A multi-proxy peat study of Holocene vegetation history,bog development,and carbon accumulation on northern Vancouver Island,Pacific coast of Canada

Abstract

We present a multi-proxy paleoenvironmental study on a 10,400 cal. yr peat sequence from an ombrotrophic bog in coastal British Columbia, Canada. Pollen, non-pollen palynomorph, plant macrofossil, and physicochemical analyses (bulk density, %C, %N, δ¹³C, and δ¹⁵N isotopes) were used to document changes in vegetation, bog development, and carbon and nitrogen accumulation. Pollen assemblages indicate that regional vegetation in the warm, relatively dry early Holocene was mixed coniferous forest with scattered Pseudotsuga menziesii. Herbaceous peat with a C:N of ~28, combined with Nuphar microfossils and relatively high %N, suggests the presence of a herb-dominated peatland with standing water and/or bog pools. Carbon and nitrogen accumulation were at their highest during this early Holocene period at mean rates of 30.6 and 1.2 g/m²/cal. yr, respectively. By 8000 cal. yr BP and under a cooler, wetter climate, northern Vancouver Island supported Tsuga heterophylla rainforest similar to today. Decreasing relative water table, inferred from testate amoebae and fungal remains, facilitated the establishment of a Sphagnum bog by 8000 cal. yr BP with abundant ericaceous shrubs after 5000 cal. yr BP. Temporal variation in carbon accumulation rates corresponds with changes in plant functional types and hydrological conditions: rates were lowest in the early to mid-Holocene during accumulation of Sphagnum peat (7.1 g/m²/cal. yr) and increased in late Holocene ligneous peat (12.4 g/m²/cal. yr). Our multi-proxy approach not only demonstrates the overarching control of climate on bog development and carbon and nitrogen accumulation, with seasonality likely playing a major role, but also highlights the strong influence of autogenic processes at a local scale.

Keywords

C:N ratio carbon and nitrogen accumulation non-pollen palynomorphs peat pollen Procrustes analysis

Introduction

Peatlands provide excellent records of ecological and environmental change on long timescales, as slow rates of decay relative to primary production promote the accumulation of peat and exceptional preservation of fossil material (Charman, 2002; Rydin and Jeglum, 2013). Peatlands are also important long-term sinks for carbon dioxide, storing more carbon (C) per unit area than any other terrestrial ecosystem, and major sources of atmospheric methane (Baird et al., 2009; Frolking and Roulet, 2007; Gorham, 1991; Korhola et al., 2010; Limpens et al., 2008). Climate is the dominant macroscale control on peatland development and long-term C accumulation, with both temperature and precipitation playing important roles in primary production and decay (e.g. Baird et al., 2009; Clymo et al., 1998). Loisel et al. (2014) compiled a large database of Holocene carbon accumulation rates (CARs) from northern peatlands that demonstrates that, at large spatial scales, the highest CARs occurred during the early Holocene, when northern latitudes were characterized by relatively high summer insolation and greater seasonality (Berger and Loutre, 1991). In another large compilation, Charman et al. (2013) show the dominant control of climate during the last millennium: in northern peatlands, C accumulation was higher during the warm Medieval Climate Anomaly than during the ‘Little Ice Age’. At an individual site, however, local-scale factors such as topography, hydrology, species composition, and disturbance can act as principal controls on peatland dynamics and C accumulation (e.g. Magnan and Garneau, 2014; Shiller et al., 2014; Turunen and Turunen, 2003). In order to better understand the effects of climate change on peatlands and their role in global change, as well as the importance of local autogenic processes, more research is needed on past peatland dynamics, long-term C accumulation, and connections to changes in vegetation, hydrology, and climate, particularly in regions where peat studies are lacking.

Peatlands are common in Pacific Canada, particularly on the hypermaritime north coast of British Columbia (BC), where wetlands cover up to 75% of the terrestrial landscape (National Wetlands Working Group, 1988). Most of these are peatlands that form on flat to moderately sloping terrain. Mild temperatures, high precipitation, and abundant fog sustain these wetlands, reduce decomposition, and promote the accumulation of peat (Asada and Warner, 2005). Paleoecological studies conducted in coastal BC bogs (e.g. Banner et al., 1983; Brown and Hebda, 2002; Hebda, 1983; Heusser, 1960) have focused primarily on reconstructing Holocene vegetation dynamics from fossil pollen assemblages. Recently, Huntley et al. (2013) combined pollen and non-pollen palynomorphs (NPPs) to assess the impact of climate change and recent human activity on vegetation and bog development over the last 1800 cal. yr on Haida Gwaii (Queen Charlotte Islands, BC). Turunen and Turunen (2003) provide the only study in coastal BC that documents changes in %C and %N in a Holocene peat record and infers long-term rates of C accumulation. Based on n = 12 samples spanning the last 8500 cal. yr, they found a relatively low mean rate of C accumulation (8.6 g/m²/cal. yr, as reported by Loisel et al., 2014), with the highest CARs occurring in the early Holocene. Given the abundance of wetlands and specifically peat-accumulating bogs in Pacific Canada, there is a need for more research aimed at understanding bog development and peat C accumulation in this maritime region. The coastal perspective is important as it offers an opportunity to examine whether peatlands on the north Pacific coast have lower C accumulation relative to continental sites, as has been addressed elsewhere (Yu et al., 2009, 2014). Despite abundant precipitation and mild temperatures year-round that result in long growing seasons, oceanic bogs generally accumulate less peat and store less C than continental peatlands because the high primary production is offset by high decomposition (Asada and Warner, 2005; Malmer and Wallén, 1993). In fact, Gorham et al. (2003) noted a strong negative relationship between long-term rates of peat accumulation and precipitation in North American peatlands. Thus, we expect sites in coastal BC to have relatively low rates of peat accumulation and C sequestration compared to regions where seasonality in temperature is greater and precipitation is less abundant.

Here, we present a multi-proxy study of a peat sequence from an ombrotrophic bog on northern Vancouver Island, BC, aimed at understanding how changes in ecological, hydrological, and environmental conditions over the Holocene relate to long-term rates of C and N accumulation. Fossil pollen and plant macrofossils are used to document regional and local changes in vegetation communities and infer past environmental conditions. Bog development, edaphic conditions, and relative water table depth are also inferred from NPPs, that is, testate amoebae, fungal and algal remains, and aquatic plant microfossils. Physicochemical proxy records (i.e. bulk density, %C and %N content, δ¹³C and δ¹⁵N isotopes) are used to further document the bog’s history including changes in plant functional groups, nutrient status, and peat decomposition. Finally, long-term rates of C and N accumulation are compared to vegetation and hydro-climatic changes over the Holocene. This multi-proxy approach allows for a more detailed record of bog dynamics from coastal BC than achieved in previous studies as well as the opportunity to evaluate concordance between different peat-based proxy records. This study also helps to clarify the relationship between long-term rates of peat accumulation and precipitation and aids in the understanding of how climate, hydrology, and vegetation interact to shape peatlands and C accumulation in oceanic settings.

Materials and methods

The study site (50°34′21″N, 127°04′20″W; 93 m a.s.l.) is a flat ombrotrophic bog located near the town of Port McNeill on northeastern Vancouver Island (Figure 1). Port McNeill Bog (PMB; informal name) lies within the Squash Basin, an area of particularly low relief (<350 m) within extensive lowlands that span northern Vancouver Island, the mainland coast of BC, and northeastern Haida Gwaii (Queen Charlotte Islands). The region supports temperate coniferous rainforest with bog-forest complexes under a cool, humid climate. Mean summer temperature is 13.7°C, and mean winter temperature is 4.1°C (Port Hardy Airport weather station; Environment Canada, 2014). Mean annual precipitation is 1908 mm/yr, with most precipitation falling as rain between September and April.

Figure 1.

Location of Port McNeill Bog (PMB) on northern Vancouver Island, British Columbia, Canada, and other sites mentioned in the text. 1: Tiny Lake (Galloway et al., 2009); 2: Two Frog Lake (Galloway et al., 2007); 3: Woods Lake (Stolze et al., 2007); 4: Cook Bank (Lacourse et al., 2003); 5: Bear Cove Bog (Hebda, 1983); 6: Misty Lake (Lacourse, 2005); 7: Pyrola Lake (Hebda and Haggarty, 1997); 8: Kalmia Lake (Hebda and Haggarty, 1997).

The summer water table depth at PMB is ~20 cm (this study; Howie and van Meerveld, 2013), which approximates the current depth of the acrotelm–catotelm interface (Clymo, 1984). The plant community is in advanced succession, dominated by ericaceous shrubs (i.e. Rhododendron groenlandicum (formerly Ledum groenlandicum), Empetrum nigrum, Vaccinium uliginosum, and Kalmia microphylla ssp. occidentalis) and two species of Sphagnum moss (S. capillifolium and S. fuscum). Other common species include Rubus chamaemorus, Vaccinium oxycoccos, Sphagnum papillosum, Sphagnum pacificum, and Pleurozium schreberi as well as stunted Pinus contorta var. contorta trees. Gaultheria shallon, Eriophorum chamissonis, Cornus unalaschkensis, Drosera rotundifolia, and Juncus ensifolius occur infrequently. Open Pinus. contorta bog-woodland with Tsuga heterophylla and Thuja plicata surround the bog.

