Ra

Ta

b

a

ARRAA

KBEIGMPR

1

soSee

a

(

h1

Perspectives in Plant Ecology, Evolution and Systematics 20 (2016) 46–55

Contents lists available at ScienceDirect

Perspectives in Plant Ecology, Evolution and Systematics

jo ur nal ho me p age: www.elsev ier .com/ locate /ppees

ange limits and population dynamics of non-native plants spreadinglong elevation gradients

im Seipela,b,∗, Jake M. Alexanderb, Peter J. Edwardsb, Christoph Kuefferb

Department of Land Resources and Environmental Sciences, Montana State University, Bozeman, MT 59717, USAInstitute of Integrative Biology, ETH Zurich, Universitätsstrasse 16, CH-8092 Zürich, Switzerland

r t i c l e i n f o

rticle history:eceived 7 July 2015eceived in revised form 1 April 2016ccepted 6 April 2016vailable online 7 April 2016

eywords:iodiversity monitoringuropean Alpsnvasive specieseographic range shiftsetapopulation

opulation turnoverange margin

a b s t r a c t

Monitoring the elevation limits of non-native species is a potentially sensitive means of detecting effectsof environmental change on invasion dynamics and species ranges. The aim of this study was to inves-tigate temporal changes in the distribution of non-native plant species along elevation gradients in theSwiss Alps by repeating, in 2009, a regional survey from 2003 of 230 sites ranging in elevation from 200to 2400 m a.s.l. We also studied the fine-scale spatiotemporal population structure of two of the non-native species – Erigeron annuus and Solidago canadensis – along an elevation gradient in a heterogeneouslandscape.

Most non-native species in the Swiss Alps rapidly decline in probability of occurrence as elevationincreases. We found little change in the elevation ranges limits of species in time, suggesting that mostspecies are not rapidly expanding at their high elevation range limits. For most species, populationswere more dynamic (colonizations and extinctions) at the upper range limit where occurrence rapidlydeclined. Population turnover was negatively correlated with probability of occurrence at the regionaland local scale. At low elevations, where probability of occurrence was higher, the number of individualsin a population was also greater. At the local and regional scales, E. annuus and S. canadensis had similarrange limits. At the local scale, propagule production of both E. annuus and S. canadensis was greater in thecore of their distributions at lower elevations, and distance to nearest neighbor increased as occurrencedecreased.

Our data demonstrate that range limits of non-native species at high elevation are associated with

high population turnover, which results in a transition zone characterized by source-sink dynamics.Populations within this zone exhibit reduced probability of occurrence, and smaller patches. This resulthas important implications for the monitoring of spreading species along environmental gradients. Tounderstand these limits and predict range expansion, multi-year monitoring and demography data thatincludes information on colonization and extinction events will be needed.

© 2016 Elsevier GmbH. All rights reserved.

. Introduction

There is a great need to understand and monitor changes inpecies richness and individual species distributions in the facef global change (Gottfried et al., 2012; Hellmann et al., 2008;

tohlgren et al., 2000). An increasing proportion of species arexpanding their ranges beyond historical distributions because ofnvironmental change, including non-native species, and some of

∗ Corresponding author at: Leon Johnson Hall 325, Department of Land Resourcesnd Environmental Sciences, Montana State University, Bozeman, MT 59717, USA.

E-mail addresses: tim.seipel@gmail.com, timothy.seipel@montana.eduT. Seipel).

ttp://dx.doi.org/10.1016/j.ppees.2016.04.001433-8319/© 2016 Elsevier GmbH. All rights reserved.

these species can threaten biodiversity and ecosystem functioning(Caplat et al., 2013; Carey et al., 2012; Pysek et al., 2012). Shiftsin distribution of plants to higher elevations have been demon-strated for many native species in the European Alps (Grabherret al., 1994; Lenoir et al., 2008; Pauli et al., 2012), apparently inresponse to warmer conditions, but species in other parts of theworld have also shifted in response to water availability (Crimminset al., 2011). It is expected that climate change will also triggerrapid expansion of non-native plants into mountain ecosystems,which may become an important threat to species-rich and unique

mountain floras (Diez et al., 2012; Kueffer et al., 2013; Pauchardet al., 2009; UNEP, 2009). Monitoring species distributions to detectchanges in species’ range distributions is therefore an important

gy, Ev

clu

u(TpttMfWcfiicst

d2(tmia2trdqdH2

tsfhstisnslstrttwcapl

ebfpg2s

T. Seipel et al. / Perspectives in Plant Ecolo

omponent of a proactive management strategy to protect theseeast invaded habitats at high elevations from future invasions ofpwards moving species (McDougall et al., 2011b).

Species distribution modeling has been an important tool fornderstanding the current and potential distributions of speciesGuisan and Thuiller, 2005; Guisan and Zimmermann, 2000;huiller et al., 2008), and is therefore used to make assessments ofotential changes in a species distribution in time and in relationo different environmental controlling factors; including of eleva-ional distributions of non-native species (Petitpierre et al., 2016).

onitoring changes in populations has been an important toolor detecting changes in species occupancy regionally and locally.

hile species distribution models are often fit with coarse dataovering a large area, population monitoring is more usually per-ormed at a fine scale and small extent. Yet, despite the obviousmportance of combining these scales if we are to understand shiftsn distributions, there have been few attempts to integrate studiesonducted at regional and local scales. The integration of locallycaled monitoring of populations across a regional scale will helpo understand how range limits form (Schurr et al., 2012).

Species range limits are ultimately caused by varying birth,eath, and dispersal rates across space and time (Holt and Keitt,000; Oborny et al., 2009). Examining the turnover of populationsi.e. colonization and extinctions) can provide information abouthe dynamics of range margins because there should be an approxi-

ate balance between extinction and colonization if the range limits stable, whereas colonization should exceed extinctions if rangesre expanding (Anderson et al., 2009; MacArthur, 1972; Travis,004). Also possible are variable ratios of colonization to extinctionhat affect probability of occurrence and abundance within the cur-ent range. Colonization is affected by the ability of populations toisperse to suitable habitat and extinction rates affected by habitatuality, hence population turnover gives general insight into rangeynamics (Brown and Kodric-Brown, 1977; Holt and Keitt, 2000;olt et al., 2005; Kawecki, 2008; Lennon et al., 1997; Schurr et al.,012; Sorte, 2013).

Many non-native invasive plant species occupy disturbed habi-ats (Alpert et al., 2000; Seipel et al., 2012), and spread of thesepecies depends upon the capacity of their propagules to disperserom one suitable habitat patch to the next. However, because theseabitat patches tend to be temporary, local populations are oftenhort-lived and the persistence of a species depends upon its abilityo colonize new sites. Under favorable conditions, a high density ofndividuals may develop in a suitable patch, making it likely thatome of the seeds produced will disperse to unoccupied patchesearby. Under these conditions, we would expect most suitableites in the region to be occupied (Hanski, 1997). However, underess favorable conditions – for example, towards the margin of apecies’ fundamental or Hutchinsonian niche – local seed produc-ion may be much lower, reducing the probability that seeds willeach unoccupied patches nearby. For this reason, we would expecto observe a marked decline in the occupancy of sites towardshe edge of a species niche; indeed the potential range margin,here growth and reproduction are still possible, may be unoc-

upied simply because too few seeds reach these patches (Brownnd Kodric-Brown, 1977; Hanski, 1997). Therefore an interaction ofopulation dynamics and fundamental niche gives rise to the range

imit of species (Pulliam, 2000; Schurr et al., 2012).In this study we use non-native plant species spreading along

levation gradients in mountains to understand the relationshipsetween distribution limits and population-level processes thatorm species’ ranges. In most mountainous regions, non-native

lant species richness peaks in the lower third of the elevationradient and declines rapidly as elevation increases (Becker et al.,005; McDougall et al., 2011a; Seipel et al., 2012). A global analy-is suggests that high elevation sites are rarely invaded because

olution and Systematics 20 (2016) 46–55 47

non-native species are normally introduced to lowland areasfrom which they may subsequently spread to higher elevations(Alexander et al., 2011; McDougall et al., 2011b; Seipel et al., 2012).This makes the distribution of non-native species along elevationgradients an excellent model system for studying the dynamicsof species distributions in response to global change, because thespread and environmental gradient is unidirectional from low tohigh elevations and are the result of contemporary environmentalfactors and biotic interactions. There is some evidence that longerestablished non-native species reach higher elevations (Beckeret al., 2005; Haider et al., 2010), which has been interpreted as theresult of long-term expansion possibly involving local adaptation(Haider et al., 2012). On the other hand, non-native species oftenshow similar distributional limits to those in their native range(Alexander and Edwards, 2010; Alexander et al., 2009), suggestingthat they have reached the limits of their climatic niche and thuselevation range limits may be stable as long as the climate does notchange.

