Patterns and gradients of diversity in South Patagonianombrotrophic peat bogsaec_2003 1..12

TILL KLEINEBECKER,* NORBERT HÖLZEL AND ANDREAS VOGELInstitute of Landscape Ecology, University of Münster, Robert-Koch-Straße 26, 48149 Münster,Germany (Email: tillhan@uni-muenster.de)

Abstract Many north-hemispherical mires seemingly untouched by drainage and cultivation are influenced bya diffuse sum of man-made environmental changes, such as atmospherical nitrogen deposition that mask generalpatterns in species richness and functional group responses along resource gradients. To obtain insights intonatural diversity-environment relationships, we studied the vegetation and the peat chemistry of pristine bogecosystems in southern Patagonia along a west–east transect across the Andes. The studied bog ecosystemscovered a floristic gradient from hyperoceanic blanket bogs dominated by cushion building vascular plants viaa transitional mixed type to Sphagnum-dominated raised bogs east of the mountain range. To test the influenceof resource availability on diversity patterns, species richness and functional groups were related to environ-mental variables by calculating general regression models and generalized additive models. Species richnessshowed strong linear correlations to peat chemical features and the general regression model resulted in threemajor environmental variables (water level, total nitrogen, NH4Cl soluble calcium), altogether explaining 76% ofvariance. Functional group response illustrated a clear separation along environmental gradients. Mosses domi-nated at the low end of a nitrogen gradient, whereas cushion plants had their optimum at intermediate levels,and graminoids dominated at high nitrogen contents. Further shifts were related to NH4Cl soluble calcium andwater level. The models documented partly non-linear relationships between functional group response andtrophical peat properties. Within the three bog types, the calculated models differed remarkably illustrating thescale-dependency of the explanatory factors. Our findings confirmed several general patterns of species richnessand functional shifts along resource gradients in a surprisingly clear way and underpin the significance ofundisturbed peatlands as reference systems for testing of ecological theory and for conservation and ecologicalrestoration in landscapes with strong human impact.

Key words: biogeochemistry, cushion bog, functional group, ombrotrophic, species richness, Sphagnum.

INTRODUCTION

Ombrotrophic peat bogs of the temperate and borealregions are characterized by harsh environmental con-ditions that constrain plant life to an evolutionaryadaptation to cold and permanently water saturatedsoils, high soil acidity and a low supply of essentialmacro-nutrients. All of these stress imposing factorslimit the total number of species found in this ecosys-tem (Rydin & Jeglum 2006). Ombrotrophic peatlandsreceive all of their nutrients from atmospherical inputso that the environmental conditions only represent aportion of the potential range of resource conditions(Huston 1994). Considering the hump-backed modelby Grime (1973), ombrotrophic peat bogs are locatedat the left side slope of the hump-backed curve, whereunder relatively extreme environmental conditions apositive relation between species richness and produc-tivity can be found.

Within ombrotrophic bogs and related peatlands thewater table gradient, the base cation supply, and theavailability of N and P are among the most oftenmentioned environmental gradients affecting floristiccomposition and diversity patterns (Wheeler &Proctor 2000). As mire ecosystems represent the wetend of a potential water level gradient, species richnessusually increases with decreasing soil moisture (White-house & Bayley 2005). Typically, ombrotrophic bogecosystems are characterized by a microtope zonationfrom wet pools and hollows to relatively dry hum-mocks, which is often associated with a clear floristicgradient (Wheeler 1993). A decrease of species rich-ness with increasing moisture has not been explicitlydescribed for bog vegetation so far, but can be pre-sumed from phytosociological studies (e.g. Dierssen1982; Wells 1996). Given the particularly acidic andnutrient poor conditions of ombrotrophic bogs,species richness was often found to be positively cor-related with increasing soil-pH, base cation supply,and the availability of N and P (Vermeer & Berendse1983; Glaser et al. 1990; Gunnarsson et al. 2000). In

*Corresponding author.Accepted for publication November 2008.

Austral Ecology (2010) 35, 1–12

© 2009 The Authors doi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

the extreme habitat of an ombrotrophic bog otherfactors such as dispersal limitation affecting diversitypatterns (Pärtel & Zobel 2007) are of minorimportance.

Diversity patterns along resource gradients areoften more distinctly indicated by shifts in functio-nal types than by changes in overall species-richness (Pausas & Austin 2001). This is probablyparticularly relevant in species-poor systems such asombrotrophic bogs. For example, Weltzin et al.(2000) documented a divergent functional groupresponse to resource gradients in mires. On the con-trary, Hájek et al. (2002) reported that cryptogamsand vascular plants respond similarly to environmen-tal gradients in spring fens.

Ombrotrophic peatlands in South Patagonia showa clear floristic differentiation from pacific blanketbogs dominated by cushion-forming vascular plantsto more continental Sphagnum-dominated raisedbogs (Pisano 1983; Boelcke et al. 1985; Krisai 2005).A third main type can be found in a transition zone,where cushion bogs and Sphagnum bogs intermingle.The vegetation of many north-hemispherical types ofmire seemingly untouched by drainage and cultiva-tion is mostly influenced by a diffuse sum of man-made environmental changes such as a lowering ofthe regional ground water level (Malmer & Wallén1999) as well as acidification and nitrogen inputdue to atmospherical pollution (Bobbink et al. 1998;Lee 1998). In contrast, South Patagonian peat bogsprovide the unique opportunity to study diversitypatterns under almost pristine conditions with nearlyzero input of atmospherical pollutants (Godoy et al.2001). Hidden patterns of underlying processes alongenvironmental gradients (Wilson 2007) may bebetter observed in those pristine regions. There is aneed for data with pristine mire ecosystems thatcould also provide important information for conser-vation and ecological restoration (Couwenberg &Joosten 2005).