In July 2012, a 3.24-m peat core was collected from PMB using a side-filling ‘Russian’ sampler. Overlapping sections were retrieved from two parallel boreholes located 50 cm apart. After retrieval, the peat was stored at 4°C at the University of Victoria. Seven AMS radiocarbon ages (¹⁴C yr BP) were obtained, each of which was on a single, small fragment of wood (Table 1), and calibrated to calendar years (cal. yr BP) using the IntCal13 calibration data set from Reimer et al. (2013). A chronology based on these ages and −62 cal. yr BP for the top of the core was fit using a Stineman (1980) interpolation with the ‘stinepack’ (Johannesson and Bjornsson, 2012) and ‘clam’ (Blaauw, 2010) packages in R (R Core Team, 2014). Stineman interpolation is similar to simple linear interpolation in that the age–depth model is forced through the best available estimates of age, that is, the calibrated ¹⁴C ages (Telford et al., 2004), but it stabilizes slopes near ¹⁴C ages, providing a more robust estimate of mean accumulation rates over long timescales. Thus, the Stineman interpolation helps to avoid the unrealistic assumptions of simple linear interpolation that accumulation is constant between ages and that accumulation rates change abruptly and precisely at dated depths in a sequence. A Bayesian approach to age–depth modeling using the ‘Bacon’ package (Blaauw and Christen, 2011) was also considered. The ‘Bacon’ age model is nearly identical to the interpolation methods for most time intervals, and the Stineman model sits entirely within the 95% confidence intervals of the ‘Bacon’ model; however, in our case, ‘Bacon’ forces the age–depth model toward the edge of the 2σ calibration range for the ¹⁴C age at 140 cm, producing a lower change in accumulation rate for a portion of the peat core where ash-free bulk density (AFBD) measurements suggest a larger change in accumulation rate. Changing the default prior information in ‘Bacon’ results in an even more problematic age–depth model: the ¹⁴C age at 140 cm is ignored, and accumulation rates between 124 and 235 cm, which account for ~3700 cal. yr, are more or less constant, despite important changes in peat type.

Table 1.

AMS radiocarbon and calibrated calendar ages on wood from Port McNeill Bog, Vancouver Island, British Columbia.

Depth (cm)	Radiocarbon age (¹⁴C yr BP ± 1σ)	δ¹³C (‰)	Calendar age (cal. yr BP)^a	Laboratory no.
30–31	740 ± 30	−24.8	683 (660–730)	Beta-365555
65–66	2850 ± 30	−28.8	2961 (2880–3060)	Beta-337811
109–110	4110 ± 30	−25.9	4654 (4520–4810)	Beta-337812
124–125	4690 ± 30	−24.2	5412 (5320–5580)	Beta-365556
140–141	6240 ± 40	−24.4	7157 (7010–7260)	Beta-335265
234–235	8160 ± 40	−24.5	9109 (9010–9250)	Bets-335266
321–322	9220 ± 40	−23.9	10,380 (10,260–10,500)	Beta-335267

Weighted average of the probability distribution and 2σ age range rounded to the nearest 10 years.

Peat subsamples (1–2 cm³) were prepared for pollen and NPP analyses using 10% HCl, 10% KOH and acetolysis solution, and then mounted in 2000 cs silicone oil. Samples were not treated with hydrofluoric acid. One tablet of 18,684 ± 371 Lycopodium spores (batch no. 177745) was added to each sample to estimate pollen and NPP concentrations. Before chemical treatment, each sample was sieved at 150 µm and the larger fraction was retained for peat composition analysis. A minimum of 350 terrestrial pollen and spores was identified in each sample using dichotomous keys (e.g. Kapp et al., 2000; McAndrews et al., 1973) and reference material at the University of Victoria. Alnus pollen were differentiated into Alnus rubra and Alnus viridis types following May and Lacourse (2012). Identification of Ericaceae pollen was aided by Warner and Chinnappa (1986) and local reference material; however, most Ericaceae pollen from this site could not be differentiated below the family level due to overlapping pollen morphologies (Lacourse, unpublished data; Warner and Chinnappa, 1986). NPPs including testate amoebae tests, fungal spores and hyphopodia, aquatic plant microfossils, and algal remains were identified on the same slides used for pollen analysis, using various sources including van Geel (1978), Pals et al. (1980), Kuhry (1997), Charman et al. (2000), Marrotte et al. (2012), and Payne et al. (2012).

Pollen percentages were calculated using the sum of all terrestrial pollen and spores, excluding pollen from obligate aquatic species and Sphagnum spores. Numerical zonation of the pollen and NPP data was based on taxa that exceeded 1% of the sum and conducted using optimal splitting by information content (Bennett, 1996), although binary splitting and constrained cluster analysis produced identical results. Statistical significance of the resulting zones was tested using broken stick models. Procrustes analysis (Peres-Neto and Jackson, 2001) was used to assess the degree of congruence in the major variation of the pollen and NPP data sets. Procrustes was conducted on the sample scores of independent principal component analyses of the two data sets using the ‘vegan’ package (Oksanen et al., 2013) in R. The significance of the Procrustes correlation statistic (r), which quantifies the similarity of the two matrices, was tested using 9999 permutations.

To estimate the major macrofossil components of the peat and relative changes with depth, we analyzed the >150 µm fraction of each 1–2 cm³ pollen sample using an approach similar to the quadrat method of Barber et al. (1994) and Mauquoy et al. (2010). The >150 µm fraction was poured into gridded Petri dishes, and all macrofossils in 15 randomly selected 1 cm × 1 cm quadrats were differentiated into major peat components (i.e. herbaceous material, moss stems/leaves, Ericaceae roots and leaves, and unidentifiable organic material (UOM)) and are expressed as percentages of the total count from those quadrats. Seeds, Nuphar sclereids (>150 µm), and charcoal fragments (>150 µm) are presented as the total number encountered in those same quadrats.

Loss-on-ignition (LOI) was used to estimate organic matter content (%) of the peat (Dean, 1974; Heiri et al., 2001). Peat samples of 2 cm³ were taken at 2-cm intervals along the peat core (n = 163), dried at 105°C for 20 h, and then ignited at 550°C for 4 h. AFBD was calculated as the dry weight (g) divided by the wet volume (cm³) and then multiplied by % organic matter content. For C and N analyses, an additional 2 cm³ from the same depths (n = 163) was dried to constant mass at 70°C for 20 h and ground to a fine powder (<125 µm) using a ball mill. C and N contents and δ¹³C and δ¹⁵N isotopes were determined on 3–5 mg of dried homogenized peat in 5 mm × 8 mm tin capsules using a Costech ECS 4010 thermal combustion elemental analyzer attached to a Thermo Finnigan DELTA^Plus Advantage isotope ratio mass spectrometer at the University of Victoria. Replicate analyses were performed on 19 of the 163 samples. C and N isotope data are reported in conventional δ (‰) notation with reference to Vienna Pee Dee Belemnite (VPDB) carbonate and atmospheric nitrogen (air), respectively. Analytical reproducibility of the mass spectrometer was assessed using acetanilide (71.09% C and 10.36% N) and caffeine (−42.22‰ δ¹³C and −0.94‰ δ¹⁵N) standards included in every run. Relative uncertainties calculated from analyses of these standards are better than ±0.1% and ±1.5% for %C and %N, respectively, and ±0.27‰ and ±0.09‰ for δ¹³C and δ¹⁵N, respectively. CAR (g/m²/cal. yr) and nitrogen accumulation rate (NAR; g/m²/cal. yr) were calculated using %C and %N, respectively, along with AFBD (g/cm³) and modeled deposition times (cal. yr/cm). Mean CAR and NAR are weighted by the deposition time of each sample.

Results

Chronology, peat stratigraphy, and plant macrofossils

The age–depth model for the PMB peat core (Figure 2) predicts an age of 10,413 cal. yr BP for the base of the core, that is, at 324 cm. Modeled accumulation rates vary between 0.01 and 0.08 cm/cal. yr, with a mean deposition time of 32 cal. yr/cm. Deposition times are notably higher during two intervals (140–124 and 64–34 cm), indicating slower net peat addition at ~7100–5400 and ~2900–800 cal. yr BP, respectively. The overall age–depth relationship is convex with decreasing peat accumulation rates through time. This is in contrast to the concave pattern of peat accumulation in other oceanic settings (Clymo, 1984).

Figure 2.

Age–depth model for the Port McNeill Bog core with 95% confidence intervals shown as gray bands. The model is based on a Stineman interpolation fit using ‘stinepack’ (Johannesson and Bjornsson, 2012) and ‘clam’ (Blaauw, 2010).