The aim of this study was to investigate the distribution andpopulation turnover of non-native plant species along elevationgradients in the Swiss Alps. To do this we utilize two datasetsto analyze elevation distributions of non-native species, and theirassociated population dynamics. At the scale of the complete SwissAlps we repeated a survey performed six years earlier of 230 rud-eral sites ranging in elevation from 200 m to 2400 m a.s.l. (Beckeret al., 2005). At a local scale we monitored population dynamicsof two model species – Erigeron annuus and Solidago canadensis– in the upper Rhine river drainage in the Swiss Alps over threeyears. By combining these two datasets we can examine both pop-ulation dynamics regionally at disturbed sites (i.e. Swiss Alps), andlocally within a heterogeneous landscape (i.e. varying habitat andland use). Our specific research questions were:

1 How stable are the upper elevation range limits of non-nativespecies?

2 Do population dynamics differ along the elevation range a speciesoccupies?

3 How are elevation range limits and population-level processesinterrelated?

2. Materials and methods

2.1. Species richness and population turnover at a regional scale

In 2003, Becker et al. (2005) investigated the elevation distri-bution of non-native plant species in Switzerland using a sampleof 230 sites of 0.5-1.0 km2 in size located at roadsides or at rail-way stations that ranged in elevation between 200 and 2400 mabove sea level. We resurveyed these sites in 2009, visiting themat the same time of year (individual sites were recorded betweenJuly and September) and following exactly the same procedures.Most species recorded in 2003 were recorded in 2009; for newlyrecorded species, origin (native vs. non-native) was determinedusing Flora Helvetica (Lauber and Wagner, 2001). At each site, twotrained botanists searched the surveyed area for 30 min and esti-mated the abundance of all non-native species using a four pointscale: 0 (absent), 1 (1–10 individuals), 2 (10–100 individuals) and3 (greater than 100 individuals). The records of the two botanistswere then combined to yield one list for each site, and the abun-

dance was estimated based on consensus between the botanists.To estimate the error in recording species, we revisited three sitesat low, middle and high elevations and made a detailed study of thespecies present (Supplementary material Appendix A Fig. A1).

4 gy, Evolution and Systematics 20 (2016) 46–55

2a

satdassiitt1

at9tvl(Btrcsatas

eftgdrpvpipmamptb

2

te(sputmgcf

Table 1Differences in frequency of the 29 most common non-native species recorded at230 sites spread across the Swiss Alps in 2003 and 2009. Sites were stratified byelevation and were located along roadsides or at train stations between 200 and2400 m a.s.l. Species are ranked in descending order based on frequency in 2009.

Species Change in rankfrom 2003

Frequency 2009 Frequency 2003

Matricaria discoidea 0 122 156Bromus inermis 0 110 127Conyza canadensis 0 104 111Erigeron annuus 2 84 89Trifolium hybridum −1 77 103Euphorbia peplus −1 70 90Eragrostis minor 8 62 54Solidago canadensis 2 59 76Cerastium tomentosum 17 55 23Galinsoga ciliata −1 55 81Portulaca oleracea 2 51 62Buddleja davidii 2 50 57Amaranthus retroflexus −2 47 68Oxalis fontana −5 46 80Robinia pseudoacacia 5 45 49Senecio rupestris 1 43 53Oenothera biennis −1 35 53Impatiens parviflora 0 34 52Aurinia saxatilis 9 33 22Lepidium ruderale −1 28 51Echinochloa crus-galli −9 27 63Panicum capillare −1 26 31Reynoutria japonica −1 24 29Lupinus polyphyllus 0 23 25Veronica persica −18 22 86Artemisia verlotiorum −1 20 24Diplotaxis tenuifolia −1 20 24

8 T. Seipel et al. / Perspectives in Plant Ecolo

.2. Population distribution and structure of two model species at local scale

We selected two non-native species – E. annuus and S. canaden-is – as model species to test how well the regional-scale data with

coarse spatial and temporal resolution reflects fine-scale spa-iotemporal patterns, and to investigate the relationship betweenistributional limits and population turnover with population char-cteristics (patch size and density, distance between patches, plantize, and reproductive output). We chose E. annuus and S. canaden-is because they are two of the most common non-native speciesn Switzerland and Europe, have contrasting traits (S. canadensiss a rhizomatous perennial, and E. annuus is a tap-rooted annualo short-lived perennial), and similar and sharp elevation limits inhe Swiss Alps (Becker et al., 2005; Meyer and Schmid, 1999; Meyer,999; Trtikova et al., 2011).

In 2008, and 2010, we recorded all patches of E. annuusnd S. canadensis, along the two upper tributaries that mergeo form the Rhine River in southeastern Switzerland (46◦49′N◦24′E). The patches were located by cycling from the tops ofhe Lukmanier and Oberalp passes through the Vorderrhein riveralley, and from Ausserfera through the Hinterrhein river val-ey, to their confluence at Bonaduz, and from there to Sargans∼200 km in total length; see Supplementary material Appendix

Fig. A1). The transect followed marked bicycle routes andhe routes were chosen because they passed through areas rep-esentative of intermountain valleys of Switzerland, with largehanges in elevation and declining occurrence of non-nativepecies. The elevation of the transect varied from 480 m to 2050 mbove sea level, and the entire route was recorded as a GPSrack. In each year, the observations were made during Augustnd September, which is near the end of the growing sea-on.

Every time a patch of either S. canadensis or E. annuus wasncountered up to 50 m on either side of the bicycle route, theollowing parameters were recorded: GPS location at the patch cen-er, rectangular area covered (where a gap between individuals ofreater than 20 m was considered a separate patch), and averageensity of ramets within 1 m-square sample plots. We aimed toecord density of ramets in at least ten 1 m-square plots at everyatch. When a patch was greater than 10 m in length; plots or indi-iduals were randomly chosen, without replacement. In the largestatches (i.e. greater than 50 m long), we sampled density of ramets

n 1 m-square plots that totaled at 20% of the total length of theatch. In patches less than 10 m in length, a plot was placed everyeter. In 2010, plant height and the average number of capitula (E.

nnuus) or flowering branches (S. canadensis) per ramet were alsoeasured on a minimum of 5 and maximum of 20 individuals per

atch. Individuals were chosen by selecting a random point alonghe transect, going a random distance away from the transect lineut within the patch, and locating the nearest mature individual.

