Portal de Periódicos da CAPES

Variation in CO 2 exchange over three summers at microform scale in a boreal bog, Eastmain region, Québec, Canada

2011; American Geophysical Union; Volume: 116; Issue: G3 Linguagem: Inglês

10.1029/2011jg001657

2156-2202

Luc Pelletier, Michelle Garneau, Tim R. Moore,

Botany and Plant Ecology Studies

Journal of Geophysical Research: BiogeosciencesVolume 116, Issue G3 Free Access Variation in CO2 exchange over three summers at microform scale in a boreal bog, Eastmain region, Québec, Canada L. Pelletier, L. Pelletier [email protected] Peatland Ecosystems Dynamics and Climatic Change, Center for Research in Isotopic Geochemistry and Geochronology and Department of Geography, Université du Québec à Montréal, Montréal, Québec, Canada Now at Department of Natural Resource Sciences, McGill University, Sainte-Anne-de-Bellevue, Québec, Canada.Search for more papers by this authorM. Garneau, M. Garneau Peatland Ecosystems Dynamics and Climatic Change, Center for Research in Isotopic Geochemistry and Geochronology and Department of Geography, Université du Québec à Montréal, Montréal, Québec, CanadaSearch for more papers by this authorT. R. Moore, T. R. Moore Department of Geography, McGill University, Montréal, Québec, CanadaSearch for more papers by this author L. Pelletier, L. Pelletier [email protected] Peatland Ecosystems Dynamics and Climatic Change, Center for Research in Isotopic Geochemistry and Geochronology and Department of Geography, Université du Québec à Montréal, Montréal, Québec, Canada Now at Department of Natural Resource Sciences, McGill University, Sainte-Anne-de-Bellevue, Québec, Canada.Search for more papers by this authorM. Garneau, M. Garneau Peatland Ecosystems Dynamics and Climatic Change, Center for Research in Isotopic Geochemistry and Geochronology and Department of Geography, Université du Québec à Montréal, Montréal, Québec, CanadaSearch for more papers by this authorT. R. Moore, T. R. Moore Department of Geography, McGill University, Montréal, Québec, CanadaSearch for more papers by this author First published: 18 August 2011 https://doi.org/10.1029/2011JG001657Citations: 23AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Abstract [1] We examine variability in carbon dioxide (CO2) exchange as a function of meteorological conditions using static chambers across a peatland microtopographic gradient (high hummock, low hummock, lawn and hollow microforms) between June and August 2006, 2007, and 2008. Total June to August precipitation increased during the study period with 285 mm in 2006, 320 mm in 2007, and 369 mm in 2008, whereas monthly average temperature was similar among years, with the exception of August 2008, which was 2.4°C warmer than previous years. Significantly different relationships between photosynthetic photon flux density and net ecosystem exchange (NEE) were observed in 2008 on three of the four microforms corresponding with the different meteorological conditions. The controls on CO2 exchange varied among microforms: a water table closer to the peat surface increased NEE on the high and low hummocks by increasing maximum rates of photosynthesis (PSNmax) but reduced NEE on the hollow microform by flooding the surface vegetation and reducing PSNmax. Water table position was a significant control on ecosystem respiration (ER) only on the lawn microform. We examine relationships between water table position and PSNmax and ER, and propose a theoretical relationship, which supports results from other studies examining water table position and peat accumulation. Key Points Important interannual variation in NEE related to water table position Hummocks max photosynthesis increased when water table was closer to the surface Water table position significantly controlled R on the lawn microform 1. Introduction [2] Boreal and subarctic peatlands cover only 3% of Earth's land surface and store the equivalent of 50% of carbon (C) found in the atmosphere as carbon dioxide (CO2) [Turunen