The comparison of the three main ombrotro-phic bog types in Patagonia allowed us to test thefollowing hypothesis: (i) the three main bog typeswill differ in patterns of a-, b-, and g-diversity (allof these diversity measures are supposed to behighest in the ecotonal mixed type) (Lloyd et al.2000); (ii) according to the hump-backed model(Grime 1973), species-richness will increase withincreasing pH-value, base cations (Ca, Mg), nitro-gen and phosphorus supply and decreasing anoxicstress due to lower soil water levels; and (iii) differ-ent functional groups will show a different responseto the most important environmental gradients:mosses and dwarf shrubs will dominate under thepoorest nutrient conditions, whereas graminoidsand herbs will increase with improving nutrientsupply.

METHODS

Study area

The investigated mire complexes are situated in south-ernmost Chile with most sites located northwest ofPunta Arenas, the capital of the XIIth Region (Fig. 1).The whole study area belongs to the geological unit ofthe Andean Cordillera that can be divided from west toeast into three units (Pisano 1977): Coastal, Centraland Marginal Cordillera. Although the genesis is quitedifferent, the bedrock of the whole study area is moreor less uniformly acidic and base-poor and its weath-ering products result in similar substrate nutrientconditions.

The southern Andes create a steep climatic gradientthat is reflected most significantly by the precipitationregime (Endlicher & Santana 1988). Hyperoceanicwestern parts have up to 10 000 mm annual precipi-tation near the highest elevations of the mountainrange, whereas at the east side of the Andes precipita-tion decreases within a short distance to less than500 mm (Schneider et al. 2003). Although meanannual temperatures change scarcely along the gradi-ent (~6°C), the inter-annual amplitude increases sig-nificantly towards the continental parts (Pisano 1977).Consequently, regular frost events in winter as well ashigher summer temperatures could be measured in theeast, whereas the absence of frost periods and coolersummer temperatures characterize the hyperoceanicconditions in the west (Tuhkanen 1992).This climaticgradient is well expressed by the variation inombrotrophic peatland vegetation. With increasingdistance from the Pacific Ocean, hyperoceanic blanketbogs built up by cushion-forming vascular plants suchas Donatia fascicularis and Astelia pumila are graduallyreplaced by Sphagnum magellanicum that exist exclu-sively under more continental climatic conditions(Kleinebecker et al. 2007). Both types interminglewithin a transition zone.

Sampling and biogeochemical analysis

In a first subset, vegetation of six bog complexes wassampled with two sites representing one of the threemajor mire types, respectively (Fig. 1): cushion bog,Sphagnum-cushion mixed bog and Sphagnum bog. Ourwesterly study areas represented cushion bog com-plexes and were situated at the entrance to the LagoMuños Gamero (MG, 52°42′S, 73°14′W) and at thewest bank of the Bahía Bahamondes (BB, 52°47′S,72°57′W). The two study areas representing the tran-sitional Sphagnum-cushion mixed type were locatedwest of the Río Azogar (RA, 52°32′S, 72°21′W) andwest of the Río Pinto at the north shore of the Bahía

2 T. KLEINEBECKER ET AL.

© 2009 The Authorsdoi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

Williams (BW, 52°31′S, 72°07′W). The easternSphagnum-dominated peatlands were situated nearEstancia Kerber in the Río Rubens valley (RR,52°04′S, 72°02′W) and north of the Estancia Skyring,east of the Río Perez (RP, 52°28′S, 71°54′W).

A total of 263 relevés were made and distributedalmost equally over the six study areas.Vegetation wassampled at 1 m2, a plot size that is routinely used inbogs (e.g. Dierssen 1982). Cover-abundance data wererecorded following the Braun-Blanquet approach(Westhoff & van der Maarel 1973) and translated ina medial percentage cover scale. According to theirlife form and their taxonomic affiliation (Chapin et al.1996), all recorded species were categorized into thefollowing functional groups: cushion plants (CP),other herbs (H), other graminoids (G), mosses (M),liverworts (LW), lichens (Li), dwarf-shrubs (DS) andother woody plants (WP). The nomenclature of plantspecies mentioned in the text refers to the followingsources: vascular plants (Moore 1983) and mosses (He1998).

The second subset contains 44 samples of the firstsubset and an additional 38 samples of other peatlandcomplexes (Fig. 1). For these 82 samples, additionallyto the vegetation, volumetric surface samples of peatwere taken in a depth of 5–10 cm. Samples were dried

at 50°C and screened through a sieve with a 2 mm meshwire.The following chemical analyses were carried outin the laboratory: pH (CaCl2); total contents of C andN: Element autoanalyzer (Carlo Erber NA 1500); ash-content: ashing overnight in a muffle furnace at 550° C;CAL-soluble P (Schüller 1969): photometrical mea-surement (Perkin Elmer) with 578 nm; extraction ofexchangeable fractions of K, Na, Ca, Mg with NH4Cl(Meiwes et al. 1984): determination with an AAS(Perkin Elmer). As additional ecological information,the edaphic moisture regime was described using anordinal scale reaching from 0 = ground water at soilsurface to 7 = groundwater > 40 cm below soil surfaceby digging a hole. It has to be considered that the watertable is variable and can show large seasonal differences(Wheeler & Proctor 2000). However, in previousstudies this method provided sufficiently precise resultsin the studied bog systems (Kleinebecker et al. 2007,2008).

Data analysis

a-diversity was measured as the average number ofspecies per sample for each of the six geographicregions. The Shannon-Wiener diversity index and the

Fig. 1. Locations of studied peatland complexes. BB, Bahía Bahamondes; BW, BahíaWilliams; MG, Muñoz Gamero; RA, RíoAzogar; RP, Río Perez; RR, Río Rubens.

DIVERSITY IN SOUTH PATAGONIAN BOGS 3

© 2009 The Authors doi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

evenness were calculated to incorporate the relativeabundance of the different species. b-diversity as ameasure of among site variability was expressed by theBray Curtis dissimilarity index. g-diversity was esti-mated as total species richness calculated over allsamples for each of the six study areas. g-diversity wasalso calculated for the different functional groups.

