms within the Albany River region.

hydrogeology/water chemistry

The Albany River peatlands have dilute acidic surface waters that can be classified into different trophic groups on the basis of pH and calcium (Fig. 13). The bog waters have a pH lower than 4.2 and calcium concentrations less than 2 mg L−1. The fen waters, in contrast, tend to cluster into three different groupings typical of poor fen, moderately rich fen and extremely rich fen (Fig. 13). Only a few internal water tracks had a pH of 4.3–4.7, whereas most of the fen surface waters had pH values between 5 and 6.7 and calcium concentrations from 2.5 to 30.8 mg L−1. The surface water chemistry was similar to that obtained from the top of the water table (0 cm depth) with piezometers at the same site (Table 3), but most of these pore-water samples had large positive errors in their mineral-ion balance (CBE) greater than 99%. The pH of these samples generally decreased with increasing DOC, except for those samples with DIC concentrations greater than 10 mg C L−1. The organic-acid anion concentrations estimated by the Oliver et al. (1983) model ranged from 20% to 67% of the total anion equivalents, with a mean contribution of 44%, indicating that the dissociation of organic acids is an important source of H+.

Figure 13.

The relationship between pH and calcium concentrations in the surface waters of the Albany River peatlands.

Table 3.  Water chemistry for pore-water samples collected from the top of the water table. Units for cations and anions are mg L−1. Also listed are the charge-balance errors with no organic anion (CBE NoA), the amount of organic acid calculated using Olivers model (OM A), and the charge-balance error when the organic acids calculated with Oliver's Model are included (CBE (OM)). Values below the detection limit are indicated by nd. The peat vegetation/landforms sampled were bogs (B), fens (F) and spring-fen channels (SFC). Values below detection limit are listed as ND
8502B 6.746.8− 0.1
9201B 6.152.1−0.3
9207B 2.640.6−0.1
9208B 2.733.9 0.1
9214B 5.142.5 0.3
9204B 4.941.0− 0.3
9202F 1.520.8−0.1
9203F 4.533.5 0.1
9205F 0.4 2.031.0 0.1
9216F 0.1 2.629.1 0.1
9209F 0.013.832.3 0.0
9215F 0.3 1.844.8 0.2
9216F 0.1 2.629.1 0.1
9211SFC30. 0.1
9212SFC29. 0.1
9213SFC20. 0.2

Vertical hydraulic-head gradients were generally slight across the study area during the summer of 1992. Hydraulic head decreased with depth by 2–5 cm on raised bogs, indicating areas where surface waters moved downwards and recharged the underlying groundwater system (Table 1). However, other bogs had zones of overpressure at 2 m depth that were probably generated by trapped gas, or zones of underpressure indicating either insufficient recovery times or converging flow-systems (sensuRomanowicz et al. 1993; Glaser et al. 1997; Reeve et al. 2000). Zones of overpressure were also found in some internal water tracks, although other water tracks showed no head gradients, indicating lateral flow. Hydraulic head slightly increased with depth by several centimetres in the spring-fen channels and reticulate fens, indicating groundwater discharge. Dissolved inorganic solutes increased in concentration with depth on all landform types (Fig. 14) with the chemical depth profiles typical of recharge or discharge zones (e.g. Siegel & Glaser 1987).

Figure 14.

Depth concentrations for Ca in pore waters from bogs, fens and spring fens of the Albany River region.


Large peat basins are unusual areas where a combination of climate, landscape and biota produce the high water tables necessary for the regional expansion of peatlands. Peatlands, for example, spread across the nearly flat landscape of former Glacial Lake Agassiz in Minnesota after 5000 bp following a shift towards a cooler moister climate (Janssen 1968; Heinselman 1970). Alternatively, beaver dams (Sjörs 1963) or small peatlands (Kulczynski 1949) were proposed as agents that promote widespread paludification by obstructing local drainage. Although Kulczynski (1949) was the first to emphasize the role of regional geological processes in the development of large peat basins, these processes operate too slowly in most regions to produce noticeable effects on the modern landscape.

