Substitutes for grazing in semi-natural grasslands--do mowing or mulching represent valuable alternatives to maintain vegetation structure?

Substitutes for grazing in semi-natural grasslands – do mowing

or mulching represent valuable alternatives to maintain

vegetation structure?

Romermann, Christine1�; Bernhardt-Romermann, Markus

2; Kleyer, Michael

3&

Poschlod, Peter4

1Institute of Physical Geography, University of Frankfurt, D-60438 Frankfurt, Germany; 2Department Silviculture and

Forest Ecology of the Temperate Zones, Faculty of Forest Sciences and Forest Ecology, Georg-August University

Gottingen, D-37077 Gottingen, Germany; E-mail [email protected]; 3Landscape Ecology

Group, Institute of Biology and Environmental Sciences, University of Oldenburg, D-26111 Oldenburg, Germany; E-mail

[email protected]; 4Institute of Botany, University of Regensburg, D-93040 Regensburg, Germany;

E-mail [email protected];�Corresponding author; E-mail [email protected]

Abstract

Question: Which management treatments are suitable toreplace historically applied grazing regimes? How and whydoes vegetation structure change following changes in man-agement?

Location: Semi-natural calcareous dry grasslands insouthwest Germany.

Methods: We analysed changes in floristic and functionalcomposition induced by different management treatments(grazing, mowing, mulching, succession) in long-term experi-mental sites. First, floristic and functional distances betweenthe initial conditions and the following years were deter-mined. Second, we used RLQ analyses to include data onabiotic conditions, vegetation composition and functionaltraits in one common analysis. Finally, we applied clusteranalyses on RLQ species scores to deduce functional groups.

Results: In contrast to the historical management regimeof grazing, all alternative management treatments led tochanges in floristic and functional composition, dependingon their intensity with respect to biomass removal. Thedistance analyses showed that mulching twice per year andmowing did not lead to strong changes in floristic orfunctional composition. However, RLQ analysis clearlyprovided evidence that only the grazed sites are in equili-brium, indicating that vegetation change still goes ahead.

Conclusions: The current study clearly shows that RLQ isa powerful tool to elucidate ongoing processes that mayremain hidden when separately analysing floristic andfunctional data. Alternative management treatments arenot appropriate to sustain the typical disturbance dy-namics of species-rich semi-natural grasslands. The lessfrequent an alternative management treatment is withrespect to biomass removal, the less the floristic andfunctional structure can be maintained.

Keywords: cross compliance; functional traits; RLQ ana-lyses; succession; vegetation response.

Nomenclature: Wisskirchen & Haeupler (1998)

Abbreviations: GLM5 generalised linear model; GR5

grazing; MO5mowing; 2M5mulching twice per year;M15mulching once per year; M25mulching every sec-ond year; SLA5 specific leaf area; US5 succession

Introduction

Semi-natural grasslands that have been used at lowintensities for centuries are extremely valuable for nat-ure conservation due to their high diversity in floraand fauna (WallisdeVries et al. 2002). During the lastdecades, such grassland sites have been increasinglyabandoned or, alternatively, management treatmentshave changed because extensive land-use practices,such as grazing, did not prove to be of economic valuein the actual agricultural landscape (Watkinson & Or-merod 2001; Poschlod&WallisDeVries 2002; Poschlodet al. 2005). Typical plant species of these habitats arespecifically adapted to the management history (Dutoitet al. 2005; Poschlod et al. 2008) and require regulardisturbance management, including nutrient removalto avoid becoming overgrown by more competitivespecies. Hence, a major goal of nature conservation isto find ameans to keep these sites open and tomaintainthe typical species composition using cost-effectivealternative management treatments.

An EU Directive regulates the ‘‘minimal’’ man-agement required to keep abandoned grassland sitesin good agricultural and ecological condition

Journal of Vegetation Science 20: 1086–1098, 2009& 2009 International Association for Vegetation Science

(CC – Cross Compliance): either mulching once ayear (5mowing without removal of the phytomass)or mowing every second year must be applied. Itis, however, not clear whether these treatmentsare appropriate to maintain the typical vegetationstructure of semi-natural grasslands. For example,Moog et al. (2002) showed that mulching might leadto auteutrophication rather than conservation ofspecies-rich grasslands due to litter accumulationand decomposition.

In the current study, we used long-term data toanalyse whether mowing or mulching allow preserva-tion of the vegetation structure of formerly grazedsites. In addition, we included a site with succession,which is often the fate of abandoned grasslands oneconomically unprofitable sites. Our overall goal is toreveal and understand the mechanisms and processesthat determine the vegetation structure induced bydifferent management regimes using functional ana-lyses. The functional approach is based on theassumption that plants with similar ecologically re-levant trait attributes will respond to environmentalchanges in comparable ways (McIntyre et al. 1995).Hence, if a habitat maintains its functional character-istics we can conclude that the vegetation structuredid not change due to altered management (called‘‘functional equilibrium’’ and ‘‘floristic equilibrium’’in the following). Alternatively, if the functionalcharacteristics change we can conclude that the ap-plied management is not appropriate to maintain theoriginal vegetation structure.