A general regression model (GRM) with backwardelimination was used to find the best predictors ofspecies richness among the measured environmentalvariables. When variables were strongly inter-correlated, the variable with the best correlation withspecies richness was chosen. Functional groupresponse curves to the best predicting environmentalvariables based on a cover matrix were created bygeneralized additive models (GAMs; Hastie & Tibshi-rani 1990) using Poisson distribution. The fittedmodel was compared with the null model, and onlyfunctional groups with significant response wereconsidered. Smooth term complexity was selectedusing the Akaike information criterion (AIC; Lepš &Šmilauer 2003).

Not specified taxa (mostly liverworts) were notincluded in the calculation of diversity patterns, butwere included with their cover values in the GAM. Forall statistical calculations, variables (except pH) were

log-transformed. STATISTICA 7.1 software (http://www.statsoft.com) was used for common correlationanalysis and calculating the GRM.The CANOCO 4.5software package (ter Braak & Šmilauer 2002) wasused for GAM modelling.

RESULTS

a-, b- and g-diversity

Mean species richness per sample-plot (a-diversity)was lowest in pure Sphagnum bog complexes (~10species per plot) and significantly higher in theSphagnum-cushion mixed bogs (~15 species per plot)and in the pure cushion bogs (~21 species per plot).b-diversity was lowest in continental Sphagnum bogs.The hyperoceanic cushion bogs showed slightly highervalues and highest mean b-diversity was calculated forSphagnum-cushion mixed bogs (Table 1).

A total of 118 species were recorded within the sixstudy sites. g-diversity was lowest for continental Sph-agnum bogs (51 species) and conspicuously higher forSphagnum-cushion mixed bogs and pure cushion bogs(70–75). In terms of species richness, liverworts

Table 1. a-, b- and g diversity measurements

Study site

SB SCB CB

AllRR RP BW RA BB MG

n 42 40 45 46 46 44 263Mean species richness (species*m-2) 9.4A 10.5A 14.5B 16.8B 21.4C 21.4C –Minimum (species*m-2) 2 4 5 4 13 10 –Maximum (species*m-2) 20 20 21 27 33 32 –Standard deviation 4.6 4.8 3.2 5.6 4.6 4.5 –Bray–Curtis dissimilarity (D) 0.57a 0.51b 0.71c 0.75d 0.69e 0.61a –Minimum d 0.002 0.047 0.088 0.071 0.125 0.055 –Maximum d 1.000 1.000 1.000 1.000 0.996 0.995 –Standard deviation 0.35 0.31 0.22 0.21 0.20 0.20 –Total

Species per site 51 51 70 74 70 75 118Vascular plants 23 24 35 40 41 53 64cryptogams 28 27 35 34 29 22 54CP 0 1 5 5 5 8 8DS 5 5 7 6 7 7 9G 6 6 7 9 8 10 14H 8 8 11 13 14 17 21Li 7 10 9 8 7 2 14LW 14 10 19 20 18 15 29M 7 7 7 6 4 5 11WP 4 4 5 7 7 11 12

Study sites: BB, Bahía Bahamondes; BW, Bahía Williams; CB, cushion bog; MG, Muñoz Gamero; RA, Río Azogar; RP, RíoPerez; RR, Río Rubens; SB, Sphagnum bog; SCB, Sphagnum-cushion mixed bog. Functional groups: CP, cushion plants; DS,dwarf-shrubs; G, other graminoids (excluding cushion plants); H, other herbs (excluding cushion plants); Li, lichens; LW,liverworts; M, mosses; WP, other woody plants (seedlings, saplings, shrubs). Different letters indicate significant differences(P < 0.05) after parametric Tukey test (A, B, C) and after non-parametric Mann–Whitney U-test (a, b, c, d, e; Bonferronicorrected).

4 T. KLEINEBECKER ET AL.

© 2009 The Authorsdoi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

represented the most important functional group,although species number was definitely underesti-mated due to difficulties in differentiation anddetermination. A remarkable incline in species rich-ness from eastern Sphagnum-dominated bogs towardscushion bogs in the west was recorded for cushionplants, graminoids and herbs, whereas the oppositewas true for lichens and mosses (Table 1).

Species richness along environmental gradients

Species richness in South Patagonian peat bogs wasclearly related to environmental variables. Stronglinear correlations were found for nitrogen, calcium,and the ash-content (r2 > 0.5) whereas the correlationwith the pH was weak. No significant linear relation-ship was found for the water level (Fig. 2). The GRMof the overall dataset including samples of the threemajor bog types resulted in an explained variance of76% (Table 2). The selected environmental variableswere total N (Beta = 0.545), NH4Cl soluble Ca(Beta = 0.365) and water level (Beta = 0.223). NH4Clsoluble Mg (Beta = 0.658) and water level (Beta =0.327) explained almost 80% of the variation ofspecies richness within pure Sphagnum bogs. InSphagnum-cushion mixed bogs total N (Beta = 0.725)and water level (Beta = 0.285) were chosen in theGRM and explained 54% of the variance, whereas inpure cushion bogs the only remaining variable was theash-content (Beta = 0.685) explaining 45% of the vari-ability of species richness (Table 2).

Response of functional groups to environmentalgradients

Response curves of functional groups along the waterlevel, total nitrogen and calcium gradients showed aclear differentiation to their ecological niches (Figs 3–5). In the overall dataset, mosses showed increasingcover values from relatively dry towards damp siteconditions (Fig. 3a). On the contrary, abundance ofdwarf-shrubs, liverworts and lichens increased towardsthe dry end. Cushion plants showed a peak at a waterlevel of 10–15 cm below the surface. Considering onlythe Sphagnum-cushion mixed bog samples, similar pat-terns could be found (Fig. 3c). In pure Sphagnumbogs, mosses dominated the vegetation from thewettest parts up to a water level of 30 cm belowthe surface (Fig. 3b). At the dry end of the gradientdwarf-shrubs, liverworts and lichens became moreimportant. In cushion bogs, cushion plants had a waterlevel optimum near the surface where liverworts hadminimum cover values (Fig. 3d). Contrary to the otherbog types, the cushion bogs response values of mossesincreased to the dry end of the moisture gradient.

Dwarf-shrubs and other woody plants also showed apositive response to drier site conditions (Fig. 3).