.3. Data analysis

All analyses were performed in R (R Core Team, 2013), usinghe lme4 package and MASS package for mixed models (Batest al., 2013; Venables and Ripley, 2002), the ggplot2 packageWickham, 2009) for graphics, the fields package for calculatingpatial distance (Nychka et al., 2013), and the vegan package forlant species composition dissimilarity (Oksanen et al., 2013). Wesed correlogram functions in the ncf package (Bjornstad, 2013)o examine patterns in spatial autocorrelation of residuals of all

odels in response to increasing distance based on the Swiss-rid coordinates of each site. When necessary we added spatialorrelation structure to mixed effect models using the glmmPQLunction in MASS (Dormann et al., 2007). Models were then reex-

Helianthus annuus 1 18 20Oxalis corniculata −6 16 26

amined to determine if spatial autocorrelation was reduced. Wealso examined dispersion in all the Poisson and binomial models,by comparing the Pearson correlation of squared residuals to thenumber of degrees of freedom.

2.3.1. Species richness patternsTo describe non-native species richness along elevation and

compare it between years we calculated three different general-ized linear mixed models. To control for the repeated sampling ofthe same sites in 2003 and 2009 all models include site as a ran-dom effect to account for the grouping. The three models of speciesrichness differed in how year was included (as a fixed effect): onlyelevation, elevation and year, and elevation and year plus the inter-action of these factors; all models had a Poisson distribution andlog link. We compared the different nested models using likelihoodratio tests. If the best model contained elevation, year and an inter-action, a significant difference exists in richness and the shape curvein response to elevation was different between years.

2.3.2. Probability of occurrence and population turnoverTo investigate whether the probability of occurrence (incidence)

of individual species along the elevation gradient changed between2003 and 2009, we fitted three nested generalized linear modelswith binomial distribution and logit link for each of the 29 mostcommon species (those found in at least 10% of plots, Table 1). Theresponse variable was presence-absence of a species, and the pre-dictor variables for each of three nested models were elevation,elevation and sampling year, and elevation, sampling year and theirinteraction. First we compared the models of elevation and eleva-

tion and year, then elevation and sampling year with elevation andsampling year including the interaction, using likelihood ratio tests.When year was not an important parameter in the model then theprobability of occurrence to be the same in between years. If year,

T. Seipel et al. / Perspectives in Plant Ecology, Evolution and Systematics 20 (2016) 46–55 49

500 1000 1500 2000 2500

0

10

20

30

40(a) All species

2003

2009

500 1000 1500 2000 2500

0

5

10

15

20

(b) Common species

Elevation m a.s.l .

Non

-nat

ive

richn

e ss

Fig. 1. Fitted generalized mixed-effects models of richness of non-native species in response to elevation and year of the survey (2003 and 2009) across 230 sites in theSwiss Alps for all species (a) and only common species found in at least 10% of plots (b). Gray shading indicates the 95% confidence interval for the coefficient estimates. Bothmodels used a poisson distribution, and identity of each site was fit as a random factor in the models to control for repeated measures. The Poisson regression equation fort Pearsor evatioo

eoe

otttedslapm

tpdWoo

2

scteydhrcu

pesileq

he all species (a) is ln(Yrichness) ∼ 3.87 − (0.00131 × Elevation) − (0.034 × Year), the

egression equation for the common species (b) is ln(Yrichness) ∼ 3.39 − (0.00141 × Elf freedom = 450.

levation and their interaction significantly predicted the responsef a species then the shape and location of the curve along thelevation gradient differed between years.

To investigate patterns of population turnover (local extinctionr colonization) of the 29 species, we calculated for each specieshe frequency of populations establishing or going extinct alonghe elevation gradient. To do this we first assigned a value of 1 tohose populations that turned-over between 2003 and 2009 (i.e.ither established or went extinct) and a value of 0 to those thatid not (i.e. that were present in both years). At sites where thepecies did not occur in either year we assigned NA. A generalizedinear model was then fitted to these data for each species using

binomial error distribution and logit link using elevation as theredictor variable. We compared each model to an intercept-onlyodel using likelihood ratio tests.To test the relationship of population turnover and abundance

o probability of occurrence we fitted turnover and abundance torobability of occurrence in response to elevation, using a Gaussianistribution for abundance, and binomial distribution for turnover.e also compared the correlation of abundance with probability

f occurrence, and the correlation of turnover and probability ofccurrence.

.3.3. Species compositionWe investigated changes in species composition at the 230 sites

pread across the Swiss Alps in two ways. The first was to measurehanges in species composition (beta-diversity changes in time) athe 230 Swiss-wide sites and to investigate possible regional andlevation trends. To do this we calculated dissimilarity betweenears for each site using the Jaccard (�j) and Simpson’s beta-issimilarity (�sim) indices (Koleff et al., 2003). The latter indexas the advantage of not being sensitive to differences in speciesichness (Koleff et al., 2003). We investigated changes in speciesomposition in response to elevation and river drainage of the sitesing permutational analysis of variance (Oksanen et al., 2013).

To investigate changes in species richness and species com-osition at individual sites, we summed the number of speciesstablishing or going extinct between 2003 and 2009 at a site (i.e.pecies turnover per site), and related these values to the changes

n species richness at sites between the sampling years. We fitted ainear mixed effects model to examine the population turnover atach site, in which the response variable was turnover at sites, anduartiles of the elevation range the fixed effect. To concentrate on

n residual squared = 215.5 and the residual degrees of freedom = 450. The Poissonn) − (0.0397 × Year), the Pearson residual squared = 199.8 and the residual degrees

the turnover of populations at a site and to account for differencesin species richness among sites between years (which was greaterat low elevation), site richness in 2009 was fit as a random effect. Tojudge how well the model with the fixed effect of elevation quar-tiles characterized the data we compared it to an intercept-onlymodel using a likelihood ratio test.

2.3.4. Population distribution and structure of two model speciesat a local scale

To generate presence-absence of the two species (S. canadensis,and E. annuus) along the elevation gradient, we converted the GPStrack of the route sampled into a point vector, where points werespaced at 200 m intervals giving 2785 points. If a study speciesoccurred within 100 m of a point then this point was assigned avalue of ‘present’, otherwise it was treated as ‘absent’. We spacedpoints at 200 m so as not to exceed the spatial resolution of the GISdata used in the analysis of occurrence.

To test whether the isolation of patches increased along the ele-vation gradient, we calculated the average distance from each patchof either E. annuus or S. canadensis to its four intra-specific nearest-neighbors using Euclidean distance. We used the average distanceto the four nearest populations as the response variable in a gener-alized linear mixed effect model fit via penalized quasi likelihood(glmmPQL) with elevation of each site as the predictor to determineif isolation increases with elevation. We used a Gaussian spatialautocorrelation structure and Gaussian distribution. We relied onthe Wald t-test to test significance.

To determine patch colonization and extinction we used theoccupancy data from 2008 and 2010, scoring patches present in2010 but not 2008 as colonization, and patches present in 2008but not 2010 as extinction. Patches that were extant in both yearswere assigned a zero. We fitted separate models of turnover, col-onization, and extinction with glmmPQL with the binary variablesas the response and elevation as the predictor variable to test if therelationships of colonization and extinction significantly differedalong the elevation gradient. Each model had a quasi-binomial dis-tribution and accounted for spatial correlation in the Gaussian formusing Swiss-grid coordinates. The significance of elevation in pre-dicting turnover, extinction and colonization were judged using

Wald t-tests.

