63 Abundance and composition of biomass as

potential controls for mire net ecosytem CO2 exchange

Anna M. Laine, Jill Bubier, Terhi Riutta, Mats B. Nilsson, Tim R. Moore, Harri Vasander, and Eeva-Stiina Tuittila

Abstract: We compared the amount and composition of different aboveground biomass (BM) fractions of four mires with their net ecosystem CO2 exchange (NEE) measured by eddy covariance. We found clear differences in response of green bi- omass (GBM) of plant functional types (PFTs) to water table (WT), which resulted in larger spatial variation in GBM within a mire than variation between mires. GBM varied between mires from 126 ± 7 to 336 ± 16 g·m–2 (mean ± SE), while within mire variation at largest was from 157 ± 17 to 488 ± 20 g·m–2 (mean ± SE). GBM of dominant PFTs appeared to be better in explaining the peak growing season NEE than the total BM or GBM of a mire. The differences in photosyn- thetic capacity between PTFs had a major role, and thus a smaller GBM with different species composition could result in higher NEE than larger GBM. GBM, especially that of sedges, appeared to have a high impact on NEE. Eleven PFTs, defined here, appeared to capture well the internal variation within mires, and the differences in GBM between communities were explained by the water table response of PFTs. Our results suggest the use of photosynthesizing BM, sep- arated into PFTs, in modelling ecosystem carbon exchange instead of using just total BM. Key words: , , microtopography, net ecosystem exchange, peatland, plant functional type, water table. Résumé : L’auteure a comparé la quantité et la composition des différentes fractions de la biomasse (BM) épigée dans qua- tre tourbières, incluant la mesure de l’échange net de CO2 par l’écosystème (NEE), à l’aide de la covariance de fluctuation. Elle a observé des différences nettes de réaction de la biomasse verte (BMV) des types fonctionnels de plantes (TFP) à la nappe phréatique, lesquelles conduisent à une variation plus importante de la BMV à l’intérieur d’une tourbière qu’entre les tourbières. La BMV varie entre les tourbières de 126 ± 7 à 336 ± 16 g·m–2 (moyenne ± SE), alors qu’àl’intérieur des tour- bières la plus grande va de 157 ± 17 à 488 ± 20 g·m–2 (moyenne ± SE). La BMV du TPF dominant semble la meilleure pour expliquer le pic de le NEE de la saison de croissance, que la BM ou la BMV d’une tourbière. Les différences de capa-

For personal use only. cité photosynthétique entre les TFP jouent un rôle majeur, et ainsi une plus petite BMV avec différentes espèces peut conduire à des résultats de NEE plus élevés qu’une BMV plus grande. La BMV des plantes vasculaires, surtout pour bien capter la variation interne dans une tourbière, et les différences de BMV entre les communautés s’expliquent par la réaction de la nappe phréatique des TPF. Les résultats suggèrent d’utiliser la BM photosynthétique, distribuée selon les TPF, pour la modélisation de l’échange de carbone par les écosystèmes, plutôt que la seule BM. Mots‐clés : tourbière ombrotrophe, tourbière minérotrophe, microtopographie, échange écosystémique net, tourbière, type fonctionnel de plantes, nappe phréatique. [Traduit par la Rédaction]

Introduction (CO2) in photosynthesis and release CO2 and methane (CH4) Mires have a dual effect on atmospheric greenhouse gas in respiration. Mires have stored a significant amount of car- concentrations. They assimilate atmospheric carbon dioxide bon into over the last millennia (Gorham 1991; Turunen

Received 9 February 2011. Accepted 6 September 2011. Published at www.nrcresearchpress.com/cjb on 16 December 2011. A.M. Laine. Department of Biology, University of Oulu, P.O. Box 3000, FI-90014 University of Oulu, Finland; Peatland Ecology Group,

Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 Department of Forest Science, University of Helsinki, P.O. Box 27, FI 00014 University of Helsinki, Finland. J. Bubier. Environmental Studies Program, Mount Holyoke College, Clapp Laboratory 50 College Street, South Hadley, MA 01075, USA. T. Riutta. Peatland Ecology Group, Department of Forest Science, University of Helsinki, P.O. Box 27, FI 00014 University of Helsinki, Finland; University of Oxford, School of Geography and the Environment, Environmental Change Institute, South Park Road, OX1 3QY, UK. M.B. Nilsson. Department of Forest Ecology and Management, SLU, SE-901 83 Umeå, Sweden. T.R. Moore. Department of Geography, McGill University, 805 Sherbrooke Street West, Montreal, QC H3A 2K6, Canada. H. Vasander. Peatland Ecology Group, Department of Forest Science, University of Helsinki, P.O. Box 27, FI 00014 University of Helsinki, Finland. E.-S. Tuittila. Peatland Ecology Group, Department of Forest Science, University of Helsinki, P.O. Box 27, FI 00014 University of Helsinki, Finland; School of Forest Sciences, University of Eastern Finland, PL 111, 80101 Joensuu, Finland. Corresponding author: Anna M. Laine (email: [email protected]).

Botany 90:63–74 (2012) doi:10.1139/B11-068 Published by NRC Research Press 64 Botany, Vol. 90, 2012

et al. 2002) and during the Holocene they have had a net et al. 2003; Leppälä et al. 2008; Riutta et al. 2007b), quality cooling effect on the atmosphere radiation balance, as the ef- of litter (Moore et al. 2007; Szumigalski and Bayley 1996), fect of long-term carbon accumulation overrules the transient and roles in CH4 dynamics (Joabsson et al. 1999; Shannon warming caused by the CH4 emissions (Frolking et al. 2006). et al. 1996). In addition, their responses to changing environ- Carbon gas exchange has been studied in several mire eco- mental conditions may vary (Walker et al. 2006; Weltzin et systems during the past decades (e.g., Alm et al. 1997; Lund al. 2001). Owing to their close linkage to mire ecosystem et al. 2010; compilation by Saarnio et al. 2007), and since function, PFTs offer a useful tool for describing vegetation the 1990s the increased use of eddy covariance (EC) method abundance and structure as a part of ecosystem models. Re- has enabled long-term landscape level measurements of CO2 cently Frolking et al. (2010) presented a peatland develop- and to a lesser extent CH4 exchange (Baldocchi 2003). To- ment model (Holocene peat(land) model, HPM) where 12 day, long-term data on mire–atmosphere CO2 exchange exist PFTs are responsible for peat growth over the Holocene time from a very limited number of sites, the most extensive ones scale. HPM is a first modelling attempt to quantitatively link being Kaamasensuo, northern Finland, since 1997 (Aurela et carbon accumulation process to vegetation composition and al. 1998, 2004), and Mer Bleue, southern Canada, since 1998 dynamics. A close link between vegetation composition and (Lafleur et al. 2001; Roulet et al. 2007; Teklemariam et al. CH4 emission exists, related to the close relationship between 2010). To estimate the regional and global carbon gas ex- plant distribution and WT and to the abundance of aeren- change between mires and the atmosphere, the long-term chymatous species (Bubier 1995; Dias et al. 2010; Granberg measurement data is used for parameterization of models et al. 1997; Koelbener et al. 2010; Nilsson et al. 2001). How- and photosynthesizing biomass (BM) is a prerequisite for ever, such a relationship between net ecosystem CO2 ex- CO2 assimilation (Frolking et al. 2002; Schubert et al. 2010). change (NEE)) and PFTs BM has not been established. Within mires, plant communities and corresponding BM In this paper, we compare the amount and composition of may vary to a large extent (Kosykh et al. 2008; Reinikainen BM of four mires, at which NEE has been measured by EC. et al. 1984; Vasander 1982, 1987). The dominant controls on We hypothesize that (i) there is larger spatial variation in the species composition and BM are water table (WT) position BM within a mire than variation between mires that results and nutrient status. Nutrient-poor that are isolated from from a strong relationship between vegetation and WT, and groundwater influence are characterized by dwarf and (ii) NEE is related to the BM of the dominant PFTs. mosses, while sedges and herbs are more abun- dant in that receive groundwater from the catchment. Material and methods Most mires in the northern hemisphere have a distinct sur- face pattern of microforms. While the variation in vegetation Study sites between mires is largely caused by nutrient status, the plant The study was carried out in four northern mires located in communities in microforms vary due to the WT (Bragazza Finland, Sweden, and Canada. The sites are part of the Carbo and Gerdol 2002). Spatial diversity varies between mires; Europe and Canadian Carbon Program networks, which coor-