Along a gradient of total N contents in the peatsubstrate, within the overall dataset mosses reachedhigh cover values exclusively at the extremely nutrient-poor end of the gradient (Fig. 4a). Cushion plantsshowed an optimum at 2 g N*l-1. Cover of dwarf-shrubs and lichens decreased with increasing total N.The opposite could be measured for graminoids, liv-erworts, other woody plants and herbs.Within the pureSphagnum bogs, most functional groups did not showany significant response to the generally extremely lowN content (Fig. 4b). In the Sphagnum-cushion mixedbog samples, a clear separation of the mosses at theN-poor end of the gradient and the cushion plants onthe other side could be found (Fig. 4c). In the purecushion bogs, graminoids dominated on N-rich sites,whereas the cushion plants occupied the other end ofthe gradient (Fig. 4d).

Focussing on NH4Cl soluble Ca as a measure for basecation supply in the overall dataset, again mosses andcushion plants clearly differed in response (Fig. 5a). Atthe base-rich end of the gradient cushion plants werethe dominating functional group, whereas mossestogether with dwarf-shrubs occupied the poor end.Again, very similar patterns with mosses and cushionplants dominating at the corresponding ends of thegradient could be found in the Sphagnum-cushionmixed bog samples (Fig. 5c). Within pure Sphagnumbogs, only dwarf-shrubs dominating at comparativelyhigh Ca contents and liverworts had a significantresponse curve (Fig. 5b), whereas in the pure cushionbog samples no significant response of any functionalgroup could be found.

DISCUSSION

Ombrotrophic bogs are species-poor ecosystems dueto a rather small within habitat variability and theharsh environmental conditions, which allow only afew adapted species to grow. We found 118 species,which is even lower than species numbers recorded forBritish (Wheeler 1993) or Canadian ombrotrophicbogs (Warner & Asada 2006). Species richness of vas-cular plants seems to be only slightly lower (Glaser1992), whereas South Patagonian bogs showed a sig-nificantly lower diversity of moss species, which is bestexpressed by the occurrence of only two Sphagnumspecies. In contrast, various Sphagnum species arefrequent constituents of north hemispherical bogs(e.g. Gunnarsson et al. 2000; Locky & Bayley 2006).Although our investigation could not provide a com-plete list of all occurring species, liverworts were themost important group within the cryptogams. Even inthe well studied north hemispherical peatlands, hepat-ics are scarcely included in ecological investigations.

DIVERSITY IN SOUTH PATAGONIAN BOGS 5

© 2009 The Authors doi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

The few publications available (e.g. Såstad & Moen1995; Wells 1996; Gunnarsson et al. 2000, 2002) rec-ognized few (mostly less than 10) liverwort species.Thus, we suppose South Patagonian bogs, especiallythose of more oceanic regions, to be characterized by ahigher hepatic diversity than their north hemisphericalcounterparts. Lichens mostly characterize relatively

dry communities in bog complexes (e.g. Dierssen1982; Wells 1996). This pattern is similar in SouthPatagonian bogs (Kleinebecker et al. 2007); above all,most species have a bipolar distribution. Comparedwith phytosociological studies on the northern hemi-sphere (e.g. Wells 1996), lichen species richness ofSouth Patagonian bogs appear to be similar.

–10 0 10 20 30 40 50

Water level below surface [cm]

0

5

10

15

20

25

30

35S

peci

es p

er p

lot (

1 m

2 )

0 1 2 3 4 5

Total nitrogen [g/L]

0

5

10

15

20

25

30

35

Spe

cies

per

plo

t (1

m2 )

0 10 20 30 40

NH4 soluble Ca [cmolc /L]

0

5

10

15

20

25

30

35

Spe

cies

per

plo

t (1

m2 )

0 10 20 30

NH4 soluble Mg [cmolc/L]

0

5

10

15

20

25

30

35

Spe

cies

per

plo

t (1

m2 )

0 5 10 15 20 25

Ash-content [g/L]

0

5

10

15

20

25

30

35

Spe

cies

per

plo

t (1

m2 )

2.5 3.0 3.5 4.0 4.5

pH (CaCl2)

0

5

10

15

20

25

30

35

Spe

cies

per

plo

t (1

m2 )

R2: 0.630, P < 0.001R2: 0.004, P = 0.611

R2: 0.480, P < 0.001R2: 0.619, P < 0.001

R2: 0.227, P < 0.001R2: 0.631, P < 0.001

Fig. 2. Simple linear regression of species richness and selected environmental variables. Statistics were calculated withlog-transformed data.

6 T. KLEINEBECKER ET AL.

© 2009 The Authorsdoi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

Table 2. General regression model (GRM) results of an overall dataset (n = 82), samples of pure Sphagnum bogs (n = 18),samples of Sphagnum-Cushion mixed bogs (n = 34) and Cushion bogs (n = 30)

Corr r2 DF F P

Overall 0.761 78 86.949 0.000Remaining variables after backwards elimination Beta t P

WL 0.223 3.967 0.000N 0.545 6.701 0.000Ca 0.365 4.472 0.000

Sphagnum bog 0.797 15 29.501 0.000Remaining variables after backwards elimination Beta t P

WL 0.327 2.246 0.040Mg 0.658 4.521 0.000

Sphagnum-Cushion mixed bog 0.539 31 20.279 0.000Remaining variables after backwards elimination Beta t P

WL 0.285 2.398 0.023N 0.725 6.103 0.000

Cushion bog 0.450 28 24.756 0.000Remaining variable after backwards elimination Beta t P

AC 0.685 4.976 0.000

AC, ash-content; Ca, Mg, NH4Cl soluble calcium and manganese; N, total nitrogen; WL, water level.