et al., 2002; Intergovernmental Panel on Climate Change, 2007]. These ecosystems have accumulated C over the Holocene because of their positive balance between net primary production (NPP) and peat decomposition. Peatlands absorb CO2 through surface vegetation photosynthesis and release C through autotrophic respiration (CO2), peat decomposition (CO2 and methane [CH4]), and export of dissolved organic carbon (DOC). The high water table creates anoxic conditions and thermal properties, keeping peat cool favors biomass production over decomposition. Long-term average C accumulation rates range between 8 and 38 g m−2 yr−1 [Gorham, 1991, 1995; Turunen et al., 2002; Yu et al., 2010]. Although peatlands are a long-term sink for C, the interannual variability obtained from contemporary continuous measurements and integration of CO2 net ecosystem exchange (NEE), CH4 fluxes, and DOC export reveal that they can also be sources of C during certain years [Roulet et al., 2007]. [3] Bog peatlands are characterized by a succession of microtopographic forms from hollows, where the water table is near or above the surface, to hummocks where the water table is tens of centimeters below the surface, with lawns in between, creating variations in vascular plant and moss distribution, which affect photosynthetic capacity. Waddington and Roulet [1996] found important differences in CO2 flux at the microform scale in a Swedish peatland, with hummocks being a net sink for CO2, whereas hollows were a net source. The NEE of bogs is generally smaller than fens, the latter having greater aboveground biomass of Cyperaceae, which leads not only to more CO2 uptake but also higher respiration rates than bogs [Bellisario et al., 1998; Frolking et al., 1998]. [4] Interannual variability in growing season CO2 fluxes has been strongly linked to changes in precipitation patterns. Using eddy covariance CO2 exchange data from four fen and two bog sites, Sulman et al. [2010] reported that wetter summer conditions at the fens led to smaller gross ecosystem photosynthesis (GEP) and ecosystem respiration (ER), whereas at the bogs this led to larger GEP and ER; there was no distinct change in NEE with water table in either fen or bog sites. Bubier et al. [2003a, 2003b] measured chamber NEE during wet and dry summers in a bog and a mineral-poor fen and found an increase in respiration during the drier summer. This increase in respiration explained most of the differences in NEE between these 2 years in the bog and fen. In the fen, Bubier et al. [2003b] also measured an increase in overall uptake from evergreen and deciduous shrubs, and an earlier senescence of Cyperaceae during the dry year. The gross photosynthesis in the bog was not affected by the drought, but the net CO2 uptake was reduced by increased respiration [Bubier et al., 2003a]. In the same temperate bog over a 6 year study, Roulet et al. [2007] found that summers-autumns with low water tables resulted in less CO2 uptake that, combined with loss of C as CH4 and DOC, resulting in a net annual loss of 13.5–0.8 g C m−2 yr−1. [5] Field manipulations have been used to examine the effect of lowering and raising the water table on photosynthesis, respiration, and overall productivity. Riutta et al. [2007] and Strack and Waddington [2007] both looked at the effects of water table drawdown on CO2 dynamics in fens with different results. Riutta et al. [2007] measured an increase in ER and a decrease in GEP, resulting in a smaller NEE following water table drawdown, with Sphagnum mosses being the most affected by the lowered water table. Strack and Waddington [2007], in contrast, measured no significant differences in CO2 exchange in a poor fen 3 years after lowering the water table initially by 20 cm, owing to peat subsidence, although there were differences in the response of the microforms at the site. Hollow respiration and GEP both increased following water table drawdown, probably due to the observed increase in vascular plant