For personal use only. some fens have a rather uniform surface, while most bogs dinate and facilitate regional and global analysis of observa- and the northerly fens have a conspicuous pattern of high tions from micrometeorological tower sites. Climatic ridges/hummocks and open water pools/flarks (Ruuhijärvi information for the sites is given in Table 1. 1983). The optimal WT conditions vary between different Kaamasensuo (Aurela et al. 1998, 2001, 2002, 2004) is a plant species, with shrubs generally occurring in drier habi- mesotrophic flark fen located in Kaamanen, Finland, in the tats than most sedge species (Lumiala 1944). Vascular northern boreal zone (Ahti et al. 1968; Kalela 1961) with a tend to have a wider tolerance for the WT than Sphagnum subarctic climate. Typical of northern fens, also called aapa species (Lumiala 1944; Väliranta et al. 2007). Different mires, the surface forms a distinctive pattern of strings and Sphagnum species replace each other along the hummock– flarks. On the flarks, sedges such as and Trichopho- hollow gradient and therefore, as a genus, are able to inhabit rum species and brown mosses mostly belonging to Drepa- the wide range of WT conditions met in the mires (Andrus et nocladus species sensu lato dominate, while the plant al. 1983; Laine et al. 2009; Rydin and McDonald 1985). Ow- communities on the strings are composed of dwarf shrubs ing to differences in WT and plant communities, all compo- such as Ledum palustre L. and Calluna vulgaris (L.) Hull, nents of carbon balance vary substantially between and feather mosses such as Pleurozium schreberi (Brid.) microforms and plant communities, and the C balance of mi- Mitt. For more details on the site see Maanavilja et al. croforms within a mire may vary from net uptake to net loss (2010). (Alm et al. 1999; Laine et al. 2007; Waddington and Roulet Siikaneva (Aurela et al. 2007) is an oligotrophic fen lo-

Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 2000; Sullivan et al. 2008). cated in Ruovesi, Finland, close to the border of the southern Because plant species composition varies largely between and middle boreal zones (Ahti et al. 1968; Kalela 1961). The mires, BM at the species level may be too detailed for mod- Siikaneva mire consists of relatively homogenous lawn vege- ellers when examining feedbacks between vegetation, WT, tation dominated by sedges (Carex spp., spp.), nutrients, and ecosystem carbon dynamics. Plant functional Scheuchzeria palustris L., and Sphagnum mosses. Hollows types (PFTs) that combine species with similar responses to are covered by Sphagnum species adapted to the wettest con- environmental factors are more useful for predicting vegeta- ditions and somewhat higher hummocks by dwarf veg- tion responses to the environment (Chapin et al. 1996). Typi- etation occur among the lawn level vegetation. For more cal mire PFTs such as shrubs, sedges, forbs, and mosses, details on the site see Riutta et al. (2007a). show different seasonal growth dynamics (Leppälä et al. Degerö Stormyr (Nilsson et al. 2008; Sagerfors et al. 2008) 2008), photosynthetic capacities and respiration rates (Bubier is an oligotrophic fen located in northern Sweden in the mid-

Published by NRC Research Press Laine et al. 65

Table 1. Study site descriptions.

Kaamasensuo Siikaneva Mer Bleue Degerö Stormyr Location 69°08′N, 27°17′E 61°49′N, 24°11′E 45°41′N, 75°48′W 64°11′N, 19°33′E Precipitation (mm) 441 713 943 567 Average temperature in July (°C) 13.1 15.5 20.9 14.8 Average temperature in January (°C) –13.9 –7.4 –10.8 –11.2 Average annual temperature (°C) –1.1 3.0 6.0 1.2 Note: The climate data is 30 year averages for the time period 1971–2000. Reference for Finnish sites Kaamasensuo and Siikaneva: Dreps et al. (2002); for the Canadian site Mer Bleue: Canadian Climate Normals 1971–2000 (Canada’s National Climate Archive (http://climate.weatheroffice.ec.gc. ca/climate_normals)); and for the Swedish site Degerö Stormyr: The Swedish Meteorological and Hydrological Institute.

dle boreal zone. The climate of the site is defined as cold, rostrata Stokes dominated lawns (CR, n = 8); and Carex la- temperate, and humid. The part of the mire complex used siocarpa Ehrh. dominated lawns (CL, n = 9), hollows (HO, for the NEE measurements is dominated by lawn and hollow n = 9), low hummocks (HU, n = 9), minerotrophic hum- plant communities. The vascular plant community is domi- mocks (MI, n = 9), and ombrotrophic hummocks (OM, n = nated by Eriophorum vaginatum L., Trichophorum cespito- 5). At Mer Bleue the samples were collected in early Septem- sum (L.) Hartm., oxycoccos L., Andromeda ber 2003 from the EC tower footprint (n = 31) and from two polifolia L., and L., with both Carex li- additional locations, i.e., Poor Fen (n = 32) and Blue Dome, mosa L. and S. palustris occurring more sparsely. The bot- which is a slightly wetter bog area (n = 28) (Bubier et al. tom layer of the hollows (approximate average WT level 2006). All samples at Degerö Stormyr were located at lawn during the growing season, 0–10 cm below the mire surface) level vegetation from the EC tower footprint and the sam- is dominated by Sphagnum majus (Russ.) C. Jens. and of the pling was carried out in August 2008. lawns (approximate average WT level during the growing Vascular plant BM was first separated by species. Woody season, 10–25 cm below the mire surface) is dominated by species were further separated into leaf and stem BM. Sam- Sphagnum balticum (Russ.) C. Jens. and Sphagnum lindber- ples were oven dried (at 60 °C) and weighed. In most analy- gii Schimp. ex Lindb. For more details on the site see Nils- ses we use green biomass (GBM) that combines the leaf BM son et al. (2008) and Eppinga et al. (2010). of woody vascular plants, all aboveground BM of herbaceous Mer Bleue (Lafleur et al. 2001; Roulet et al. 2007) is an vascular plants, and the upper part of mosses. To define BM ombrotrophic bog located near Ottawa, Canada, in the south- variation using more general terms than species, the plant ern boreal zone. The mire area can be divided into clearly species were divided into 11 PFTs (Frolking et al. 2010), ombrotrophic and poor fen areas. The EC tower is located in namely minerotrophic sedges, ombrotrophic sedges, minero- the bog area that has a relatively homogenous hummocky trophic forbs, ombrotrophic forbs, deciduous shrubs, ever-

