Submitted 1 June 2016Accepted 16 June 2017Published 31 July 2017

Corresponding authorTommaso Sitziatommasositziaunipdit

Academic editorNigel Yoccoz

Additional Information andDeclarations can be found onpage 17

DOI 107717peerj3552

Copyright2017 Sitzia et al

Distributed underCreative Commons CC-BY 40

OPEN ACCESS

Landscape metrics as functional traitsin plants perspectives from a glacierforelandTommaso Sitzia12 Matteo Dainese13 Bertil O Kruumlsi4 and Duncan McCollin2

1Department of Land Environment Agriculture and Forestry Universitagrave degli Studi di Padova Legnaro (PD)Italy

2 Landscape amp Biodiversity Research Group The University of Northampton Northampton United Kingdom3Department of Animal Ecology and Tropical Biology Universitaumlt Wuumlrzburg Wuumlrzburg Germany4 School of Life Sciences and Facility Management Zuumlrich University of Applied Science WaumldenswilSwitzerland

ABSTRACTSpatial patterns of vegetation arise from an interplay of functional traits environmentalcharacteristics and chance The retreat of glaciers offers exposed substrates which arecolonised by plants forming distinct patchy patterns The aim of this study was tounravel whether patch-level landscape metrics of plants can be treated as functionaltraits We sampled 46 plots each 1 m times 1 m distributed along a restricted range ofterrain age and topsoil texture on the foreland of the Nardis glacier located in theSouth-Eastern Alps Italy Nine quantitative functional traits were selected for 16 ofthe plant species present and seven landscape metrics were measured to describe thespatial arrangement of the plant speciesrsquo patches on the study plots at a resolutionof 1 cm times 1 cm We studied the relationships among plant communities landscapemetrics terrain age and topsoil texture RLQ-analysis was used to examine trait-spatial configuration relationships To assess the effect of terrain age and topsoiltexture variation on trait performance we applied a partial-RLQ analysis approachFinally we used the fourth-corner statistic to quantify and test relationships betweentraits landscape metrics and RLQ axes Floristically-defined releveacute clusters differedsignificantly with regard to several landscape metrics Diversity in patch types and sizeincreased and patch size decreased with increasing canopy height leaf size and weightMoreover more compact patch shapes were correlated with an increased capacity forthe conservation of nutrients in leaves Neither plant species composition nor any ofthe landscape metrics were found to differ amongst the three classes of terrain age ortopsoil texture We conclude that patch-level landscape metrics of plants can be treatedas species-specific functional traitsWe recommend that existing databases of functionaltraits should incorporate these type of data

Subjects Biogeography Ecology Ecosystem Science Plant ScienceKeywords Ecological process Plant composition Life-history trait Pioneer plant Landscapepattern Spatial self-organisation Spatial pattern Landscape heterogeneity Patch size Patch shape

INTRODUCTIONHeterogeneity of vegetation pattern has long been a subject of debate in ecology (egGreig-Smith 1979 Macfadyen 1950) One area of focus on vegetation patterns is toward

How to cite this article Sitzia et al (2017) Landscape metrics as functional traits in plants perspectives from a glacier foreland PeerJ5e3552 DOI 107717peerj3552

spatial self-organization regular pattern formation arising as an emergent property of localbiotic interactions in combination with large-scale physical agents especially in relativelysimple ecosystems such as deserts (Rietkerk amp VandeKoppel 2008 Sole amp Bascompte 2006)or glacier forelands (Koumlrner 2003) Such interactions between biotic and prevalent physicalagents such as directional water drainage in peatlands or tidal currents in shallow marinebeds can generate striking spatial patterns such as stripes or polygons (Jiang et al 2012Van de Koppel Bouma amp Herman 2012)

Spatial patterns are frequently associated with primary succession on glacier forelandswhere the barren soil is gradually colonised by plants mostly characterised by clonal growth(Koumlrner 2003) Small-scale interactions between organisms and physical processes canbreak the symmetry of bare glacial deposits to initiate pattern formation in terms of aconcentration or aggregation of individuals in clusters After glacier retreat vegetationand soil usually develop rapidly (Chapin et al 1994 Matthews 1999) with changes oftendriven more by allogenic than by autogenic mechanisms during the very early stages ofsuccession (Matthews amp Whittaker 1987)

Successional rates are non-linear with the fastest increases in total vegetation coverduring the first 15ndash20 years after deglaciation (Raffl et al 2006) and significant localvariability in species composition at the youngest sites Accordingly terrain age mightnot be the principal factor in explaining present-day variation in species compositionon glacier forelands (Rydgren et al 2014) Depending on the time range consideredas well as the sampling and analytical procedures employed (Rydgren et al 2014)speciesrsquo response may depend on factors other than time since deglaciation (Matthewsamp Whittaker 1987 Těšitel et al 2014) Amongst these it has been shown that simplemorphological and physiological attributes (ie life-history or functional traits) combinedwith suitable establishment and environmental conditions are the most important driversof colonization success (Erschbamer Niederfriniger Schalg amp Eckart 2008) Such attributesare often helpful for identifying successional stages (Caccianiga et al 2006 Erschbamer ampMayer 2011 Nascimbene et al 2017 Schwienbacher et al 2012) Indeed the integrationbetween the available trait databases and databases that combine the abundance of specieswith environmental information can help to identify the traits that respond influence orinteract with environmental factors and ecological processes (Suding et al 2008) a majorfield of functional ecology (McGill et al 2006)

Early stage succession should be mainly driven by neutral or trait-driven dispersalprocesses while competition filtering should show stronger responses in later stages In thecase of herbaceous plants competitive effects may be found over only very small distancesnot exceeding a few centimetres (Schleicher Peppler-Lisbach amp Kleyer 2011) The study ofthese competitive effects on glacier forelands may be helped by a high resolution landscapeecological approach where vegetation patterns are linked to plant trait characteristics withthe aim of revealing interactions between spatial patterns and ecological processes

Previous analyses of vegetation patterns during glacier foreland primary successionsuggest a sequence of wave-like replacements of groups of species largely in order ofincreasing size (Reiners Worley amp Lawrence 1971)Matthews amp Whittaker (1987) observedthat an early peak in mean diameter of Poa alpina L and Oxyria digyna (L) Hill clumps

Sitzia et al (2017) PeerJ DOI 107717peerj3552 222

during succession is followed by a rapid decline in development and then in size andcover values from around 50 y terrain age In contrast shrubs but also clumps of certainherbaceous species eg Saxifraga oppositifolia L have a tendency to increase in size andnumber of flowering individuals with increasing terrain age (Těšitel et al 2014)

Recently the combination of a three-table ordination (RLQ analysis) (Doleacutedec etal 1996) and the fourth-corner method the latter an approach to test the directcorrelation between a single trait and a single environmental variable (Legendre Galzinamp HarmelinVivien 1997) has been proposed to assess trait responses to environmentalvariation (Dray et al 2014) Here we apply these techniques along with cluster analysisand ANOVA to elucidate how plant species and life trait composition as well as the spatialconfiguration of plant species patches are related on recently deglaciated substrates

Plant growth development following glacier retreat should lead to spatial organization inthe form of discs rings or fragmenting clusters very slowly moving across the landscapelike in tussock graminoids or isolated ramets forming widely spread nets of a single genetor stoloniferous or rhizomatous forbs (Koumlrner 2003) Such patterns can be measured usingmetrics derived from landscape ecology eg numbers size shape type and the spatialarrangement of plant species patches (Forman amp Godron 1981)