The basal portion of the peat core (324–171 cm; ~10,410–7960 cal. yr BP) consists of fibrous herbaceous peat (Figure 3) with Cyperaceae leaves and achenes and abundant Nuphar sclereids (>150 µm), most likely derived from Nuphar polysepala. UOM accounts for ~20% of the peat composition. Sphagnum and Ericaceae remains and fern sporangia were encountered infrequently in this portion of the core. Nuphar sclereids (>150 µm) decrease dramatically after 8000 cal. yr BP, although they continue to be present in trace amounts until ~4400 cal. yr BP. The herbaceous peat is sharply overlain by Sphagnum peat (171–122 cm; ~7960–5270 cal. yr BP) that is dominated by Sphagnum leaves and stems and includes minor components of Ericaceae roots and leaves and less overall UOM. Juncus seeds first appear at ~7900 cal. yr BP and continue to be present in low numbers for the remainder of the record. There is a gradual transition to Sphagnum-ligneous peat (122–97 cm; ~5270–4150 cal. yr BP), where Sphagnum leaves/stems and Ericaceae roots each account for ~35% of the peat composition. Ligneous peat (sensu Troels-Smith, 1955) that consists primarily of Ericaceae roots and UOM, with occasional peaks in Sphagnum remains, occurs from 97 to 30.5 cm (~4150–680 cal. yr BP). The uppermost 30.5 cm is composed of a mixed moss-ligneous peat, marked by an increase in the relative abundance of brown mosses, primarily Drepanocladus cf. exannulatus. No distinct charcoal layers or horizons are present in the PMB core; however, macroscopic charcoal (Figure 3) is abundant between 121 and 32 cm (~5220–730 cal. yr BP). Charcoal concentrations are highest between 68 and 44 cm (~3070–1510 cal. yr BP).

Figure 3.

Peat stratigraphy, peat component percentages and plant macrofossils for Port McNeill Bog, northern Vancouver Island.

Pollen and spore assemblages

Fossil pollen and spores were identified in 66 peat samples with a mean temporal resolution of 160 cal. yr between samples. Numerical zonation of the pollen and spore percentage data returned two major pollen and spore assemblage zones (Figure 4). The major pollen zone boundary at ~8000 cal. yr BP corresponds with the stratigraphic change from herbaceous peat to Sphagnum peat at 171 cm. Overall trends in pollen and spore percentages correspond closely with changes in taxa-specific pollen concentrations with a few exceptions, which are noted below.

Figure 4.

Pollen and spore percentages for Port McNeill Bog, northern Vancouver Island, showing major taxa only and 10× exaggeration of infrequent taxa. Circles within the plot of Sphagnum spores denote the stratigraphic position of spores of the parasitic fungus, Tilletia sphagni (NPP type 27).

Pollen Zone 1 (~10,410–7960 cal. yr BP) is characterized by high percentages of A. rubra, T. heterophylla, and Picea sitchensis, with A. rubra decreasing through Pollen Zone 1b (Figure 4). T. heterophylla increases in relative abundance from ~20% to 50% of the pollen sum; however, T. heterophylla pollen concentrations remain relatively constant at about 19,000 grains/cm³. Pinus contorta type is present at <10%, and P. menziesii, Abies and Cupressaceae each account for <2%. Sphagnum spores and Ericaceae pollen first appear at ~10,000 and 9200 cal. yr BP, respectively. Pollen from herbaceous taxa are infrequent throughout the record with Cyperaceae accounting for ~1% in Pollen Zone 1 and remaining relatively constant at <2% for the entire record. Fern spores including Pteridium aquilinum and Athyrium filix-femina make up ~5% of the sum in this basal zone. Nuphar pollen is at its highest relative abundance (i.e. 2–5%) with a notable decrease in concentration after 9600 cal. yr BP.

In Pollen Zone 2 (7960 cal. yr BP to present), T. heterophylla is the dominant pollen type (~40–70%) and has a higher mean pollen concentration (~35,000 grains/cm³) than in the preceding zone. Pinus sitchensis, A. rubra type, and P. contorta type continue to be present at 5–10% each. Abies and Cupressaceae are higher than in Pollen Zone 1 but remain at <5% throughout the entire record. Ericaceae pollen including Ledum type increase to ~5–25% in Pollen Zone 2a and then become co-dominant with T. heterophylla in Pollen Zone 2b, with Ericaceae pollen accounting for 20–60% of the total sum. The abundance of Sphagnum spores varies between 2% and 30% with the largest increase occurring at 3000 cal. yr BP, when spore concentrations reach 52,000 spores/cm³. Pteridophyte spores including Polypodium account for <2%. There is a single tetrad of Drosera pollen at ~5200 cal. yr BP. Alnus rubra type increases to almost 40% in the uppermost sample, but its concentration (~21,000 grains/cm³) is far less than in Pollen Zone 1a (~53,000 grains/cm³). Pinus contorta–type concentrations between 3500 cal. yr BP and the present are similar to those in the earliest Holocene (Pollen Zone 1a).

Testate amoebae, fungal remains, and other NPPs

A total of 19 different NPPs were identified on the same slides used for pollen analysis. Procrustes analysis demonstrated a significant correlation (r = 0.637, p < 0.0001) between the pollen and NPP data sets. The strength of the correlation is driven primarily by highly correlated variation on Procrustes axis 1 (r = 0.814) and less so by variation on axis 2 (r = 0.461). Numerical zonation of the NPP data set identified three statistically significant biostratigraphic zones (Figure 5).

Figure 5.

Concentrations of major non-pollen palynomorphs for Port McNeill Bog. Note changes in scale on the x-axes. Hyalosphenia subflava reaches a maximum of 162,000/cm³ (off scale) at ~400 cal. yr BP. Circles within the plot of Assulina muscorum denote the stratigraphic position of Assulina seminulum tests. Numbers in brackets refer to Pals et al.’s (1980) NPP types.

Nymphaeaceae leaf hair basal cells are the most abundant NPP in NPP Zone 1 (~10,410–7600 cal. yr BP) with a concentration of 65,000 cells/cm³ near the base of the record, followed by decreasing concentrations after 9600 cal. yr BP (Figure 5). Nuphar sclereids (<150 µm) follow the same general trend, decreasing through this zone until no longer present after 8250 cal. yr BP. Single occurrences of Mougeotia and Closterium idiosporum algal zygospores occur in a few samples in this zone. Testate amoebae are minor components of the NPP record until the middle of NPP Zone 1, when Arcella hemisphaerica appears at ~9200 cal. yr BP and then increases to ~3500 tests/cm³. Amphitrema flavum appears in the record at the same time and increases in concentration immediately before NPP Zone 2. Fungal remains are at low concentrations in NPP Zone 1 with Gaeumannomyces hyphopodia consistently present. Gelasinospora, cf. Entophlyctis lobata, Helicoon pluriseptatum, and Microthyriaceae are present in the upper portion of NPP Zone 1.

NPP Zone 2 (7600–4480 cal. yr BP) is dominated by A. flavum, which reaches a maximum of 12,700 tests/cm³ at ~4900 cal. yr BP before disappearing at the boundary to NPP Zone 3. Hyalosphenia subflava tests are present in low concentrations (<2000 tests/cm³). Assulina muscorum is also common, and single occurrences of A. seminulum occur in two samples in this zone. Gaeumannomyces and Gelasinospora continue to be present and cf. E. lobata, H. pluriseptatum, and Microthyriaceae increase relative to NPP Zone 1.

Hyalosphenia subflava testate amoebae dominate NPP Zone 3 (4480 cal. yr BP to present), with a mean concentration of 12,000 tests/cm³ that excludes the peak at ~400 cal. yr BP, where the concentration reaches 162,500 tests/cm³. Assulina muscorum is present in the uppermost samples, and Cyclopyxis arcelloides type, most likely belonging to Cyclopyxis due to its smooth-rimmed and invaginated aperture, appears in the top two samples, reaching 4500 tests/cm³. Gaeumannomyces hyphopodia and Gelasinospora ascospores are at their highest concentrations in this zone, although Gelasinospora decreases to trace amounts in the last 1500 cal. yr. Microthyriaceae, cf. E. lobata, and H. pluriseptatum are also present in NPP Zone 3, except between ~4000 and 2000 cal. yr BP when Gelasinospora ascospores are abundant.

LOI, bulk density and C and N analyses

The mean temporal resolution of the physicochemical records is 64 cal. yr between samples with a range of 14–275 cal. yr. Water content of the PMB core is 90%, on average, with organic matter content accounting for nearly all of the remaining material, that is, 95.6–99.9% LOI (Figure 6). Mean AFBD is 0.09 g/cm³ with values ranging between 0.05 and 0.16 g/cm³ (Figure 6). Variation in AFBD corresponds with stratigraphic changes in peat type (Figure 3), with relatively high bulk density (~0.11 g/cm³) in late to mid-Holocene ligneous peat, low bulk density (~0.07 g/cm³) in mid- to early Holocene Sphagnum peat, and intermediate values (~0.09 g/cm³) in the earliest Holocene herbaceous peat, in agreement with values for different peat types compiled by Loisel et al. (2014).