cover [Strack and Waddington, 2007]. In an Alaskan-rich fen, Chivers et al. [2009] examined the effect of an experimentally raised water table on ecosystem CO2 fluxes without distinguishing among microforms. Their results show increased gross primary production (GPP) in the raised water table treatment, which was mainly driven by an early spring GPP attributed to mosses. In their bog mesocosm experiment, Bridgham et al. [2008] measured an increase in productivity and C accumulation following a rise in the water table, whereas the fen mesocosms showed a less pronounced increase. [6] In this study, we present the results of CO2 exchange measured between June and August 2006–2008 in a bog located in midboreal Québec, Canada. The research is part of the Eastmain-1 (EM-1) project (www.Eastmain1.org), which aims at determining the net greenhouse gas emissions from a hydroelectric reservoir by comparing prior and post impoundment terrestrial and aquatic fluxes. Measurements of CO2 exchange presented in this article represent part of the work undertaken to quantify C dynamics from peatlands in the region. We examine NEE variations at the microform scale, examining relationships between environmental variables and maximum rates of photosynthesis (PSNmax) and ER. We hypothesize that the NEE response of microforms to variations in precipitation will differ across the microtopographic gradient because of the different effect water table position has on photosynthesis and respiration at the microform level. Several studies have looked at NEE response to a decrease in water table position [e.g., Strack and Waddington, 2007; Strack and Price, 2009], but none has documented the effect of naturally raised water table position on the microform NEE response. 2. Methods Study Area, Peatland Description, and Climate [7] The study area is located about 180 km inland from the east coast of James Bay in the humid midboreal wetland region at the northern limit of the closed boreal forest [National Wetlands Working Group, 1988] (Figure 1). Peatlands cover approximately 7% of the region with bogs, representing the main peatland feature (98% bogs and 2% fens [Grenier et al., 2008]). The bedrock is mainly composed of gneiss and granite covered by lodgement/melt till, as well as fluvioglacial and aeolian deposits left after the retreat of the ice sheet water and wind activity [Eakins et al., 1968; D. Brosseau, unpublished data, 2008]. Figure 1Open in figure viewerPowerPoint Study area and location of Lac le Caron (LLC) peatland. [8] Lac le Caron (LLC) bog (52°17′N, 75°32′W; altitude: 248 m) covers ∼247 ha and is bordered on the west side by a 40 m escarpment and a small river flowing southward on its east side (Figure 1). This ombrotrophic peatland is characterized by a hummock-lawn-hollow-pool pattern. The average peat thickness is 2.63 m, with a maximum of 5.35 m [Dallaire and Garneau, 2008], and basal dates indicate that peat started to accumulate 7520 years BP [van Bellen et al., 2011]. [9] There are no long-term temperature and precipitation data available for the EM-1 region. The nearest stations with a 30 year average are located near Radisson (200 km north) and Chibougamau (280 km south). Monthly precipitation and temperature between June and August 2006–2008 obtained from an onsite meteorological tower are presented in Figure 2. Total June to August precipitation increased during the study period with 285 mm in 2006, 320 mm in 2007, and 369 mm in 2008. The average temperature for June and July were within 1.5°C for all 3 years, with the most important difference in monthly average temperature being in August 2008 with 15.3°C compared to 12.9°C and 12.8°C in August 2006 and 2007, respectively. Figure 2Open in figure viewerPowerPoint Monthly average precipitation and temperature for May−September 2006–2008 at Petit Opinaca meteorological station. Experimental Setup [10] In June 2006, eight collars (diameter 25 cm) were installed in the LLC peatland. Sets of two collars