Our main research question here concerns how landscape metrics correlate withmeasures relating to plant species composition and functional traits We restrict samplingto terrain ages of gt15 years which would not include the very early pioneer stageand lt70 years before the establishment of woody species Within these constraintswe hypothesise that differences in spatial configuration would combine with plant traitpatterning while terrain age and environmental variability intentionally reduced shouldhave minimal effects

Landscape filters on functional traits have received increasing attention in recent yearson both animals and plants (eg Deckers et al 2004 Duflot et al 2014 Gaacutemez-Virueacuteset al 2015) In general studies often agree that landscape-scale effects could dominatecommunity-level filtering on functional traits (Gaacutemez-Virueacutes et al 2015) Landscapemetrics are usually determined at large scale resolutions for example within a certaindistance from the target habitats eg hedgerows (Deckers et al 2004) or grasslands(Gaacutemez-Virueacutes et al 2015) or from land-cover maps in 1 km2 landscape replicates (Duflotet al 2014) Here we adapt this approach often used over the scales of several kilometresto a species-level approach over scales of centimetres to metres

We aim to provide insight into the role of plant traits in spatial organization on glacialforelands at high resolutions but also in general to test hypotheses regarding the linkagebetween plant trait distributions and species-level spatial organization For instancemetricsquantifying landscape complexity may be related to a higher frequency of morphologicaltraits related to competitive ability while metrics related to patch compactness may berelated to traits that indicate conservation of acquired resources The patch-level landscapemetrics of plant individuals could be treated as functional traits themselves

Sitzia et al (2017) PeerJ DOI 107717peerj3552 322

Figure 1 (A) Location of the study area (triangles) (B) image of the glacier foreland in 2011 and (C)a map of how the glacier has retreated from 1945 (external isochrones) to 2006 (date of the digital ter-rain model used as a base map) with the positions of the 46 1 mtimes 1 m-plots (black dots) Coordinatesare reported according to the geographic coordinate reference systems (A) WGS84 (EPSG 4326) and (C)ETRS 89UTM zone 32N (EPSG 25832)

MATERIALS AND METHODSStudy areaField work was performed on the Nardis glacier foreland (4612prime14

primeprime

N 10 40prime21primeprime

E)located in the Adamello-Presanella group (Rhaetian Alps southern sector of the ItalianCentral Alps) on the southern slope of Presanella Peak (3556 m asl) (Fig 1A) Theglacier has a surface area of approximately 167 km2 (SAT 2007) and its tongue extendsdown to an altitude of 2720 m asl (Fig 1B) The bedrock consists primarily of acidicgranitoid material The geology is characterized by the large Adamello-Presanella-MonteRe di Castello batholith (294ndash41 Ma) consisting of tonalite an igneous plutonic intrusiverock Available climatic data taken from a nearby weather station (4625prime33

primeprime

N 1041prime51primeprime

E)located at the same altitude indicate a mean summer temperature of 57 C and a meanannual precipitation of 897 mm The study area where the sampling took place wasapproximately 7 hectares in size and corresponded to the zone in front of the glaciertongue where the glacier was still present in 1945 (Fig 1C)

Sitzia et al (2017) PeerJ DOI 107717peerj3552 422

Figure 2 Schematic representation of the plant species survey To each 1 m2 sampling plot (A) a virtual1 cm-grid is superimposed to produce the final species-patch map (B)

Data collectionWe sampled 46 randomly distributed points in the study area Using the closest individualmethod (Krebs 1999) from each of these sampling points we selected the closest 1mtimes 1msample plot which was (i) safe from landslides and flat (lt5) and (ii) without boulders(d gt 256 mm) This selection procedure was done to avoid marked differences in siteconditions (Vetaas 1997) even if it were not possible to completely control for a certainamount of variability in topsoil texture Using historical maps and aerial photographsfrontlines of the glacier tongue ie lines of equal terrain age (isochrones) were establishedfor 1945 1954 1970 and 1996 Terrain ages (ta) of each sample plot were then classifiedas follows (ta1) between 15 and 41 years (n= 11) (ta2) between 41 and 57 years (n= 17)and (ta3) between 57 and 66 years (n= 18) (Fig 1C) Each plot was subdivided into 100001 cmtimes 1 cm-grid cells In August 2011 vascular plant species distribution intersecting thecentral axes of each 1 cm times 1 cm-cell were then mapped and digitised using ESRI ArcGIS93 yielding for each species the number of 1 cm2 cells occupied per 1 m2 (Fig 2) Thesame was done to classify the topsoil texture since amongst the environmental variablestopsoil texture was considered to be a major physical influence on patch development As ameasure of topsoil texture we used the cover of cobbles which according to theWentworthscale (Wentworth 1922) have a diameter gt64 mm a size (gt32 cm2) comparable to theobserved mean size of plant patches (47 cm2) Then we distinguished three classes ofcobble cover (ts1) lt7 (n= 15) (ts2) between 7 and 14 (n= 15) and (ts3) gt14(n= 16)

Data analysisLandscape metrics were calculated according to the procedure of Teixido et al (2007) Foreach 1 mtimes 1 m-plot we calculated seven metrics (Table 1) Calculations were made usingthe Patch Analyst 50 extension for ArcGIS 93 (Rempel Kaukinen amp Carr 2012) adoptinga four-neighbour rule to identify the patches A synthetic description of each patch metricis reported in Table 1 Two metrics patch size and shape can be computed both for everypatch in the landscape and as an average for all patch metrics at the plot level (see also

Sitzia et al (2017) PeerJ DOI 107717peerj3552 522

Table 1 Landscape metrics used to quantify plant patches patterns on a glacier foreland and plant species traits used to correlate them to thephysiological characteristics of plant species

Abbreviation Unit Variable name Description

Landscape metricsMPS cm2 Mean patch size Mean size of all patchesPSCV Patch size coefficient of variation Variability in patch size relative to mean patch size PSCV=

0 when all patches are the same size or when there is only 1patch

TE cm Total edge Total length of edge of all patch boundariesNP None Number of patches Total number of patchesMSI None Mean shape index Mean shape index of all patches MSI= 1 when a patch is

maximally compact (ie a square) and increases withoutlimit as patch shape becomes more irregular (Patton 1975)

PR None Patch type richness Number of different patch types (ie plant species)SHDI None Shannonrsquos diversity index minus1 times the sum across all patch types of the proportional

abundance of each patch type multiplied by thatproportion SHDI= 0 when PR= 1 and increases withoutlimit as PR increases andor the proportional distribution ofarea among patch types becomes more equitable

Functional traitsCH mm Canopy height Aspects of competitive abilityLDMC Leaf dry matter content Resistance to physical hazards long life-span relative

growth rateLS cat (1ndash6) Lateral spread Aspects of competitive abilityLDW mg Leaf dry weight Growth indexSLA mm2 mgminus1 Specific leaf area Internal resistance to CO2 movement nitrogen mass

fraction Rubisco specific activity relative growth rateLNC Leaf nitrogen content Assimilation capacityLA mm2 Leaf area Leaf energy and water balanceLFW mg Leaf fresh weight Growth indexLCC Leaf carbon content Photosynthetic rate