Figure 6.

Peat stratigraphy and physicochemical records from Peat McNeill Bog, northern Vancouver Island. See Figure 3 for peat stratigraphy legend.

More or less synchronous changes are observed in %C and %N, corresponding with changes in peat type (Figure 6). Mean C content is 55.5% with variation between 47.7% and 61.5% (Figure 6); these values agree with or exceed those found in northern peatlands (Loisel et al., 2014). In most samples, %C exceeds the mean C content of peat organic matter in western Canada (i.e. 51.8%; Vitt et al., 2000) that is routinely used to infer peat C accumulation from bulk density (e.g. Yu et al., 2014). The early to mid-Holocene Sphagnum peat and the uppermost moss-ligneous peat are characterized by relatively low %C (52% on average), whereas the basal herbaceous peat and the mid-Holocene ligneous peat have ~57% C. Mean N content is 1.4% and is relatively high (~2.1%) in early Holocene herbaceous peat and low (~0.7%) in the early to mid-Holocene Sphagnum peat with increasing values (~1%) toward the present. Accordingly, C:N varies considerably: mean values in the herbaceous, Sphagnum, and ligneous peats are 28, 90, and 67, respectively.

As expected, changes in δ¹³C and δ¹⁵N isotopes also follow stratigraphic changes in peat type (Figure 6). δ¹³C is highest in the Sphagnum and moss-ligneous peats with mean values of −26.6‰ and lowest in the ligneous and herbaceous peats, where mean values are −27.6‰. There is a weak positive correlation (r = 0.38, p < 0.001) between δ¹³C and C:N, suggesting that diagenesis may have played some role in altering δ¹³C values, particularly in the mid- to early Holocene Sphagnum peat where the correlation is stronger (r = 0.57, p < 0.001). δ¹⁵N reaches up to 3.7‰ in the late to mid-Holocene ligneous peat and is low in the preceding Sphagnum peat (~−3.9‰) and intermediate in the early Holocene herbaceous peat (~−2.3‰). There is no correlation between δ¹⁵N and C:N (r = −0.09, p = 0.254).

CARs are 16.1 g/m²/cal. yr, on average, and vary between 2.4 and 48.7 g/m²/cal. yr, following changes in peat stratigraphy (Figure 7). CAR is highest in the early Holocene herbaceous peat (mean = 30.6 g/m²/cal. yr), gradually decreasing to low rates at ~7000–5500 cal. yr BP during accumulation of Sphagnum peat. In the mid- to late Holocene, CAR increases with accumulation of ligneous peat, although rates are somewhat lower at ~2000–900 cal. yr BP. CAR is marked by an increase in the uppermost peat that is likely related to low decomposition in the acrotelm. The time-weighted mean NAR is 0.41 g/m²/cal. yr. NAR is highest (mean = 1.15 g/m²/cal. yr) in the earliest Holocene and lowest (mean = 0.11 g/m²/cal. yr) during accumulation of Sphagnum peat in the mid-Holocene (Figure 7). Because both CAR and NAR are calculated using peat accumulation rates (Figure 2) and AFBD (Figure 6), NAR follows the same temporal trend as CAR (Figure 7).

Figure 7.

Summary peat composition, C:N ratios, carbon accumulation rate (CAR), and nitrogen accumulation rate (NAR) at Port McNeill Bog, Vancouver Island. Relatively wet conditions are inferred from herbaceous, brown moss and unidentifiable organic material percentages and relatively dry conditions from Sphagnum and Ericaceae percentages, following Loisel and Yu (2013). Also shown are pollen-inferred mean annual precipitation (MAP) for southwest British Columbia (Mathewes and Heusser, 1981), C₃₇ alkenone–inferred sea surface temperature (SST) from immediately west of Vancouver Island (Kienast and McKay, 2001), and January and July insolation anomaly at 50°N (Berger and Loutre, 1991).

Discussion

Holocene vegetation history on northern Vancouver Island

Pollen records from northern Vancouver Island (Hebda, 1983; Hebda and Haggarty, 1997; Lacourse, 2005), the adjacent mainland coast (Galloway et al., 2007, 2009; Stolze et al., 2007) and the exposed continental shelf at Cook Bank (Lacourse et al., 2003) demonstrate that the region supported open P. contorta woodland soon after deglaciation, which occurred at ~16,500 cal. yr BP in this region (Hebda, 1983; Howes, 1981; Luternauer et al., 1989). With increasing temperatures (Heusser et al., 1985), open Pinus-dominated communities were followed by mixed coniferous forests that included P. sitchensis, T. mertensiana, and T. heterophylla, species with higher relative shade tolerance, longer lifespans, and slower relative growth rates (Lacourse, 2009). At some sites, the cool Younger Dryas is marked by increases in the abundance of T. mertensiana (e.g. Lacourse, 2005; Mathewes, 1993; Stolze et al., 2007).

At PMB, the pollen record begins at ~10,400 cal. yr BP (Figure 4) during mixed coniferous forest that characterized much of coastal BC in the early Holocene (e.g. Galloway et al., 2009; Hebda, 1983; Heusser, 1960; Stolze et al., 2007). Until about 8000 cal. yr BP, forests on northern Vancouver Island were composed primarily of T. heterophylla, P. sitchensis, and A. rubra and were likely somewhat more open than at the present, as indicated by abundant Polypodiaceae ferns including P. aquilinum. Alnus pollen can be overrepresented in modern pollen assemblages compared to its abundance in nearby plant communities (e.g. Allen et al., 1999). Therefore, it is possible that the abundant A. rubra pollen at PMB in the early Holocene may reflect, at least in part, regional pollen transport linked to high pollen production and effective wind dispersal; however, modern pollen assemblages similarly dominated by T. heterophylla, P. sitchensis, and A. rubra have been retrieved from lowland lakes in P. sitchensis–dominated forest to the south in coastal Washington (Gavin et al., 2005; Heusser, 1978). Early Holocene pollen assemblages at PMB are nearly identical to those at Misty Lake (Lacourse, 2005), located 14 km northwest of PMB (Figure 1), indicating that pollen assemblages from peat deposits at PMB provide a similar record of regional forest composition as those derived from nearby lake sediments.

Pseudotsuga menziesii was at its highest abundance in the PMB pollen record between 10,000 and 8250 cal. yr BP, reaching a maximum of 4% (Figure 4). Because P. menziesii produces less pollen relative to other conifers and its pollen has a short dispersal distance (Tsukada, 1982), P. menziesii pollen greater than 1–2% indicates its local presence in surrounding forests (Hebda, 1983; Heusser, 1978). Its local presence is confirmed by P. menziesii cone scales and seed wings in nearby lacustrine sediments that date to ~9300 cal. yr BP (8300 ± 70 ¹⁴C yr BP; GSC-2435 (Howes, 1981)). Today, Pseudotsuga is the dominant tree in forests on southeastern Vancouver Island, where the climate is significantly warmer and drier, but it is uncommon on northern Vancouver Island. Warm early Holocene summers permitted P. menziesii to migrate northward on Vancouver Island (this study; Hebda, 1983; Lacourse, 2005) and the adjacent mainland (Galloway et al., 2007). Once climate began to cool, the northern limit of P. menziesii retracted (Gugger and Sugita, 2010); by about 7200 cal. yr BP, P. menziesii more or less disappears from the PMB pollen record. A decline in P. menziesii was observed at about the same time in the pollen record from nearby Misty Lake (Lacourse, 2005).

By 8000 cal. yr BP, T. heterophylla became the dominant tree species near PMB (Figure 4) and throughout the region (Galloway et al., 2007, 2009; Hebda, 1983; Lacourse, 2005; Stolze et al., 2007), forming forests along with P. sitchensis, Abies, and Cupressaceae. Increases in Abies and Cupressaceae, likely mostly derived from Thuja plicata, indicate an increasingly closed forest canopy, given their high shade tolerance (Lacourse, 2009). Climate was cooler and wetter on northern Vancouver Island than in the earliest Holocene, due to lower summer insolation and intensification of the Aleutian Low pressure system that allowed near-modern precipitation regimes to establish (Brown et al., 2006; Heusser et al., 1985). This cooler, wet climate facilitated the establishment of T. heterophylla–dominated rainforest that continue to occupy much of the northeast Pacific coast today. The dramatic increase in A. rubra–type pollen in the uppermost sample from PMB is likely associated with disturbance in surrounding forests, with commercial logging in the region beginning in the 1930s.