Specifically, we focus on the following ques-tions. (1) Following the hypothesis that traditionallygrazed sites are in functional equilibrium, we askwhich management treatment will yield commu-nities whose floristic and functional composition aremost similar to the traditional grazing regime. Weexpect that changes in functional composition (andthus also floristic composition) will occur on thosesites where the effect of the newly established man-agement treatment on vegetation dynamics differsmost from the original management treatment.Similarity analyses are used to separately analysefloristic and functional changes over the last 30years. We analyse plant functional traits related topersistence and regeneration (see also Westoby et al.2002). (2) We want to understand how and why ve-getation structure has developed following changesin management. As changes in environmental con-ditions are correlated to changes in trait responses,we used RLQ analyses. This method simultaneouslycombines data on species composition, species traitsand environmental parameters in one commonmultivariate analysis (Doledec et al. 1996). RLQ

provides a covariance matrix that reveals which traitattributes are positively or negatively related to thetime gradient and subsequently to changes in en-vironmental conditions. We are dealing with a timeseries of vegetation data covering 30 years, whichlacks corresponding information on soil nutrientsand moisture for the whole period. Thus, as directmeasurements are missing, we used weighted Ellen-berg indicator values as substitutes (Ellenberg et al.2001), although this may increase the risk of circu-larity (but compare Ozinga et al. 2005, 2007).Evidence for the accuracy of these indicator values isprovided by several studies that report a close cor-relation between indicator values and correspondingmeasurements of environmental variables on largegradients (Schaffers & Sykora 2000; Diekmann2003). As suggested by Wamelink et al. (2002), weused Ellenberg indicator values for comparisonsonly within the same vegetation type. Furthermore,the application of these values allows us to demon-strate the usefulness of the applied methods forwhich information on environmental parameters isrequired. As additional factors, cover of moss andlitter and the time gradient were also included.

The presented framework presents a novelmethodological contribution to analyse functionaldevelopment over time: shifts in trait compositionprovide an indication of the underlying processes ofvegetation dynamics under different managementregimes, such as competitive exclusion of speciesthat are less adapted to the management history.These processes may remain hidden when focussingonly on floristic composition. From an applied pointof view, the study offers a protocol on how to eval-uate the effect of alternative management regimeson semi-natural grassland vegetation and may pro-vide support for management recommendations forabandoned sites.

Material and Methods

Data

Long-term vegetation dataWe used long-term vegetation data covering

30 years (1974-2004) for a calcareous grassland(Gentiano-Koelerietum) in St. Johann in the Swa-bian Alb in southwest Germany (760ma.s.l., annualprecipitation 1000mm, mean temperature 7.61C).Until 1974, the site was used as low-intensity sheeppasture. When grazing ceased, the management re-gimes used were mulching once per year (in late

- Substitutes for grazing in semi-natural grasslands - 1087

summer), twice per year (early and late summer) orevery second year (in late summer). In addition, onepart of the site was left fallow (succession). Grazingcontinued on a remnant of this site. Mowing wasonly introduced in 1981 on a formerly abandonedsite. The original aim of the experiment (set up byProf. Karl-Friedrich Schreiber in 1974) was to studythe suitability of different management techniquesto retain these grasslands open and to maintain theirhigh biodiversity.

Per site and per management treatment, onelong-term quadrat of 5m�5m was established in1974 and monitored regularly (every 3-5 years) usingpercentage cover following the Schmidt scale(Schmidt 1974). Comprising only one plot of 25m2

per treatment, the experimental design established inthe 1970s does not reflect the spatial variabilitywithin these sites (Schreiber 1997; Moog et al. 2002).However, this 30-year time sequence is much moreimportant for the focus of this study; several ‘‘snap-shots’’ in space focusing on one or a very few yearsonly would not have revealed clear results.

Appendix S1 gives an overview of the occurringspecies found in the different management treatments.

Environmental conditionsWe calculated mean weighted indicator values

for nutrient status (N), moisture (F), soil reaction(R), light (L) and temperature (T) for each plot,based on the cover of the species present in thecommunity. The Ellenberg indicator values are spe-cies-specific scores ranging from 1 to 9 (or 1-12 formoisture) and estimate the optimum ecologicaloccurrence of species along environmental gradients(Ellenberg et al. 2001). Moreover, the thicknesses ofthe moss and litter layers were included.

Life-history traits of plant speciesLife-history traits were extracted from the

LEDA Traitbase (Knevel et al. 2003; Kleyer et al.2008; http://www.leda-traitbase.org), from BioPop(Jackel et al. 2006), BiolFlor (Klotz et al. 2002) andCloPla3 (Klimes & Klimesova 1999; http://clopla.butbn.cas.cz). The traits and their assignment to theprocesses of regeneration and persistence are sum-marised in Table 1.

As these databases contain many entries perspecies and trait, the following aggregation ruleswere used. Per species, the numeric trait values wereaggregated by taking the means of all values presentin the databases. If several different entries for cate-gorical traits were present, we aggregated andranked them according to appropriate rules (seeTable 2, e.g. lateral spread has the categorieso0.01,0.01-0.25, 40.25m; if there were several entries perspecies, the maximal lateral spread was used).

Only those species for which a complete traitmatrix was available were included in the analyses(per management treatment 480% of all species).

Data analyses

Floristic and functional distancePer management treatment, floristic and func-

tional developments were calculated using Bray-Curtisdistance:

D1j ¼Xjðxi1 � xijÞj=

Xðxi1 þ xijÞ

where xi1 is abundance of species i or trait i in the firstyear, and xij is abundance of species i or trait i in jthyear. Floristic distances between initial and sub-sequent years were calculated based on raw data. Toinclude management-related community-aggregated

Table 1. Life-history traits related to regeneration and persistence and their trait stages. Traits were extracted from the LEDAtraitbase (Kleyer et al. 2008), BioPop (Jackel et al. 2006), CloPla (Klimes & Klimesova 1999) and BioFlor (Klotz et al. 2002).

Process Trait Categorisation Species-level aggregation

Persistence Branching No – yes Yes4noCanopy height Metric MeanSLA Metric MeanSeed mass Metric MeanLeaf distribution Rosette, semi-rosette- regularly n.a.Seed shedding start Metric MeanSeed shedding end Metric MeanSeed shedding duration Metric MeanGrowth form Geophytes, therophytes,

hemicryptophytes, chamaephytesn.a.

Regeneration Bud bank depth o� 10 cm, � 10–0 cm,40 cm The deeper the betterNo. of buds Metric MeanLateral spread o0.01m, 0.01mo 0.25m,40.25m The further the better

Vegetative dispersules

n.a.5 not applicable.