McGarigal amp Marks 1994) In the first case the metrics may be related to each species iespecies-specific traits

We did not sample plant traits in the field because this would have been too time-consuming and too destructive to the vegetation When available plant traits were takenfrom Cerabolini et al (2010) Data for Salix herbacea L and partially Sedum alpestre Villdata were taken from Carbognani (2011) and Pierce et al (2007) Leaf nitrogen and carboncontents of S alpestre were calculated as a mean of the values of Sedum acre L and Sedumalbum L from Cerabolini et al (2010) Saxifraga stellaris L Polygonum viviparum L andAthyrium distentifolium Tausch ex Opiz due to their negligible frequencies and coverwere omitted from further analyses A brief description of each selected trait is provided inTable 1 (see also Cornelissen et al 2003 Gross Suding amp Lavorel 2007 Wilson Thompsonamp Hodgson 1999) The plant traits database used here is reported in Table 2

The 46 vegetation releveacutes were subjected to agglomerative cluster analysis using theBrayndashCurtis coefficient of dissimilarity and Wardrsquos clustering method Next ANOVA

Sitzia et al (2017) PeerJ DOI 107717peerj3552 622

Table 2 Frequency plant traits database (from different sources) and landscape metrics measured in the field on 46 1 mtimes 1 m plots CA isthe total cover in cm2 Standard deviation (plusmn) are reported only for metrics which can be computed for every patch Definitions abbreviations andunits of measurement of the landscape metrics are reported in Table 1 See Fig 5 for abbreviations of species

Species

Frequency

()(N=46)

Plant traits Landscape metrics

CH

LDMC

LS LDW

SLA

LNC

LA LFW

LCC

MPS

CA

MSI

NP

TE

AdeLeu 2 358 171 3 1165 214 32 24402 7016 467 11 11 105 1 14CarRes 11 24 260 2 37 162 36 582 144 454 9plusmn 4 70 109plusmn 007 8 100CerUni 43 16 162 3 22 376 34 842 139 418 55plusmn 151 2318 123plusmn 023 42 1188GeuRep 46 75 274 6 674 83 30 5664 2498 475 38plusmn 41 2717 119plusmn 010 72 1928GnaSup 7 12 373 4 07 251 19 175 20 477 11plusmn 7 128 116plusmn 007 12 172HieAlp 4 16 162 3 80 152 16 1218 324 460 15plusmn 15 30 121plusmn 005 2 34LeuAlp 70 33 164 4 24 182 26 424 152 447 40plusmn 237 6880 172plusmn 020 173 4018LuzAlp 22 152 179 4 139 291 29 3882 773 469 110plusmn 222 2086 125plusmn 020 19 754OxyDig 59 76 100 4 123 295 47 3535 1239 454 30plusmn 38 6206 126plusmn 023 209 5454PoaAlp 96 36 343 3 296 111 13 3286 866 458 24plusmn 30 7476 117plusmn 012 310 6456RanGla 13 46 142 2 188 141 17 2621 1320 443 10plusmn 7 177 11plusmn 007 17 228SalHer 15 14 332 6 76 152 20 1650 229 474 307plusmn 458 3682 134plusmn 035 12 1042SaxBry 72 5 324 4 02 153 18 28 07 440 95plusmn 238 30378 128plusmn 027 320 13324SaxOpp 39 6 255 4 03 157 18 40 10 468 69plusmn 114 5702 125plusmn 017 83 3006SedAlp 33 15 78 2 04 142 11 57 51 415 7plusmn 6 644 113plusmn 010 87 1010VerAlp 54 46 180 2 21 286 16 579 119 474 15plusmn 14 2013 117plusmn 013 134 2260

was applied to verify for differences in landscape metrics between (i) the three classes ofterrain age and (ii) topsoil texture distinguished and (iii) the three floristic releveacute clustersidentified

To relate plant traits to spatial configuration taking into account species cover inthe plots we applied RLQ-analysis a tool to assess how the environment filters certainspecies traits (Doleacutedec et al 1996 Dray Chessel amp Thioulouse 2003) The RLQ procedureperforms a double inertia analysis of an environmental-variables-by-samples (R-table) anda species-by-traits (Q-table) matrix with a link expressed by a species-cover-by-samplesmatrix (L-table) RLQ-analysis combines three unconstrained separate ordinationscorrespondence analysis of L-table and centred normed principal component analysesof Q- and R- tables to maximise the covariance between environmental factors and traitdata by the use of co-inertia analysis (Bernhardt-Roumlmermann et al 2008) Here we studiedthe joint structure of three data tables namely (i) a plot-by-landscape metrics data tableR-table (ii) a plot-by-species table containing the abundances of the plant species presentin our set of 46 plots table (L-table) and (iii) a species-by-trait data Q-table

This RLQ analysis (basic-RLQ) was followed by a partial-RLQ with the aim of checkingthe effects of the covariates terrain age and cobble cover ie to verify the need to partitionthe variation related to these factors This type of analysis is a special case of RLQ where

Sitzia et al (2017) PeerJ DOI 107717peerj3552 722

the covariable represents a partition of samples into groups If the percentage of co-inertiaexplained by the most representative axis of partial-RLQ were to be much higher thanin the basic-RLQ this would mean that the influence of the covariate is relevant Thesame approach was followed byWesuls et al (2012) to partition the response of plant traitsto grazing-related environmental parameters from other environmental and temporalvariations

A permutation method was used to compare the hypothesis H0 X = 0 (trait andlandscape are unrelated) against H1 X 6= 0 (trait and landscape are related) where X isthe fourth corner a trait-by-landscape table whose parameters cross the traits (Q-table)to the landscape variables (R-table) via the abundance table L-table (Legendre Galzin ampHarmelinVivien 1997) The null hypothesis consists of three null joint hypotheses bothR and Q are linked to L (Lharr Q Lharr R) only R is linked to L (L = Q Lharr R) onlyQ is linked to L (Lharr Q L = R) The overall null hypothesis is rejected when both nullhypotheses are rejected (L = Q and L = R) Dray amp Legendre (2008) proposed to setthe alpha argument to α=

radic005 but recently it has been shown that α should be 005

instead (Ter Braak Cormont amp Dray 2012) Given the limited power of this test with fewspecies (Ter Braak Cormont amp Dray 2012) as in the present study we show the resultsaccording to both the Dray amp Legendre (2008) and the Ter Braak Cormont amp Dray (2012)alpha argument settings A multivariate permutation test was applied to evaluate the globalsignificance of the traits-spatial configuration relationships implemented by the functionlsquorandtestrsquo of the package ade4 (Dray amp Dufour 2007) Next we tested the associations ofspatial configuration and trait variables with the axes of the basic-RLQ The strength of theassociation of landscape metrics and plant traits was measured with the D2 statistic (Drayet al 2014) All tests were performed using the combined fourth-corner statistic (Dray etal 2014) with 49999 permutations

All statistical analyses were performed using the open source R software (R Core Team 2013) We used the library vegan (Oksanen et al 2011) for the cluster analysis the librarystats (R Core Team 2013) for ANOVA and the library ade4 (Dray amp Dufour 2007) for theRLQ analysis