Local bog development near Port McNeill

The base of the PMB sequence is characterized by rapidly accumulating herbaceous peat (Figures 2 and 3) with relatively high C content (~56%) and C:N ratios of ~25 (Figure 6) that are indicative of a terrestrial environment as opposed to a lake (Meyers and Teranes, 2001). However, Nuphar pollen, sclereids, and basal cells (Figures 3 –5) indicate the presence of standing water and/or bog pools necessary to support these aquatic plants. Analogous peat-accumulating wetlands with N. polysepala occur in saturated low-lying areas in Pacific Canada (National Wetlands Working Group, 1988), and N. polysepala, often with erect aerial leaves as opposed to floating lily pads, occurs in shallow bog hollows and pools on Vancouver Island today (Golinski, 2004). Etnier and Villani (2007) showed that sclereids in Nymphaea odorata (Nymphaeaceae) confer structural support to aerial leaves and occur in higher frequency in the epidermis (rind) of aerial leaf petioles, compared to floating lily pads. The high abundance of Nuphar sclereids (Figures 3 and 5) in the basal peat at PMB may very well reflect the presence of such aerial leaves. The relatively high %N and low δ¹³C (Figure 6) of this herbaceous peat also suggest the presence of shallow water and/or bog pools: peat that accumulates in aquatic settings tends to have higher %N and lower δ¹³C due to the presence of aquatic plants and/or algal proteins and lipids that are typically N-rich and ¹³C-depleted (Jones et al., 2010; Meyers and Teranes, 2001; Talbot, 2001).

The transition to a classic bog ecosystem begins with the appearance of Sphagnum spores and Ericaceae pollen (Figure 4), as well as associated macrofossils (Figure 3), shortly after 10,000 cal. yr BP. An increase in the abundance of A. hemisphaerica, a wet peatland testate amoeba (Markel et al., 2010; Payne et al., 2012), after 9200 cal. yr BP and a corresponding increase in C:N ratios (Figure 6) as well as decreases in Nuphar remains (Figures 3 –5) suggest shallowing of standing water and/or decreases in the extent of bog pools through the earliest Holocene. Increasing concentrations of fungal groups, which require oxic conditions to be major decomposers in peatlands (Wieder and Vitt, 2006; Rydin and Jeglum, 2013), also reflect the lowering of relative water table depths after ~8500 cal. yr BP (Figure 5). It is difficult to uncouple climatic and non-climatic drivers of change at PMB during this interval, and it is likely that both macroscale climate and local eco-hydrological processes played a role. The decreasing relative water table depth would have been facilitated by warm and relatively dry climate during the early Holocene as well as the rapid accumulation of peat, allowing bog vegetation, that is, Sphagnum mosses and Ericaceae shrubs to colonize the surface as the wetland dried up. A similar early Holocene transition from a shallow water wetland with Nuphar to a Sphagnum bog was also recorded at nearby Bear Cove Bog (Figure 1; Hebda, 1983). However, in bogs on the north coast of BC (e.g. Banner et al., 1983; Turunen and Turunen, 2003), paludification was the primary mode of development, with succession from wet minerotrophic coniferous forest to blanket bog-woodland ecosystems in the early to mid-Holocene. These developmental pathways contrast with the swamp-fen-bog sequence typical of many boreal peatlands (Wieder and Vitt, 2006; Rydin and Jeglum, 2013).

A Sphagnum-dominated bog was fully established by 8000 cal. yr BP, as shown by the abrupt stratigraphic change to Sphagnum peat (Figure 3), the abundance of Sphagnum leaves/stems, and increases in Sphagnum spores (Figure 4). A rapid transition to lower C and N content (~50% and 0.5%, respectively) but C:N ratios >100 (Figure 6) coincides with the change to Sphagnum peat and dominance by Sphagnum mosses, which, although resistant to decay, have lower C and N content than vascular plants (Loisel et al., 2014). Furthermore, δ¹³C values become more positive, reflecting peat accumulation above the water table, changes in surface vegetation, and perhaps increased decomposition (Andersson et al., 2012; Jones et al., 2010).

Increases in Ericaceae roots and leaves (Figure 3) and pollen including Ledum type (Figure 4) indicate that ericaceous shrubs also colonized the bog surface by 8000 cal. yr BP. Fungi including H. pluriseptatum, Microthyriaceae and cf. E. lobata, which are often associated with Sphagnum peat and to some extent ericaceous shrubs (e.g. Kuhry, 1997; Yeloff et al., 2007), also appear in the PMB record at about the same time (Figure 5), and changes in their concentrations correlate well with the varying abundance of Sphagnum remains (Figure 3). Testate amoebae during this time suggest variable hydrological conditions and/or the development of hummock-hollow surface topography. Amphitrema flavum, by far the most abundant testate amoeba between 7900 and 4700 cal. yr BP, is one of the most common species in ombrotrophic peatlands and is generally found in wet conditions (Charman et al., 2000). A. muscorum, which can tolerate a wide range of conditions but is most abundant in intermediate to dry peatlands (Booth and Zygmunt, 2005; Charman et al., 2000; Payne et al., 2012), is also present. Hyalosphenia subflava, a testate amoebae indicative of dry conditions when in great abundance (Charman et al., 2000; Payne et al., 2012), also occurs in the mid-Holocene, although it is far more abundant after 4000 cal. yr BP, when ericaceous shrubs dominate the bog’s vascular plant community, and the water table continued to lower relative to the bog surface.

Around 5200 cal. yr BP, Ericaceae macrofossils increase in relative abundance, and by 4000 cal. yr BP, PMB was dominated by ericaceous shrubs, as shown by the transition to ligneous peat (Figure 3) and the increase in Ericaceae pollen (Figure 4). Pinus contorta–type pollen percentages increase around 3500 cal. yr BP, and given a coincident increase in Pinus pollen concentrations, this likely reflects colonization of the bog surface by this species. As expected, bulk density and C and N contents (Figure 5) increase and C:N ratios decrease during accumulation of this ligneous peat and dominance by ericaceous shrubs, suggesting increased peat decomposition at this time. δ¹⁵N transitions to more positive values after 4000 cal. yr BP, reflecting a lower water table that allowed aerobic decay and ¹⁵N enrichment (Andersson et al., 2012; Jones et al., 2010). Lowering of the water table is also suggested by substantial increases in H. subflava (Payne et al., 2012) after 4000 cal. yr BP (Figure 5). δ¹³C decreases with the transition to an Ericaceae-dominated bog, which is likely related to a number of interacting factors including the change in plant functional types and increased microbial decomposition and summer moisture stress related to lowering of the water table. Given the cool, wet climate of the mid- to late Holocene, the increase in ericaceous shrubs and relative lowering of the water table appear to have been driven primarily by succession and autogenic processes rather than changes in climatic conditions.

The abundance of macroscopic charcoal increases fairly abruptly starting 5200 cal. yr BP with the highest concentrations occurring at 3000–1500 cal. yr BP (Figure 5). The concentration of Gelasinospora ascospores, which are often associated with dry conditions and peat containing abundant charcoal (e.g. Chambers et al., 2011; Kuhry, 1997; Yeloff et al., 2007), follows the same general trend through time (Figure 5). Brown and Hebda (2002) suggest that because regional climate was cool and moist in coastal BC during the mid- to late Holocene, higher charcoal concentrations on southern Vancouver Island at this time may reflect the use of fire as a landscape management tool by local indigenous peoples. It is possible that small, low-severity fires were used at PMB to increase berry production in ericaceous shrubs, similar to that documented on Vancouver Island and elsewhere in BC during the historical period (Turner, 1999). Since no distinct charcoal layers were observed in the peat core from PMB, it is unlikely that fires were sufficiently severe to combust significant amounts of peat and/or the entire surface vegetation.

The mixed moss-ligneous peat in the uppermost portion of the PMB sequence (0–30.5 cm) may reflect changes in environmental conditions over the last 700 cal. yr; however, because much of this peat sits above the current depth of the water table (20 cm) and therefore corresponds with lower decomposition in the acrotelm, we limit our interpretation of changes in the various proxies for this portion of the core. The uppermost peat is marked most notably by the increased relative abundance of Drepanocladus cf. exannulatus, a brown moss common in wet BC bogs that suggests an increase in surface wetness, potentially associated with recent land use changes adjacent to the bog. The increase in Drepanocladus is likely responsible for the decrease in %C in the uppermost peat, as brown mosses typically contribute less C to peat than vascular plants (Loisel et al., 2014). With the increased abundance of brown moss in the uppermost peat and presumably a higher water table to support these mosses, δ¹³C values increase to levels that are generally higher than in the mid-Holocene. Assulina muscorum and Cyclopyxis arcelloides type, testate amoebae that can tolerate varying levels of moisture (Booth and Zygmunt, 2005; Charman et al., 2000; Payne et al., 2012), also appear in the uppermost peat (Figure 5).