001

0R

esp

onse

Li***Li***

01

00

Li***

05–5 –5

–5 –5

Water level (cm)

CP***DS***

Li***

LW**

M*

WP*

05Water level (cm)

001

0R

esp

onse

DS***

G*

Li***

LW*

M*

WP**

05Water level (cm)

001

0R

esp

onse

CP*** DS***

LW***

M*

05Water level (cm)

Res

pon

se

CP***

DS***

LW*

M*

WP**

(b)(a)

(d)(c)

Fig. 3. Generalized additive model (GAM) response curves to water level in (a) all peatland samples (n = 82); (b) pureSphagnum bog samples (n = 18); (c) Sphagnum-cushion mixed bog samples (n = 34); and (d) pure cushion bog samples(n = 30). Only functional groups with significant response (*P < 0.05, ** P < 0.01, *** P < 0.001) and not eliminated by theAkaike information criterion (AIC) were considered.The response is scaled as mean percentage cover per plot (1 m2). Functionalgroups: CP, cushion plants; DS, dwarf-shrubs; G, other graminoids (excluding cushion plants); H, other herbs (excludingcushion plants); Li, lichens; LW, liverworts; M, mosses; WP, other woody plants (seedlings, saplings, shrubs).

DIVERSITY IN SOUTH PATAGONIAN BOGS 7

© 2009 The Authors doi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

Our results showed a significant decrease ofa-diversity from hyperoceanic cushion bogs viaSphagnum-cushion mixed bogs to more continentalpure Sphagnum bog complexes. Glaser (1992) foundsimilar patterns for eastern Canadian ombrotrophicpeatlands. He mainly traced this fact back to climaticthresholds along a gradient of continentality that leadsto a regional differentiation of bog types and diversitypatterns. Total species richness of our studied bogcomplexes as a measure of g-diversity was low forSphagnum-dominated bogs and distinctly higherfor Sphagnum-cushion mixed and pure cushionbog complexes. The relatively high g-diversity ofSphagnum-cushion mixed bogs might be caused by thecharacter as an ecotone (Walker et al. 2003), wheretypical species of Sphagnum bogs as well as typicalspecies of western cushion bogs intermingle(Kleinebecker et al. 2007). This assumption is con-firmed by highest b-diversity values, which indicate amosaic of communities of continental Sphagnum bogsas well as of hyperoceanic cushion bogs and concurs

with general properties of species richness in ecotonespresented by Walker et al. (2003).

Species richness along various gradients often showsdifferent patterns in taxonomic and functional groups(Pausas & Austin 2001). Most of the differentiatedfunctional groups of our study presented a continuousincrease in species number from eastern Sphagnumbogs towards hyperoceanic cushion bogs. This is bestexpressed by cushion-forming vascular plants beingalmost absent in continental bogs and various speciesbeing present exclusively in oceanic bog complexes(Pisano 1983; Boelcke et al. 1985; Kleinebecker et al.2007). Similar patterns were described by Warner andAsada (2006), who found higher species richness formost functional groups in more oceanic peatland com-plexes in Canada. In contrast, mosses and lichensshowed opposite patterns with highest species richnessin continental bogs.

Water level and peat chemical characteristics areamong the most important determinants of speciesdistribution patterns in mires, as indicated by numer-

50

001

0R

esp

onse CP***

DS*

G***

H***Li**

LW**

M***

WP***

Nt [g*L-1] Nt [g*L-1]

Nt [g*L-1] Nt [g*L-1]

H***

LW**

5.10

001

0R

espo

nse

CP***

H**

M***

5.10

CP***

H**

50

00

10

Re

spon

se

CP*G**

H*

6.00

00

10

Res

pon

se

LW**

(b)(a)

(d)(c)

Fig. 4. Generalized additive model (GAM) response curves to total nitrogen content (Nt, g*l–1) in (a) all peatland samples(n = 82); (b) pure Sphagnum bog samples (n = 18); (c) Sphagnum-cushion mixed bog samples (n = 34); and (d) pure cushion bogsamples (n = 30). Only functional groups with significant response (*P < 0.05, ** P < 0.01, *** P < 0.001) and not eliminatedby the Akaike information criterion (AIC) were considered. The response is scaled in mean percentage cover per plot (1 m2).Functional groups: CP, cushion plants; DS, dwarf-shrubs; G, other graminoids (excluding cushion plants); H, other herbs(excluding cushion plants); Li, lichens; LW, liverworts; M, mosses; WP, other woody plants (seedlings, saplings, shrubs). Pleasenote the different scaling of the x-axes.

8 T. KLEINEBECKER ET AL.

© 2009 The Authorsdoi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

ous studies from the northern hemisphere (e.g. Vittet al. 1990; Malmer et al. 1992; Wheeler & Proctor2000). In South Patagonia, climatic conditions wereconsidered to be responsible for the strong floristicchange across the Andes, although a recent studyemphasized biogeochemical parameters also having aneffect on floristic composition (Kleinebecker et al.2008). The typical ‘hump-backed’ curve presented byGrime (1973) shows low species richness at lowresource levels, increases to an optimum at intermedi-ate levels and declines at high resource levels.We foundlinear relationships between species richness and mostenvironmental gradients resulting in a GRM withthree major environmental variables (total N, NH4Clsoluble Ca and water level) predicting species richnesswith an explained variance of 76 %. CAL soluble Pshowed only a weak correlation with species richness(data not shown) and was not among the major envi-ronmental variables affecting plant diversity. Thus, wesuppose that in contrast to mire ecosystems thatreceive high atmospherical nitrogen input (Venterink

et al. 2003; Wassen et al. 2005), P availability has aminor effect on species competition and diversity pat-terns in pristine South Patagonian bogs.

Although there was no significant linear relationshipbetween species richness and the water level, the mois-ture gradient seems to explain an important part of theremaining variance not explained by biogeochemicalvariables. Ca supply as a measure of base cation supplyand N contents as a measure of peat decompositionand nutrient supply showed a highly significant linearrelationship to species richness. A linear increase ofrichness across defined trophic gradients was alsofound by Vermeer and Berendse (1983) and theyhypothesized that at low productivity levels only theinitial rise of the optimum curve becomes evident.Thisis confirmed by our findings, due to the generally lowresource levels in the studied ombrotrophic peatlands.