For personal use only. surface dominated by a dense ericaceous dwarf shrub cover green shrubs, hollow sphagna, lawn sphagna, hummock ( calyculata (L.) Moench., Ledum groenlandi- sphagna, brown mosses, and feather mosses (including li- cum Oeder, Kalmia angustifolia L., and Vaccinium myrtil- chens). Plant species were classified into PFTs based on the loides Michx.) accompanied by sparse clusters of literature according to their ecological niche optima along E. vaginatum. The bottom layer is a dense moss carpet fertility and WT gradients. In the logic of PFTs classification, formed by Sphagnum capillifolium (Ehrh.) Hedw. and Sphag- minerotrophic plant types are those that need minerotrophic num magellanicum Brid., interspersed with Polytrichum stric- conditions to form a substantial BM. Ombrotrophic plant tum Menzies ex Bridel. For more details on the site see types, on the other hand, are capable of growing also in min- Bubier et al. (2006). erotrophic conditions, but in more nutrient-rich conditions such as in mesotrophic fens they are outcompeted by minero- Vegetation sampling trophic species. In nutrient-poor minerotrophic conditions, The aboveground BM of vascular plants was collected namely in oligotrophic fens, the competitional superiority of from circular 15 cm radius plots, except in Mer Bleue, where minerotrophic species is weak and ombrotrophic plant types the plots were 50 cm × 50 cm square. Bryophytes and li- can form large parts of the communities. The division of chens were inventoried from a one-quarter section of the total moss PFTs is based on their ecological optimum; species plot. In Degerö Stormyr, all the vegetation was inventoried have their highest abundance in those conditions where the from a one-quarter section of the circular 15 cm plot. Only environment is optimal or close to optimal and where they

Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 the capitula were collected for Sphagnum mosses for consis- are able to successfully compete with other species. Outside tency in comparing BM data across sites. For other mosses of the optimum, species abundance decreases as conditions the uppermost 2 cm was harvested. The number of sample become suboptimal and at the end of their tolerance range plots was 36, 59, 93, and 32 in Kaamasensuo, Siikaneva, they disappear (Austin 2005). In addition to nutrient require- Mer Bleue, and Degerö Stormyr, respectively. ments, Sphagnum species have an optimum along the micro- At Kaamasensuo the sampling was done in August 2006 topographic gradient (Väliranta et al. 2007). Hollow sphagna and sample plots were divided into two vegetation types, have the narrowest tolerance and can generally grow only in flarks (n = 18) and strings (n = 18). At Siikaneva the sam- wet conditions, while the tolerance of lawn sphagna is the pling was also conducted in August 2006 and sample plots largest and they can be found along the whole gradient. were stratified to inventory seven different plant commun- Hummock sphagna are usually outcompeted from wet hol- ities: E. vaginatum dominated lawns (EV, n = 10); Carex lows. Brown mosses and feather mosses are divided by the

Published by NRC Research Press 66 Botany, Vol. 90, 2012

, convergent effect of fertility and moisture to rich–wet and , poor–dry PFTs, respectively. For the classification of vascu- lar plant species into PFTs we used Malmer (1962), Eurola et al. (1995), and Gignac et al. (2004), and for the mosses Carex

, we used Lumiala (1944), Andrus et al. (1983), Glaser et al. Sphagnum , (1990), and Laine et al. (2009) and references therein. Spe- cies included in each PFTs are given in Table 2.

Ledum groenlandicum To estimate the average BM for the footprint of EC tower , Cladonia rangiferina

, we used area-weighted BM for different plant communities in Kaamasensuo and Siikaneva. In Kaamasensuo, on average,

Carex rotundifolia flarks covered 70% and strings covered 30% of the mire area ,

Sphagnum russowii (Maanavilja et al. 2010); the area proportions for Siikaneva , are given in Riutta et al. (2007a). In Mer Bleue and Degerö

, Stormyr, the sampling next to EC tower represented variation

Cladonia cornuta in the footprint and we therefore used the arithmetic average , . Minerotrophic plant types here are those that need

Carex rostrata of the sample plots to estimate the average BM for the foot- ,

lerance range, they disappear. Brown mosses and feather print. sion of PFTs of mosses is based on their ecological In Siikaneva and Kaamasensuo WT was monitored during

Warnstrorfia exannulata the growing season from permanent sample plots represent- ,

Sphagnum papillosum ing the plant communities from which BM was collected. , Carex livida Kalmia angustifolia , , An average WT was calculated for each plant community

Polytrichum strictum based on these measurements. In Mer Bleue and Degerö Vaccinium vitis-idaea , , Stormyr WT was measured adjacent to each sample plot over the growing season when BM was collected. We express WT as relative to mire surface, i.e., negative when it is below Carex limosa

, the soil surface. Empetrum nigrum Scorpidium scorpioides , , Maianthemum trifolium

, Data analysis Sphagnum magellanicum

, To quantitatively compare the plant compositional varia- Utricularia intermedia Pleurozium schreberii

, tion in the tower footprints of each mire we applied multi- Vaccinium uliginosum Vaccinium microcarpum , variate methods. We performed detrended correspondence , spp., Carex lasiocarpa

, analysis (DCA) using Hill's scaling and detrending by seg-

spp. ments (ter Braak and Smilauer 2002a) separately for each For personal use only. dry PFTs, respectively. mire. The resulting length of the first axis is a measure of – Sphagnum fallax Limprichtia revolvens , , the largest compositional variation in the data sets (ter Braak Sphagnum rubellum Scheuchzeria palustris , Tricophorum cespitosum , Melampyrum Chamaedaphne calyculata

Cladonia and Prentice 1988). For we used only data , , , , collected from the tower site representing the EC tower foot- Hylocomnium splendens wet and poor

Vaccinium oxycoccus print. Carex chordorrhiza – Vaccinium myrtilloides , , , , To study the general pattern in mire vegetation, we related BM of PFTs to WT using DCA and generalized additive models (GAM) available in CanoDraw 4.12 (ter Braak and Sphagnum balticum Sphagnum fuscum ,

, Smilauer 2002b). Here we also used the additional BM data Cladonia gracilis Limprichtia intermedia Rubus chamaemorus , ,

, from Mer Bleue bog, namely from Poor Fen and Blue Dome Menyanthes trifoliata , Pinus sylvestris Sphagnum platyphyllum Eriophorum vaginatum ,

, sites. We first performed a DCA analysis with detrending by , , Carex buxbaumii Dicranum bergerii Eriophorum angustifolium , second-order polynomials based on the BM of PFTs includ- , Vaccinium myrtillus

, ing sample plots from all four studied mires. WT was in- spp., cluded in the analysis as a supplementary environmental variable. In GAM, response variables are related to the pre- dictor using semiparametric smoothing functions that do not Ledum palustre subsecundum oligosperma Cladonia arbuscula Andromeda glauca Sphagnum angustifolium Sphagnum capillifolium Betula nana Sphagnum majus longifolia Carex pauciflora Drosera rotundifolia Carex aquatilis Dicranum Campylium stellatum have fixed shapes. We set the number of degrees of freedom Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 to four to allow response curves to be highly flexible (Guisan et al. 2002). We used log-link function with Poisson distribu- tion for model residuals (Lepš and Smilauer 2003).

Results Composition of GBM in EC footprint Classification of species occurring in the study sites into plant functional types (PFTs).

Plant species were classified into PFTs based on the literature according to their ecological niche optima along fertility and water level gradients The amount of variation in plant species GBM differed greatly between the EC tower footprints of the four mires as

Note: indicated by the lengths of the first gradient (the longest dis- minerotrophic conditions to formoptimum considerable along biomass. the Ombrotrophic microtopographical plant gradient. types While are conditions also become capable suboptimal, of species growing abundance in decreases minerotrophic and conditions. finally, The in divi the end of their to mosses are divided by the convergent effect of fertility and moisture to rich PFTsEvergreen shrub Species Lawn sphagna Hummock sphagna Deciduous shrub Hollow sphagna Minero forb Ombro sedge Ombro forb Minero sedge Feather moss and Brown moss Table 2. tance between the sample plot scores in the first axis) in

Published by NRC Research Press Laine et al. 67

Fig. 1. Ordination of sample plots and plant species green biomass for (A) Kaamasensuo and (B) Siikaneva (MI, minerotrophic hummocks; OM, ombrotrophic hummocks; HU, low hummocks; EV, Eriophorum vaginatum dominated lawns; CL, Carex lasiocarpa dominated lawns; CR, Carex rostrata dominated lawns; and HO, hollows). (C) Mer Bleue (including sample plots around EC tower) and (D) Degerö Stormyr. Ordination was done by detrended correspondence analysis that was detrended by segments; therefore the length of the first axis describes the maximal amount of compositional variation. The longest distance between the sample plot scores in the first axis that is a measure of the largest compositional variation in the data sets was 7.2, 5.8, 1.7, and 1.4 for Kaamasensuo, Siikaneva, Mer Bleue, and Degerö Stormyr, re- spectively. For the list of species present in the figure see Table 2. For personal use only.