were installed on different microforms, following the moisture/vegetation gradient: high hummock, low hummock, lawn, hollow (Table 1). Wood planks were installed on the peat surface to minimize disturbance during measurements. PVC tubes were inserted in the peat to measure water table position next to each microform where gas flux was measured. CO2 fluxes were measured once a month over 1 (2006 and 2008) or 2 weeks (2007) between June and August. The remoteness of the site, accessible only by helicopter, explains the frequency of sampling and the number of replicate collars per microform. Table 1. Microforms, Vegetation Description, Aboveground and Green Biomass, and Water Table Position, 2006–2008a Microform Vascular Plants Bryophytes Total Aboveground Biomass (g m−2) Green Biomass (g m−2) Water Table (cm) HH Kalmia angustifolia, Chamaedaphne calyculata, Kalmia polifolia, Andromeda glaucophylla, Rubus chamaemorus, Vaccinium oxycoccus Sphagnum fuscum, Sphagnum angustifolium, Mylia anomala 431 225 2006 −33.4 2007 −39.1 2008 −26.1 LH K. angustifolia, C. calyculata, K. polifolia, A. glaucophylla, R. chamaemorus, Carex spp., Eriophorum spissum, V. oxycoccus S. fuscum, S. angustifolium 182 127 2006 −16.3 2007 −22.6 2008 −9.1 Ln C. calyculata, K. polifolia, V. oxycoccus, Carex oligosperma, Carex spp., Eriophorum spp., Drosera rotundifolia Sphagnum fallax, M. anomala 100 90 2006 −0.1 2007 −7.8 2008 2.3 Hw Carex spp., Eriophorum spp. Sphagnum cuspidatum, Cladopodiella fluitans 36 33 2006 0.2 2007 −1.4 2008 2.3 a HH, high hummock; LH, low hummock; Ln, lawn; Hw, hollow. CO2 Flux Measurements [11] Net ecosystem exchange CO2 measurements were made using a closed system comprising three components: (1) a PP Systems EGM-4 (Amesbury, Massachusetts, USA) infrared gas analyzer (IRGA), (2) a static chamber, and (3) a cooling system to keep the temperature inside the chamber close to ambient. The Plexiglas chamber material transmits approximately 96% of the incoming light. The cooling system consisted of a heat exchanger in the chamber and a water pump submerged in a bucket filled with cold water. For each sampling, the chamber was installed on the collar for 3 min, with CO2 concentration recorded every 10 s during the first minute and every 30 s for the last 1.5 min. For each sampling of a collar, four measurements were made under different light conditions, simulated by covering the chamber with shrouds. The chamber had no shroud for the first measurement and the second, third, and fourth included shrouds allowing 54%, 27%, and 0%, respectively, of the light to enter. Between each run, the chamber was ventilated to reach ambient CO2 concentration. The fluxes were calculated by linear regression using the CO2 concentration change during the 2.5 min period. The IRGA was calibrated prior to each field campaign using a CO2 standard of 400 ppm by volume (ppmv). Photosynthetic photon flux density (PPFD) in μmol m−2 s−1 was measured with a PAR-1 light sensor (PP Systems, Amesbury, Massachusetts). Water table position and peat temperature were recorded at each microform at time of measurement. Peat temperature was measured at 5, 10, 20, 30, and 40 cm depths using a Hannah HI935005 K-thermocouple thermometer probe. Data Analysis [12] The relationship between NEE and PPFD was examined using a rectangular hyperbola curve (equation (1)). The sign convention is negative for CO2 uptake by the peatland and positive for CO2 release to the atmosphere: where α is the initial slope of curve, Pmax is the maximum gross productivity, and ER is the dark respiration value. [13] Since NEE = GPP + ER, water table position and peat temperature controls on overall NEE were assessed by looking at GPP and ER separately. Because Pmax assumes infinite PPFD, maximum rates of photosynthesis (PSNmax) were calculated from all GPP with PPFD > 1000 μmol m−2 s−1 [Bubier et al., 2003a]. Statistical differences in the rectangular hyperbola parameters (α, Pmax, and NEEcap) between years were determined from confidence intervals, whereas one-way analyses of variance (ANOVAs) were performed on PSNmax and ER to assess differences between years. Vegetation [14] At the end of the 2008 growing season, aboveground vegetation in 50 × 50 cm plots representative of each microform was collected and stored in plastic bags. For the Sphagnum spp., the capitula mass was estimated from a 10 × 10 cm area within the plot. Vegetation was identified and divided into leaves and stems, dried in an oven at 70°C for 48 h and weighed (Table 1). 