Within the three major peatland types we calculateddifferent GRMs, but all resulted in environmental vari-ables belonging more or less to the three major gradi-ents calculated for the overall dataset: water level, base

001

0

040 Ca [cmolc/L] Ca [cmolc/L]

Ca [cmolc/L]

Res

pon

se

CP***

DS*

G**

H***

LW***

M***

WP***

5.70

001

0R

esp

onse

DS*

LW*

210

001

0R

esp

onse

CP***

H**

M***

(b)(a)

(c)

Fig. 5. Generalized additive model (GAM) response curves to NH4Cl soluble calcium (cmolc*l–1) in (a) all peatland samples(n = 82); (b) pure Sphagnum bog samples (n = 18); and (c) Sphagnum-cushion mixed bog samples (n = 34). Only functionalgroups with significant response (*P < 0.05, ** P < 0.01, *** P < 0.001) and not eliminated by the Akaike information criterion(AIC) were considered. In pure cushion bogs no significant response of any functional group could be measured.The responseis scaled in mean percentage cover per plot (1 m2). Functional groups: CP, cushion plants; DS, dwarf-shrubs; G, othergraminoids (excluding cushion plants); H, other herbs (excluding cushion plants); Li, lichens; LW, liverworts; M, mosses; WP,other woody plants (seedlings, saplings, shrubs). Please note the different scaling of the x-axes.

DIVERSITY IN SOUTH PATAGONIAN BOGS 9

© 2009 The Authors doi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

cation availability and N contents. In continentalSphagnum bogs, water level and NH4Cl soluble Mgexplained 80% of the variability in species richness ofour samples. In this case, base cation availability wasdocumented by Mg as the best predicting variable.The water level represents the hummock-hollowgradient. N contents were generally extremely low(Kleinebecker et al. 2008) and had no significantimpact on the variation of species richness within thecontinental Sphagnum bogs.Within Sphagnum cushionmixed bogs, the N content was a well predicting vari-able reflecting the mosaic of more diverse and N-richcushion plant-dominated patches and N-poor Sphag-num dominated communities with lower species-richness. The ash-content as the only predictor ofspecies-richness in the pure cushion bog samples canbe seen as an integrating environmental variable thatreflects the effects of higher minerotrophy and betterdrainage in sloped positions within the glacially-eroded landscape of the South Chilean Channels.Generally, drainage conditions are considered to be amajor environmental factor affecting species composi-tion in hyperoceanic cushion bogs (Moore 1979).

A reduction in species number along a gradient ofsoil acidity, as observed from Glaser et al. (1990) andGunnarsson et al. (2000, 2002), could also be con-firmed by our results, although the correlation wasweak. At pH values below 4.5 the correlation betweenCa concentrations and pH is low to insignificant (Sjörs& Gunnarsson 2002). In accordance with Teneb andDollenz (2004), our samples showed little variation insoil acidity (pH mostly within a range of 3 to 4) andconsequently pH had no predominant effect onspecies composition and richness patterns.

Besides species richness patterns our results clearlyestablished divergent responses of different functionalgroups along environmental gradients. Responsecurves along a water level gradient of the overalldataset showed a clear separation of mosses andcushion plants on the wet end of the gradient andincreasing importance of dwarf-shrubs, liverwortsand lichens to the dry end. Within pure Sphagnumbogs and Sphagnum-cushion mixed bogs, theresponses of the functional groups, were roughly thesame. This concurs with phytosociological descrip-tions of north hemispherical Sphagnum-dominatedbogs (e.g. Dierssen 1982; Wells 1996), where dwarf-shrubs and lichens benefit from drier site conditionsat the cost of bryophytes. Among bryophytes, Sphag-num species play the most important role in mirevegetation and show a clear niche differentiationalong a moisture gradient (Hájková & Hájek 2004).In South Patagonian bogs there are only two Sphag-num species (S. magellanicum and S. cuspidatum) inombrotrophic mires, but they also show a clear sepa-ration along the hummock-hollow moisture gradient(Kleinebecker et al. 2007). In hyperoceanic cushion

bogs, mosses showed the opposite behaviour withincreasing response values towards drier site condi-tions. This is obviously caused by the absence of S.magellanicum in the west (Pisano 1983; Boelcke et al.1985; Kleinebecker et al. 2007) and a stronger occur-rence of moss species such as Dicranoloma robustumand Racomitrium lanuginosum at edaphically driermicro-sites.

Our results suggest that there is a significantresponse for both, the nutritional status and the basecation supply for bryophytes as well as for vascularplants. Mosses dominated at the low ends of Ca and Ngradients, whereas vascular plant functional groups,except dwarf-shrubs, showed high response values atintermediate or high nutrient contents. This concurswith Kleinebecker et al. (2008) who found higher basecation contents in hyperoceanic cushion bogs, wherecushion plants and graminoids are the most importantconstituents. This trend also became obvious withinSphagnum-cushion mixed bogs, where the influence ofsea spray (Kleinebecker et al. 2008) that is consideredto be a possible cause of different nutritional condi-tions should not change significantly at short distanceswithin a single bog complex. We suppose that betterrecycling properties of cushion plants and graminoids,which lead to an enrichment of nutrients affect thebiogeochemical peat characteristics and consequentlythe floristic composition.

Ombrotrophic peatlands receive all of their nutrientsfrom the atmosphere and are particularly sensitive toair-borne N input. In Europe, N enrichment via air-borne pollutants has increased nutrient availability,leading to competitive exclusion by nitrophiliousspecies at the cost of low competitive stress-toleratingplants (Bobbink et al. 1998). In contrast, the input inSouth Patagonian ecosystems is extremely low (Godoyet al. 2001). Tomassen et al. (2003) reported stimu-lated growth of woody species and graminoids dueto atmospherical nitrogen input. We found a similarpattern under pristine conditions along a natural gra-dient of nutrient availability. This also concurs withRuthsatz and Villagran (1991) who described a higherdominance of graminoids under better trophical siteconditions in a Chilean cushion bog complex.