RESULTSSpecies compositionWe recorded a total of 19 plant species (including S stellaris P viviparum andA distentifolium which were omitted from further analyses) Each species frequencyand mean patch-level metrics are reported in Table 2 This table shows that the mostfrequent species wereP alpina Saxifraga bryoides L Leucanthemopsis alpina (L)HeywoodO digyna and Veronica alpina L present in more than half of the plots Mean patch sizetotal cover and mean shape index of species did not necessarily reflect their frequencyConsidering only the species where the number of sampled patches was higher than 30the most compacted shapes belonged to S alpestre the most irregular to L alpina withthe largest patches to S bryoides and the smallest to S alpestre The dispersion of patch sizevalues around the mean was much larger than for patch shape The shape of the frequency

Sitzia et al (2017) PeerJ DOI 107717peerj3552 822

1 2 3

05

1015

2025

30

(A)

F243=345

p=0041

1 2 3

050

010

0015

0020

0025

0030

00

(B)

F243=181p=0176

1 2 3

020

040

060

080

010

00

(C)

F243=3304

plt0001

1 2 3

050

010

0015

0020

0025

0030

00

(D)

F243=3048

plt0001

1 2 3

020

040

060

080

0

(E)

F243=824

p=0006

1 2 3

010

020

030

040

0

(F)

F243=467p=0015

Spe

cies

cov

er p

er s

quar

e m

eter

(cm

2 m2 )

Figure 3 Change in cover (cm2m2) of (A) Cardamine resedifolia (B) Leucanthemopsis alpina (C)Oxyria digyna (D) Saxifraga bryoides (E) Saxifraga oppositifolia (F) Veronica alpina among the threereleveacute clusters (1 Saxifraga-cluster CL1 2 Leucanthemopsis-cluster CL2 3Oxyria-cluster CL3) Fand p-values were obtained by ANOVA

distribution of patch sizes was usually positively skewed (Fig S1) while patch shape valuesshowed more frequently symmetrical distributions (Fig S2)

Three clusters (CLs) resulted from the analysis of the 46 vegetation releveacutes characterisedby differences in the cover of six plant species (Fig 3) In the Saxifraga-cluster (CL1)S oppositifolia was much more abundant and V alpina much less abundant than inthe other two clusters while in the Leucanthemopsis-cluster (CL2) L alpina was moreabundant Finally the Oxyria-cluster (CL3) was characterised by relatively high coversof O digyna and Cardamine resedifolia L (Fig 3) ANOVA showed that the floristicallydefined releveacute clusters differed significantly with regard to certain landscape metrics TheShannonrsquos index of diversity and the patch type (=species) richness were significantlyhigher in the Leucanthemopsis-cluster (CL2) than in the Saxifraga-cluster (CL3) and the

Sitzia et al (2017) PeerJ DOI 107717peerj3552 922

Table 3 Mean (plusmn95 confidence intervals) of landscape metrics in relation to releveacute clusters Releveacute clusters (i) Saxifraga-cluster (CL1)(ii) Leucanthemopsis-cluster (CL2) and (iii) Oxyria-cluster (CL3) See Table 1 for abbreviations of landscape metrics P-values were obtained byANOVA Statistically significant differences amongst groups are indicated in bold

MPS PSCV TE NP MSI SHDI PR

CL1 744plusmn 101 1903plusmn 172 1008plusmn 63 320plusmn 36 123plusmn 001 082plusmn 010 53plusmn 04CL2 369plusmn 92 1632plusmn 221 807plusmn 82 358plusmn 30 119plusmn 001 138plusmn 007 65plusmn 05CL3 417plusmn 52 1186plusmn 339 798plusmn 158 261plusmn 47 128plusmn 003 111plusmn 012 57plusmn 07F 5108 2059 241 1328 9946 1139 1915p 0010 014 0102 0275 0006 lt0001 0160

Table 4 Eigenvalues and percentage of variance () explained by the first five axes of the basic RLQ(sum of eigenvalues 1175) and the partial RLQs (sum of eigenvalues 0930 and 113 respectively usingterrain age and cobble cover as covariates)

Axis Basic RLQ Partial RLQ terrain age Partial RLQ cobble cover

Eigenvalue Eigenvalue Eigenvalue

1 1056 898 0802 862 1023 9052 0075 63 0085 91 0059 523 0029 24 0027 29 0027 234 0010 08 0012 13 0011 105 0005 04 0002 02 0009 08

Oxyria-cluster (CL1) (Table 3) Also the mean size of the vegetation patches differedsignificantly among the three releveacute clusters being greatest in the Saxifraga-cluster (CL1)and smallest in the Leucanthemopsis-cluster (CL2)

As expected none of the landscape metrics differed amongst the three classes of terrainage that is 15ndash41 y 41ndash57 y and 57ndash66 y and of cobble cover that is lt7 between 7and 14 and gt15 (results reported in Table S1) Correlations among landscape traitsare reported in Table S2

Relationships between plant traits and landscape metricsThe percentages of total co-inertia explained by the first two axes of the basic-RLQ was961 while those of the two partial-RLQs were 953 and 957 respectively adoptingterrain age and cobble cover as covariates The first axis of the basic-RLQ explained898 co-inertia while the percentage explained by the first axis of the two partial-RLQswas lower or almost identical (862 and 905 same order above) meaning that thespatial configuration gradient along the first axis of both the partial-RLQs was not morepronounced than the basic-RLQ (Table 4) Moreover the ordination diagrams of thebasic-RLQ did not show any grouping of plots according either to the terrain age nor tothe cobble cover factor (Fig 4)

The test for the model (H1 LharrR) showed that the distribution of species with fixedtraits was influenced by the spatial configuration (p= 0004) while the test for the model(H1 LharrQ) showed that species composition of plots with fixed spatial configurationwas not influenced by the species traits (p= 0590) This means that the traits-spatialconfiguration relationships were not globally significant

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1022

Figure 4 Sample scores (46 vegetation plots) of the first two axes of the basic-RLQ The colours showthe terrain age (A) and the cobble cover (B) classes of the sample plots The eigenvalues are reported inTable 4 The values of d give the grid size

Figure 5 Ordination diagrams of the first two axes of the RLQ-analysis displaying the (A) landscapemetrics scores (B) plant trait scores (C) species scores Abbreviations for landscape metrics andplant traits are reported in Table 1 Abbreviations for species CarRes Cardamine resedifolia CerUniCerastium uniflorum GeuRep Geum reptans GnaSup Gnaphalium supinum HieAlp Hieraciumalpinum LeuAlp Leucanthemopsis alpina LuzAlp Luzula alpino-pilosa OxyDig Oxyria digyna PoaAlpPoa alpina RanGla Ranunculus glacialis SalHer Salix herbacea SaxBry Saxifraga bryoides SaxOppSaxifraga oppositifolia SedAlp Sedum alpestre VerAlp Veronica alpina Adenostyles leucophylla was notreported to avoid excessive gathering of points its position is approximately (20 5) The values of d givethe grid size

The first basic-RLQ axis was significantly and negatively correlated with mean patchsize and patch size coefficient of variation and positively to Shannonrsquos diversity Thesecond basic-RLQ axis was negatively correlated to mean shape index Among the planttraits canopy height leaf fresh and dry weight and leaf area showed a positive significantcorrelation with the first basic-RLQ axis (Table 5)

In summary the first basic-RLQ axis represented a gradient of increasing diversity as aresponse to the presence of larger plant species patches (Fig 5A) Moreover as can be seen