3. Results Spatial Variability [15] The 1084 NEE measurements made between June and August 2006–2008 revealed variable NEE-PPFD relationships among the microforms (Figure 3 and Table 2). The largest NEEcap (uptake), NEE calculated for PPFD = 1800 μmol m−2 s−1, and ER (emission) rates were measured on the high hummock, whereas the hollow had the smallest NEE CO2 (uptake) and ER rates (Figure 3). The low hummock and lawn microforms had a similar range for Pmax (−16 to −31 g CO2 m−2 d−1 for low hummock and −16 to −24 g CO2 m−2 d−1 for lawn), but ER was generally larger on the low hummock microform (Table 2). Our results support the findings of other studies [Frolking et al., 1998] that CO2 uptake saturates at a PPFD of ∼1000 μmol m−2 s−1. However, PPFD appeared to have a minor influence on NEE at the hollow microform and the rectangular hyperbola relationship fitted poorly in 2006 and 2007 (r2 < 0.55) and could not be fitted in 2008. Figure 3Open in figure viewerPowerPoint Relationship between net ecosystem exchange (NEE) and photosynthetic photon flux density (PPFD) at four microforms, for 2006, 2007, and 2008, fitted with a rectangular hyperbola equation. Table 2. Rectangular Hyperbola Curve Fit Parameters for Each Microform in Lac le Caron (LLC) Peatland, June–August 2006–2008a Microform Year n α Pmax NEEcap r2 PSNmax ER High hummock 2006 60 −0.09 (0.01) (a)b 38.5 (2.16) (a) −20.2 (2.34) (a) 0.96 −18.8 (0.82) (a) 10.8 (1.03) (a) 2007 153 −0.09 (0.01) (a) −30.1 (1.57) (b) −17.4 (1.70) (b) 0.91 −17.2 (0.75) (a) 8.55 (0.66) (a) 2008 47 −0.09 (0.02) (a) −49.9 (3.56) (c) −30.7 (3.70) (c) 0.95 −29.3 (1.52) (b) 8.13 (0.97) (a) Low hummock 2006 51 −0.09 (0.03) (a) −16.0 (1.73) (a) −6.43 (1.91) (a) 0.87 −6.49 (0.96) (a) 8.07 (1.06) (a) 2007 137 −0.06 (0.01) (a) −19.2 (1.10) (b) −9.41 (1.22) (b) 0.88 −9.13 (0.66) (a) 6.77 (0.48) (a) 2008 36 −0.07 (0.01) (a) −31.6 (2.55) (c) −19.9 (2.54) (c) 0.96 −18.8 (1.73) (b) 5.61 (0.45) (a) Lawn 2006 52 −0.07 (0.02) (a) −16.4 (1.30) (a) −11.0 (1.42) (a) 0.91 −11.3 (0.80) (a) 3.75 (0.78) (b) 2007 142 −0.07 (0.01) (a) −24.3 (1.11) (b) −14.7 (1.18) (b) 0.94 −13.9 (0.51) (b) 6.09 (0.44) (a) 2008 48 −0.04 (0.01) (b) −21.8 (1.57) (c) −13.8 (1.58) (c) 0.96 −13.2 (0.60) (a,b) 3.13 (0.50) (b) Hollow 2006 48 −0.01 (0.01) (a) −1.26 (0.51) (a) −0.16 (0.42) (a) 0.40 −0.58 (0.19) (a,b) 0.99 (0.22) (b) 2007 142 −0.01 (0.01) (a) −2.34 (0.37) (b) −0.99 (0.45) (b) 0.55 −1.02 (0.22) (a) 0.94 (0.13) (b) 2008 40 – – – – 0.00 (0.00) (b) 1.98 (0.39) (a) a Note that the rectangular hyperbola curve did not fit the hollow microform in 2008. Standard error is given in parentheses. NEEcap is NEE calculated for PPFD = 1800 μmol m−2 s−1. See section 2.4 for symbol and abbreviation descriptions. Values of α are in g CO2 m−2 d−1/μmol m−2 s−1. Pmax, PSNmax, NEEcap, and ER values are in g CO2 m−2 d−1. b Parameters within microforms are significantly different if they have no letters in common. Statistical differences in α, Pmax, and NEEcap between years were determined from confidence intervals, whereas one-way ANOVA were performed on PSNmax and R to assess differences between years. Interannual Variability [16] The NEE-PPFD relationships varied significantly among the 2006–2008 growing seasons and the magnitude of the variations differed between microforms (Figure 3 and Table 2). The largest year-to-year variations were measured on the high and low hummocks with NEEcap ranging from −17.4 to −30.7 and −6.43 to −19.9 g CO2 m−2 