To determine the reproductive output of each population of E.annuus and S. canadensis, and to assess whether reproductive out-put decreased in response to increasing elevation we calculated

50 T. Seipel et al. / Perspectives in Plant Ecology, Evolution and Systematics 20 (2016) 46–55

Matricaria discoidea

0.0

0.2

0.4

0.6

0.8

1.0P= 0.001

Bromus inermis

P= 0.11 6

Conyza canadensisP= 0.187 *

Erigeron annuus

P= 0.34 6

Trifolium hybridum

P= 0.01

Euphorbia peplus

0.0

0.2

0.4

0.6

0.8

1.0P= 0.007

Eragrostis minor

P= 0.346

Solidago canadensis

P= 0.017

Cerastium tomentosum

P= <0.001

Galinsoga ciliata

P= 0.001

Portulaca oleracea

0.0

0.2

0.4

0.6

0.8

1.0P= 0.097

Budd leja davidii

P= 0.217

Amaranthu s retroflexusP= 0.006

Oxali s fon tana

P= <0.001

Rob inia pseudo acacia

P= 0.489

Senecio rupestris

0.0

0.2

0.4

0.6

0.8

1.0P= 0.276

Oenothera biennis

P= 0.009

Impatiens parviflora

P= 0.018

Aurinia saxatilis

P= 0.118

Lepidium ruderale

P= <0.001 *

Echino chloa crus-galli

0.0

0.2

0.4

0.6

0.8

1.0

P= <0.001

Panicum capill are

P= 0.383

Reynou tria japon ica

P= 0.4

Lup inu s po lyph yllus

P= 0.794

Veron ica persica

P= <0.001

Artemisia verlotiorum

0.0

0.2

0.4

0.6

0.8

1.0

500 1500 2500

P= 0.449 *

Diplotaxis tenu ifolia

500 15 00 25 00

P= 0.46 6

Heli anthu s annuu s

500 150 0 250 0

P= 0.696

Oxali s corniculata

500 15 00 25 00

P= 0.06 1

Elevation m a.s.l.

Pro

babi

lity

ofoc

cure

nce

20092003

Probabili ty of Population t urnoverPopulation with >100 individuals

Fig. 2. Probability of occurrence of 29 non-native plant species that occurred in at least 10% of sites in 2003 and 2009 (dashed and solid lines, respectively) at 230 sites acrossthe Swiss Alps. The panels are arranged in decreasing relative species abundance of species in 2003. P-values indicate whether the probability of occurrence differs betweeny yeart ts ind

tttt(

ears and asterisks after P-values indicate a significant interaction of elevation andurnover (extinction and colonization) along the elevation gradient, with black poin

he reproductive output (fecundity) of each population. Reproduc-

ive output was calculated as the product of the area occupied byhe population (m2), the average density of plants (ramets/m2) andhe average number of capitula (E. annuus) or flowering branchesS. canadensis).

based on likelihood ratio tests. Gray points indicate the probability of populationicating populations of greater than 100 individuals.

3. Results

3.1. Species richness and population turnover at a regional scale

Across sites, the rank abundance of species, particularly of the20 most abundant species, was almost identical between years

gy, Evolution and Systematics 20 (2016) 46–55 51

(pswubt��y�ydvm

3

2wmammno(

vfataaeqqiuaOiFd(tapa

3

(p–ftsUtwcp(

Fig. 3. The response of population turnover, defined as either a local extinction orcolonization (a), and abundance (b) to probability of occurrence for the 29 mostcommon non-native species presented in Fig. 2 and indicated with gray lines. Toidentify the individual species see Supplementary material Appendix A Fig. A2. Inboth panels the thick black line indicates the overall response of all species as fit bygeneralized mixed-effects models. The gray shading around the responses to prob-

T. Seipel et al. / Perspectives in Plant Ecolo

Table 1). However, fewer populations of most species were sam-led in 2009 than in 2003. Thus, 103 species were recorded at fewerites, while only 58 species were more frequent, and 17 speciesere equally frequent. At the site level, the smaller number of pop-lations was reflected in fewer species per site in 2009 (Fig. 1),ut the shape of the elevational richness curve did not change;hus, despite differences between years (all species: likelihood ratio2

1 = 45.86, P < 0.001; 29 most common species: likelihood ratio2

1 = 45.12, P < 0.001), there was no significant interaction betweenear and elevation for all or only common species (likelihood ratio2

1 < 0.001, P = 1). The average site difference in richness betweenears was 2.7 species, but was largest at low elevations. The meanetection error of species at a site was 20% of richness at low ele-ation and declined with increasing elevation (see Supplementaryaterial Appendix A Fig. A1).

.2. Probability of occurrence and population turnover

Eighteen of the 29 most common species were less frequent in009 than in 2003, with an average decline in occurrence of 10% thatas greatest at lower to mid elevations (Fig. 2). Of the 29 most com-on species, some had an almost equal probability of occurrence

cross the elevation gradient (e.g. Matricaria discoidea, Bromus iner-is, Lupinus polyphyllus). The other 24 declined steeply towards aid-elevation range limit (Fig. 2). For six species, there was a sig-

ificant interaction of year and elevation, indicating a change notnly in frequency but also in the shape of the elevation distributionFig. 2; e.g. Conyza canadensis, Lepidium ruderale).

The turnover in species occupancy of sites between the two sur-ey dates was remarkably high, ranging from an average of 28%or Senecio rupestris to 80% for Helianthus annuus, with an aver-ge value across all species of 52%. For most species, populationurnover increased with elevation, with the increase often beingssociated with a sharp decline in species occurrence (Fig. 2). Theverage ratio of populations which turned over (colonized or wentxtinct) to those which remained extant was greatest in the thirduartile of the elevation gradient (ratio: q1 = 1.0, q2 = 1.5, q3 = 2.0,4 = 0.61), where most species reached their high elevation lim-

ts. The linear mixed-effect model of population turnover basedpon elevation quartiles fitted the data significantly better thann intercept-only model (likelihood ratio �2

3 = 12.81, P = 0.005).verall population turnover decreased gradually as the probabil-

ty of occurrence increased (r = −0.28, t = −13.5, df = 2098, p < 0.001;igs. 2 and 3a), although population turnover of some speciesecreased drastically as the probability of occurrence increasede.g. S. rupestris and L. polyphyllus; see Fig. 2 and the Supplemen-ary material Appendix A Fig. A2). Overall, abundance recorded on

scale of 1–10, 10–100, and 100+ was positively correlated withrobability of occurrence (r = 0.19, t = 9.13, df = 2098, p < 0.001), butlso varied among species (Fig. 3b).

.3. Species composition of sites

The species composition of sites varied between 2003 and 2009Supplementary material Appendix A Figs. A3 and A4). Four speciesresent in 2003 were not recorded in 2009, while three species

Caragana arborescens, Carpobrotus edulis, and Eschscholzia cali-ornica – were recorded for the first time in 2009. Sites that hadhe largest differences in composition of common species werepread across the entire study area and were not spatially clustered.sing the Simpson’s beta-dissimilarity index, which is insensi-

ive to richness differences, sites were 34.9% (s.d. = 23.2) dissimilar

hen considering all species, and 21.5% (s.d. = 22.1) different when

onsidering only the most common species. The species turnoverer site was greatest in the first and second elevation quartilesq1 = 7.7, q2 = 7.6,q3 = 4.8, and q4 = 3.7); elevation quartiles pre-

ability of occurrence is the 95% interval of the coefficient estimate. The abundanceclass is: 1 is 1–10 individuals, 2 is 10–100 individuals and 3 is greater that 100individuals.

dicted turnover at sites significantly better than an intercept-onlymodel (likelihood ratio �2

3 = 108.51, P < 0.001).