Among the largest peat basins, the Hudson Bay region is distinguished by an exceptionally high rate of isostatic adjustment to glacial loading (Bostock 1970; Andrews & Peltier 1989). The Albany River study area is located near the centre of the former Laurentide Ice sheet, which reached its maximum thickness about 20 000 bp (Dyke et al. 1989; Dredge & Cowan 1989). The weight of the ice mass depressed the lithosphere, and deglaciation led to rapid flooding of the Lowlands by the Tyrell Sea around 8000 years ago. Subsequently, the lowlands have been rising isostatically, with an uplift dome centred around the mouth of James Bay (Andrews & Peltier 1989). Isobases for vertical displacement relative to sea level for the past 8000 years are greater than 300 m within the centre of the dome, with decreasing isobases outward. The land surface is still rising rapidly in the James Bay region, with a rate of 1.5 m century−1 at the mouth of the Albany River (Hunter 1970; Webber et al. 1970). Peatlands rapidly covered this emerging land surface according to peat cores that show sharp basal contacts with the mineral substratum, which is unaltered by weathering.

The rising landscape creates a regional chronosequence as new land continues to emerge from the sea. Subsequently, peat development is subject to intense isostatic forcing, which has raised the land surface in the study area by as much as 120 m in 6000 years. The rate and magnitude of this change provides an unusual opportunity for testing some basic assumptions on vegetation patterning in large peat basins. If the primary drivers for peatland succession are biotic processes, then the vegetation patterns should change sequentially across the region as a function of time. Bogs will replace fens on older land surfaces as the peat cover thickens and blocks the supply of inorganic solutes and bases to the surface waters (Klinger & Short 1996). Alternatively, groundwater has been proposed as the primary determinant for the distribution of bogs, fens and spring fens in the Glacial Lake Agassiz region of northern Minnesota (Siegel 1983; Glaser 1987a; 1992; Glaser et al. 1997). According to this hypothesis the distribution of the various peatland types should be determined by geological factors that control the groundwater flow systems and maintain complex associations of bog and fen landforms on mature land surfaces.

local patterning

These hypotheses can be directly tested by an analysis of the vegetation/landform patterns because the principal species assemblages (i) recur across the study area, (ii) respond predictably to the major environmental gradients, and (iii) form vegetation/landform patterns identifiable from remote-sensing imagery. Both the DCA ordinations and the vegetation tables indicate that the species assemblages within this area respond most closely to local changes in water level, water chemistry and peat microtopography. However, a close relationship also exists between the species assemblages and the various peat landform patterns at both the micro- (100–500 m2) and meso- (0.5–200 km2) scale and such relationships have been interpreted as evidence for an underlying biotic control on peatland development, as first emphasized by Weber (1902).

The Albany River assemblages conform to the same trophic types originally defined on the basis of water levels, pH and calcium concentration in Sweden (Fig. 4; Sjörs 1948, 1950, 1952; Malmer 1962) and later applied to other regions such as the Hudson Bay Lowlands (Sjörs 1963; Jeglum & Cowell 1982; Sims et al. 1982). Although recent studies also stress the importance of nitrogen, phosphorus and potassium in solid-phase peat (Bridgham et al. 1996, 1998), the effects of these nutrients are less apparent in the Albany River peatlands in which the surface waters are exceptionally dilute. Changes in pH and calcium, however, are associated with floristic changes that define the bog–fen boundary and also with vegetational changes related to the shift from poor fen to moderately rich and extremely rich fen.

However, these four trophic groupings cannot account for the full differentiation of the 11 discrete vegetation noda that were distinguished on the phytosociological tables and DCA ordinations. Most of the variation in the bog and fen vegetation is expressed along the first DCA axis, which closely corresponds to changes in water level and microtopography (Fig. 2a,b). Changes in water level are primarily responsible for the five bog noda, which otherwise share the same narrow range in water chemistry. Moreover, the microtopographic gradient from pool to forest produces separate sets of noda in moderately rich and extremely rich fens. The higher variability in the string assemblages may indicate their polygenetic origin and lack of stasis.