1088 Romermann, C. et al.

plant traits, trait values were weighted by speciesabundances using

traitagg ¼X

xi � traiti

where xi is the abundance of species i, and traiti is thetrait value of species i (Garnier et al. 2007). We choseto use weighted trait values instead of un-weightedvalues because differently managed sites have a verylarge set of species in common; therefore, the use ofun-weighted values would result in nearly identical si-milarity curves for the different treatments. Moreover,the different treatments led to strong shifts in speciesabundances, but only in some cases to species exclu-sions or new arrivals. Grime (1998) and Cingolaniet al. (2007) suggest that, in a given plant assemblage,dominant species should have trait values better fittedto environmental conditions than minor species; theirimportance should, therefore, not be down-weightedby using only presence/absence data.

To perform computations for the smoothingline, the function lowess was used, which uses locallyweighted polynomial regression (Cleveland 1981).The initial fit was done using weighted least squares.In an iteration process, further weighted fits weredone using the product of the weights from the

proximity of the x values and case weights derivedfrom the residuals at the previous iteration.

The impact of different management treatmentson functional and floristic development was in-vestigated using generalised linear models (GLM) totest for the direction (and significance) of the slope.Positive slopes indicate that the management treat-ment induced floristic and functional changes overtime, while non-significant slopes indicate equili-brium states.

Changes in functional trait composition over time

To combine floristic, functional and environ-mental data in one common analysis (for a descriptionof included data see above), three table ordinations(RLQ, Doledec et al. 1996) were used. Prior to ordi-nation analyses, percentage cover values of thespecies�site matrices were arcsine-square root trans-formed. The sites�environment matrices werestandardised by adjusting the values to zero mean andunit deviance to make parameters of different scalescomparable (weighted Ellenberg indicator values andcover of moss and litter). We extracted the best subsetof environmental variables that should be includedin one common ordination analyses by using the

Table 2. Covariance between the time gradient (environmental table R) and the traits (table Q) for the different managementtreatments. Covariance increases with the intensity of the relationship between the time gradient and the species traitsthrough the species abundance table L. Different symbols indicate significant differences from random distribution with���Po0.001, ��Po0.01, �Po0.05,1Po0.1. Abbreviations in trait list: LD5Leaf distribution.

Grazing Mowing Mulching twiceper year

Mulching onceper year

Mulching every2nd year

Succession

Number of buds � 0.055 � 0.038 � 0.008 � 0.061 0.144��� 0.201���

No branching 0.354� 0.022 0.085 � 0.051 � 0.1581 � 0.204�

Branching � 0.043� � 0.002 � 0.018 0.006 0.013 0.020Bud banko–10 cm � 0.0371 0.010 � 0.004 � 00.053 0.030 0.048Bud bank � 10–0 cm 0.504�� � 0.355��� � 0.061 � 0.185��� � 0.1771 � 0.207���

Bud bank40 cm � 0.057 0.192 0.092 0.3251 � 0.2081 � 0.463��

Canopy height 0.044 � 0.115��� 0.023 0.005 0.178��� 0.236���

Chamaephytes � 0.027 0.232 0.016 � 0.510� � 0.107 0.115Geophytes � 0.170 0.128 � 0.200 0.481 � 0.001 � 0.704�

Hemicryptophytes 0.006 � 0.016 0.006 0.0271 0.012 0.016Therophytes � 0.015 � 0.209 � 0.228 � 0.4641 � 0.281 � 0.572�

Grasses 0.057 � 0.050 � 0.009 0.062 0.0611 0.109Herbs � 0.058� 0.076 0.007 � 0.071 � 0.073 � 0.148Lat. spread:o0.01m � 0.003 � 0.061 � 0.039 0.019 � 0.200��� � 0.137��

Lat. spread: 0.01–0.25m � 0.025 0.049 0.050 � 0.037 0.157 0.165Lat. spread40.25m 0.079 0.986��� 0.126 0.233�� 0.381��� � 0.487���

LD: regular 0.016 0.025 0.099� 0.0621 0.276�� 0.393���

LD: semi-rosette � 0.017 � 0.051 � 0.063 � 0.133� � 0.214� � 0.223��

LD: rosette 0.011 0.2811 0.085 0.2651 � 0.345��� � 0.527���

Seed mass � 0.036 � 0.0681 � 0.041 � 0.043 � 0.025 0.028Seed shedding duration � 0.093 � 0.040 � 0.077� � 0.008 � 0.075� � 0.208���

Seed shedding end � 0.082� � 0.039 � 0.124�� � 0.143�� � 0.267��� � 0.390���

Seed shedding start � 0.029� � 0.050 � 0.024 � 0.0721 0.044 0.253���

SLA 0.010 0.101� 0.0761 0.048 0.184��� 0.101�

Non-woody 0.0181 � 0.018 0.007 0.041 � 0.011 0.009Woody � 0.091 0.198��� � 0.062 � 0.543��� 0.070 0.098�

- Substitutes for grazing in semi-natural grasslands - 1089

maximum rank correlation with community dissim-ilarities. This procedure is explained in Clarke &Ainsworth (1993). The species�trait matrices wereprepared using log-transforming seed mass. Further-more, we excluded interrelated traits from theanalyses (threshold value for interrelated traits:Kendall’s t40.5, Po0.05).

RLQ represents a double co-inertia analysis oftwo matrices R (sites�environment) and Q (spe-cies�traits), with a link expressed by a contingencytable L (species�sites). The ‘‘raw’’ matrices do notgo into RLQ analyses but duality diagrams receivedfrom a cluster analysis (CA) on the L table (CA-L),a constrained mixed ordination (Hill Smith) on theQ table using the column scores of the CA-L asconstraints (MO-Q/L) and a constrained mixed or-dination on the R table using the row scores of theCA-L (MO-R/L). Within the RLQ procedure, twoco-inertia analyses are carried out: one between theMO-R/L and CA-L and one between MO-Q/L andCA-L. An iterative process follows to find the RLQaxes that display maximum covariance between theQ and the R axes (co-variance analyses betweenthe axes from the two co-inertia analyses). For moredetails, including a graphical illustration of themethod, see Doledec et al. (1996).