DCA (Fig. 1). The compositional variation was largest in the Abundance of BM footprint of the northern fen Kaamasensuo, where the string Total area weighted average aboveground BM within the

Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 and flark sample plots had only one species (E. vaginatum) EC tower footprint was 303 ± 31, 253 ± 25, 543 ± 27, and in common (Fig. 1A). In the Siikaneva footprint, the sample 141 ± 45 g·m–2 (±SE) in Kaamasensuo, Siikaneva, Mer plots formed a gradient from wet hollows through lawns that Bleue, and Degerö Stormyr, respectively; while the GBM are dominated by different sedge species (E. vaginatum in (including vascular plant leaves and the topmost part of the EV lawns, C. lasiocarpa in CL lawns, and C. rostrata in mosses) was 236 ± 28, 228 ± 23, 336 ± 16, and 126 ± CR lawns) to hummocks (Fig. 1B). In the footprint of Mer 7g·m–2 (±SE) in Kaamasensuo, Siikaneva, Mer Bleue, and Bleue bog, the variation was small and all sample plots be- Degerö Stormyr, respectively (Fig. 2). The differences in to- longed to a similar hummock community (Fig. 1C). In the tal and GBM between the mires were significant (p <0.001, footprint of Degerö Stormyr, the gradient was short as well Kruskal–Wallis Test). Similar to the composition of plant and all sample plots represented similar lawn level vegeta- species GBM, there were differences in the amount of inter- tion (Fig. 1D). nal variation in GBM within the sites. In Kaamasensuo, the

Published by NRC Research Press 68 Botany, Vol. 90, 2012

Fig. 2. (A) Total aboveground biomass (BM) divided into mosses, PFTs BM distribution in relation to WT green parts of vascular plants, and woody parts of vascular plants; We found both similarities and differences in the composi- aboveground green biomass (GBM) of (B) vascular plants and (C) tion of GBM of PFTs between the mires that were related to mosses in four northern mires. In Mer Bleue only the sample plots the WT (Fig. 4). The strings in Kaamasensuo, the hummocks around the eddy covariance tower (Tower site) are included. For in Siikaneva, and the bog vegetation in Mer Bleue had simi- Sphagnum mosses only capitulum BM was included. For other lar PFTs BM compositions; and the lawns in Degerö Stormyr mosses the uppermost 2 cm was harvested. and Siikaneva and the fen vegetation in Mer Bleue were also similar. The flarks in Kaamasensuo and the hollows in Siika- neva differed from all other plots (Fig. 4). The average WT of Kaamasensuo strings was –65 cm (SD ± 9), of Mer Bleue bog (Tower Site and Blue Dome) was –37 cm (SD ± 8), and of Siikaneva hummocks was –19 cm (SD ± 5). The average WT in Mer Bleue Poor Fen site was –29 cm (SD ± 7), in Siikaneva lawns was –12 cm (SD ± 2), and in Degerö Stor- myr was –8 cm (SD ± 3). Siikaneva hollows had an average WT of 0 cm (SD ± 3), while at Kaamasensuo flarks it was +5 cm (SD ± 5). The GBM of PFTs showed clear differences in response to the WT (Figs. 5A, 5B). Shrubs in general had their highest BM at dry sites; evergreen shrubs had, however, a wide range along the WT. Minerotrophic sedges had two BM peaks, first at the intermediate WT and then in the wettest sites. Ombrotrophic sedges occurred in all microforms, but were most abundant in lawns, with the WT around –10 cm (Fig. 5A). Feather mosses dominated the bottom layer in the driest sites, while hummock sphagna had a BM peak at the WT of –30 cm; both groups were missing from sites with the WT higher than –10 cm below surface. Lawn sphagna occurred throughout the gradient, but had its highest BM at the WT around –10 cm. Hollow sphagna occurred in sites with WT near the surface, while the brown mosses were re- stricted to flarks of Kaamasensuo (Fig. 5B).

For personal use only. Relationship between different BM compartments and NEE We compared the different BM compartments with the average July NEE, which represents the maximal capacity of the vegetation and is less dependent on the length of the growing season than annual NEE (Fig. 6). Total BM or GBM was not always related to the NEE of the peak growing season (Figs. 6A, 6B). However, the NEE was clearly higher in mires with high vascular plant GBM, namely in Kaama- flarks and strings differed strongly with GBM of 160 ± 20 sensuo and Mer Bleue (Fig. 6C). Siikaneva and Degerö Stor- – and 415 ± 44 g·m 2 (±SE), respectively (Fig. 3). In flarks, myr had nearly identical July NEE, despite the difference in the BM was composed of brown mosses and minerotrophic total BM. The difference in BM at these two mires is formed sedges, while in hummocks feather mosses and evergreen by the Sphagnum mosses, and therefore, vascular plants ap- shrubs dominated. In Siikaneva, the C. rostrata lawns had peared to drive the growing season maximum NEE (Figs. 6C, – the highest (488 ± 20 g·m 2) and hollows had the lowest 6D). Kaamasensuo had equally high July NEE as Mer Bleue, – (157 ± 14 g·m 2) GBM (Fig. 3). Sphagnum mosses com- despite the smaller vascular plant BM. The higher proportion posed the largest share of the GBM in all plant communities. of sedge BM in Kaamasensuo was able to compensate in the Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 In Mer Bleue bog, each of the three sites was relatively ho- NEE: the higher total vascular BM of Mer Bleue, which is mogeneous, and the amount and composition of GBM var- formed mostly by shrubs (Figs. 6E, 6F). ied only little between the sites being 336 ± 16, 238 ± 12, – and 348 ± 17 g·m 2 in the Tower, Blue Dome, and Poor Fen Discussion sites, respectively. Shrubs dominated the vascular vegetation in all Mer Bleue sites. In the Poor Fen, the sedges were Abundance of BM mostly minerotrophic Carex species, while in the Tower and The total aboveground BM of all four studied mires (from Blue Dome sites E. vaginatum was the most common sedge. 0.1 to 0.5 kg·m–2) was within the range measured from other In Degerö Stormyr, the BM was composed of evergreen treeless mires (Kosykh et al. 2008; Reinikainen et al. 1984; shrubs, ombrotrophic sedges, and Sphagnum mosses Vasander 1982, 1987). In comparison with other ecosystems, (Fig. 3). the total aboveground BM of treeless mires is clearly smaller

Published by NRC Research Press Laine et al. 69

Fig. 3. Aboveground green biomass (GBM) of (A) vascular plants and (B) mosses in four northern mires pooled to plant community level. KA_F, Kaamasensuo flarks and KA_S, Kaamasensuo strings. SN relates to Siikaneva, with plant communities: MI, minerotrophic hummocks; OM, ombrotrophic hummocks; HU, low hummocks; EV, Eriophorum vaginatum dominated lawns; CL, Carex lasiocarpa dominated lawns; CR, Carex rostrata dominated lawns; HO, hollows. DS is Degerö Stormyr. MB refers to Mer Bleue, with sites: PF, Poor Fen; BD, Blue Dome; and TW, bog around EC tower. For Sphagnum mosses only capitulum biomass was included. For other mosses the uppermost 2 cm was harvested. The number in parentheses following the site name refers to the average water table in relation to the soil surface. For personal use only.