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1122

Table 5 Percentage contribution to total inertia of the basic RLQ and Pearson correlations of spatialconfiguration and plant functional traits with first two basic RLQ axes Projected inertia by each axisis reported in parentheses Associations significantly correlated with RLQ axes are shown in bold (p ltradic005) as proposed by Dray amp Legendre (2008) or with an asterisk (p lt 005) as proposed by Ter Braak

Cormont amp Dray (2012) Abbreviations are reported in Table 1

Variables Contribution tototal inertia ()

Axis 1 (898) Axis 2 (63)

Spatial configurationMPS 231 minus026 minus001PSCV 262 minus017 minus008NP 32 minus009 000TE 105 minus001 minus007MSI 529 minus006 minus015SHDI 531 039 minus002PR 310 018 minus009Plant traitsCH 230 030 minus003LDMC 133 minus007 008LS 189 000 minus010LDW 270 027 007SLA 160 minus011 minus008LNC 437 016 minus014LA 326 034 005LFW 222 029 003LCC 32 minus012 000

in Fig 5B and Fig 5C the first basic-RLQ axis represented a gradient of increasing cover oftaller species with larger and softer leaves on the right hand side like Luzula alpino-pilosa(Chaix) Breistr O digyna Geum reptans L and Adenostyles leucophylla (Willd) Rchb totiny species with smaller heavier and harder leaves on the left hand side like saxifrages andGnaphalium supinum L The second basic-RLQ axis even if less pronounced than the firstaxis represented a gradient of increasing conservation of acquired resources that is fromspecies with higher specific leaf area leaf nitrogen content and tendency for lateral spreadlike Cerastium uniflorum Clairv and O digyna to species with larger dry matter contentlike saxifrages (Fig 5C) Interestingly this axis also represents a gradient of increasingpatch compactness

There were eleven significant associations between plant traits and landscape metricsMean patch size was negatively correlated with canopy height leaf dry and fresh weightand leaf area the latter two of which were also negatively correlated with the coefficientof variation of patch size Among the spatial metrics the Shannonrsquos index of diversity wasmost frequently and positively associated with plant traits namely with canopy heightleaf dry and fresh weight and leaf area Finally patch richness was positively associatedwith canopy height (Fig 6) We should stress that the majority of the tests were significantonly when using a significance level of

radic005 The complete list of correlations between

landscape and plant traits is reported in Table S3

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1222

CH

LDMC

LS

LDW

SLA

LNC

LA

LFW

LCC

MP

S

PS

CV

TE

NP

MS

I

SH

DI

PR

Figure 6 Schematic representation of the association between individual plant traits and landscapemetrics Red cells correspond to positive significant relationships while blue cells correspond to negativesignificant relationships The strength of the association was measured with the D2 statistic and testedthrough a fourth-corner analysis (Dray et al 2014) P-values were corrected by a sequential procedure(49999 repetitions) which leads to significant associations if the maximum p-value was lower than α =radic005 as proposed by Dray amp Legendre (2008) Abbreviations are reported in Table 1

DISCUSSIONPrior work on glacier forelands has documented that whilst terrain age may appear tobe one of the most obvious factors affecting plant colonisation processes (Caccianiga ampAndreis 2004) spatial patterns of plant species and their constituent assemblages are oftenheterogeneous and complex and not always directly related to the time factor (Burga et al2010Matthews amp Whittaker 1987 Vetaas 1994)

Here we applied a grid-sampling technique where the variables of interest are surveyedon a regular lattice grid at the appropriate scale This approach which has many possibleapplications such as habitat suitability assessment (Sitzia et al 2014a) and trail alignment

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1322

(Sitzia et al 2014b) has been possible since in the type of habitat studied here plant speciesform discrete patches in the form of small cushions tussocks or rosettes separated bybare substrate

Species composition and landscape metricsThe terrain age and plant composition observed here correspond to early-successionalstages of glacier foreland colonization it does not include any mid- or late-successionalspecies according to Caccianiga et al (2006) apart from S bryoides However this speciesaccording to same authors is present in the same proportion in early- andmid-successionalsites The composition also includes several ubiquitous species like P alpina L alpino-pilosa and Hieracium alpinum L These species are typical wind-dispersed ruderals Thespectrumof plant traits encountered is limited and only some species namelyA leucophyllaP alpina Ranunculus glacialis L G reptans L alpino-pilosa and O digyna exhibitedrelatively strong departures from the average characteristics of the main group of species

We have shown that releveacute clusters differ in landscape metrics For instance S bryoidesa species with small leaves but relatively high nitrogen content showed the highest patchsizes amongst the frequent species and characterised a cluster with a correspondinglyhigher mean patch size V alpina and L alpina both medium-sized species were presentwith many patches and accordingly characterised a cluster with a higher Shannonrsquosdiversity Finally O digyna a relatively low compact species with medium lateral spreadand nitrogen content was the most frequent in the cluster with the highest mean patchshape

Spatial configuration as a correlate of plant traitsThe RLQ-analysis showed that plant species with efficient conservation of nutrientswhich have a higher dry matter content in leaves (Pierce et al 2013 Wilson Thompson ampHodgson 1999) also have lower values of the shape index or in other words their patchesare more compact This pattern maintained due to processes of intra- and interspecificcompetition for space and nutrients ensures efficient acquisition-conservation trade-offsin plants characterized by slow growth as in P alpina which among the common speciesis second in patch compactness only to S alpestre Second we found that Shannon diversityincreased with increasing cover of upright-growing plant species characterised by largerand heavier leaves A possible explanation for this lies in the way plants with these traitscompete directly andor indirectly and how they modify one anotherrsquos biotic and abioticenvironment thereby generating a more equitable distribution of patch sizes combinedwith a higher number of species

The observed correlation of landscape metrics with species composition together withthe correlations with specific life-form traits seem to indicate some level of life-form orspecies-based spatial self-organisation Self-organization does not imply specific causalitiesbetween vegetation patterns and the environment but is induced by internal variationindependent of external drivers (Bolliger Sprott amp Mladenoff 2003) Initial establishmentof any particular species in this microhabitat depends on successful seed establishment Thefirst selection of species is mainly trait-driven because only species with a wind dispersal

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1422

strategy are selected from the geographical species pool (Keddy 1992) However thelater establishment is dependent on random abiotic factors such as wind-aided dispersaland by small-scale variation of the soil surface characteristics such as texture aidinggermination Stochastic factors thus potentially lead to a high degree of heterogeneity inseedling distribution due to the variability of seed rain the soil seed bank germinationmortality rates of the seedlings as well as other factors (Erschbamer Kneringer amp Schlag2001 Marcante Winkler amp Erschbamer 2009) Large-scale successional stages usuallycontain a wide array of different microhabitats where species composition is mainly drivenby habitat conditions and local disturbances (eg floods rock falls and avalanches)(Burga et al 2010)

Here we tried to maintain at a minimum level these potential environmental filters byselecting a restricted range of terrain ages slope and topsoil texture This has probablyfavoured us in finding the observed relationships between landscape patch metrics andfunctional traits