Taken together, our findings underpin the scaledependency (gradient length) and partly non-linearrelationship between environmental variables anddiversity patterns in the studied bog ecosystems.Species richness showed a linear relationship to trophi-cal peat properties reflecting the initial rise of a poten-tial hump-backed curve (Vermeer & Berendse 1983),whereas functional group response was partlyunimodal. Our results also show that South Patago-nian peatlands provide an excellent opportunity tostudy ecological relationships without overlappinganthropogenic influences. Such data are of particularvalue not only for the testing of ecological theory, but

10 T. KLEINEBECKER ET AL.

© 2009 The Authorsdoi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

also as a reference for conservation and restoration inlandscapes highly affected by human interferences(Couwenberg & Joosten 2005).

ACKNOWLEDGEMENTS

We would like to thank Rolf Kilian (University of Trier,Germany), chief of the Gran Campo Nevado Project,and Christoph Schneider (University of Aachen,Germany) for logistic support. Orlando Dollenz (Insti-tuto de la Patagonia, Punta Arenas, Chile) gave helpfulhints to find good study sites and networked withpersons knowing the study area. Verena Möllenbeck(University of Münster) provided helpful comments ona former version of this manuscript and Lillian Harrisassisted with English language editing.

REFERENCES

Bobbink R., Hornung M. & Roelofs J. G. M. (1998) The effectsof air-borne nitrogen pollutants on species diversity innatural and semi-natural European vegetation. J. Ecol. 86,717–38.

Boelcke O., Moore D. M. & Roig F. A., eds (1985) TransectaBotánica de la Patagonia Austral. CONICET Argentina, Inst.Patagonia Chile. Royal Society Gran Bretaña, Buenos Aires.

ter Braak C. F. J. & Šmilauer P. (2002) CANOCO ReferenceManual and User’s Guide to Canoco forWindows: Software forCanonical Community Ordination (Version 4.5). Microcom-puter Power, Ithaca, NY.

Chapin F. S., Bret-Harte M. S., Hobbie S. E. & Zhong H. (1996)Plant functional types as predictors of transient responses ofarctic vegetation to global change. J.Veg. Sci. 7, 347–58.

Couwenberg J. & Joosten H. (2005) Self-organization in raisedbog patterning: the origin of microtope zonation and meso-tope diversity. J. Ecol. 93, 1238–48.

Dierssen K. (1982) Die Wichtigsten Pflanzengesellschaften DerMoore NW-Europas. Editions des Conservatoire et JardinBotaniques de la Ville de Genève, Genève.

Endlicher W. & Santana A. (1988) El clima del sur de la Pat-agonia y sus aspectos ecológicos. Un siglo de medicionesclimatológicas en Punta Arenas. An. Inst. Patagonia 18,57–86.

Glaser P. H. (1992) Raised bogs in eastern North-America –regional controls for species richness and floristicassemblages. J. Ecol. 80, 535–54.

Glaser P. H., Janssens J. A. & Siegel D. I. (1990) The response ofvegetation to chemical and hydrological gradients in the LostRiver Peatland, Northern Minnesota. J. Ecol. 78, 1021–48.

Godoy R., Oyarzún C. & Gerding V. (2001) Precipitation chem-istry in deciduous and evergreen Nothofagus forests ofsouthern Chile under a low-deposition climate. Basic Appl.Ecol. 2, 65–72.

Grime J. P. (1973) Competitive exclusion in herbaceousvegetation. Nature 242, 344–7.

Gunnarsson U., Malmer N. & Rydin H. (2002) Dynamics orconstancy in Sphagnum dominated mire ecosystems? A40-year study. Ecography 25, 685–704.

Gunnarsson U., Rydin H. & Sjörs H. (2000) Diversity and pHchanges after 50 years on the boreal mire SkattlosbergsStormosse, Central Sweden. J.Veg. Sci. 11, 277–86.

Hastie T. J. & Tibshirani R. J. (1990) Generalized Additive Models.Chapman and Hall, London.

He S. (1998) A checklist of the mosses of Chile. J. Hattori Bot.Lab. 85, 23–109.

Huston M. A. (1994) Biological Diversity. The Coexistence ofSpecies on Changing Landscapes. Cambridge UniversityPress, Cambridge.

Hájek M., Hekera P. & Hájková P. (2002) Spring fen vegetationand water chemistry in the western Carpathian flysch zone.Folia Geobot. 37, 364–

Hájková P. & Hájek M. (2004) Bryophyte and vascular plantresponses to base-richness and water level gradients inwestern Carpathian Sphagnum-rich mires. Folia Geobot. 39,335–51.

Kleinebecker T., Hölzel N. & Vogel A. (2007) Gradients ofcontinentality and moisture in South Patagonian peatlandvegetation. Folia Geobot. 42, 363–82.

Kleinebecker T., Hölzel N. & Vogel A. (2008) South Patagonianombrotrophic bog vegetation reflects biogeochemical gradi-ents at the landscape level. J.Veg. Sci. 19, 151–60.

Krisai R. (2005) Mires of Tierra del Fuego. Stapfia 85, 373–7.Lee J. A. (1998) Unintentional experiments with terrestrial eco-

systems: ecological effects of sulphur and nitrogenpollutants. J. Ecol. 86, 1–12.

Lepš J. & Šmilauer P. (2003) Multivariate Analysis of EcologicalData Using CANOCO. Cambridge University Press,Cambridge.

Lloyd K. M., McQueen A. A. M., Lee B. J., Wilson R. C. B.,Walker S. & Wilson J. B. (2000) Evidence on ecotone con-cepts from switch, environmental and anthropogenicecotones. J.Veg. Sci. 11, 903–10.

Locky D. A. & Bayley S. E. (2006) Plant diversity, composition,and rarity in the southern boreal peatlands of Manitoba,Canada. Can. J. Bot. 84, 940–55.

Malmer N., Horton D. G. & Vitt D. H. (1992) Element concen-trations in mosses and surface waters of western Canadianmires relative to precipitation chemistry and hydrology.Ecography 15, 114–28.

Malmer N. & Wallén B. (1999) The dynamics of peat accumu-lation on bogs: mass balance of hummocks and hollows andits variation throughout a millennium. Ecography 22, 736–50.