The abundance of Sphagnum spores (Figure 4) varies considerably after the bog established in the early Holocene and does not correspond with changes in the abundance of Sphagnum macro-remains or overall peat type (Figure 3). For example, Sphagnum spores reach a maximum abundance of ~30% between 3200 and 2700 cal. yr BP during accumulation of ligneous peat as opposed to Sphagnum peat. Similarly, S. capillifolium and S. fuscum are abundant on the bog surface today, but Sphagnum spores are less than 4% in the uppermost peat. These differences suggest that the Sphagnum spore record from PMB does not accurately reflect changes in the abundance of Sphagnum moss at the site. Instead, Sphagnum spore frequencies likely reflect changes in sporophyte production and spore release through time with peaks corresponding with environmental conditions that were ideal for sexual reproduction. Water table depth is a major determinant of reproductive mode, and even if sporophytes are produced, sexual reproduction may be unsuccessful if sporophytes are inundated or desiccate before maturation, both of which limit spore production and dispersal (Sundberg, 2002). It is also possible that temporal variability in the frequency of Sphagnum spores reflects changes in the abundance of different Sphagnum species with disparate spore production and/or changes in the microtopography of the bog surface, as hummocks and hollows typically host different Sphagnum species (Sundberg, 2002).

Procrustes analysis demonstrates that the pollen and NPP data sets share similar overall structure with biostratigraphic changes that are comparable in timing and magnitude. The significant congruence (r = 0.637, p < 0.0001) is noteworthy given that the two proxy records provide information about different aspects of environmental change at different spatial scales: the pollen assemblages primarily reveal regional vegetation dynamics, whereas the NPP data provide a record of local bog development and changing edaphic and hydrological conditions. The congruence of the pollen and NPP data sets suggests the importance of macroscale climate as a shared long-term driver of change in regional forest composition and local bog development. However, relatively high residuals in the Procrustes analysis, which indicate dissimilarity between the two proxies, occur between 5600 and 7500 cal. yr BP and underscore the importance of local-scale factors in peatland dynamics. During this interval, increases in A. flavum and A. muscorum testate amoebae reflect local changes in hydrological conditions, whereas pollen assemblages show relatively stable regional forest composition through much of the Holocene.

Long-term C and N accumulation

The mean CAR at PMB (16.1 g/m²/cal. yr) is almost double that of the only other Holocene C accumulation record available from coastal BC: at a slope bog on BC’s north coast (Turunen and Turunen, 2003), the long-term rate of C accumulation is only 8.6 g/m²/cal. yr (as reported by Loisel et al., 2014). At peatlands further north along the coast in southern Alaska, mean C accumulation is similar to these two records with rates between 8.9 and 18.2 g/m²/cal. yr (Jones and Yu, 2010; Loisel et al., 2014; Nichols et al., 2014). C accumulation at these oceanic sites in the northeast Pacific is lower than Loisel et al.’s (2014) estimate for northern peatlands (22.9 g/m²/cal. yr) and significantly lower than rates in continental fens in western Canada (32.5 g/m²/cal. yr; Yu et al., 2014) and in peatlands on the Tibetan Plateau (31.1 g/m²/cal. yr; Zhao et al., 2014). Continental peatlands tend to accumulate more C, on average, than oceanic bogs for a number of reasons: higher bulk density of peat, greater seasonality in temperature which promotes primary production in the summer and reduces decomposition in the winter, and lower likelihood of drought if groundwater sources are present (Asada and Warner, 2005; Yu et al., 2014). Although oceanic bogs tend to have higher net primary production than continental peatlands, likely due to a protracted growing season and abundant precipitation, this is accompanied by higher decomposition that acts as a counterbalance, leading to lower overall peat accumulation and C sequestration (Asada and Warner, 2005; Malmer and Wallén, 1993). Gorham et al. (2003) found a strong negative correlation (r = −0.76, p < 0.05) between mean annual precipitation and long-term rates of peat accumulation in North American peatlands (n = 21 peat cores), but their study did not include sites with 1400–2900 mm/yr of precipitation. PMB with 1900 mm/yr and a long-term peat accumulation rate of 29.3 g/m²/cal. yr helps to fill that sampling gap. Our results from northern Vancouver Island are in line with Gorham et al.’s (2003) predicted negative relationship between precipitation and peat accumulation, lending further support to the notion that long-term rates of peat accumulation are inversely related to precipitation.

Although C accumulation at PMB and other sites in the northeast Pacific are, on average, lower than in continental settings, the temporal trend in C accumulation over the Holocene is similar to that at many peatlands in the Northern Hemisphere (Loisel et al., 2014; Yu et al., 2009). At PMB, C accumulation is highest (~31 g/m²/cal. yr) during the earliest Holocene, when temperatures were higher, precipitation was lower, and seasonal cycle of insolation was amplified relative to the present (Figure 7). As in many northern peatlands (Loisel et al., 2014), maximum CARs coincide with the timing of the Holocene thermal maximum. Yu et al. (2014) point out that greater seasonality in the early Holocene (i.e. low winter insolation and high summer insolation) would have favored primary production in summer and reduced decomposition during winter, leading to high C accumulation. Through the mid- to late Holocene, when climate was generally cooler, wetter, and less seasonal than in the early Holocene (Figure 7), C accumulation at PMB is markedly lower (14 g/m²/cal. yr, on average), as it is in most northern peatlands.

N accumulation in peatlands has received far less attention, and there is a paucity of studies on N accumulation in Pacific North America peatlands to compare to our record from northern Vancouver Island. At PMB, the time-weighted mean NAR is 0.4 g/m²/cal. yr, and this compares well with Loisel et al.’s (2014) estimate of 0.5 g/m²/cal. yr for northern peatlands over the Holocene. Overall, N accumulation at PMB follows C accumulation, as would be expected given that both are calculated using the same bulk density measurements and accumulation rates and both are related to primary productivity and decomposition. N accumulation is relatively high (~1.2 g/m²/cal. yr) in the warm and drier early Holocene and low (~0.2 g/m²/cal. yr) after 7500 cal. yr BP under cooler, wetter, and less seasonal climate. The correlation between Holocene changes in climate, in particular seasonality, and C and N accumulation underscores the importance of macroscale climate as a dominant control on long-term C and N accumulation in peatlands.

Although Holocene trends in C and N accumulation follow long-term changes in climate, changes in plant functional groups and local hydrological conditions also play an important role in driving fluctuations in accumulation rates. At PMB, C and N accumulation are high during accumulation of early Holocene herbaceous peat, when C:N ratios are relatively low and multiple proxies suggest the presence of standing water and/or bog pools. Despite lower precipitation, wet conditions at PMB would have lowered overall decay and facilitated higher rates of C and N accumulation in the early Holocene. C and N accumulation decrease abruptly at ~8000 cal. yr BP, when the water table lowered relative to the surface and the site became a Sphagnum bog. By about 5000 cal. yr BP, C and N accumulation show notable increases with further lowering of the water table, the transition to an Ericaceae-dominated bog and the associated accumulation of ligneous peat. The decrease in C accumulation at PMB between ~3000 and 1000 cal. yr BP coincides generally with not only Neoglacial cooling and glacial advances in the adjacent Coast Mountains of BC (e.g. Coulthard et al., 2013) but also an interval of high charcoal concentrations (Figure 3). It is likely that disturbance from fire removed some portion of the bog’s surface vegetation and enhanced decomposition. Slower net peat addition during this time likely reflects an increase in decomposition rather than a substantial decrease in productivity.

Conclusion

In the warm, relatively dry early Holocene, PMB was a herb-dominated peatland with Nuphar in standing water and/or bog pools, set within mixed coniferous forest with scattered P. menziesii. C and N accumulation was highest during this interval with mean rates of 30.6 and 1.2 g/m²/cal. yr, respectively. By about 8000 cal. yr BP, regional forests transitioned under a cooler, wetter climate to T. heterophylla–dominated rainforests, similar to those that still occupy the region today. On a local scale, lowering of relative water table depth facilitated the development of a Sphagnum bog by 8000 cal. yr BP, with an increasing abundance of ericaceous shrubs and further lowering of the water table after 5000 cal. yr BP. CAR and NAR were low during accumulation of Sphagnum peat, particularly between ~7000 and 5500 cal. yr BP, and then increased during accumulation of ligneous peat in the mid- to late Holocene. Temporal changes in the frequency and concentration of Sphagnum spores do not correlate well with stratigraphic changes in Sphagnum macro-remains or peat type, suggesting that Sphagnum spore records may not always reflect the abundance of Sphagnum moss accurately.

This is the first multi-proxy peat study in coastal BC that spans the full Holocene, and our results underscore the importance of using a multi-proxy approach in peat-based paleoenvironmental studies. Procrustes analysis is an effective numerical tool for revealing similarities and differences in proxy records; however, physicochemical records are essential for a more complete understanding of bog development and peat C and N accumulation. Our study shows that changes in plant functional groups, local hydrological conditions, and macroscale climate interact to affect C and N accumulation in bog ecosystems on long ecological timescales. Peat C accumulation on northern Vancouver Island was significantly higher during the early Holocene, but the overall rate of accumulation is significantly lower than in continental peatlands. This suggests that differences in seasonality play an important role in driving C accumulation, both on long temporal scales and at large spatial scales. Additional research is needed from other sites on the northeast Pacific coast in order to better understand how long-term changes in climate, vegetation, and hydrology have shaped peatland development and C and N accumulation in this oceanic setting.