To find management-related trait changes overtime, we extracted the cross matrix generated by RLQanalyses (the covariance between R and Q) carriedout separately for every management treatment butusing the same trait and environmental sets. The cov-ariance increases with the intensity of the relationshipbetween the environmental and the species traits ta-bles through the species abundance table (Ribera et al.2001). The global significance of this relationship wastested using random permutation tests, where the nullhypothesis was independence of the two tables R andQ (compare also Dray et al. 2002). Covariance valueswere tested for significance using permutation tests:RLQ analyses were done on 1000 randomised traittables. The randomly generated cross matrices weretested against the real cross matrix.

To compare the effect of different managementtreatments on functional composition with respectto environmental changes, deduced covariances be-tween traits and time were compared between allstudied management treatments using Kendall’s tcorrelation analyses.

Changes in functional groups over timeWe applied cluster analyses on RLQ species

scores to deduce functional groups. For each man-agement treatment, RLQs were re-calculated usingonly optimal trait sets to maximise RLQ fits. Opti-

mal trait sets were found using iterative RLQ, whichselects those traits from a wider set of traits thatare optimal for describing the ecosystem responseon a given environmental gradient (Bernhardt-Romermann et al. 2008).

We used k means clustering, choosing the opti-mal numbers of clusters k with respect to clusterstability. Cluster stability was assessed using non-parametric bootstrap methods (Hennig 2007). Weassessed the stability of a cluster of the initial clus-tering with respect to a new clustering using theJaccard similarity coefficient. Bias and variation canbe expressed by the maximum Jaccard coefficientbetween the set of all the points and the most similarcluster in the actually obtained clustering. The meanmaximal Jaccard coefficient g[0, 1] can be inter-preted as indicating the stability of the originalcluster (with g40.75 for stable clusters, go0.5 fordissolved clusters).

Plant functional groups (here: clusters) werecharacterised by mean values of numerical traits anddominating categories of categorical traits. We testedfor statistical differences using randomisation tests fornominal values and chi2 tests for categorical traits.

Spatial and temporal autocorrelation may occurwhen analysing long-term vegetation data of severalsites. Autocorrelation tends to inflate the actual num-ber of degrees of freedom in statistical tests, leading tothe acceptance of some hypotheses that should be re-jected. However, RLQ analysis does not includestatistical tests and does not test null hypotheses,therefore a correction for autocorrelation is not ne-cessary (compare also Thuiller et al. 2006).

All calculations were made using the librariesADE4 (Chessel et al. 2004), vegan (Oksanen et al.2007) and fpc (Hennig 2006) in the open-source Rsoftware version 2.5.1 (R Development Core Team2004).

Results

Environmental, floristic and functional distance

The floristic development of the treatments,grazing, mulching twice per year and mowing, didnot change over time (floristic equilibrium), and theslopes of their regression lines were not significant(Fig. 1). The management treatments mulching onceper year and every second year and the successionled to floristic changes over time, where the slopeswere positive and highly significant (Po0.001).

Similarly, the functional trait composition ofthe treatments, grazing, mowing and mulching twice

1090 Romermann, C. et al.

per year, did not significantly differ between the firstand subsequent years (non-significant slopes; func-tional equilibrium). Less frequent mulching and thesuccession treatment led to significant changes infunctional composition (positive regression slopeswith Po0.05).

Changes in functional trait composition over time

Table 2 provides an overview on the importanceof trait changes over time for the different manage-ment treatments: presented covariances (trait-timecovariances) increased with the intensity of the re-lationship between the environmental and the speciestraits tables through the species abundance table.These relationships were highly significant for allmanagement treatments (permutation test:Po0.001).

For the grazing treatment, most traits onlyshowed low (for most traits the absolute values ofcovariance, i.e. |cov|o0.1) and non-significant covar-iance values, indicating that traits did not stronglychange over time. Only the trait attributes ‘‘nobranching’’, and ‘‘bud bank between � 10 and 0 cm’’showed a clear significant increase over time. Also, on

sites with mowing and mulching twice per year, veryfew traits showed clear significant changes over time(when assuming |covariance|40.1 as threshold). Inthe mowing treatment, species with high capacity forlateral spread (40.25m) noticeably increased overtime, while the importance of low canopy heights andintermediate bud bank depths decreased. Mulchingtwice per year led to a shift towards an earlier seed-shedding period.

The less intensive management treatments,mulching once per year, mulching every second yearand succession, provoked clear changes in severaltraits over time. Again, using |covariance|40.1 as athreshold value gives the following results: withmulching once per year, a total of eight traits clearlychanged over time; with mulching every secondyear, 12 traits changed; and with succession 16 traitschanged. For a detailed overview on trait responsesover time see Table 2.

Table 3 shows the results from analyses corre-lating the trait-time covariances of the differentmanagement treatments with each other. Strongdifferences in changes in trait values occurredover time. Trait-time covariances of the treatments

1975 1980 1985 1990 1995 2000 2005

0.0

0.2

0.4

0.6

0.8

floristic development

years

Bra

y−C

urtis

dis

sim

ilarit

y

grazingmowingmulching twicemulching oncemulching every2nd yearsucession

1975 1980 1985 1990 1995 2000 2005

0.00

0.02

0.04

0.06

0.08

0.10

functional development

years

Bra

y−C

urtis

dis

sim

ilarit

y

grazingmowingmulching twicemulching oncemulching every2nd yearsucession

Fig. 1. Floristic and functional development of the differently managed sites in St. Johann. Bray-Curtis dissimilarity valuesgive the floristic or functional distance for the different years compared to the initial floristic or functional composition in1975 (mowing started in 1982). Asterisks (�) indicate significant slopes (deduced from GLM analyses).