Meiwes K. J., König N., Khana P. K., Prenzel G. & Ulrich B.(1984) Chemische Untersuchungsverfahren für Mineral-boden, Auflagehumus und Wurzeln zur Charakterisierungund Bewertung der Versauerung von Waldböden. Berichtedes Forschungszentrums Waldökosysteme/Waldsterben (Göttin-gen) 7, 1–67.

Moore D. M. (1983) Flora of Tierra del Fuego. Anthony Nelson,London.

Moore D. M. (1979) Southern oceanic wet-heathlands (includ-ing magellanic moorland. In: Heathlands and Related Shru-blands (ed. R. L. Specht) pp. 489–97. Ecosystems ofthe World 9A, Elsevier Scientific Publishing Company,Amsterdam.

Pausas J. G. & Austin M. P. (2001) Patterns of plant speciesrichness in relation to different environments: an appraisal.J.Veg. Sci. 12, 153–66.

Pisano E. (1977) Fitogeografía de Fuego-Patagonia Chilena –comunidades vegetales entre las latitudes 52 y 56° S. An.Inst. Patagonia 8, 121–250.

Pisano E. (1983) The magellanic tundra complex. In: Mires:Swamp, Bog, Fen and moor, B. Regional Studies (ed. A. J. P.Gore) pp. 295–329. Ecosystems of the world 4, ElsevierScientific Publishing Company, Amsterdam.

DIVERSITY IN SOUTH PATAGONIAN BOGS 11

© 2009 The Authors doi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia

Pärtel M. & Zobel M. (2007) Dispersal limitation may result inthe unimodal productivity-diversity relationship: a newexplanation for a general pattern. J. Ecol. 95, 90–4.

Ruthsatz B. & Villagran C. (1991) Vegetation pattern and soilnutrients of a Magellanic moorland on the Cordillera dePiuchué, Chiloé Island, Chile. Rev. Chil. Hist. Nat. 64, 461–78.

Rydin H. & Jeglum J. (2006) The Biology of Peatlands. OxfordUniversity Press, New York.

Schneider C., Glaser M., Kilian R., Santana A., Butorovic N. &Casassa G. (2003) Weather observations across the south-ern Andes at 53°S. Phys. Geogr. 24, 97–119.

Schüller H. (1969) Die CAL-Methode, eine neue Methode zurBestimmung des pflanzenverfügbaren Phosphates imBoden. Z. Pflanz. Bodenkunde 123, 48–63.

Sjörs H. & Gunnarsson U. (2002) Calcium and pH in north andcentral Swedish mire waters. J. Ecol. 90, 650–7.

Såstad S. & Moen A. (1995) Classification of mire localities andmire species in central Norway by vegetational regions,Ellenberg species indicator values and climatic data. Gun-neria 70, 177–98.

Teneb E. & Dollenz O. (2004) Distribución especial de la Floravascular, la humedad y el pH en un turbal de esfagno(Sphagnum magellanicum Brid.), Magallanes, Chile. An. Inst.Patagonia 32, 5–12.

Tomassen H. B. M., Smolders A. J. P., Lamers L. P. M. & RoelofsJ. G. M. (2003) Stimulated growth of Betula bubescens andMolinia caerulea on ombrotrophic bogs: role of high levels ofatmospheric nitrogen deposition. J. Ecol. 91, 357–70.

Tuhkanen S. (1992) The climate of Tierra del Fuego from avegetation geographical point of view and its ecoclimaticcounterparts elsewhere. Acta Bot. Fenn. 145, 1–64.

Venterink H. O., Wassen M. J., Verkroost A.W. M. & de Ruiter P.C. (2003) Species richness-productivity patterns differbetween N-, P-, and K-limited wetlands. Ecology 84,2191–9.

Vermeer J. G. & Berendse F. (1983) The relationship betweennutrient availability, shoot biomass and species richness ingrassland and wetland communities. Vegetatio 53, 121–6.

Vitt D. H., Horton D. G., Slack N. G. & Malmer N. (1990)Sphagnum-dominated peatlands of the hyperoceanic BritishColumbia coast: patterns in surface water chemistry andvegetation. Can. J. Forest Res. 20, 696–711.

Walker S., Wilson J. B., Steel J. B. et al. (2003) Properties ofecotones: evidence from five ecotones objectively deter-mined from a coastal vegetation gradient. J. Veg. Sci. 14,579–90.

Warner B. G. & Asada T. (2006) Biological diversity of peatlandsin Canada. Aquat. Sci. 68, 240–53.

Wassen M. J., Venterink H. O., Lapshina E. D. & Tanneberger F.(2005) Endangered plants persist under phosphoruslimitation. Nature 437, 547–50.

Wells E. D. (1996) Classification of peatland vegetation in Atlan-tic Canada. J.Veg. Sci. 7, 847–78.

Weltzin J. F., Pastor J., Harth C., Bridgham S. D., Updegraff K.& Chapin C. T. (2000) Response of bog and fen plantcommunities to warming and water-table manipulations.Ecology 81, 3464–78.

Westhoff V. & van der Maarel E. (1973) The Braun-Blanquetapproach. In: Ordination and Classification of Communities(ed. R. H. Whittaker) pp. 617–737. Junk, Den Haag.

Wheeler B. D. (1993) Botanical diversity in British mires.Biodivers. Conserv. 2, 490–512.

Wheeler B. D. & Proctor M. C. F. (2000) Ecological gradients,subdivisions and terminology of north-west Europeanmires. J. Ecol. 88, 187–203.

Whitehouse H. E. & Bayley S. E. (2005) Vegetation patterns andbiodiversity of peatland plant communities surroundingmid-boreal wetland ponds in Alberta, Canada. Can. J. Bot.83, 621–37.

Wilson M.V. (2007) Measuring the components of competitionalong productivity gradients. J. Ecol. 95, 301–8.

12 T. KLEINEBECKER ET AL.

© 2009 The Authorsdoi:10.1111/j.1442-9993.2009.02003.xJournal compilation © 2010 Ecological Society of Australia