The vegetation noda can also be distinguished independently by their ratio of vascular plant to bryophyte taxa, indicating a very close coupling between these disparate plant groups. Previous studies beginning with Weber (1902) showed how interactions between vascular plants and bryophytes could produce the different microtopographic units on raised bogs (e.g. Wallén 1980; Tallis 1983; Glaser 1987a; Clymo 1991; Belyea & Clymo 2001; Pastor et al. 2002; Malmer et al. 2003). The conservative nature of these interactions is further indicated by the nearly linear decline in the ratio of vascular plant to bryophyte taxa with rising water levels in the Albany River region (Appendices S3 and S4). Only on some of the drier landforms is the balance disrupted by shading from Picea mariana and competition from lichens.

peat landforms

Peat accumulation alters the local hydrological setting by creating three-dimensional landforms composed largely of dead organic matter. Weber (1902) stressed how the morphology of a landform shapes the water levels, water chemistry and vegetation patterns of peatlands, thus providing a basis for all future studies of landform morphology. He intuitively recognized that bog landforms enclose water table mounds that separate the bog vegetation from geogenous waters, a concept that was later formalized mathematically by Ingram (1982) and Siegel (1983). These water table mounds are maintained partly by the capacity of the peat mass to impede drainage.

This effect probably accounts for the distribution patterns of trees and pools on bogs across the study area. Detailed levelling of the Albany River bog (8501), for example, shows that tree cover is restricted to the moderately sloping bog margin where the porous upper peat promotes surface drainage (Fig. 10). The bog forest is abruptly replaced by pool networks where drainage is inhibited by the nearly level slope and firmer, more compact peat. Similar results were reported by Sjörs (1963) on the Attawapiskat River to the north. This close relation between tree growth and peat slopes provides a basis for interpreting bog landform patterns from aerial photographs.

Bog landforms in the study area seem to fit the various stages of bog development proposed by Glaser & Janssens (1986). This model predicts that rapidly growing bogs initially have (i) a forested crest but, as the rate of peat growth slows from the interior of a landform outwards, the bog morphology changes to (ii) a semi-forested mound, followed by (iii) a non-forested plateau with pools. The second stage first appears as an open rift along a forested crest and then expands outwards, spreading apart the trees into ribbons that are progressively restricted to the outer sloping margins of an expanding central plateau (Fig. 7, 1a and 1b, and Fig. 8, black arrow). In time bogs will be completely covered by concentric networks of pools, such as Belec Lake (Fig. 4: 1).

Fen landforms, in contrast, develop into aprons or troughs that channel peatland runoff towards some drainage outlet (Fig. 1). The absence of orderly surface patterns on the swamp forests may be related to their topographic positions at the extreme down-gradient edges of local catchments, where drainage is promoted by a steeper hydraulic gradient. Similarly, the absence of pool patterns in the relatively short and narrow spring-fen channels may be related to their steeper slopes along the edges of the minor moraines (Fig. 2). As the slope becomes more gradual spring-fen channels tend to grade into patterned fens with their transverse networks of pools and peat ridges.

Peat accumulation produces gentler slopes in many areas and the resulting sluggish drainage is an important factor in the development of pool patterns. Weber (1902) and Boatman (1983) observed that mire pools form by local flooding parallel to the contour interval. This process is also apparent in the Albany River region, particularly where a series of bogs constricts the flow within the adjacent water tracks (Figs 3 and 7). Many of these fen pools are deep and seem to be expanding by decomposition of the pool bottom as suggested by Sjörs (1963) and supported by carbon-flux measurements (Hamilton et al. 1994; Roulet et al. 1994). Local flooding may also be responsible for the local fragmentation of bogs into streamlined islands of various sizes (e.g. Fig. 7).

water chemistry

The nearly continuous peat cover in the Albany River region has greatly modified the chemistry of the surface waters. Although the underlying mineral substratum is calcareous, the surface waters have relatively low concentrations of calcium and even lower concentrations of the other major cations (Table 3). In addition, regional acidification is indicated by (i) the unexpectedly low pH of many bog and extremely rich fen waters, and (ii) the position of many fen data points below the theoretical calcium saturation line for solutions in equilibrium with atmospheric CO2 (Fig. 13). The bog and fen waters, moreover, have large negative-charge deficits in their inorganic ionic balances according to the Oliver et al. (1983) model, indicating that organic acids provide the anions necessary to balance charge (Table 3). These findings support previous work that indicates organic acids are important contributors to the acidity of peatlands and lakes (Gorham et al. 1985; Hemond 1990; Driscoll et al. 1994).