Table 3. Correlation between the time-related trait responses (R-Q covariance) between the different management treatments.Kendall’s tau is shown throughout. Different symbols indicate significance at ���Po0.001, ��Po0.01, �Po0.05,1Po0.1, n.s.not significant. xMowing was not directly applied after grazing has ceased, but after 5 years of succession.

Grazing Mowing� Mulching twice per year Mulching once per year Mulching every 2nd year Succession

Grazing 1Mowingx � 0.13 n.s. 1Mulching twice per year 0.36�� 0.44�� 1Mulching once per year 0.17 n.s. 0.271 0.39�� 1Mulching every 2nd year 0.10 n.s. 0.16 n.s. 0.30� 0.21 n.s. 1Succession 0.08 n.s. � 0.06 n.s. 0.22 n.s. � 0.09 n.s. 0.58��� 1

- Substitutes for grazing in semi-natural grasslands - 1091

grazing and mulching twice per year were sig-nificantly related, hence traits responded similarly(i.e. in the same direction) over time. In contrast,trait-time covariances between grazing andmowing, mulching once per year, every second yearor succession were not significantly correlated,indicating that functional development was notcomparable.

Changes in functional groups over time

The cluster analyses of species scores deducedfrom the iterative RLQ analyses showed that thegrazing sites are in functional equilibrium (Fig. 2A):there were no changes in trait groups over time buttwo functional groups co-occurred over the wholetime gradient (indicated by two clusters located in

A Grazing (original treatment)

RLQ axis 1

RLQ

axi

s 2

Cluster 1Cluster 2

time

B Mowing

RLQ axis 1

RLQ

axi

s 2

Cluster 1Cluster 2Cluster 3Cluster 4

time

C Mulching twice per year

RLQ axis 1

RLQ

axi

s 2

Cluster 1Cluster 2Cluster 3

time

D Mulching once per year

RLQ axis 1

RLQ

axi

s 2

Cluster 1Cluster 2Cluster 3

time

E Mulching every second year

RLQ axis 1

RLQ

axi

s 2

Cluster 1Cluster 2Cluster 3

time

F Succession

RLQ axis 1

RLQ

axi

s 2

Cluster 1Cluster 2Cluster 3

time

Fig. 2. Results of the three-table ordination RLQ (R- sites�environment, L- species�sites, Q- species�traits). Shown arespecies locations in an environmental trait space of the respective management regime (A-F). Different symbols representspecies assignment to different clusters according to kmeans clustering with subsequent bootstrapping (for all clusters, meanmaximal Jaccard coefficient40.72). Additionally, the time vector is shown. [Correction added after online publication 25September 2009: corrected figure 2 (with missing headers) inserted]

1092 Romermann, C. et al.

parallel to the time vector in Fig. 2A; for both clus-ters, g40.75). Cluster characterisations are given inTable 4: while cluster one consists of small-seededspecies with a higher seed bank longevity, rosettespecies and species with shallow bud banks, thespecies of the second cluster can be characterised bycompetitive species, i.e. species with larger seeds,leaves regularly distributed along the stem, a deeperbud bank and transient seed bank.

All other management treatments provokedclear shifts in functional groups over time (Fig. 2B-F;for all. g40.75), i.e. species clusters were clearly se-parated along the time gradient, distinct clusters(5 species) could be observed along the time vector.See Table 4 for cluster characterisations.

Discussion

The current study demonstrates that a combina-tion of floristic and functional approaches is useful toanalyse floristic and functional development over timeand to derive management recommendations from theresponse of trait attributes under different manage-ment regimes. Furthermore the study emphasisedthe importance of analysing floristic, functional andenvironmental data together, as changes in environ-mental conditions are directly correlated to changes intrait responses. Up to now, few studies have integratedthese data in one common analysis (but compareBarbaro et al. 2000; Thuiller et al. 2006; Garnier et al.2007; Bernhardt-Romermann et al. 2008).

The results clearly support our hypothesis thatgrazed sites are in functional and floristic equilibrium.This was supported by both the separate considera-tion of floristic and functional analyses and by thejoint analyses of species, traits and environment.

In contrast to the historical management re-gime, all alternative management treatments led tomore or less strong changes in floristic and func-tional composition, depending on their intensitywith respect to biomass removal. Mulching twiceper year and mowing did not lead to strong changesin floristic or functional composition; however,combining floristic, functional and environmentaldata in the RLQ analysis clearly provided evidencethat these sites are not in equilibrium, indicatingthat vegetation change still goes ahead. Hence,RLQ is a strong method that can unravel processesthat remain hidden when analysing the different da-tasets separately. From the RLQ analyses, one canconclude that even the most intensive managementtreatments considered in this study are not appro-priate to replace the traditional management when

aiming to maintain the specific disturbance dy-namics of these historically grazed sites.

The less intensive management treatments,mulching once per year or every second year andsuccession, led to obvious changes in functional andfloristic composition, irrespective of the type of databeing considered in the different analysis techniques.Most of the trait attributes changed over time, in-dicating a strong change in vegetation structure anddynamics since the set-up of these managementtreatments. Although the functional distance valueswere extremely low, we are convinced that theseslight changes are ecologically meaningful: low va-lues can be attributed to the fact that trait attributeschanged but certainly did not disappear over time.In contrast, species disappearances or new arrivals(though rare events) strongly influenced the floristicdistances, leading to higher values compared tofunctional distances.