Footnotes

Acknowledgements

We thank D Canil for field assistance; D Peteet for use of laboratory facilities; S Mazumder for help with isotope analyses; T Johannesson, DG Gavin, and S Goring for sharing R code; and R Hebda for discussion. We appreciate the comments of two anonymous reviewers and editorial handling of the manuscript by Z Yu.

Funding

Funding was provided through research grants from the Natural Sciences and Engineering Research Council of Canada (No. 342003) and Canadian Foundation for Innovation (No. 17214) to T Lacourse and through a Marie Tharp Fellowship from Lamont-Doherty Earth Observatory at Columbia University to T Lacourse.

References

Allen

Brown

Hebda

(1999) Surface pollen spectra from southern Vancouver Island, British Columbia, Canada. Canadian Journal of Botany 77: 786–799.

Andersson

Meyers

Hornibrook

. (2012) Elemental and isotopic carbon and nitrogen records of organic matter accumulation in a Holocene permafrost peat sequence in the East European Russian Arctic. Journal of Quaternary Science 27(6): 545–552.

Asada

Warner

(2005) Surface peat mass and carbon balance in a hypermaritime peatland. Soil Science Society of America Journal 69: 549–562.

Baird

Belyea

Comas

. (eds) (2009) Carbon Cycling in Northern Peatlands: AGU Monograph 184. Washington, DC: American Geophysical Union.

Banner

Pojar

Rouse

(1983) Postglacial paleoecology and successional relationships of a bog woodland near Prince Rupert, British Columbia. Canadian Journal of Forest Research 13: 938–947.

Barber

Chambers

Maddy

. (1994) A sensitive high-resolution record of the late Holocene climatic change from a raised bog in northern England. The Holocene 4(2): 198–205.

Bennett

(1996) Determination of the number of zones in a biostratigraphical sequence. New Phytologist 132: 155–170.

Berger

Loutre

(1991) Insolation values for the climate of the last 10 million of years. Quaternary Science Reviews 10: 297–317.

Blaauw

(2010) Methods and code for ‘classical’ age-modelling of radiocarbon sequences. Quaternary Geochronology 5: 512–518.

10.

Blaauw

Christen

(2011) Flexible paleoclimate age-depth models using an autoregressive gamma process. Bayesian Analysis 6: 457–474.

11.

Booth

Zygmunt

(2005) Biogeography and comparative ecology of testate amoebae inhabiting Sphagnum-dominated peatlands in the Great Lakes and Rocky Mountain regions of North America. Diversity and Distributions 11: 577–590.

12.

Brown

Hebda

(2002) Origin, development, and dynamics of coastal temperate conifer rainforests of southern Vancouver Island, Canada. Canadian Journal of Forest Research 32: 353–372.

13.

Brown

Fitton

Schoups

. (2006) Holocene precipitation in the coastal temperate rainforest complex of southern British Columbia, Canada. Quaternary Science Reviews 25: 2762–2779.

14.

Chambers

van Geel

van der Linden

(2011) Considerations for the preparation of peat samples for palynology, and for the counting of pollen and non-pollen palynomorphs. Mires and Peat 7: Article 11 (14 pp.).

15.

Charman

(2002) Peatlands and Environmental Change. Chichester: John Wiley & Sons.

16.

Charman

Hendon

Woodland

(2000) The Identification of Testate Amoebae (Protozoa: Rhizopoda) in Peats. QRA technical guide no. 9. London: Quaternary Research Association.

17.

Charman

Beilman

Blaauw

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

18.

Clymo

(1984) The limits to peat bog growth. Philosophical Transactions of the Royal Society of London Series B: Biological Sciences 303: 605–654.

19.

Clymo

Turunen

Tolonen

(1998) Carbon accumulation in peatland. Oikos 81: 368–388.

20.

Coulthard

Smith

Lacourse

(2013) Dendroglaciological investigations of mid- to late-Holocene glacial activity in the Mt. Waddington area, British Columbia Coast Mountains, Canada. The Holocene 23(1): 93–103.

21.

Dean

(1974) Determination of carbonate and organic matter in calcareous sediments and sedimentary rocks by loss on ignition: Comparison with other methods. Journal of Sedimentary Petrology 44: 242–248.

22.

Environment Canada (2014) Normales climatiques au Canada, 1981–2010 [Canadian climate normals, 1981–2010]. Meteorological Service of Canada, Environment Canada. Available at: http://climate.weatheroffice.ec.gc.ca/climate_normals/index_e.html

23.

Etnier

Villani

(2007) Differences in mechanical and structural properties of surface and aerial petioles of the aquatic plant Nymphaea odorata subsp. tuberosa (Nymphaeaceae). American Journal of Botany 94: 1067–1072.

24.

Frolking

Roulet

(2007) Holocene radiative forcing impact of northern peatland carbon accumulation and methane emissions. Global Change Biology 13: 1–10.

25.

Galloway

Doherty

Patterson

. (2009) Postglacial vegetation and climate dynamics in the Seymour-Belize Inlet Complex, central coastal British Columbia, Canada: Palynological evidence from Tiny Lake. Journal of Quaternary Science 24: 322–335.

26.

Galloway

Patterson

Doherty

. (2007) Multi-proxy evidence of postglacial climate and environmental change at Two Frog Lake, central mainland coast of British Columbia, Canada. Journal of Paleolimnology 38: 569–588.

27.

Gavin

Brubaker

McLachlan

. (2005) Correspondence of pollen assemblages with forest zones across steep environmental gradients, Olympic Peninsula, Washington, USA. The Holocene 15(5): 648–662.

28.

Golinski

(2004) Mires of Vancouver Island, British Columbia: Vegetation classification and differences between disturbed and undisturbed mires. PhD Dissertation, University of Victoria.

29.

Gorham

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

30.

Gorham

Janssens

Glaser

(2003) Rates of peat accumulation during the postglacial period in 32 sites from Alaska to Newfoundland with special emphasis on northern Minnesota. Canadian Journal of Botany 81: 429–438.

31.

Gugger

Sugita

(2010) Glacial populations and postglacial migration of Douglas-fir based on fossil pollen and macrofossil evidence. Quaternary Science Reviews 29: 2052–2070.

32.

Hebda

(1983) Late-glacial and postglacial vegetation history at Bear Cove Bog, northeast Vancouver Island, British Columbia. Canadian Journal of Botany 61: 3172–3192.

33.

Hebda

Haggarty

(1997) Brooks Peninsula: An Ice Age Refugium on Vancouver Island (Occasional paper no. 5). Victoria: British Columbia, Ministry of Environment, Lands and Parks.

34.

Heiri

Lotter

Lemcke

(2001) Loss on ignition as a method for estimating organic and carbonate content in sediments: Reproducibility and comparability of results. Journal of Paleolimnology 25: 101–110.

35.

Heusser

(1960) Late Pleistocene Environments of North Pacific North America. New York: American Geographical Society.

36.

Heusser

(1978) Modem pollen rain of Washington. Canadian Journal of Botany 56: 1510–1517.

37.

Heusser

Mathewes

(1985) Late Quaternary climatic change on the American North Pacific Coast. Nature 315: 485–487.

38.

Howes

(1981) Late Quaternary sediments and geomorphic history of north-central Vancouver Island. Canadian Journal of Earth Sciences 18: 1–12.

39.

Howie

van Meerveld

(2013) Regional and local patterns in depth to water table, hydrochemistry and peat properties of bogs and their laggs in coastal British Columbia. Hydrology and Earth System Sciences 17: 3421–3435.

40.

Huntley

MJW

Mathewes

Shotyk

(2013) High-resolution palynology, climate change, and human impact on a Late Holocene peat bog on Haida Gwaii, British Columbia, Canada. The Holocene 23(11): 1572–1583.

41.

Johannesson

Bjornsson

(2012) stinepack: Stineman, a consistently well behaved method of interpolation (R package version 1.3). Available at: http://CRAN.R-project.org/package=stinepack

42.

Jones

(2010) Rapid deglacial and early Holocene expansion of peatlands in Alaska. Proceedings of the National Academy of Sciences of the United States of America 107: 7347–7352.

43.

Jones

Peteet

Sambrotto

(2010) Late-glacial and Holocene δ¹⁵N and δ¹³C variation from a Kenai Peninsula, Alaska peatland. Palaeogeography, Palaeoclimatology, Palaeoecology 293: 132–143.

44.

Kapp

Davis

King

(2000) Pollen and Spores. 2nd Edition. Dallas, TX: American Association of Stratigraphic Palynologists Foundation.

45.