3.4. Population distribution and structure of two model species ata local scale

We recorded a total of 131 S. canadensis patches and 74 patchesof E. annuus, with both species occurring predominantly below1000 m a.s.l. The probability of occurrence of both species wasgreatest at low elevations, but patches were present in only asmall proportion of the available 2785 points, making the maxi-mum probability of occupancy 2% for Erigeron and 6% for Solidago.Distance to nearest neighbors of Erigeron and Solidago increasedtoward the high elevation range margin, but this pattern was no

longer significant for Erigeron after the most isolated patch at1800 m was excluded from the model (Supplementary materialAppendix B Fig. A2). The maxima of reproductive output per popu-lation were greatest at low elevations for both Solidago and Erigeron

52 T. Seipel et al. / Perspectives in Plant Ecology, Evolution and Systematics 20 (2016) 46–55

600 1000 1400 1800

0.0

0.2

0.4

0.6

0.8

1.0Erigeron annu us

turnoverextinctioncolonization

(a)

600 800 1000 1200 1400

0.0

0.2

0.4

0.6

0.8

1.0Soli dago canadensis

(b)

600 1000 14 00 180 0

1

2

3

4

5

6

7

(cap

itula

orflo

wer

ing

bran

ches

)

log1

0(C

umm

ulat

ive

o utp

utpe

rpa t

ch)

(c)

0

20

40

60

80

100

600 800 1000 1200 1400

1

2

3

4

5

6 (d)

0

20

40

60

80

100

Elevation a.s.l.

Bin

ned

perc

e nta

geof

max

imum

outp

ut

Pro

babi

lity

Fig. 4. Models of probability of turnover, extinction, and colonization between 2008 and 2010 for Erigeron annuus (a) and Solidago canadensis (b) along the local elevationg rodur each

c nd sta

(vtpps

tlrAetdFbpvsOtwcw

4

4r

t

radient in the Vorder and Hinter Rhine valleys of Switzerland. The cumulative repamets × fecundity (flowering branches for Solidago and capitula for Erigeron) andumulative reproductive output of all populations summed within elevation bins a

Fig. 4c and d). The cumulative reproductive output at lower ele-ations was much greater than at higher elevations as a result ofhe higher number of populations, and higher density and occu-ied area per population (Fig. 4c and d). The cumulative outputer elevation band declined along the elevation gradient for bothpecies.

For both species, climate, topography, and land use were impor-ant for predicting the species occurrence, although models hadow explanatory power, and predictor variables were highly cor-elated with elevation (Supplementary material Appendix B Table1). For E. annuus we recorded 16 colonization and 10 extinctionvents between 2008 and 2010, with no significant effect of eleva-ion upon turnover (t = 0.60, df = 72, p = 0.51), colonization (t = 0.31,f = 72, p = 0.75) or extinction (t = 0.42, df = 72, p = 0.67; Fig. 4a).or S. canadensis we recorded 26 colonizations and 10 extinctionsetween 2008 and 2010, with both turnover (t = 3.75, df = 129,

< 0.001) and probability of patch extinction increasing with ele-ation (t = 30.5, df = 129, p = 0.003); colonization, however, was notignificantly affected by elevation (t = 1.5, df = 129, p = 0.13; Fig. 4b).n average, it was the smaller populations that went extinct. Thus,

he E. annuus patches that disappeared covered on average 2 m2

hile the mean patch size was 16 m2; similarly, the patches of S.anadensis that went extinct occupied on average 0.5 m2, comparedith an overall mean of 182 m2.

. Discussion

.1. Stable elevation limits and patterns of non-native species

ichness

Monitoring the elevation range limits of plants species in moun-ain regions is important for detecting the effects of global change

ctive output (c and d) for each population in 2008 is equal to the area × density ofpopulation is represented by open black circles. Gray bars indicate the percent ofndardized to the bin with the largest cumulative output.

(Gottfried et al., 2012; Pauli et al., 2007). At a regional scale, mostspecies in the Swiss Alps had similar range limits in 2003 andin 2009, and most species declined sharply in occurrence withincreasing elevation (Fig. 2). As a consequence, the number of non-native species per site also declined strongly with elevation inboth 2003 and 2009, with very few species being recorded fromsites above 1500 m a.s.l. At low elevations we found fewer speciesin 2009 than 2003, while there was no difference at higher ele-vation. Similar high elevation range limits of the most commonnon-native species (Fig. 2), and hence patterns of non-native rich-ness along elevation gradients in the Swiss Alps (Fig. 1), suggestthat high elevation range limits were stable, at least over the six-year period of our study. This implies that most non-native speciesin the study area have expanded to near their niche boundaries,regionally, and therefore the patterns of declining occurrence andspecies richness with elevation are – at least not entirely – related toa non-equilibrium situation (Pysek et al., 2011). However, assum-ing the limits of species, at least regionally, are set by climate, thethreat of increased invasions into mountain areas may drasticallyincrease with climate change (Kueffer et al., 2013; Pauchard et al.,2009; Petitpierre et al., 2016).

4.2. Turnover: greater dynamics at the range margins

The balance between population extinction and colonizationhas long been recognized to influence the distributions of species(Brown and Kodric-Brown, 1977; Levins, 1969; Pulliam, 1988;Tilman and Kareiva, 1997). Although range limits and patterns

of richness of non-native species seem to be stable at a regionalscale in the Swiss Alps, the populations of most non-native specieswere very dynamic, with over half of all populations recorded ineither 2003 or 2009 not being recorded on the other survey date.

gy, Ev

Iassachsvsrjsippos

cw(vtwpthttL

rcmorria1srt2tl(1eaehlctoce

4

moc

T. Seipel et al. / Perspectives in Plant Ecolo

t was striking that the highest percentage of population turnovermong common species occurred towards the range limits of mostpecies, in the third quartile of the elevation range, which corre-ponds to a zone of rapid decline in occurrence (e.g. C. canadensisnd E. annuus). The commonest species in the fourth quartile wereold-tolerant species such as B. inermis, M. discoidea, and Trifoliumybridum, which had lower temperature indicator values than mostpecies (mean indicator value T = 3 for species in fourth quartileersus T = 4.3 for all species; Landolt, 2010). In contrast to most otherpecies, these species showed an increasing probability of occur-ence with elevation over the elevational range studied (Fig. 2). Butust like species that declined in occurrence as elevation increased,pecies that increased in probability of occurrence with increas-ng elevation also exhibited decreasing population turnover asrobability of occurrence increased (Fig. 3a), indicating that theiropulations were less stable towards the range limit. As probabilityf occurrence increased, populations became both larger and moretable (Fig. 3b).

The high population turnover at the range limit observed for S.anadensis at the regional scale was also detected at a local scale,ith much higher rates of extinction towards the range margin

Fig. 4b). In contrast, turnover of E. annuus populations did notary significantly with elevation (Fig. 4a). However, the popula-ion structure of both species changed toward the range limits,ith populations becoming more isolated from each other andropagule production declining precipitously (Fig. 4; Supplemen-ary material Appendix B Fig. A1). E. annuus and S. canadensisad similar range limits between 1100 and 1300 m a.s.l at bothhe regional and local scale, an elevation that corresponds withhe upper limit of the montane zone (typified by Fagus sylvatica;andolt and Urbanska, 2003).

The most probable reason for more dynamic populations at theange limit is that plants growing towards the range margin arelose to their physiological limits and populations are thereforeore vulnerable to fluctuations in abiotic conditions and stochastic

r extreme events resulting in population extinction, which givesise to greater turnover (Holt and Keitt, 2000; Lande, 1993). Theesult is a metapopulation structure of source-sink dynamics thats dominated by regional processes (Alexander et al., 2011; Doxfordnd Freckleton, 2012; Freckleton and Watkinson, 2002; Pulliam,988). Thus, for most lowland species, the range margin or inva-ion front might represent an equilibrium between extinction ande-establishment from propagules originating from core popula-ions at lower elevation (Brown and Kodric-Brown, 1977; Gaston,003; Holt and Keitt, 2000, 2005; Kawecki, 2008). Dispersal fromhe core (areas of high probability of occurrence) toward the rangeimit may promote a dispersal-induced stability of the range marginAbbott, 2011; Brown and Kodric-Brown, 1977; Carter and Prince,981; Carter and Prince, 1988; Gotelli, 1991). There is also geneticvidence suggesting that turnover of Erigeron populations is greatert the upper range limit than at lower elevations. Thus, Trtikovat al. (2011) found that genetic diversity among populations wasigher at higher elevations, while diversity within populations was

ower than in low elevation populations, suggesting more frequentolonization and extinction at higher elevations. The turnover athe margin may also explain why higher population growth rates,r at least more variable growth rates, can occur at the margin whenompared to the more stable core of the distribution (see Thuillert al., 2014).