According to our analyses, if at all, mulchingtwice per year would be most similar to the grazingtreatment with respect to vegetation structure. How-ever, when having a closer look at trait responsesover time, differences between the two managementtreatments become apparent. Under the grazing re-gime, only the importance of the trait attributes budbanks located between � 10 cm and the surface andnon-branched growth forms increased; similar pat-terns were described in Klimesova et al. (2008). Formulching twice per year, different traits responded;here, we especially found a decrease in the end of theseed shedding period as an adaptation to manage-ment: the first mulching takes place in early summer(end of June/beginning of July) leading to the loss ofspecies shedding seeds later in the year. Comparablepatterns were described in Poschlod et al. (2009).Furthermore, species with leaves distributed reg-ularly along the stem and with higher SLA valuesslightly increased, indicating that more competitivespecies may become more important under thismanagement treatment (compare also Kahmen, Pos-chlod & Schreiber 2002). But this trend is less obviouscompared to the other mulching treatments that wereless frequently applied.

We cannot recommend introducing mowing onformerly grazed sites to maintain the specific dis-turbance dynamics. Mowing generated comparablefloristic and functional development over time com-pared to grazing, but results should be treated withcaution: after grazing ceased and before mowingstarted, the site underwent 5 years of succession,leading especially to changes in abiotic conditions.Hence, ignoring abiotic conditions led to the as-sumption that floristic and functional development

- Substitutes for grazing in semi-natural grasslands - 1093

Table 4. Characterisation of the clusters presented in Fig. 1. Only traits differing significantly between clusters are shownwith ���Po0.001, ��Po0.01, �Po0.05,1Po0.1. Per line, significant differences are characterised by small letters a and b.

Treatment Traits Cluster 1 Cluster 2 Cluster 3 Cluster 4 Test value

Grazing Seed mass 0.7910.11a 2.5110.57b �

Bud bank depth 7.93�

o� 10 cm 14 26� 10–0 cm 2 44 0 cm 5 0

Leaf distribution: 15.47���

Regularly 12 16Semi-rosette 2 14Rosette 7 0

Mowing Canopy height 0.6010.09b 0.2910.04a 0.2210.02a 0.2610.10a �

Seed mass 1.5610.43a 1.8310.36a 0.7710.12b 0.4510.16b �

SLA 22.5512.13ab 20.4911.36a 24.5411.43b 27.8913.29b �

Budbank layer max 2.6910.16b 2.5110.13ab 2.4210.17ab 1.8810.59b �

Leaf distribution 31.09���

Regularly 10 3 14 2Semi-rosette 0 13 2 0Rosette 0 1 4 1

Bud bank depth 12.431

o� 10 cm 7 14 16 2� 10–0 cm 3 3 0 04 0 cm 0 0 4 1

Lateral spread 79.98���

o0.01 0 12 5 00.01–0.25m 6 5 15 040.25m 0 0 0 3

Veg. diaspores 4 0 0 0Woodiness 10.931

Woody 1 0 3 0Semi-woody 0 1 0 1Non-woody 9 16 17 2

Mulching twice per year Seed mass 0.9210.19a 2.1410.31b 1.9210.48ab �

SLA 24.5410.72a 17.5310.40b 23.7910.76a �

Seed shedding end 11.1510.13a 10.6910.23a 9.53010.34b �

Bud bank depth 24 12 9 22.61���

o� 10 cm 2 4 1� 10–0 cm 0 0 74 0 cm

Branching 9.51��

Yes 26 11 12No 0 5 5

Leaf distribution 65.30 ���

Regularly 24 1 7Semi-rosette 0 0 9Rosette 2 15 1

Lateral spread 26.93 ���

o0.01 5 12 50.01–0.25m 17 4 840.25m 0 0 4Veg. diaspores 4 0 0

Mulching once per year Seed shedding start 6.8010.13a 6.1610.12a 5.1910.25b �

Seed shedding end 11.3510.12a 9.9510.25b 10.241.030b �

Seed shedding dur. 5.5510.24a 5.0510.32a 6.9510.41b �

Budbank layer max 2.6110.05a 2.4610.08b 2.2110.12b �

Bud bank depth 17.62��

o� 10 cm 4 4 0� 10–0 cm 16 15 144 0 cm 0 0 7

Branching 15.66���

Yes 19 10 20No 1 9 1

Growth form 12.63�

Chamae 4 0 0Geo 0 0 1Hemi 15 19 20Thero 1 0 0

Herbs/grasses 15.11���

Herbs 20 10 10Grasses 0 9 11

1094 Romermann, C. et al.

is comparable to those of the grazing treatment.When also including information on abiotic condi-tions, these similarities in changes in trait attributesover time disappeared.

All other treatments (mulching once per yearand every second year and succession) are not ap-propriate at all to maintain the specific disturbancedynamics of historically grazed sites. With decreas-ing management frequency/intensity, rosette speciesand non-branched growth forms decreased, whereastaller species, species with high SLA values andhence growth rates (Cornelissen et al. 2003) andspecies with leaves distributed regularly along thestem increased, indicating a shift towards strongeraboveground competition. Similarly, other studiesnoted that less intensive management regimes leadto an increase in competitive species, especiallywhen competition was for light, or as a cause of au-teutrophication (Pywell et al. 2003; Kahmen &Poschlod 2004; Eler et al. 2005; Louault et al. 2005;Garnier et al. 2007; Quetier et al. 2007; Romermannet al. 2008). Furthermore, species with higher num-bers of vegetative buds became more important overtime, indicating that clonal growth plays an im-portant role in less frequent management treatments

that lead mostly to nutrient-rich conditions (seeKleyer 2002). After infrequent injury to plants, thepresence of resprouting organs is advantageous(Klimesova & Klimes 2007). Furthermore, locationof the bud bank becomes deeper, perhaps as anadaptation to the management treatment: underfrequent mulching regimes only those species sur-vive that are able to directly resprout near the placeswhere they are damaged (i.e. at/above the soil sur-face; Klimes & Klimesova 2005). Also, lateralspread becomes important under less intensivemanagement regimes. Under succession, clonalgrowth itself might be important (compare Kleyer1999 and Kuhner & Kleyer 2008), but not the dis-tances that can be achieved by vegetative organs, asindicated by a decrease in vegetative lateral spread.