than in temperate and boreal forests (21 and 6 kg·m–2, re- Minerotrophic sedges include several Carex spp. with con- spectively) and about the same as in temperate grasslands trasting response to the WT. Firstly, they were common in (∼0.25 kg·m–2) and in arctic tundra, ∼0.25 kg·m–2 (Chapin relatively dry poor fen communities (Mer Bleue and Siika- et al. 2002). The GBM of vascular plants and mosses was neva), and secondly, their GBM was high in the inundated quite similar in Kaamasensuo and Siikaneva, which are very flarks of Kaamasensuo. Sedges are deep-rooting species with different types of mires. Kaamasensuo is a northern aapa aerenchymous cell tissue, which enables them to habit a wide mire with a remarkably distinct surface pattern of high ridges range of sites. Forbs constituted little of the BM; of the om- and open water flarks and Siikaneva is a fen with continuous brotrophic forbs R. chamaemorus inhabited the dry sites, Sphagnum cover. Mer Bleue that had 1.5× larger GBM than while Maianthemum trifolium (L.) Sloboda, Menyanthes tri- Kaamasensuo and Siikaneva is a rather dry bog with abun- foliata L., and S. palustris occurred in wetter lawns and hol- dant shrub cover. Although Degerö Stormyr is visually simi- lows. Earlier findings show that mosses have narrower lar to Siikaneva, the GBM was much lower, and especially tolerance to the WT than vascular plants (Väliranta et al. the moss GBM was only one third of that in Siikaneva. 2007). Our results seem to support this idea to some extent, and the wider ranges are due to pooling species into groups. Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 Composition of BM Feather mosses occur in the driest hummocks and strings. As expected, the GBM of PFTs responded differently to Hummock sphagna (S. capillifolium in Mer Bleue and WT. Shrubs had their peak abundance in dry sites. However, Sphagnum fuscum (Schimp.) Klinggr. in Siikaneva) occurred evergreen shrubs, which had by far the highest proportion of in relatively dry hummocks, having GBM optima at ∼–30 cm. vascular BM, occurred throughout the WT gradient, although Lawn sphagna had clearly the highest proportion of all moss to a smaller extent in the wetter sites. The group includes GBM and occurred throughout the sites; the optima at species V. oxycoccos and A. polifolia that are known to have –12 cm was still quite distinct. Hollow sphagna (S. majus) a wide WT tolerance (Lumiala 1944). Also sedge BM had a was restricted to the hollows in Siikaneva, while the brown wide tolerance for the WT. Ombrotrophic sedges, mainly mosses occupied the minerotrophic flarks of Kaamasensuo. E. vaginatum, occurred in all sites apart from Kaamasensuo, One needs to bear in mind that WTs used here represent but their BM peaked at lawns with the WT around –10 cm. only the period of the sampling time, while the vegetation

Published by NRC Research Press 70 Botany, Vol. 90, 2012

Fig. 4. Ordination of sample plots and plant functional types bio- Fig. 5. Green biomass (GBM) of (A) vascular plant and (B) moss mass from four northern mires: Kaamasensuo (KA), Siikaneva plant functional types in relation to the water table (WT). We use a (SN), Mer Bleue (MB), and Degerö Stormyr (DS). Ordination was negative sign for WT below the soil surface and a positive sign carried out by detrended correspondence analysis (DCA) that was when the WT is above soil surface. Response curves are created detrended by second-order polynomials. Eigenvalue of the first and using generalized additive models (GAM). In GAM, response vari- second axis were 0.767 and 0.326, respectively. The two axes to- ables are related to the predictor using semiparametric smoothing gether explained 50.6% of the variation in the data. Water table is functions that do not have fixed shapes. included as a supplementary variable. We use a negative sign for water table below the soil surface and a positive sign when the water table is above the soil surface; conditions become wetter in the direction of the arrow. For personal use only.

maximal capacity of the vegetation. Vascular plant GBM was more closely connected to NEE than total BM or GBM. July NEE was clearly higher in mires with high vascular response may relate to the WT during the life history of plant GBM, namely in Kaamasensuo and Mer Bleue. Further, plants. The long-term WT may vary from those used here. the different photosynthetic capacity of PFTs appeared to matter and thus the smaller BM with different species com- BM as potential control for mire NEE position may result in similar or higher CO2 flux rates. In all studied sites, multiyear biosphere–atmosphere CO2 Sphagnum mosses are known to be less efficient photosyn- Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 exchange has been measured by the EC method. The average thesizers under all light levels than vascular plants and of annual NEE is 22 ± 20 g C·m–2·a–1 in Kaamasensuo (Aurela vascular plants both ombrotrophic and minerotrophic sedges et al. 2004), 56 ± 18 g C·m–2·a–1 in Siikaneva (Aurela et al. have higher capacity to photosynthesize than shrubs, when 2007), 40.2 ± 40.5 g C·m–2·a–1 in Mer Bleue (Roulet et al. comparing the efficiency per green area unit (Leppälä et al. 2007), and 55 ± 7 g C·m–2·a–1 in Degerö Stormyr (Sagerfors 2008; Riutta et al. 2007b). In our study, the composition of et al. 2008). NEE represents the difference in photosynthesis BM varied between and within the mires more than the total and ecosystem respiration and the length of the growing sea- BM. Siikaneva and Degerö Stormyr have nearly identical son is important in determining to what extent the photosyn- July NEE, despite the difference in total BM. The difference thetically active BM can contribute to the annual NEE (Lund in BM at these two mires is formed by the Sphagnum et al. 2010). Therefore, we compared the different BM com- mosses, and therefore, vascular plants appear to drive the partments with the average July NEE, which represents the growing season maximum NEE. Furthermore, the vascular

Published by NRC Research Press Laine et al. 71

–2 –1 Fig. 6. Average cumulative July net ecosystem CO2 exchange (NEE) (g C·m ·month ±SE) of each site related to (A) total aboveground biomass (BM), (B) green biomass (GBM), (C) GBM of vascular plants, (D) GBM of mosses, (E) GBM of evergreen shrubs, and (F) GBM of minerotrophic sedges. NEE represents the difference in photosynthesis and ecosystem respiration. The July NEE is an average (±SE) of 6, 2, 7, and 12 years from sites Kaamasensuo, Siikaneva, Degerö Stormyr, and Mer Bleue, respectively. Reference for NEE data for Kaamasensuo (Aurela et al. (2004)); Siikaneva (Aurela et al. (2007)); Mer Bleue (Roulet et al. (2007) and Humphreys, personal communication); and De- gerö (Sagerfors et al. (2008), Nilsson et al. (2008), and Nilsson, unpublished data). For personal use only. Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12

plant composition is rather similar in these two mires. The nearly identical July NEE, despite the difference in total BM. significance of the vascular PFTs composition is most ob- The difference in BM at these two mires is formed by the vious when comparing Kaamasensuo and Mer Bleue. The to- Sphagnum mosses, and therefore, vascular plants appeared to tal BM was clearly lower in Kaamasensuo than in Mer Bleue. drive the growing season maximum NEE. However, the higher proportion of sedge BM resulted in The dominant impact of sedge GBM over the NEE is evi- equally high NEE in Kaamasensuo than in Mer Bleue. In dent also when the internal GBM variation within mires is Mer Bleue, BM was formed mostly of shrubs. In our study, observed. The amount of internal variation in plant species the composition of BM varied between and within the mires GBM was highest in Kaamasensuo and lowest in Degerö more than the total BM. Siikaneva and Degerö Stormyr have Stormyr. The plant communities in Kaamasensuo, strings