.3. Detecting changes in the elevation limits of species

Monitoring the elevation limits of species can be a sensitiveeans of detecting effects of environmental change. Upward shifts

ver short time periods have been reported for many native herba-eous species in the European Alps (Grabherr et al., 1994; Lenoir

olution and Systematics 20 (2016) 46–55 53

et al., 2008; Pauli et al., 2007). Most of the non-native species inves-tigated in this study have even shorter life cycles, being mostlyannuals and short-lived perennials. Our results show that for suchspecies, elevation limits are not defined by a sharp boundary buta transition zone over which populations become increasinglyunstable (Gaston, 2003). An important methodological question,therefore, is whether (and if so, how) data from repeated surveysof short-lived species can provide information about distribu-tional changes. Our results show that despite high turnover ofpopulations, especially towards their upper limits, the elevationdistributions of most species changed little. Thus, their range dis-tributions appear to be stable, and reflect an underlying dynamicequilibrium of local extinction and colonization (Brown and Kodric-Brown, 1977; Hanski, 1997; Levins, 1969; MacArthur, 1972). Suchmetapopulation process-based shifts in demographic rates mayboth accelerate or slow down future changes in species distri-butions in response to climate and land use changes (Andersonet al., 2009; Holt and Keitt, 2000; Kirkpatrick and Barton, 1997;Schurr et al., 2012). In particular, because species boundaries arezones of declining abundance characterized by high populationturnover, merely recording the highest occurrence of a popula-tion does not provide reliable information about elevation limits.Instead, a relatively extensive sample of populations, including esti-mating the ratio of colonizations to extinctions of local populations,spanning the transitional zone is needed to be able to detect possi-ble directional changes in this zone (Normand et al., 2014; Schurret al., 2012). Given an adequate sampling design, however, mon-itoring of short-lived plants offers a potentially sensitive methodfor detecting changes in environmental conditions and forecastingthe impacts of global change (Pagel and Schurr, 2012).

Marginal habitats characterized by more extreme abiotic condi-tions such as mountain ecosystems are often less invaded (Alpertet al., 2000; Chytry et al., 2008; Pauchard et al., 2009). However, thisresistance to plant invasions may weaken with climate change, andchanging land use and composition of introduced floras and theirintroduction pathways (Chytry et al., 2012; Kueffer, 2010; Waltheret al., 2009). In particular, the current resistance might have lessto do with harsh climatic conditions at the site of invasion andmore to do with the fact that most non-native species have beenintroduced to anthropogenic areas that are separated by steep cli-matic gradients from less invaded habitats (Alexander et al., 2011).Effective proactive management against plant invasion risks in theleast invaded habitats such as mountains must therefore extendto the surrounding areas (McDougall et al., 2011b), and accountfor source-sink dynamics when controlling populations (Schreiberand Lloyd-Smith, 2009). An important component of such a proac-tive management strategy is the monitoring of species distributionsalong environmental gradients stretching from sites of introduc-tion to potential sites of invasion. We have shown in this studythat such monitoring must cover complete gradients rather thanonly invasion fronts because of high population turnover at species’ecological margins.

Acknowledgements

We thank Aud Halbritter, and Sonja Hassold, for helping withfield sampling. T.S. was support by an ETH Zurich doctoral grant.

Appendix A. Supplementary data

Supplementary data associated with this article can be found,in the online version, at http://dx.doi.org/10.1016/j.ppees.2016.04.001.

5 gy, Ev

R

A

A

A

A

A

A

B

B

B

B

C

C

C

C

C

C

R

C

D

D

D

F

G

G

G

G

G

4 T. Seipel et al. / Perspectives in Plant Ecolo

eferences

bbott, K.C., 2011. A dispersal-induced paradox: synchrony and stability instochastic metapopulations. Ecol. Lett. 14, 1158–1169.

lexander, J.M., Edwards, P.J., 2010. Limits to the niche and range margins of alienspecies. Oikos 119, 1377–1386.

lexander, J.M., Edwards, P.J., Poll, M., Parks, C.G., Dietz, H., 2009. Establishment ofparallel altitudinal clines in traits of native and introduced forbs. Ecology 90,612–622.

lexander, J.M., Kueffer, C., Daehler, C.C., Edwards, P.J., Pauchard, A., Seipel, T.,MIREN Consortium, 2011. Assembly of nonnative floras along elevationalgradients explained by directional ecological filtering. Proc. Natl. Acad. Sci. U.S. A. 108, 656–661.

lpert, P., Bone, E., Holzapfel, C., 2000. Invasiveness, invasibility and the role ofenvironmental stress in the spread of non-native plants. Perspectives in PlantEcology. Evol. Syst. 3, 52–66.

nderson, B.J., Akcakaya, H.R., Araujo, M.B., Fordham, D.A., Martinez-Meyer, E.,Thuiller, W., Brook, B.W., 2009. Dynamics of range margins formetapopulations under climate change. Proc. R. Soc. B: Biol. Sci. 276,1415–1420.

ates, D., Maechler, M., Bolker, B., Walker, S., 2013. lme4: Linear mixed-effectsmodels using Eigen and S4. R package version 1. 0–4. http://CRAN.R-project.org/package=lme4.

ecker, T., Dietz, H., Billeter, R., Buschmann, H., Edwards, P.J., 2005. Altitudinaldistribution of alien plant species in the Swiss Alps. Perspect. Plant Ecol. Evol.Syst. 7, 173–183.

jornstad, O.N., 2013. ncf: spatial nonparametric covariance functions. R packageversion 1. 1–5. http://CRAN.R-project.org/package=ncf.

rown, J.H., Kodric-Brown, A., 1977. Turnover rates In insular biogeography—effectof immigration on extinction. Ecology 58, 445–449.

aplat, P., Cheptou, P.O., Diez, J., Guisan, A., Larson, B.M.H., Macdougall, A.S.,Peltzer, D.A., Richardson, D.M., Shea, K., van Kleunen, M., Zhang, R., Buckley,Y.M., 2013. Movement, impacts and management of plant distributions inresponse to climate change: insights from invasions. Oikos 122, 1265–1274.

arey, M.P., Sanderson, B.L., Barnas, K.A., Olden, J.D., 2012. Nativeinvaders—challenges for science, management, policy, and society. Front. Ecol.Environ. 10, 373–381.

arter, R.N., Prince, S.D., 1981. Epidemic models used to explain biogeographicaldistribution limits. Nature 293, 644–645.

arter, R.N., Prince, S.D., 1988. Distribution limits from a demographic viewpoint.In: Davy, A.J., Hutchings, M.J., Watkins, A.R. (Eds.), Plant Population Ecology:the 28th Symposium of the British Ecological Society. Blackwell ScientificPublications, Sussex, UK, pp. 165–184.

hytry, M., Maskell, L.C., Pino, J., Pysek, P., Vilà, M., Font, X., Smart, S.M., 2008.Habitat invasions by alien plants: a quantitative comparison amongMediterranean, subcontinental and oceanic regions of Europe. J. Appl. Ecol. 45,448–458.

hytry, M., Wild, J., Pysek, P., Jarosík, V., Dendoncker, N., Reginster, I., Pino, J.,Maskell, L.C., Vilà, M., Pergl, J., Kühn, I., Spangenberg, J.H., Settele, J., 2012.Projecting trends in plant invasions in Europe under different scenarios offuture land-use change. Global Ecol. Biogeogr. 21, 75–87.