Concluding remarks

The analyses clearly showed that grazed sites arein functional and floristic equilibrium and that alter-native management treatments are not appropriate tosustain the typical disturbance dynamics. If grazingcannot be maintained on species-rich, semi-naturalgrasslands, we would suggest applying at least

Table 4. (Continued).

Treatment Traits Cluster 1 Cluster 2 Cluster 3 Cluster 4 Test value

Leaf distribution: 47.60 ���

Regularly 19 3 12Semi-rosette 1 15 1Rosette 0 1 8

Lateral spread 12.67�

o0.01 4 12 70.01–0.25m 1 0 340.25m 12 7 10Veg. diaspores 3 0 1

Mulching every 2nd year Canopy height 0.2110.02a 0.1110.02a 0.4410.03b �

SLA 20.1410.68a 25.8410.76b 24.8110.79b �

Seed shedding end 11.0410.18a 10.7510.13ab 9.9610.29b �

Lateral spread 74.53���

o0.01 19 0 40.01–0.25m 0 4 040.25m 10 0 17Veg. diaspores 0 0 4

Succession Canopy height 0.2310.02a 0.4610.04b 0.2610.02a �

SLA 26.6310.83a 23.0810.80a 18.5410.58b �

Seed shedding end 11.4310.12a 9.8310.35b 10.7910.23ab �

Seed shedding start 5.3610.29a 6.5010.24b 6.1610.13ab �

Branching 7.09�

Yes 14 15 12No 0 3 7

Leaf distribution: 28.34���

Regularly 10 17 2Semi-rosette 3 1 12Rosette 1 5 5

Lateral spread 39.32���

o0.01 1 2 150.01–0.25m 4 0 040.25m 6 15 4Veg. diaspores 3 1 0

- Substitutes for grazing in semi-natural grasslands - 1095

mulching twice per year, as this would provoke moresimilar floristic and functional development to thegrazing regime compared with less frequent manage-ment treatments or even succession. Of course, thissuggestion is only valid when aiming to replace man-agement treatments in warm and dry temperateclimates on grasslands on shallow soils. Studies inoceanic climates or in moist habitats have shownthat mulching leads to auteutrophication even insemi-natural, nutrient-poor grasslands and to a redis-tribution of species (Bakker 1989; Briemle 2005).Even though the EU Directive (Cross-Compliance)suggests mulching once per year or mowing everysecond year as suitable management tools to preserveecosystems of an ecological valuable status, there is alack of knowledge about the consequences of suchmanagement regimes. At least for the semi-naturalgrasslands analysed in the current study, the above-mentioned management regimes are likely to fail toachieve the goals of the EU Directive.

Acknowledgements. We acknowledge Karl-Friedrich

Schreiber for setting up this great experiment and thank all

contributors who carried out vegetation monitoring over the

last 30 years. KirstenMitlacher answeredmany questions on

the experimental design. We acknowledge all members and

contributors of the LEDA Traitbase (http://www.leda-trait

base.org), Biopop, CloPla and BiolFlor for providing species

trait data. The comments of Helge Bruelheide and three

anonymous reviewers considerably improved the manu-

script. Research was financially supported by Deutsche

Forschungsgemeinschaft DFG- EuroDIVERSITY 040,

‘‘Assemble’’, PO 491/5-1, and the Federal Agency for Nat-

ure Conservation (BfN, F1E project GAP-CC).

References

Anon. (R Development Core Team) 2004. A language and

environment for statistical computing. R Foundation

for Statistical Computing, Vienna, Austria.

Bakker, J.P. 1989. Nature management by grazing and

cutting. On the ecological significance of grazing and

cutting regimes applied to restore former species-rich

grassland communities in the Netherlands. Geobotany

14: 1–400.

Barbaro, L., Corcket, E., Dutoit, T. & Peltier, J.-P. 2000.

Reponses fonctionnelles des communautes de pelouses

calcicoles aux facteurs agro- ecologiques dans les

Prealpes francaises. Canadian Journal of Botany 78:

1010–1020.

Bernhardt-Romermann, M., Romermann, C., Nuske, R.,

Parth, A., Klotz, S., Schmidt, W. & Stadler, J. 2008.

On the identification of the most suitable traits for

plant functional trait analyses. Oikos 117: 1533–1541.

Briemle, G. 2005. Effekte einer Grunland-Mindestpflege

nach ‘‘Cross-Compliance’’. Berichte uber Landwirtschaft

83: 376–387.

Chessel, D., Dufour, A.B. & Thioulouse, J. 2004. The

ade4 package I: one-table method. R News 4: 5–10.

Cingolani, A.M., Cabido, M., Gurvich, D.E., Renison, D.

& Diaz, S. 2007. Filtering processes in the assembly of

plant communities: are species presence and

abundance driven by the same traits? Journal of

Vegetation Science 18: 911–920.

Clarke, K.R. & Ainsworth, M. 1993. A method of linking

multivariate community structure to environmental

variables.Marine Ecology Progress Series 92: 205–219.

Cleveland, W.S. 1981. LOWESS: a program for

smoothing scatterplots by robust locally weighted

regression. The American Statistician 35: 54.

Cornelissen, J.H.C., Lavorel, S., Garnier, E., Diaz, S.,

Buchmann, N., Gurvich, D.E., Reich, P.B., ter Steege,

H., Morgan, H.D., van der Heijden, M.G.A. & Pausas,

J.G. 2003. Handbook of protocols for standardised and

easy measurement of plant functional traits worldwide.

Australian Journal of Botany 51: 335–380.

Diekmann, M. (2003). Species indicator values as an

important tool in applied plant ecology – a review.