Published by NRC Research Press 72 Botany, Vol. 90, 2012

and flarks, strongly differed from each other in terms of WT, into PFTs, in modelling ecosystem carbon exchange instead species composition, and BM PFTs. Consequently, their NEE of using just total photosynthesizing plant BM. vary considerably. In Kaamasensuo the closed chamber measurements showed that flark communities with fen vege- Acknowledgements tation had higher NEE during the growing season (average We are grateful to Paavo Ojanen and other participants of and cumulative NEE) than the string tops with bog vegeta- Peatland Ecology course of University of Helsinki in August tion (Maanavilja et al. 2010). Second largest variation in 2006 for their help with BM sampling in Siikaneva and Kaa- GBM composition was observed in Siikaneva, where the hol- masensuo; Gareth Crosby and other Mount Holyoke College lows and ombrotrophic hummocks differed from the dominat- and McGill University students for assistance with BM sam- ing lawns. In Siikaneva the NEE was highest in pling at Mer Bleue; and Rose-Marie Kronberg for assistance C. lasiocarpa lawns and lowest in hollows, according to with BM sampling at Degerö Stormyr. We thank Elyn Hum- chamber measurements (Riutta et al. 2007a). Degerö Stormyr phreys and Mika Aurela for NEE data from the Mer Bleue, vegetation showed little variation, being most similar to Siikaneva, and Kaamasensuo sites. Funding from University E. vaginatum lawns of Siikaneva. Eriophorum vaginatum of Oulu to A.M.L., US National Science Foundation to J.B., lawns are the dominant plant community in Siikaneva cover- Graduate School in Forest Sciences to T.R., Swedish Re- ing 34% of the mire, which may explain the equally high search Council for Environment, Agricultural Sciences and NEE in the two mires. Our findings seem to contrast with Spatial Planning to M.B.N., Natural Sciences and Engineer- those of Shaver et al. (2007) that showed leaf area index ing Research Council of Canada to T.R.M., University of (LAI) to be the only plant parameter needed to successfully Helsinki to H.V., and Academy of Finland to E.S.T. (project predict NEE in tundra vegetation. As Shaver et al. (2007) conclude, this indicates that in spite of species composition code 218101, 140863) are acknowledged. The National Cap- any kind of vegetation over the tundra landscape produces ital Commission of Canada is thanked for access to the Mer the same functions with same leaf area unit. This high con- Bleue site. vergence, not found in our study in different peatlands, may References relate to smaller spatial variation in tundra than in peatlands. While the study plots in the tundra study had variable plant Ahti, T., Hämet-Ahti, L., and Jalas, J. 1968. Vegetation zones and communities, the main components (namely deciduous and their sections in northwestern Europe. Ann. Bot. Fenn. 5: 169– evergreen shrubs, sedges, and forbs) were present in all of 211. them with relatively high cover. That was not the case in the Alm, J., Talanov, A., Saarnio, S., Silvola, J., Ikkonen, E., Aaltonen, peatlands of this study where the differences in the plant H., Nykänen, H., and Martikainen, P.J. 1997. Reconstruction of community compositions were quite high. The other, more the carbon balance for microsites in a boreal oligotrophic pine fen, – interesting, explanation for the quite different results may be Finland. Oecologia (Berl.), 110(3): 423 431. doi:10.1007/ due to the different plant units studied. In our study, we re- s004420050177. Alm, J., Schulman, L., Walden, J., Nykänen, H., Martikainen, P.J.,

For personal use only. late NEE to BM, while Shaver et al. (2007) were linking and Silvola, J. 1999. Carbon balance of a boreal bog during a year NEE to LAI. It has already been shown that among different – dwarf shrub species there can be remarkable variation in pho- with an exceptionally dry summer. Ecology, 80(1): 161 174. doi:10.1890/0012-9658(1999)080[0161:CBOABB]2.0.CO;2. tosynthetic capacity per BM. Kulmala et al. (2008) found that Andrus, R.E., Wagner, D.J., and Titus, J.E. 1983. Vertical zonation of Vaccinium myrtillus, with its thin deciduous leaves, can have Sphagnum mosses along hummock–hollow gradients. Can. J. Bot. two or three times higher momentary photosynthesis per leaf 61(12): 3128–3139. doi:10.1139/b83-352. mass than Vaccinium vitis-idaea or Calluna vulgaris that Aurela, M., Tuovinen, J.-P., and Laurila, T. 1998. Carbon dioxide have evergreen leaves. However, per LAI the photosynthetic exchange in a subarctic peatland ecosystem in northern Europe capacities did not differ much because of the heavier leaves measured by the eddy covariance technique. J. Geophys. Res. of V. vitis-idaea. A mechanism behind the functional diver- 103(D10): 11289–11301. doi:10.1029/98JD00481. gence in vegetation might be the differentiation between Aurela, M., Laurila, T., and Tuovinen, J.-P. 2001. Seasonal CO2 plants in how they allocate BM to build leaf area. balances of a subarctic mire. J. Geophys. Res. 106(D2): 1623– 1637. doi:10.1029/2000JD900481. Conclusions Aurela, M., Laurila, T., and Tuovinen, J.-P. 2002. Annual CO2 balance of a subarctic fen in northern Europe: importance of the The results supported our first hypothesis that the spatial wintertime efflux. J. Geophys. Res. 107(D21): 4607–4618. doi:10. variation in BM within a mire can be larger than the differ- 1029/2002JD002055. ences in BM between mires. In the mires Kaamasensuo and Aurela, M., Laurila, T., and Tuovinen, J.-P. 2004. The timing of snow Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 Siikaneva, where the internal variation within plant commun- melt controls the annual CO2 balance in a subarctic fen. Geophys. ities was distinctive, the differences in the GBM between Res. Lett. 31(16): L16119–L16123. doi:10.1029/2004GL020315. mires were smaller than between the plant communities. The Aurela, M., Riutta, T., Laurila, T., Tuovinen, J.-P., Vesala, T., Tuittila, similarity of NEE and GBM of vascular PFTs in the domi- E.-S., Rinne, J., Haapanala, S., and Laine, J. 2007. CO2 exchange nant plant community types in Degerö Stormyr and Siika- of a sedge fen in southern Finland—the impact of a drought neva gave support to the second hypothesis that NEE is period. Tellus B Chem. Phys. Meterol. 59(5): 826–837. doi:10. related to BM of the dominant PFTs. Eleven PFTs appeared 1111/j.1600-0889.2007.00309.x. to capture the internal variation; the differences in BM be- Austin, M.P. 2005. Vegetation and environment: discontinuities and tween communities, which were defined using species level continuities. In Vegetation ecology. Edited by E. van der Maarel. approach, were explained by the WT response of PFTs. Our Blackwell Publishing, Malden, USA; Oxford, UK; Carlton, results suggest the use of photosynthesizing BM, separated Australia. pp. 52–84.