Core Team, 2013. R: A Language and Environment for Statistical Computing. RFoundation for Statistical Computing, Vienna, Austria.

rimmins, S.M., Dobrowski, S.Z., Greenberg, J.A., Abatzoglou, J.T., Mynsberge, A.R.,2011. Changes in climatic water balance drive downhill shifts in plant species’optimum elevations. Science 331, 324–327.

iez, J.M., D’Antonio, C.M., Dukes, J.S., Grosholz, E.D., Olden, J.D., Sorte, C.J.B.,Blumenthal, D.M., Bradley, B.A., Early, R., Ibánez, I., Jones, S.J., Lawler, J.J., Miller,L.P., 2012. Will extreme climatic events facilitate biological invasions? Front.Ecol. Environ. 10, 249–257.

ormann, C.F., McPherson, J.M., Araujo, M.B., Bivand, R., Bolliger, J., Carl, G., Davies,R.G., Hirzel, A., Jetz, W., Kissling, W.D., Kuehn, I., Ohlemueller, R., Peres-Neto,P.R., Reineking, B., Schroeder, B., Schurr, F.M., Wilson, R., 2007. Methods toaccount for spatial autocorrelation in the analysis of species distributionaldata: a review. Ecography 30, 609–628.

oxford, S.W., Freckleton, R.P., 2012. Changes in the large-scale distribution ofplants: extinction, colonisation and the effects of climate. J. Ecol. 100, 519–529.

reckleton, R.P., Watkinson, A.R., 2002. Large-scale spatial dynamics of plants:metapopulations, regional ensembles and patchy populations. J. Ecol. 90,419–434.

aston, K.J., 2003. The Structure and Dynamics of Species Ranges. OxfordUniversity Press, Oxford.

otelli, N.J., 1991. Metapopulation models—the rescue effect, the propagule rain,and the core-satellite hypothesis. Am. Nat. 138, 768–776.

ottfried, M., Pauli, H., Futschik, A., Akhalkatsi, M., Barancok, P., Benito Alonso, J.L.,Coldea, G., Dick, J., Erschbamer, B., Fernandez Calzado, M.R., Kazakis, G., Krajci,J., Larsson, P., Mallaun, M., Michelsen, O., Moiseev, D., Moiseev, P., Molau, U.,Merzouki, A., Nagy, L., Nakhutsrishvili, G., Pedersen, B., Pelino, G., Puscas, M.,Rossi, G., Stanisci, A., Theurillat, J.-P., Tomaselli, M., Villar, L., Vittoz, P.,Vogiatzakis, I., Grabherr, G., 2012. Continent-wide response of mountainvegetation to climate change. Nat. Clim. Change 2, 111–115.

rabherr, G., Gottfried, M., Pauli, H., 1994. Climate effects on mountain plants.Nature 369, 448–448.

uisan, A., Thuiller, W., 2005. Predicting species distribution: offering more thansimple habitat models. Ecol. Lett. 8, 993–1009.

olution and Systematics 20 (2016) 46–55

Guisan, A., Zimmermann, N.E., 2000. Predictive habitat distribution models inecology. Ecol. Modell. 135, 147–186.

Haider, S., Alexander, J., Dietz, H., Trepl, L., Edwards, P., Kueffer, C., 2010. The role ofbioclimatic origin, residence time and habitat context in shaping non-nativeplant distributions along an altitudinal gradient. Biol. Invasions 12, 4003–4018.

Haider, S., Kueffer, C., Edwards, P., Alexander, J., 2012. Genetically baseddifferentiation in growth of multiple non-native plant species along a steepenvironmental gradient. Oecologia 170, 89–99.

Hanski, I., 1997. Predictive and practical metapopulation models: the incidencefunction approach. In: Tilman, D.E., Kareiva, P.M.E. (Eds.), Spatial Ecology: TheRole of Space in Population Dynamics and Interspecific Interactions. PrincetonUniversity Press.

Hellmann, J.J., Byers, J.E., Bierwagen, B.G., Dukes, J.S., 2008. Five potentialconsequences of climate change for invasive species. Conserv. Biol. 22,534–543.

Holt, R.D., Keitt, T.H., 2000. Alternative causes for range limits: a metapopulationperspective. Ecol. Lett. 3, 41–47.

Holt, R.D., Keitt, T.H., 2005. Species’ borders: a unifying theme in ecology. Oikos108, 3–6.

Holt, R.D., Keitt, T.H., Lewis, M.A., Maurer, B.A., Taper, M.L., 2005. Theoreticalmodels of species’ borders: single species approaches. Oikos 108, 18–27.

Kawecki, T.J., 2008. Adaptation to marginal habitats. Annu. Rev. Ecol. Evol. Syst. 39,321–342.

Kirkpatrick, M., Barton, N.H., 1997. Evolution of a species’ range. Am. Nat. 150,1–23.

Koleff, P., Gaston, K.J., Lennon, J.J., 2003. Measuring beta diversity forpresence-absence data. J. Anim. Ecol. 72, 367–382.

Kueffer, C., McDougall, K., Alexander, J.M., Daehler, C., Edwards, P.J., Haider, S.,Milbau, A., Parks, C., Pauchard, A., Reshi, Z.A., Rew, L.J., Schroder, M., Seipel, T.,2013. Plant invasions into mountain protected areas: assessment, preventionand control at multiple spatial scales. In: Foxcroft, L.C., Pysek, P., Richardson,D.M., Genovesi, P. (Eds.), Plant Invasions in Protected Areas: Patterns, Problemsand Challenges. Springer, Dordrecht, Netherlands.

Kueffer, C., 2010. Transdisciplinary research is needed to predict plant invasions inan era of global change. Trends Ecol. Evol. 25, 619–620.

Lande, R., 1993. Risks of population extinction from demographic andenvironmental stochasticity and random catastrophes. Am. Nat. 142, 911–927.

Landolt, E., Urbanska, K.M., 2003. Our Alpine Flora. SAC Publications, Frutigen.Landolt, E., 2010. Flora Indicativa: Okologische Zeigerwerte und biologische

Kennzeichen zur Flora der Schweiz und der Alpen. Haupt Verlag, Bern,Switzerland.

Lauber, K., Wagner, G., 2001. Flora Helvetica, 3rd ed. Verlag Paul Haupt.Lennon, J.J., Turner, J.R.G., Connell, D., 1997. A metapopulation model of species

boundaries. Oikos 78, 486–502.Lenoir, J., Gegout, J.C., Marquet, P.A., de Ruffray, P., Brisse, H., 2008. A significant

upward shift in plant species optimum elevation during the 20th century.Science 320, 1768–1771.

Levins, R., 1969. Some demographic and genetic consequences of environmentalheterogeneity for biological control. Bull. Entomol. Soc. Am. 15, 237–240.

MacArthur, R.H., 1972. Geographic Ecology: Patterns in the Distribution of Species.Harper & Row, New York.

McDougall, K.L., Alexander, J.M., Haider, S., Pauchard, A., Walsh, N.G., Kueffer, C.,2011a. Alien flora of mountains: global comparisons for the development oflocal preventive measures against plant invasions. Divers. Distrib. 17, 103–111.

McDougall, K.L., Khuroo, A.A., Loope, L.L., Parks, C.G., Pauchard, A., Reshi, Z.A.,Rushworth, I., Kueffer, C., 2011b. Plant invasions in mountains: global lessonsfor better management. Mt. Res. Dev. 31, 380–387.