Landscape metrics as functional traitsLandscape-scale variation usually surveyed at larger resolutions than those used here isoften seen as a filter of plant traits rather than a plant trait itself This can be partiallyexplained by the need to study how environmental change and disturbance especiallyanthropogenic may affect the landscape mosaic and in turn filter specific traits (Gaacutemez-Virueacutes et al 2015) Another possible explanation is the difficulty to select the appropriategrain and extent to correctly detect each speciesrsquo patch boundaries According to OrsquoNeillet al (in Turner Gardner amp OrsquoNeill 2001) the grain size of a map should be two to fivetimes smaller than the spatial features being analysed and map extent should be two tofive times larger than the largest patches The mean upper quartile and maximum patchsize was 47 cm2 35 cm2 and 3104 cm2 respectively This means that the extent (1 m2) wechose was appropriate and that the grain (1 cm2) could be apparently increased Howeversome species presented very low mean patch sizes eg S alpestre (only 7 cm2) This meansthat to perform a multi-species patch-level calculation of landscape metrics on a glacierforeland with terrain ages lt70 y the grain should be maintained at least lower than 35ndash14cm2 in practice units of 30ndash10 cm2 Finally we have to stress again that the identificationof each patch could be difficult in late successional stages where their edges could not beeasily distinguished

Databases of plant functional traits such as LEDA (Knevel et al 2003) TRY (Kattgeet al 2011) or BIOPOP (Poschlod et al 2003) are very useful since they provide possibledeterminants of the response of primary producers to environmental factors and howthey can affect other trophic levels ecosystem processes services and diversity (Kattge etal 2011) Existing databases of functional traits do not include data about patch-levellandscape metrics of plant species We suggest that these databases should incorporatesuch data

One possible criticism of this proposal might be that some landscape traits might alreadybe incorporated in existing plant species attributes However this should be treated withcaution For example while patch size might be associated with the Hodgson et alrsquos (1999)

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1522

lateral spread measure included as a class parameter into the attribution of Grimersquos (1977)CSR strategy here we observed that a reduction in patch size is more clearly associatedwith canopy height and leaf traits Also the capability to perform clonal growth resultingin large and compact patches is an attribute largely included into existing databases andis well known for many alpine grasses and sedges (Groenendael et al 1996) On the otherhand patch size and shape could be linked to ontogenetic features of the individual (ieindividual age) for species with indefinite growth such as many woody species like manycushion species Individual age in turn could be related to stochastic station parameterssuch as terrain age whichmay be the case in a glacier foreland and should not be consideredas a plant trait Compactness of the patch itself could be the result of many plant attributesand thus requires care if treated as a plant trait Moreover non-clonal single individualsare rather compact and we have shown that the most compact patches are also the smallestThis could be due to the situation where the species occurs in single individuals In generalwe have shown that the distribution of species-specific landscape trait values in particularpatch size might be rather skewed meaning the mean might not be the best summarisingstatistical indicator and that care should be taken in the selection of the individuals wheremeasuring the traits (Cornelissen et al 2003) All these aspects call for further researchto collect these data and to link them with physiological attributes directly measured inthe field

CONCLUSIONSIn summary we have presented a new high-resolution approach for the detection oflandscape-level correlation with plant species traits Analyses of how species patches arespatially arranged in herbaceous glacier foreland communities indicate that diversity inpatch types and size both increases whilst patch size decreases with increasing canopyheight leaf size and weight Moreover more compact patch shapes indicate an increasedcapacity of conservation of nutrients in leaves We stress the need to perform furtheranalyses on the distribution of speciesrsquo patch sizes which here we nested inside squareplots but treating the plots as random factors could provide further insights into thedynamics of space use and niche overlap on glacier forelands Moreover some of theresults were not statistically significant because of the limited range of terrain age analysedand the few species surveyed A wider span of terrain ages and in turn of plant traits whichwould also consider the effect of time since deglaciation and environmental variabilitywould provide further information about the amount of the self-induced trait-drivenvariability in spatial configuration and that related to allogenic environmental filters thusincreasing statistical robustness

Any measurable attribute at the individual level that directly or indirectly influenceoverall fitness or performance might be regarded a functional trait (Violle et al 2007)According to this we conclude that patch-level landscape metrics of plants should betreated as functional traits Further research is needed in a range of ecosystems but weexpect this approach to open up an entirely new class of plant traits ie landscape planttraits which should be collected at patch level and might bring additional functionalmeaning as predictors of life strategies

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1622

ACKNOWLEDGEMENTSIt is our pleasure to thank Gabriele Pezzani Massimo Tondini Pietro Tonni and ThomasCampagnaro for their valuable support in the field and Bruno Michielon the editor NigelG Yoccoz and two anonymous referees for their suggestions on an earlier version of themanuscript We thank Steacutephane Dray and Dirk Wesuls for providing assistance with theRLQ analysis We also thank the cartographic centre of Carsquo Foscari Venice University forproviding crucial historical maps

ADDITIONAL INFORMATION AND DECLARATIONS

FundingRevisions on the paper were made during two visiting fellowships to DM within a 2013visiting scientist grant by the University of Padova under TS responsibility and to TS atthe University of Northampton within an European Union LLPErasmus Teaching Staffbilateral agreement The funders had no role in study design data collection and analysisdecision to publish or preparation of the manuscript

Competing InterestsThe authors declare there are no competing interests

Author Contributionsbull Tommaso Sitzia conceived and designed the study collected data in the field analyzedthe data contributed materials and analysis tools wrote the paper prepared figures andtables reviewed drafts of the paperbull Matteo Dainese collected data in the field analyzed the data reviewed drafts of thepaperbull Bertil O Kruumlsi reviewed drafts of the paperbull Duncan McCollin reviewed drafts of the paper

Data AvailabilityThe following information was supplied regarding data availability

The raw data has been supplied as a Supplementary File

Supplemental InformationSupplemental information for this article can be found online at httpdxdoiorg107717peerj3552supplemental-information

REFERENCESBernhardt-RoumlmermannM Roumlmermann C Nuske R Parth A Klotz S SchmidtW

Stadler J 2008 On the identification of the most suitable traits for plant functionaltrait analyses Oikos 1171533ndash1541 DOI 101111j0030-1299200816776x

Bolliger J Sprott JC Mladenoff DJ 2003 Self-organization and complexity in historicallandscape patterns Oikos 100541ndash553 DOI 101034j1600-0706200312109x

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1722

Burga CA Krusi B Egli MWernli M Elsener S Ziefle M Fischer T Mavris C2010 Plant succession and soil development on the foreland of the Morteratschglacier (Pontresina Switzerland) straight forward or chaotic Flora 205561ndash576DOI 101016jflora200910001

Caccianiga M Andreis C 2004 Pioneer herbaceous vegetation on glacier forelands inthe Italian Alps Phytocoenologia 3455ndash89 DOI 1011270340-269x20040034-0055

Caccianiga M Luzzaro A Pierce S Ceriani RM Cerabolini B 2006 The functionalbasis of a primary succession resolved by CSR classification Oikos 11210ndash20DOI 101111j0030-1299200614107x

Carbognani M 2011 Ecologia di due fitocenosi di valletta nivale caratteristichestrutturali e funzionali ed effetti del riscaldamento climatico Tesi di Dottorato diricerca in Biologia Vegetale Universitagrave degli Studi di Parma Parma Available athttpdspace-uniprcinecaithandle18891699

Cerabolini B Brusa G Ceriani RM De Andreis R Luzzaro A Pierce S 2010 CanCSR classification be generally applied outside Britain Plant Ecology 210253ndash261DOI 101007s11258-010-9753-6

Chapin FSI Walker LR Fastie CL Sharman LC 1994Mechanisms of primarysuccessions following deglaciation at Glacier Bay Alaska Ecological Monographs64149ndash175 DOI 1023072937039