Kienast

McKay

(2001) Sea surface temperatures in the subarctic northeast Pacific reflect millennial-scale climate oscillations during the last 16 kyrs. Geophysical Research Letters 28: 1563–1566.

46.

Korhola

Ruppel

Seppä

. (2010) The importance of northern peatland expansion to the late-Holocene rise of atmospheric methane. Quaternary Science Reviews 29: 611–617.

47.

Kuhry

(1997) The paleoecology of a treed bog in western boreal Canada: A study based on microfossils, macrofossils, and physic-chemical properties. Review of Palaeobotany and Palynology 96: 183–224.

48.

Lacourse

(2005) Late Quaternary dynamics of forest vegetation on northern Vancouver Island, British Columbia, Canada. Quaternary Science Reviews 24: 105–121.

49.

Lacourse

(2009) Environmental change controls postglacial forest dynamics through interspecific differences in life-history traits. Ecology 90: 2149–2160.

50.

Lacourse

Mathewes

Fedje

(2003) Paleoecology of late-glacial terrestrial deposits with in situ conifers from the submerged continental shelf of western Canada. Quaternary Research 60: 180–188.

51.

Limpens

Berendse

Blodau

. (2008) Peatlands and the carbon cycle: From local processes to global implications – A synthesis. Biogeosciences 5: 1475–1491.

52.

Loisel

(2013) Surface vegetation patterning controls carbon accumulation in peatlands. Geophysical Research Letters 40: 5508–5513.

53.

Loisel

Beilman

. (2014) A database and synthesis of northern peatland soil properties and Holocene carbon and nitrogen accumulation. The Holocene 24(9): 1028–1042.

54.

Luternauer

Conway

Clague

. (1989) Late Quaternary geology and geochronology of the central continental shelf of western Canada. Marine Geology 89: 57–68.

55.

McAndrews

Berti

Norris

(1973) Key to the Quaternary Pollen and Spores of the Great Lakes Region (Royal Ontario Museum Life Sciences Miscellaneous Publication). Toronto: University of Toronto Press.

56.

Magnan

Garneau

(2014) Climatic and autogenic control on Holocene carbon sequestration in ombrotrophic peatlands of maritime Quebec, eastern Canada. The Holocene 24(9): 1054–1062.

57.

Malmer

Wallén

(1993) Accumulation and release of organic matter in ombrotrophic bog hummocks – Processes and regional variation. Ecography 16: 193–211.

58.

Markel

Booth

Qin

(2010) Testate amoebae and δ¹³C of Sphagnum as surface-moisture proxies in Alaskan peatlands. The Holocene 20(3): 463–475.

59.

Marrotte

Chmura

Stone

(2012) The utility of Nymphaeaceae sclereids in paleoenvironmental research. Review of Palaeobotany and Palynology 169: 29–37.

60.

Mathewes

(1993) Evidence for Younger Dryas-age cooling on the north Pacific coast of America. Quaternary Science Reviews 12: 321–331.

61.

Mathewes

Heusser

(1981) A 12,000 year palynological record of temperature and precipitation trends in southwestern British Columbia. Canadian Journal of Botany 59: 707–710.

62.

Mauquoy

Hughes

PDM

van Geel

(2010) A protocol for plant macrofossil analysis of peat deposits. Mires and Peat 7: 6.1–6.5.

63.

May

Lacourse

(2012) Morphological differentiation of Alnus (alder) pollen from western North America. Review of Palaeobotany and Palynology 180: 15–24.

64.

Meyers

Teranes

(2001) Sediment organic matter. In: Last

Smol

(eds) Tracking Environmental Change Using Lake Sediments: Physical and Geochemical Methods, vol. 2. Dordrecht: Kluwer Academic Publishers, pp. 239–265.

65.

National Wetlands Working Group (1988) Wetlands of Canada (Ecological Land Classification Series No. 24). Ottawa: Environment Canada and Montreal: Polyscience Publications.

66.

Nichols

Peteet

Moy

. (2014) Impacts of climate and vegetation change on carbon accumulation in a south-central Alaskan peatland assessed with novel organic geochemical techniques. The Holocene 24(9): 1146–1155.

67.

Oksanen

Blanchet

Kindt

(2013) vegan: Community ecology package (R package version 2.0-10). Available at:http://CRAN.R-project.org/package=vegan.

68.

Pals

van Geel

Delfos

(1980) Paleoecological studies in the Klokkeweel Bog near Hoogkarspel (Prov. of Noord-Holland). Review of Palaeobotany and Palynology 30: 371–418.

69.

Payne

Lamentowicz

van der Knaap

. (2012) Testate amoebae in pollen slides. Review of Palaeobotany and Palynology 173: 68–79.

70.

Peres-Neto

Jackson

(2001) How well do multivariate data sets match? The advantages of a Procrustean superimposition approach over the Mantel test. Oecologia 129: 169–178.

71.

R Core Team (2014) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria (version 3.1.0.). Available at:http://www.R-project.org/

72.

Reimer

Bard

Bayliss

. (2013) IntCal13 and Marine13 radiocarbon age calibration curves, 0–50,000 years cal BP. Radiocarbon 55: 1869–1887.

73.

Rydin

Jeglum

(2013) The Biology of Peatlands. 2nd Edition. Oxford: Oxford University Press.

74.

Shiller

Finkelstein

Cowling

(2014) Relative importance of climatic and autogenic controls on Holocene carbon accumulation in a temperate bog in southern Ontario, Canada. The Holocene 24(9): 1105–1116.

75.

Stineman

(1980) A consistently well-behaved method of interpolation. Creative Computing 6(7): 54–57.

76.

Stolze

Roe

Patterson

. (2007) A record of Lateglacial and Holocene vegetation and climate change from Woods Lake, Seymour Inlet, coastal British Columbia, Canada. Review of Palaeobotany and Palynology 147: 112–127.

77.

Sundberg

(2002) Sporophyte production and spore dispersal phenology in Sphagnum: The importance of summer moisture and patch characteristics. Canadian Journal of Botany 80: 543–556.

78.

Talbot

(2001) Nitrogen isotopes in paleolimnology. In: Last

Smol

(eds) Tracking Environmental Change Using Lake Sediments: Physical and Geochemical Methods, vol. 2. Dordrecht:Kluwer Academic Publishers, pp. 401–439.

79.

Telford

Heegaard

Birks

HJB

(2004) The intercept is a poor estimate of a calibrated radiocarbon age. The Holocene 14(2): 296–298.

80.

Troels-Smith

(1955) Karakterisering af løse jordarter [Characterization of unconsolidated sediments]. Danmarks Geologiske Undersøgelse, række IV bd. 3, nr. 10. Copenhagen: Reitzel Forlag.

81.

Tsukada

(1982) Pseudotsuga menziesii (Mirb.) Franco: Its pollen dispersal and late Quaternary history in the Pacific Northwest. Japanese Journal of Ecology 32: 159–187.

82.

Turner

(1999) Time to burn: Traditional use of fire to enhance resource production by aboriginal peoples in British Columbia. In: Boyd

(ed.) Indians, Fire and the Land in the Pacific Northwest. Corvallis, OR: Oregon State University Press, pp. 185–218.

83.

Turunen

(2003) Development history and carbon accumulation of a slope bog in oceanic British Columbia, Canada. The Holocene 13(2): 225–238.

84.

van Geel

(1978) A palaeoecological study of Holocene peat bog sections in Germany and The Netherlands, based on the analysis of pollen, spores, and macro- and microscopic remains of fungi, algae, cormophytes and animals. Review of Palaeobotany and Palynology 25: 1–120.

85.

Vitt

Halsey

Bauer

. (2000) Spatial and temporal trends in carbon storage of peatlands of continental western Canada through the Holocene. Canadian Journal of Earth Sciences 37: 683–693.

86.

Warner

Chinnappa

(1986) Taxonomic implications and evolutionary trends in pollen of Canadian Ericales. Canadian Journal of Botany 64: 3113–3126.

87.

Wieder

Vitt

(2006) Boreal Peatland Ecosystems: Ecological Studies, vol.188. Berlin, Heidelberg:Springer-Verlag.

88.

Yeloff

Charman

van Geel

. (2007) Reconstruction of hydrology, vegetation and past climate change in bogs using fungal microfossils. Review of Palaeobotany and Palynology 146: 102–145.

89.

Beilman

Jones

(2009) Sensitivity of northern peatland carbon dynamics to Holocene climate change. In: Baird

Belyea

Comas

. (eds) Carbon Cycling in Northern Peatlands: AGU Monograph 184. Washington, DC: American Geophysical Union, pp. 55–69.

90.

Vitt

Wieder

(2014) Continental fens in western Canada as effective carbon sinks during the Holocene. The Holocene 24(9): 1090–1104.

91.

Zhao

Tang

. (2014) Holocene peatland initiation, lateral expansion and carbon dynamics in the Zoige Basin of the eastern Tibetan Plateau. The Holocene 24(9): 1137–1145.