The release of organic acids from decomposing peat may also be responsible for the clustering of the water samples into four discrete trophic groups (Fig. 13) buffered by different suites of organic acids (e.g. from bogs and fens) and bases (e.g. bicarbonate from the mineral substratum). Regional acidification may therefore be an important driver for the evolution of peatland types in the Albany River region. However, the pH and concentration of all major cations increased with depth in the pore waters of bogs, fens and spring fens (Fig. 14). Bogs had concave chemical profiles typical of recharge zones, whereas the spring fens had convex profiles typical of groundwater discharge. Vertical transport through the peat profile may therefore be an important control on the surface-water chemistry in the Albany River peatlands (Reeve 1996; Reeve et al. 1996, 2000, 2001). The low cation concentrations in the surface waters are probably related to slow rates of mass transport through the peat profile combined with dilution at the peat surface by precipitation.

groundwater and landscape evolution

The major vegetation zones in the James Bay region suggest a successional sequence, with coastal marshes giving way inland to featureless swamp forests and patterned bog complexes on progressively older land surfaces (Sims et al. 1979, 1982; Jeglum & Cowell 1982; Riley 1982). This hypothetical sequence was presumably initiated by land emergence and subsequently driven by the spread and thickening of peat deposits. However, the Albany River study area is dominated by fens that are intermixed with all the other peatland types without any clear relationship to landscape age or distance from the sea. These peatland patterns, in contrast, are closely adjusted to the local hydrogeological setting, indicating a close linkage between groundwater flow systems and peatland development.

The postglacial evolution of these flow systems was constrained by a landscape that was first shaped by glacial and marine processes and then altered by rapid isostatic uplift. The glacial deposition of a moraine, for example, formed a ridge of relatively porous till in the south that was capped by only a sparse veneer of clayey silt during the marine transgression (Figs 1 and 2, Table 1). As this landform was raised by postglacial uplift, a water-table mound formed within the moraine that drove surface waters downwards into the porous till under the peat (Table 1). These local recharge (downward flow) systems were promoted by both the porous properties of the till (cf. Reeve et al. 2000, 2001) and the height of the moraine above the adjacent till plain (Fig. 2). This hydrogeological setting spurred the development of the many small irregular-shaped bogs on the moraine summit despite the numerous exposures of calcareous till and small kettle lakes (Fig. 9). Downward flow disconnects the surface waters from inorganic salts in the underlying calcareous sediments and thus favours the local invasion of Sphagnum, which can acidify a site (Glaser et al. 1997).

These shallow flow systems discharge into the moraine flutes and also along the lower margins of the minor moraines, producing parallel series of spring-fen channels and patterned water tracks (Figs 1 and 5). The northern margin of the moraine apparently functions as a regional seepage face for groundwater that flushes through the large blocks of fen vegetation spreading downslope across the till plain (Figs 1 and 5). Landsat imagery shows extensive plumes of minerotrophic vegetation propagating downslope from these discharge points (Fig. 1, arrows). Similar plumes are evident in (i) the internal water tracks and downslope margins of the larger bogs, (ii) along till-plain rivers, which are lined with very short (< 0.5 km) parallel series of rills typically formed by groundwater sapping (Figs 3 and 7) and (iii) the numerous abandoned drainage valleys on the till plain.

Lateral flow systems, in contrast, should develop on the flat, subtly fluted till plain where thick layers of clayey silt were deposited during the marine transgression. Drainage was further impeded by the faster rate of postglacial uplift near the coast, which continually reduces the regional gradient, promoting rapid paludification. Reeve et al. (2000, 2001) show that lateral flow systems can maintain fen vegetation by transporting inorganic solutes upwards through the peat profile by advective dispersion. However, lateral flow was locally interrupted by the development of a dense network of parallel rivers in the till-plain flutes and by the incision of the Albany and Kenogomi rivers, which periodically changed course, leaving behind abandoned river channels. Water-table mounds form in the interfluvial divides between the rivers and abandoned channels (Freeze & Cherry 1979) driving shallow local recharge systems that will trigger the development of raised bogs (Table 1). A vivid example of this bog-forming process is provided by a triangular bog that is spreading across the divide newly formed by two headwardly eroding streams (Fig. 8). Exceptionally large raised bogs developed along the widest interfluvial divides of the till plain, but these bog complexes always grade into fens downslope where inorganic solutes are brought to the peat surface by lateral flow or groundwater discharge (Figs 1, 2 and 3).