Cornelissen JHC Lavorel S Garnier E Az S Buchmann N Gurvich DE Reich PBSteege HT Morgan HD Heij denMGAvd Pausas JG Poorter H 2003 A hand-book of protocols for standardised and easy measurement of plant functional traitsworldwide Australian Journal of Botany 51335ndash380 DOI 101071BT02124

Deckers B Verheyen K HermyMMuys B 2004 Differential environmental responseof plant functional types in hedgerow habitats Basic and Applied Ecology 5551ndash566DOI 101016jbaae200406005

Doleacutedec S Chessel D Ter Braak CJF Champely S 1996Matching species traits toenvironmental variables a new three-table ordination method Environmental andEcological Statistics 3143ndash166 DOI 101007BF02427859

Dray S Chessel D Thioulouse J 2003 Co-inertia anaylsis and the linking of ecologicaldata tables Ecology 843078ndash3089 DOI 10189003-0178

Dray S Choler P Doledec S Peres-Neto PR ThuillerW Pavoine S Ter Braak CJF2014 Combining the fourth-corner and the RLQ methods for assessing traitresponses to environmental variation Ecology 9514ndash21 DOI 10189013-01961

Dray S Dufour AB 2007 The ade4 package implementing the duality diagram forecologists Journal of Statistical Software 221ndash20

Dray S Legendre P 2008 Testing the species traits-environment relationships thefourth-corner problem revisited Ecology 893400ndash3412 DOI 10189008-03491

Duflot R Georges R Ernoult A Aviron S Burel F 2014 Landscape heterogeneity asan ecological filter of species traits Acta Oecologica-International Journal of Ecology5619ndash26 DOI 101016jactao201401004

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1822

Erschbamer B Kneringer E Schlag RN 2001 Seed rain soil seed bank seedlingrecruitment and survival of seedlings on a glacier foreland in the Central Alps Flora196304ndash312 DOI 101016S0367-2530(17)30059-2

Erschbamer B Mayer R 2011 Can successional species groups be discriminated basedon their life history traits A study from a glacier foreland in the Central Alps PlantEcology amp Diversity 4341ndash351 DOI 101080175508742012664573

Erschbamer B Niederfriniger Schalg R EckartW 2008 Colonization processeson a central Alpine glacier foreland Journal of Vegetation Science 19855ndash862DOI 1031702008-8-18464

Forman RTT GodronM 1981 Patches and structural components for a landscapeecology BioScience 31733ndash740 DOI 1023071308780

Gaacutemez-Virueacutes S Perović DJ Gossner MM Boumlrschig C Bluumlthgen N De Jong H SimonsNK Klein A-M Krauss J Maier G Scherber C Steckel J Rothenwoumlhrer C Steffan-Dewenter I Weiner CNWeisserWWerner M Tscharntke TWestphal C 2015Landscape simplification filters species traits and drives biotic homogenizationNature Communications 6Article 8568 DOI 101038ncomms9568

Greig-Smith P 1979 Pattern in vegetation Journal of Ecology 67755ndash779DOI 1023072259213

Grime JP 1977 Evidence for the existence of three primary strategies in plants andits relevance to ecological and evolutionary theory The American Naturalist1111169ndash1194 DOI 101086283244

Groenendael JMV Klimes L Klimesova J Hendriks RJJ 1996 Comparative ecologyof clonal plants Philosophical Transactions of the Royal Society of London Series BBiological Sciences 3511331ndash1339 DOI 101098rstb19960116

Gross N Suding KN Lavorel S 2007 Leaf dry matter content and lateral spread predictresponse to land use change for six subalpine grassland species Journal of VegetationScience 18289ndash300 DOI 101111j1654-11032007tb02540x

Hodgson JGWilson PJ Hunt R Grime JP Thompson K 1999 Allocating C-S-R plant functional types a soft approach to a hard problem Oikos 85282ndash294DOI 1023073546494

Jiang J DeAngelis D Smith III T Teh S Koh H-L 2012 Spatial pattern formationof coastal vegetation in response to external gradients and positive feedbacksaffecting soil porewater salinity a model study Landscape Ecology 27109ndash119DOI 101007s10980-011-9689-9

Kattge J Diacuteaz S Lavorel S Prentice IC Leadley P Boumlnisch G Garnier E WestobyM Reich PBWright IJ Cornelissen JHC Violle C Harrison SP Van BodegomPM ReichsteinM Enquist BJ Soudzilovskaia NA Ackerly DD AnandM AtkinO BahnM Baker TR Baldocchi D Bekker R Blanco CC Blonder B BondWJBradstock R Bunker DE Casanoves F Cavender-Bares J Chambers JQ ChapinIII FS Chave J Coomes D Cornwell WK Craine JM Dobrin BH Duarte L DurkaW Elser J Esser G Estiarte M FaganWF Fang J Fernaacutendez-Meacutendez F FidelisA Finegan B Flores O Ford H Frank D Freschet GT Fyllas NM Gallagher RVGreenWA Gutierrez AG Hickler T Higgins SI Hodgson JG Jalili A Jansen

Sitzia et al (2017) PeerJ DOI 107717peerj3552 1922

S Joly CA Kerkhoff AJ Kirkup D Kitajima K Kleyer M Klotz S Knops JMHKramer K Kuumlhn I Kurokawa H Laughlin D Lee TD LeishmanM Lens F LenzT Lewis SL Lloyd J Llusiaacute J Louault F Ma S MahechaMDManning P MassadT Medlyn BE Messier J Moles AT Muumlller SC Nadrowski K Naeem S NiinemetsUuml Noumlllert S Nuumlske A Ogaya R Oleksyn J Onipchenko VG Onoda Y OrdontildeezJ Overbeck G OzingaWA Patintildeos S Paula S Pausas JG Pentildeuelas J PhillipsOL Pillar V Poorter H Poorter L Poschlod P Prinzing A Proulx R RammigA Reinsch S Reu B Sack L Salgado-Negret B Sardans J Shiodera S Shipley BSiefert A Sosinski E Soussana JF Swaine E Swenson N Thompson K ThorntonPWaldramMWeiher E White MWhite S Wright SJ Yguel B Zaehle S ZanneAEWirth C 2011 TRYmdasha global database of plant traits Global Change Biology172905ndash2935 DOI 101111j1365-2486201102451x

Keddy PA 1992 Assembly and response rules two goals for predictive communityecology Journal of Vegetation Science 3157ndash164 DOI 1023073235676

Knevel IC Bekker RM Bakker JP Kleyer M 2003 Life-history traits of the NorthwestEuropean flora the LEDA database Journal of Vegetation Science 14611ndash614DOI 101111j1654-11032003tb02188x

Koumlrner C 2003 Alpine plant life functional plant ecology of high mountain ecosystems 2ndedition Berlin Heidelberg Springer-Verlag

Krebs CJ 1999 Ecological methodology 2nd edition Menlo park Addison WeseleyLongman

Legendre P Galzin R HarmelinVivienML 1997 Relating behavior to habitat solutionsto the fourth-corner problem Ecology 78547ndash562

MacfadyenWA 1950 Vegetation patterns in the semi-desert plains of British Soma-liland The Geographical Journal 116199ndash211 DOI 1023071789384