Meyer, A.H., Schmid, B., 1999. Experimental demography of rhizome populationsof establishing clones of Solidago altissima. J. Ecol. 87, 42–54.

Meyer, A.H., Schmid, B., 1999. Experimental demography of the old-field perennialSolidago altissima: The dynamics of the shoot population. J. Ecol. 87, 17–27.

Normand, S., Zimmermann, N.E., Schurr, F.M., Lischke, H., 2014. Demography as thebasis for understanding and predicting range dynamics. Ecography 37,1149–1154.

Nychka, D., Furrer, R., Sain, S., 2013. fields: Tools for spatial data. R package version6.9.1. http://CRAN.R-project.org/package=fields.

Oborny, B., Vukov, J., Csányi, G., Meszéna, G., 2009. Metapopulation dynamicsacross gradients—the relation between colonization and extinction in shapingthe range edge. Oikos 118, 1453–1460.

Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara, R.B.,Simpson, G.L., Solymos, P., Henry, M., Stevens, H., Wagner, H., 2013. vegan:Community Ecology Package. R package version 2. 0–8. http://CRAN.R-project.org/package=vegan.

Pagel, J., Schurr, F.M., 2012. Forecasting species ranges by statistical estimation ofecological niches and spatial population dynamics. Global Ecol. Biogeogr. 21,293–304.

Pauchard, A., Kueffer, C., Dietz, H., Daehler, C.C., Alexander, J.M., Edwards, P.J.,Arévalo, J.R., Cavieres, L.A., Guisan, A., Haider, S., Jakobs, G., McDougall, K.,Millar, C.I., Naylor, B.J., Parks, C.G., Rew, L.J., Seipel, T., 2009. Ain’t no mountainhigh enough: plant invasions reaching new elevations. Front. Ecol. Environ. 7,479–486.

Pauli, H., Gottfried, M., Reiter, K., Klettner, C., Grabherr, G., 2007. Signals of rangeexpansions and contractions of vascular plants in the high Alps: observations(1994–2004) at the GLORIA*master site Schrankogel Tyrol, Austria. GlobalChange Biol. 13, 147–156.

gy, Ev

P

P

P

P

P

P

S

S

S

S

Panov, V.E., Reineking, B., Robinet, C., Semenchenko, V., Solarz, W., Thuiller, W.,Vilà, M., Vohland, K., Settele, J., 2009. Alien species in a warmer world: risksand opportunities. Trends Ecol. Evol. 24, 686–693.

Wickham, H., 2009. Ggplot2: Elegant Graphics for Data Analysis. Springer, New

T. Seipel et al. / Perspectives in Plant Ecolo

auli, H., Gottfried, M., Dullinger, S., Abdaladze, O., Akhalkatsi, M., Benito Alonso,J.L., Coldea, G., Dick, J., Erschbamer, B., Fernandez Calzado, R., Ghosn, D., Holten,J.I., Kanka, R., Kazakis, G., Kollar, J., Larsson, P., Moiseev, P., Moiseev, D., Molau,U., Molero Mesa, J., Nagy, L., Pelino, G., Puscas, M., Rossi, G., Stanisci, A.,Syverhuset, A.O., Theurillat, J.-P., Tomaselli, M., Unterluggauer, P., Villar, L.,Vittoz, P., Grabherr, G., 2012. Recent plant diversity changes on europe’smountain summits. Science 336, 353–355.

etitpierre, B., McDougall, K., Seipel, T., Broennimann, O., Guisan, A., Kueffer, C.,2016. Will climate change increase the risk of plant invasions into mountains?Ecol. Appl. 26, 530–544.

ulliam, H.R., 1988. Sources, sinks and population regulation. Am. Nat. 132,652–661.

ulliam, H.R., 2000. On the relationship between niche and distribution. Ecol. Lett.3, 349–361.

ysek, P., Jarosík, V., Pergl, J., Wild, J., 2011. Colonization of high altitudes by alienplants over the last two centuries. Proc. Natl. Acad. Sci. U. S. A. 108, 439–440.

ysek, P., Jarosik, V., Hulme, P.E., Pergl, J., Hejda, M., Schaffner, U., Vila, M., 2012. Aglobal assessment of invasive plant impacts on resident species, communitiesand ecosystems: the interaction of impact measures, invading species’ traitsand environment. Global Change Biol. 18, 1725–1737.

chreiber, S.J., Lloyd-Smith, J.O., 2009. Invasion dynamics in spatiallyheterogeneous environments. Am. Nat. 174, 490–505.

churr, F.M., Pagel, J., Cabral, J.S., Groeneveld, J., Bykova, O., O’Hara, R.B., Hartig, F.,Kissling, W.D., Linder, H.P., Midgley, G.F., Schroeder, B., Singer, A.,Zimmermann, N.E., 2012. How to understand species’ niches and rangedynamics: a demographic research agenda for biogeography. J. Biogeogr. 39,2146–2162.

eipel, T., Kueffer, C., Rew, L.J., Daehler, C.C., Pauchard, A., Naylor, B.J., Alexander,J.M., Edwards, P.J., Parks, C.G., Arevalo, J.R., Cavieres, L.A., Dietz, H., Jakobs, G.,

McDougall, K., Otto, R., Walsh, N., 2012. Processes at multiple scales affectrichness and similarity of non-native plant species in mountains around theworld. Global Ecol. Biogeogr. 21, 236–246.

orte, C.J.B., 2013. Predicting persistence in a changing climate: flow direction andlimitations to redistribution. Oikos 122, 161–170.

olution and Systematics 20 (2016) 46–55 55

Stohlgren, T.J., Owen, A.J., Lee, M., 2000. Monitoring shifts in plant diversity inresponse to climate change: a method for landscapes. Biodivers. Conserv. 9,65–86.

Thuiller, W., Albert, C., Araujo, M.B., Berry, P.M., Cabeza, M., Guisan, A., Hickler, T.,Midgely, G.F., Paterson, J., Schurr, F.M., Sykes, M.T., Zimmermann, N.E., 2008.Predicting global change impacts on plant species’ distributions: futurechallenges. Perspect. Plant Ecol. Evol. Syst. 9, 137–152.

Thuiller, W., Münkemüller, T., Schiffers, K.H., Georges, D., Dullinger, S., Eckhart,V.M., Edwards, T.C., Gravel, D., Kunstler, G., Merow, C., Moore, K., Piedallu, C.,Vissault, S., Zimmermann, N.E., Zurell, D., Schurr, F.M., 2014. Does probabilityof occurrence relate to population dynamics? Ecography 37, 1155–1166.

Tilman, D.E., Kareiva, P.M.E., 1997. Spatial Ecology: The Role of Space in PopulationDynamics and Interspecific Interactions. Princeton University Press.

Travis, J.M.J., 2004. A method for simulating patterns of habitat availability at staticand dynamic range margins. Oikos 104, 410–416.

Trtikova, M., Güsewell, S., Baltisberger, M., Edwards, P.J., 2011. Distribution,growth performance and genetic variation of Erigeron annuus in the Swiss Alps.Biol. Invasions 13, 413–422.

UNEP, 2009. Climate change science compendium 2009. In: Programme.EarthPrint, Nairobi.

Venables, W.N., Ripley, B.D., 2002. Modern Applied Statistics with S, 4th ed.Springer, New York.

Walther, G.-R., Roques, A., Hulme, P.E., Sykes, M.T., Pysek, P., Kühn, I., Zobel, M.,Bacher, S., Botta-Dukát, Z., Bugmann, H., Czúcz, B., Dauber, J., Hickler, T.,Jarosík, V., Kenis, M., Klotz, S., Minchin, D., Moora, M., Nentwig, W., Ott, J.,

York.