d−1, respectively. The high and low hummock microforms had a larger NEE in 2008 than in 2006 and 2007 as soon as PPFD > 500 μmol m−2 s−1. The high hummock microform greater NEE at PPFD > 500 μmol m−2 s−1 in 2008 is explained by a significant increase in PSNmax over 2006 and 2007 (p < 0.05), combined with no significant change in ER (p > 0.05) (Table 2). On the low hummock, a combined increase in PSNmax and a decrease in ER during 2008 compared to 2006 and 2007 explained the greater NEE for PPFD > 500 μmol m−2 s−1 (Table 2). The lawn and hollow microform ranges of NEEcap were much smaller than high and low hummock, being −11.0 to −14.7 and −0.16 to −0.99 g m−2 d−1, respectively. On the lawn microform, the effect of interannual variations in PSNmax on NEE was attenuated by variations in ER because years with higher PSNmax corresponded with years with higher ER. As mentioned previously, PPFD appeared to have a minor influence on NEE at the hollow microform and the rectangular hyperbola relationship could not be fitted to the 2008 hollow data set. Environmental Correlates [17] We used individual CO2 fluxes and seasonal averages to establish the relationship with environmental variables, such as water table position, peat/air temperature, and vegetation biomass. Across the microforms, for both the individual and seasonal averages, PSNmax (GPP for PPFD > 1000 μmol m−2 s−1) was significantly (p < 0.05) related to water table (r2 = 0.28 and 0.36, respectively), with a higher water table correlated with a smaller PSNmax (Figures 4a and 4b). However, within microforms, individual PSNmax increased with a higher water table in high hummock and decreased in hollow (r2 = 0.38 and 0.56, respectively, p < 0.001; Figure 4a). The relationships between individual PSNmax and water table for the low hummock and lawn microforms were weak and not significant (r2 < 0.06, p > 0.05), although the slope for low hummock was significantly different from zero. The ∼13 cm rise in water table in 2008 over 2006 and 2007 created an increase of approximately 10 g CO2 m−2 d−1 in average PSNmax in both the high and low hummock microforms. On the hollow, the rise in water table in 2008 was enough to suppress PSNmax, so that GPP = 0 even when PPFD was greater than 1000 μmol m−2 s−1. Figure 4Open in figure viewerPowerPoint Relationship between PSNmax and water table position (WTP) for (a) individual measurement and (b) June-August averages. HH, high hummock; LH, low hummock; Ln, lawn; Hw, hollow. [18] Across the microforms, for both the individual and seasonal averages, ER was significantly (p < 0.001) related to water table (r2 = 0.42 and 0.83, respectively), with a higher water table correlated with a smaller ER (Figures 5a and 5b). Within microforms, individual measurements of ER were significantly correlated with water table at the lawn (r2 = 0.65, p < 0.001), ER decreasing with a higher water table (Figure 5a). However, the NEEcap did not vary synchronously with this control, and the rise in water table in 2008 did not result in a significant variation in NEEcap. Since ER rates are one order of magnitude smaller than PSNmax rates on the lawn microform, the variation in ER linked to water table position only slightly affected NEEcap and therefore was not a dominant control on NEE on this microform. The relationships between ER and water table position were not significant at the high and low hummock microforms, and a positive but weak significant (p = 0.03) relationship was found between water table and ER at the hollow microform. The high and low hummock microform average seasonal ER was not statistically different between years, whereas 2008 ER fluxes were significantly different from 2007 but not 2006 in the lawn microform. On the hollow, 2008 ER was significantly different from both 2006 and 2007. Figure 5Open in figure viewerPowerPoint Relationship between ER and water table position (WTP) for (a) individual measurement and (b) June-August averages. HH, high hummock; LH, low