Marcante S Winkler E Erschbamer B 2009 Population dynamics along a primarysuccession gradient do alpine species fit into demographic succession theory Annalsof Botany 1031129ndash1143 DOI 101093aobmcp047

Matthews JA 1999 Disturbance regimes and ecosystem response on recently-deglaciatedsubstrates Ecosystems of disturbed ground 17ndash37

Matthews JAWhittaker RJ 1987 Vegetation succession on the Storbreen glacierforeland Jotunheimen Norway a review Arctic and Alpine Research 19385ndash395DOI 1023071551403

McGarigal K Marks BJ 1994 Spatial pattern analysis program for quantifying landscapestructure Washington DC USDA Forest Service

McGill BJ Enquist BJ Weiher E WestobyM 2006 Rebuilding communityecology from functional traits Trends in Ecology amp Evolution 21178ndash185DOI 101016jtree200602002

Nascimbene J Mayrhofer H Dainese M Bilovitz PO 2017 Assembly patterns of soil-dwelling lichens after glacier retreat in the European Alps Journal of Biogeography441393ndash1404 DOI 101111jbi12970

Oksanen J Blanchet FG Kindt R Legendre P Minchin PR OrsquoHara RB Simpson GLSolymos P Stevens MHHWagner H 2011 Vegan community ecology package

Sitzia et al (2017) PeerJ DOI 107717peerj3552 2022

R package version 20-2 Available at httpsCRANR-projectorgpackage=vegan(accessed on 7 July 2017)

Patton DR 1975 A diversity index for quantifying habitat lsquolsquoedgersquorsquoWildlife SocietyBulletin 3171ndash173

Pierce S Brusa G Vagge I Cerabolini B 2013 Allocating CSR plant functional typesthe use of leaf economics and size traits to classify woody and herbaceous vascularplants Functional Ecology 271002ndash1010 DOI 1011111365-243512095

Pierce S Luzzaro A Caccianiga M Ceriani RM Cerabolini B 2007 Disturbanceis the principal α-scale filter determining niche differentiation coexistenceand biodiversity in an alpine community Journal of Ecology 95698ndash706DOI 101111j1365-2745200701242x

Poschlod P Kleyer M Jackel A-K Dannemann A Tackenberg O 2003 BIOPOPmdashA database of plant traits and internet application for nature conservation FoliaGeobotanica 38263ndash271 DOI 101007BF02803198

R Core Team 2013 R a language and environment for statistical computing ViennaR Foundation for Statistical Computing Available at httpwwwR-projectorg(accessed on 01 August 2015)

Raffl C MallaunMMayer R Erschbamer B 2006 Vegetation succession pattern anddiversity changes in a glacier valley Central Alps Austria Arctic Antarctic and AlpineResearch 38421ndash428 DOI 1016571523-0430(2006)38[421vspadc]20co2

ReinersWAWorley IA Lawrence DB 1971 Plant diversity in a chronosequence atGlacier Bay Alaska Ecology 5255ndash69 DOI 1023071934737

Rempel RS Kaukinen D Carr AP 2012 Patch analyst and patch grid centre for northernforest ecosystem research Thunder Bay Ontario Ministry of Natural Resources

RietkerkM Van de Koppel J 2008 Regular pattern formation in real ecosystems Trendsin Ecology amp Evolution 23169ndash175 DOI 101016jtree200710013

Rydgren K Halvorsen R Toumlpper JP Njoslashs JM 2014 Glacier foreland successionand the fading effect of terrain age Journal of Vegetation Science 251367ndash1380DOI 101111jvs12184

SAT 2007Ghiacciaio Occidentale di Nardis Comitato Glaciologico Trentino Societagravedegli Alpinisti Tridentini Trento Available at httpwwwsattnit sns17640_0htm(accessed on 07 July 2017)

Schleicher A Peppler-Lisbach C Kleyer M 2011 Functional traits during succession isplant community assembly trait-driven Preslia 83347ndash370

Schwienbacher E Navarro-Cano JA Neuner G Erschbamer B 2012 Correspondenceof seed traits with niche position in glacier foreland succession Plant Ecology213371ndash382 DOI 101007s11258-011-9981-4

Sitzia T Dainese M Clementi T Mattedi S 2014a Capturing cross-scalar vari-ation of habitat selection with grid sampling an example with hazel grouse(Tetrastes bonasia L) European Journal of Wildlife Research 60177ndash186DOI 101007s10344-013-0762-3

Sitzia et al (2017) PeerJ DOI 107717peerj3552 2122

Sitzia T Rizzi A Cattaneo D Semenzato P 2014b Designing recreational trails in aforest dune habitat using least-cost path analysis at the resolution of visitor sight dis-tance Urban Forestry amp Urban Greening 13861ndash868 DOI 101016jufug201409011

Sole RV Bascompte J 2006 Self-organization in complex ecosystems Princeton Prince-ton University Press

Suding KN Lavorel S Chapin FS Cornelissen JHC Diacuteaz S Garnier E Goldberg DHooper DU Jackson ST Navas M-L 2008 Scaling environmental change throughthe community-level a trait-based response-and-effect framework for plants GlobalChange Biology 141125ndash1140 DOI 101111j1365-2486200801557x

Teixido N Garrabou J Gutt J ArntzWE 2007 Iceberg disturbance and successionalspatial patterns The case of the shelf Antarctic benthic communities Ecosystems10142ndash157 DOI 101007s10021-006-9012-9

Ter Braak CJF Cormont A Dray S 2012 Improved testing of species traits-environment relationships in the fourth-corner problem Ecology 931525ndash1526DOI 10189012-01261

Těšitel J Těšitelovaacute T Bernardovaacute A Jankovaacute Drdovaacute E LučanovaacuteM Klimešovaacute J2014 Demographic population structure and fungal associations of plants colonizingHigh Arctic glacier forelands Petuniabukta Svalbard Polar Research 3Article 20797DOI 103402polarv3320797

Turner MG Gardner RH OrsquoNeill RV 2001 Landscape ecology in theory and practiceNew York Springer

Van de Koppel J Bouma TJ Herman PMJ 2012 The influence of local- and landscape-scale processes on spatial self-organization in estuarine ecosystems The Journal ofExperimental Biology 215962ndash967 DOI 101242jeb060467

Vetaas OR 1994 Primary succession of plant assemblages on a glacier foreland-Boslashdalsbreen southern Norway Journal of Biogeography 21297ndash308DOI 1023072845531

Vetaas OR 1997 Relationships between floristic gradients in a primary successionJournal of Vegetation Science 8665ndash676 DOI 1023073237371

Violle C Navas ML Vile D Kazakou E Fortunel C Hummel I Garnier E 2007 Let theconcept of trait be functional Oikos 116882ndash892DOI 101111j0030-1299200715559x

Wentworth CK 1922 A scale of grade and class terms for clastic sediments The Journalof Geology 30377ndash392 DOI 101086622910

Wesuls D Oldeland J Dray S 2012 Disentangling plant trait responses to livestockgrazing from spatio-temporal variation the partial RLQ approach Journal ofVegetation Science 2398ndash113 DOI 101111j1654-1103201101342x

Wilson PJ Thompson K Hodgson JG 1999 Specific leaf area and leaf dry mattercontent as alternative predictors of plant strategies New Phytologist 143155ndash162DOI 101046j1469-8137199900427x

Sitzia et al (2017) PeerJ DOI 107717peerj3552 2222