Published by NRC Research Press Laine et al. 73

Baldocchi, D.D. 2003. Assessing the eddy covariance technique for M. 1997. Sources of spatial variation in methane emission from evaluating carbon dioxide exchange rates of ecosystems: past, mires in northern Sweden: a mechanistic approach in statistical present and future. Glob. Change Biol. 9(4): 479–492. doi:10. modeling. Global Biogeochem. Cycles, 11(2): 135–150. doi:10. 1046/j.1365-2486.2003.00629.x. 1029/96GB03352. Bragazza, L., and Gerdol, R. 2002. Are nutrient availability and Guisan, A., Edwards, T., and Hastie, T. 2002. Generalized linear and acidity–alkalinity gradients related in Sphagnum-dominated peat- generalized additive models in studies of species distributions: lands? J. Veg. Sci. 13(4): 473–482. doi:10.1111/j.1654-1103.2002. setting the scene. Ecol. Model. 157(2–3): 89–100. doi:10.1016/ tb02074.x. S0304-3800(02)00204-1. Bubier, J.L. 1995. The relationship of vegetation to methane emission Joabsson, A., Christensen, T.R., and Wallén, B. 1999. Vascular plant and hydrochemical gradients in northern peatlands. J. Ecol. 83(3): controls on methane emissions from northern peatforming 403–420. doi:10.2307/2261594. wetlands. Trends Ecol. Evol. 14(10): 385–388. doi:10.1016/ Bubier, J.L., Crill, P.M., Mosedale, A., Frolking, S., and Linder, E. S0169-5347(99)01649-3. PMID:10481199. 2003. Peatland responses to varying interannual moisture condi- Kalela, A. 1961. Waldvegetationszonen Finlands und ihre klima- tions as measured by automatic CO2 chambers. Global Biogeo- tischen Paralleltypen. Arch. Soc. Vanamo, 16(Suppl): 65–83. chem. Cycles, 17(2): 1066–1080. doi:10.1029/2002GB001946. Koelbener, A., Ström, L., Edwards, P.J., and Venterink, H.O. 2010. Bubier, J.L., Moore, T.R., and Crosby, G. 2006. Fine-scale vegetation Plant species from mesotrophic wetlands cause relatively high distribution in a cool temperate peatland. Can. J. Bot. 84(6): 910– methane emissions from peat soil. Plant Soil, 326(1–2): 147–158. 923. doi:10.1139/b06-044. doi:10.1007/s11104-009-9989-x. Chapin, F.S., , III, Bret-Harte, M.S., Hobbie, S.E., and Zhong, H. Kosykh, N.P., Koronatova, N.G., Naumova, N.B., and Titlyanova, A. 1996. Plant functional types as predictors of transient responses of A. 2008. Above- and below-ground phytomass and net primary arctic vegetation to global change. J. Veg. Sci. 7(3): 347–358. production in boreal mire ecosystems of Western Siberia. doi:10.2307/3236278. Wetlands Ecol. Manage. 16(2): 139–153. doi:10.1007/s11273- Chapin, F.S., III, Matson, P.A., and Mooney, H.A. 2002. Principles of 007-9061-7. terrestrial ecosystem ecology. Springer-Verlag, New York. Kulmala, L., Launiainen, S., Pumpanen, J., Lankreijer, H., Lindroth, Dreps, A., Nordlund, A., Karlsson, P., Helminen, J., and Rissanen, P. A., Hari, P., and Vesala, T. 2008. H2O and CO2 fluxes at the floor 2002 Climatological statistics of Finland 1971–2000. In Climatic of boreal pine forest. Tellus, 60B: 167–178. statistics of Finland 2002:1. Finnish Meteorological Institute, Lafleur, P.M., Roulet, N.T., and Admiral, S.W. 2001. Annual cycle of Helsinki, Finland. CO2 exchange at a bog peatland. J. Geophys. Res. 106(D3): 3071– Dias, A.T.C., Hoorens, B., Van Logtestijn, R.S.P., Vermaat, J.E., and 3081. doi:10.1029/2000JD900588. Aerts, R. 2010. Plant species composition can be used as a proxy Laine, A., Byrne, K.A., Kiely, G., and Tuittila, E.-S. 2007. Patterns in to predict methane emissions in peatland ecosystems after land-use vegetation and CO2 dynamics along a water level gradient in a changes. Ecosystems (N. Y.), 13(4): 526–538. doi:10.1007/ lowland blanket bog. Ecosystems (N.Y.), 10(6): 890–905. doi:10. s10021-010-9338-1. 1007/s10021-007-9067-2. Eppinga, M.B., Rietkerk, M., Belyea, L.R., Nilsson, M.B., De Ruiter, Laine, J., Harju, P., Timonen, T., Laine, A., Tuittila, E.-S., Minkkinen, P.C., and Wassen, M.J. 2010. Resource contrast in patterned K., and Vasander, H. 2009. The intricate beauty of Sphagnum For personal use only. peatlands increases along a climatic gradient. Ecology, 91(8): mosses—a Finnish guide to identification. University of Helsinki 2344–2355. doi:10.1890/09-1313.1. PMID:20836456. Department of Forest Ecology, Publication 39. pp. 1–190. Eurola, S., Huttunen, A., and Kukko-oja, K. 1995. Suokasvillisuu- Leppälä, M., Kukko-oja, K., Laine, J., and Tuittila, E.-S. 2008. sopas. Oulanka Rep., 14:1–85. [In Finnish.] Seasonal dynamics of CO2 exchange during primary succession of Frolking, S., Roulet, N.T., Moore, T.R., Lafleur, P.M., Bubier, J.L., boreal mires as controlled by phenology of plants. Ecoscience, and Crill, P.M. 2002. Modeling seasonal to annual carbon balance 15(4): 460–471. doi:10.2980/15-4-3142. of Mer Bleue Bog, Ontario, Canada. Global Biogeochem. Cycles, Lepš, J., and Smilauer, P. 2003. Multivariate analysis of ecological 16(3): 1030–1050. doi:10.1029/2001GB001457. data using CANOCO. Cambridge University Press, Cambridge. Frolking, S., Roulet, N., and Fuglestvedt, J. 2006. How northern Lumiala, O.V. 1944. Uber die beziehing einiger Moorpflanzen zu der peatlands influence the Earth’s radiative budget: sustained Grundwasserhohe. C.R. Soc. Geol. Finlande, 16: 147–164. methane emissions versus sustained carbon sequestration. J. Lund, M., Lafleur, P.M., Roulet, N.T., Lindroth, A., Christensen, T.R., Geophys. Res. 111(G1): G01008. doi:10.1029/2005JG000091. Aurela, M., Chojnicki, B.H., Flanagan, L.B., Humphreys, E.R., Frolking, S., Roulet, N.T., Tuittila, E., Bubier, J.L., Quillet, A., Laurila, T., Oechel, W.C., Olejnik, J., Rinne, J., Schubert, P.E.R., Talbot, J., and Richard, P.J.H. 2010. A new model of Holocene and Nilsson, M.B. 2010. Variability in exchange of CO2 across 12 peatland net primary production, decomposition, water balance, northern peatland and tundra sites. Glob. Change Biol. 16(9): and peat accumulation. Earth Syst. Dynam. Discuss., 1(1): 115– 2436–2448. 167. doi:10.5194/esdd-1-115-2010. Maanavilja, L., Riutta, T., Aurela, M., Pulkkinen, M., Laurila, T., and Gignac, L.D., Gauthier, R., Rochefort, L., and Bubier, J.L. 2004. Tuittila, E.-S. 2010. Spatial variation in CO exchange at a Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12 2 Distribution and habitat niches of 37 peatland Cyperaceae species northern aapa mire. Biogeochemistry, 104(1–3): 325–345. doi:10. across a broad geographic range in Canada. Can. J. Bot. 82(9): 1007/s10533-010-9505-7. 1292–1313. doi:10.1139/b04-081. Malmer, N. 1962. Studies on mire vegetation in the Archaean area of Glaser, P.H., Janssens, J.A., and Siegel, D.I. 1990. The response of Southwestern Götaland (South Sweden). I. Vegetation and habitat vegetation to chemical and hydrological gradients in the Lost conditions on the Åkerhult mire. Opera Bot. 7:1–322. River peatland, north Minnesota. J. Ecol. 78(4): 1021–1048. Moore, T.R., Bubier, J.L., and Bledzki, L. 2007. Litter decomposition doi:10.2307/2260950. in temperate peatland ecosystems: the effect of substrate and site. Gorham, E. 1991. Northern peatlands: role in the carbon cycle and Ecosystems (N. Y.), 10(6): 949–963. doi:10.1007/s10021-007- probable responses to climatic warming. Ecol. Appl. 1(2): 182– 9064-5. 195. doi:10.2307/1941811. Nilsson, M., Mikkelä, C., Sundh, I., Granberg, G., Svensson, B., and Granberg, G., Mikkelä, C., Sundh, I., Svensson, B.H., and Nilsson, Ranneby, B. 2001. Methane emission from Swedish mires:

Published by NRC Research Press 74 Botany, Vol. 90, 2012

national and regional budgets and dependence on mire vegetation. cycling in a high arctic fen. Ecosystems (N.Y.), 11(1): 61–76. J. Geophys. Res. 106(D18): 20847–20860. doi:10.1029/ doi:10.1007/s10021-007-9107-y. 2001JD900119. Szumigalski, A.R., and Bayley, S.E. 1996. Decomposition along a Nilsson, M., Sagerfors, J., Buffam, I., Laudon, H., Eriksson, T., bog to rich fen gradient in central Alberta, Canada. Can. J. Bot. Grelle, A., Klemedtsson, L., Weslien, P., and Lindroth, A. 2008. 74(4): 573–581. doi:10.1139/b96-073. Contemporary carbon accumulation in a boreal oligotrophic Teklemariam, T.A., Lafleur, P.M., Moore, T.R., Roulet, N.T., and minerogenic mire – a significant sink after accounting for all C- Humphreys, E.R. 2010. The direct and indirect effects of inter- fluxes. Glob. Change Biol. 14(10): 2317–2332. doi:10.1111/j. annual meteorological variability on ecosystem carbon dioxide 1365-2486.2008.01654.x. exchange at a temperate ombrotrophic bog. Agric. For. Meteorol. Reinikainen, A., Vasander, H., and Lindholm, T. 1984. Plant biomass 150(11): 1402–1411. doi:10.1016/j.agrformet.2010.07.002. and primary production of southern boreal mire-ecosystems in ter Braak, C.J.F., and Prentice, I.C. 1988. A theory of gradient Finland. Proceedings, 7th International Peat Congress, Dublin 4. analysis. Adv. Ecol. Res. 18: 271–317. doi:10.1016/S0065-2504 pp. 1–20. (08)60183-X. Riutta, T., Laine, J., Aurela, M., Rinne, J., Vesala, T., Laurila, T., ter Braak, C.J.F., and Smilauer, P. 2002a. CANOCO Reference Haapanala, S., Pihlatie, M., and Tuittila, E.-S. 2007a. Spatial manual and CanoDraw for Windows User’s guide: software for variation in plant community functions regulates carbon gas Canonical Community Ordination (version 4.5). Microcomputer – dynamics in a boreal fen ecosystem. Tellus, 59B: 838 852. Power, Ithaca, N.Y., USA. Riutta, T., Laine, J., and Tuittila, E.-S. 2007b. Sensitivity of CO2 ter Braak, C.J.F., and Smilauer, P. 2002b. Reference manual and exchange of fen ecosystem components to water level variation. CanoDraw for Windows User’s guide: software for Canonical – Ecosystems (N. Y.), 10(5): 718 733. doi:10.1007/s10021-007- Community Ordination (version 4.5). Microcomputer Power. 9046-7. Turunen, J., Tomppo, E., Tolonen, K., and Reinikainen, A. 2002. Roulet, N.T., Lafleur, P.M., Richard, P.J.H., Moore, T.R., Hum- Estimating carbon accumulation rates of undrained mires in phreys, E.R., and Bubier, J. 2007. Contemporary carbon balance Finland—application to boreal and subarctic regions. Holocene, and late Holocene carbon accumulation in a northern peatland. 12(1): 69–80. doi:10.1191/0959683602hl522rp. Glob. Change Biol. 13(2): 397–411. doi:10.1111/j.1365-2486. Väliranta, M., Korhola, A., Seppä, H., Tuittila, E.-S., Sarmaja- 2006.01292.x. Korjonen, K., Laine, J., and Alm, J. 2007. High-resolution Ruuhijärvi, R. 1983. The Finnish mire types and their regional reconstruction of wetness dynamics in a southern boreal raised destribution. In Ecosystems of the world. Edited by A.J.P. Gore. bog, Finland, during the late Holocene: a quantitative approach. Elsevier, Amsterdam. pp. 47–67. Holocene, 17(8): 1093–1107. doi:10.1177/0959683607082550. Rydin, H., and McDonald, A.J.S. 1985. Tolerance of Sphagnum to Vasander, H. 1982. Plant biomass and production in virgin, drained water level. J. Bryol. 13: 571–578. and fertilized sites in a raised bog in southern Finland. Ann. Bot. Saarnio, S., Morero, M., Shurpali, N.J., Tuittila, E.-S., Mäkilä, M., Fenn. 19: 103–125. and Alm, J. 2007. Annual CO2 and CH4 fluxes of pristine boreal mires as a background for the lifecycle analyses of peat energy. Vasander, H. 1987. Diversity of understorey biomass in virgin and in Boreal Environ. Res. 12: 101–114. drained and fertilized southern boreal mires in eastern Fennos- – For personal use only. Sagerfors, J., Lindroth, A., Grelle, A., Klemedtsson, L., Weslien, P., kandia. Ann. Bot. Fenn. 24: 137 153. and Nilsson, M. 2008. Annual CO exchange between a nutrient- Waddington, J.M., and Roulet, N.T. 2000. Carbon balance of a boreal 2 – poor, minerotrophic, boreal mire and the atmosphere. J. Geophys. patterned peatland. Glob. Change Biol. 6(1): 87 97. doi:10.1046/j. Res. 113(G1): G01001–G01015. doi:10.1029/2006JG000306. 1365-2486.2000.00283.x. Schubert, P., Eklundh, L., Lund, M., and Nilsson, M. 2010. Walker, M.D., Wahren, C.H., Hollister, R.D., Henry, G.H.R., Ahlquist, L.E., Alatalo, J.M., Bret-Harte, M.S., Calef, M.P., Estimating northern peatland CO2 exchange from MODIS time series data. Remote Sens. Environ. 114(6): 1178–1189. doi:10. Callaghan, T.V., Carroll, A.B., Epstein, H.E., Jónsdóttir, I.S., 1016/j.rse.2010.01.005. Klein, J.A., Magnússon, B., Molau, U., Oberbauer, S.F., Rewa, S.P., Shannon, R.D., White, J.R., Lawson, J.E., and Gilmour, B.S. 1996. Robinson, C.H., Shaver, G.R., Suding, K.N., Thompson, C.C., Methane efflux from emergent vegetation in peatlands. J. Ecol. Tolvanen, A., Totland, Ø., Turner, P.L., Tweedie, C.E., Webber, P.J., 84(2): 239–246. doi:10.2307/2261359. and Wookey, P.A. 2006. Plant community responses to experi- Shaver, G.R., Street, L.E., Rastetter, E.B., van Wijk, M.T., and mental warming across the tundra biome. Proc. Natl. Acad. Sci. Williams, M. 2007. Functional convergence in regulation of net U.S.A. 103(5): 1342–1346. doi:10.1073/pnas.0503198103. PMID: CO2 flux in heterogeneous tundra landscapes in Alaska and 16428292. Sweden. J. Ecol. 95(4): 802–817. doi:10.1111/j.1365-2745.2007. Weltzin, J.F., Harth, C., Bridgham, S.D., Pastor, J., and Vonderharr, 01259.x. M. 2001. Production and microtopography of bog bryophytes: Sullivan, P.F., Arens, S.J.T., Chimner, R.A., and Welker, J.M. 2008. response to warming and water-table manipulations. Oecologia Temperature and microtopography interact to control carbon (Berl.), 128(4): 557–565. doi:10.1007/s004420100691. Botany Downloaded from www.nrcresearchpress.com by MOUNT HOLYOKE COLLEGE LIBRARY on 01/23/12

Published by NRC Research Press