The most enigmatic landforms within the study area are the internal water tracks that arise on the flanks of all bogs larger than 20 km2 (Figs 1, 2, 3 and 5). The abundance of these internal water tracks is probably responsible for the higher percentage of fen vegetation in the Albany River region (Fig. 1) as compared with the inventories compiled by Riley (1982) and Roulet et al. (1994) for neighbouring watersheds. The large number of these internal water tracks is unexpected given the high acidity of the bog waters and low alkalinity of fen waters within the study area. Surface chemical transformations are therefore unlikely to be responsible for the origin of these water tracks, as has been suggested by Sjörs (1963) and Ingram (1967). However, the location of many of these bogs in small interfluvial divides makes it also unlikely that either groundwater discharge (sensuGlaser 1987a,b) or dispersive-mixing along lateral flow paths (sensuReeve et al. 2000, 2001) transports alkalinity into tracks that arise near the bog crest. Some internal mechanism within the peat (e.g. related to the generation of overpressures) may be the ultimate cause of these internal tracks because they are ubiquitous to all larger bogs within boreal America regardless of their hydrogeological setting (Glaser 1987b).


The Albany River study area provides an exceptional setting for clarifying the interactions among peatlands, groundwater and landscape processes. Peatland development has failed to produce a continuous cover of raised bogs over this area, despite the long successional sequence, low geomorphic relief, extensive peat cover and high loadings of organic acids released from bogs. In contrast, rapid isostatic uplift continues to drive hydrogeological and geomorphic processes that maintain an array of bogs, fens and spring fens on land surfaces of similar age. The formation of groundwater mounds under the rising moraine system and the interfluves between parallel rivers spurred the development of raised bogs, whereas the discharge of groundwater around the margins of the moraine system and lower flanks of the till plain maintains large tracts of minerotrophic fen downslope.

Within a given hydrogeological setting, however, peatland succession seems to follow predictable pathways leading to the formation of a distinctive set of vegetation/landform patterns that recur across the regional study area. Peat accumulation tends to create more gentle slopes that impede drainage and raise the local water table, which in turn reduces species richness and community variability in a nearly linear manner. However, these successional sequences are continually disrupted and reset by the differential rate of uplift, which alters both surface and groundwater flow systems. In this dynamically changing landscape the striking vegetation/landform patterns therefore provide important indicators for both the biotic and physical processes that drive peatland development.


This research was supported by grants from the National Aeronautics and Space Administration and US National Science Foundation to Paul Glaser and Donald Siegel. We thank John Almendinger, Jay Gilbertson, Howard Mooers, Edwin Romanowicz, Michael Stone, Yi Ping Shen and Chris Chirmo for assistance in the field, Mark Emmonds for assistance with Landsat image processing and Paul Morin for the SRTM DEM. Helicopter support was provided by Roto-Ways Ltd and float plane service from Cochrane Air Service Ltd. H. E. Wright Jr read the manuscript and advised on the geomorphic evolution of the study area with Harvey Thorleifson. We also thank Darrel Williams, Diane Wickland, NASA's Land Cover–Land Use Change Program, and Los Alamos National Laboratory for providing financial assistance to publish the foldout figures.

Supplementary material

The following material is available from http://www.blackwellpublishing.com/products/journals/suppmat/JEC/JEC930/JEC930sm.htm

Appendix S1 Complete phytosocialogical table for relevés from raised bogs in the study area. The symbols are the same as those used in Table 2. Also provided are the pH and calcium concentration of the surface waters in each relevé and the number of bryophyte and vascular plant taxa.

Appendix S2 Complete phytosocialogical table for relevés from fens in the study area. The symbols are the same as those used in Table 2. Also provided are the pH and calcium concentration of the surface waters in each relevé and the number of bryophyte and vascular plant taxa.

Appendix S3 The decline in species number from dry-to-wet landforms on the wetter landforms of raised bogs.

Appendix S4 The decline in species number from dry-to-wet landforms on the wetter landforms of fens.