hummock; Ln, lawn; Hw, hollow. [19] Air and peat temperature were weakly related to PSNmax. Air temperature was correlated with PSNmax only on the high hummock microform (r2 = 0.18, p < 0.001; Table 3). The relationships between air/peat temperature and ER were generally stronger than for PSNmax and varied across the microtopographic gradient (Table 3). Individual ER fluxes on high and low hummock microforms were more strongly correlated with air/peat temperature at 5 cm below surface (r2 = 0.35 and 0.40, respectively; Table 3), whereas it was more strongly correlated with air/peat temperature at 40 cm below the peat surface on the lawn microform (r2 = 0.41 and 0.43; Table 3). The relationships between ecosystem individual R fluxes and air/peat temperature on the hollow microform were weaker than for the other microforms (r2 < 0.25). Table 3. Coefficient of Determination (r2) and Probability (p) Values for Relationships Between Individual PSNmax Values and Air Temperature as Well as R and Air and Peat Temperatures at 5, 10, 20, and 40 cm Depthsa Microform HH LH Ln Hw r2 p r2 p r2 p r2 p PSNmax, Air 0.18 0.000 0.01 0.474 0.00 0.625 0.02 0.317 R, Air 0.35 0.000 0.44 0.000 0.41 0.000 0.00 0.977 R, 5 cm 0.40 0.000 0.31 0.000 0.00 0.707 0.11 0.014 R, 10 cm 0.21 0.000 0.25 0.000 0.09 0.021 0.18 0.001 R, 20 cm 0.12 0.006 0.28 0.000 0.30 0.000 0.24 0.000 R, 40 cm 0.01 0.416 0.13 0.005 0.43 0.000 0.23 0.000 a Just define n and data set for these: HH, high hummock; LH, low hummock; Ln, lawn; Hw, hollow. [20] Strong stepwise regressions were found for ER fluxes and PSNmax across the microtopographic gradient. Green biomass was significantly positively related with yearly summer averages of ER and PSNmax with r2 = 0.84 (p < 0.001, n = 12) and r2 = 0.71 (p < 0.001, n = 12), respectively (Table 4). The stepwise regression was stronger for PSNmax when water table position was added to green biomass (r2 = 0.92, p < 0.001), and when water table position and peat temperature at 5 cm below the surface were added to green biomass (r2 = 0.96, p < 0.001) (Table 4). Table 4. Stepwise Regression Coefficient of Determination (r2) and Probability (p) Values for Relationships Between June and August Average PSNmax and R for Each Microform and Summer Average Water Table Position, Temperature at 5 cm Depth, and Aboveground Green Biomassa Regression Equations r2 p n R = 0.57 + 0.04GB 0.84 <0.001 12 PSNmax = 0.50 − 0.10GB 0.71 <0.001 12 PSNmax = 6.69 − 0.23GB − 0.69WTP 0.92 500 μmol m−2 s−1 on the high and low hummock microforms in 2008 were the result of an increase in PSNmax associated with a higher water table position (Figures 4a and 4b). Bridgham et al. [2008] observed that Sphagnum fuscum hummocks rapidly increased their productivity after a rise in water table position, and Strack and Price [2009] also showed that wetter Sphagnum, as a result of greater precipitation, could lead to higher C fixation, independently of water table position. Based on our measurements, the increase in PSNmax on the high and low hummock microform cannot be attributed to either vascular or bryophyte plants because no direct productivity measurements based on biomass have been made in our plots. Other studies have documented an increase in bryophyte productivity following an increase in soil moisture or a rise in water table depth [Szumigalski and Bayley, 1996; Bridgham et al., 2008], and Weltzin et al. [2000] have shown a negative relationship between shrub productivity and rise in water table. Therefore, it is reasonable to associate the increase in PSNmax on the high and low hummock microform in 2008 to increased bryophyte productivity. [23] Contrary to the high and low hummock microforms, the 2008 rise in water table had no effect on PSNmax in the lawn microform and had a negative effect in the hollow microform (Figure 4a). The latter may be explained by a higher resistance to CO2 diffusion due to higher Sphagnum water content in accord with Rice and Giles [1996], who measured, in a laboratory exper