Basic and Applied Ecology 4: 493–506.

Doledec, S., Chessel, D., ter Braak, C.J.F. & Champely, S.

1996. Matching species traits to environmental

variables: a new three-table ordination method.

Environmental and Ecological Statistics 3: 143–166.

Dray, S., Pettorelli, N. & Chessel, D. 2002. Matching data

sets from two different spatial samplings. Journal of

Vegetation Science 13: 867–874.

Dutoit, T., Forey, E., Romermann, C., Buisson, E.,

Fadda, S., Saatkamp, A., Gaignard, P. & Trivelly, E.

2005. Remanence des utilisations anciennes et gestion

conservatoire des pelouses calcicoles en France.

Biotechnologie, Agronomie, Societes et Environnment

9: 125–132.

Eler, K., Vidrih, M. & Batic, F. 2005. Vegetation chara-

cteristics in relation to different management regimes of

calcareous grassland: a functional analysis using plant

traits. Phyton-Annales Rei Botanicae 45: 417–426.

Ellenberg, H., Weber, H.E., Dull, R., Wirth, V. &Werner,

W. 2001. Zeigerwerte von Pflanzen in Mitteleuropa.

Scripta Geobotanica 18: 262.

Garnier, E., Lavorel, S., Ansquer, P., Castro, H., Cruz, P.,

Dolezal, J., Eriksson, O., Fortunel, C., Freitas, H.,

Golodets, C., Grigulis, K., Jouany, C., Kazakou, E.,

Kigel, J., Kleyer, M., Lehsten, V., Leps, J., Meier, T.,

Pakeman, R., Papadimitriou, M., Papanastasis, V.P.,

Quested, H., Quetier, F., Robson, M., Roumet, C.,

Rusch, G., Skarpe, C., Sternberg, M., Theau, J.P.,

Thebault, A., Vile, D. & Zarovali, M.P. 2007.

Assessing the effects of land-use change on plant

traits, communities and ecosystem functioning in

grasslands: a standardized methodology and lessons

from an application to 11 European sites. Annals of

Botany 99: 967–985.

1096 Romermann, C. et al.

Grime, J.P. (1998). Benefits of plant diversity to ecosystems:

immediate, filter and founder effects. Journal of Ecology

86: 902–910.

Hennig, C. 2006. fpc: Fixed point clusters, clusterwise

regression and discriminant plots. R package version

1.2-2. Available at http://www.homepages.ucl.ac.uk/

�ucakche/. Accessed February 2007.

Hennig, C. 2007. Cluster-wise assessment of cluster

stability. Computational Statistics & Data Analysis 52:

258–271.

Jackel, A.-K., Dannemann, A., Tackenberg, O., Kleyer, M.

& Poschlod, P. 2006. BioPop- Funktionelle Merkmale

von Pflanzen und ihre Anwendungsmoglichkeiten im

Arten-, Biotop- und Naturschutz. Naturschutz und

Biologische Vielfalt 32: 1–168.

Kahmen, S. & Poschlod, P. 2004. Plant functional trait

responses to grassland succession over 25 years.

Journal of Vegetation Science 15: 21–32.

Kahmen, S., Poschlod, P. & Schreiber, K.-F. 2002.

Conservation management of calcareous grasslands.

Changes in plant species composition and response of

functional traits during 25 years. Biological

Conservation 104: 319–328.

Kleyer, M. 1999. Distribution of plant functional types

along gradients of disturbance intensity and resource

supply in an agricultural landscape. Journal of

Vegetation Science 10: 697–708.

Kleyer, M. 2002. Validation of plant functional types

across two contrasting landscapes. Journal of

Vegetation Science 13: 167–178.

Kleyer, M., Bekker, R.M., Knevel, I.C., Bakker, J.P.,

Thompson, K., Sonnenschein, M., Poschlod, P., Van

Groenendal, J.M., Klimes, L., Klimesova, J., Klotz, S.,

Rusch, G., Hermy, M., Adriaens, D., Boedeltje, G.,

Bossuyt, B., Endels, P., Gotzenberger, L., Hodgson,

J.G., Jackel, A.-K., Dannemann, A., Kuhn, I.,

Kunzmann, D., Ozinga, W., Romermann, C., Stadler,

M., Schlegelmilch, J., Steendam, H., Tackenberg, O.,

Wilmann, B., Cornelissen, J.H.C., Eriksson, O.,

Garnier, E., Fitter, A. & Peco, B. 2008. The LEDA

Traitbase: a database of plant life-history traits of North

West Europe. Journal of Ecology 96: 1266–1274.

Klimes, L. &Klimesova, J. 1999. CLO-PLA2 – a database of

clonal plants in Central Europe. Plant Ecology 14: 9–19.

Klimes, L. & Klimesova, J. 2005. Clonal traits. In: Knevel,

I., Bekker, R., Kunzmann, D., Stadler, M. &

Thompson, K. (eds.) The LEDA Traitbase – Collecting

and Measuring Standards of Life-history Traits of the

Northwest European Flora. LEDA Traitbase project. pp.

66–88. University of Groningen, Community and

Conservation Ecology group, Groningen.

Klimesova, J. & Klimes, L. 2007. Bud banks and their role

in vegetative regeneration – a literature review and

proposal for simple classification and assessment.

Perspectives in Plant Ecology Evolution and Systematics

8: 115–129.

Klimesova, J., Latzel, V., de Bello, F. & van Groenendael,

J.M. 2008. Plant functional traits in studies of vegetation

changes in response to grazing and mowing: towards a

use of more specific traits. Preslia 80: 245–253.

Klotz, S., Kuhn, I. & Durka, W. 2002. BiolFlor – eine

Datenbank mit biologisch–okologischen Merkmalen

zur Flora von Deutschland. Schriftenreihe fur

Vegetationskunde 38: 1–334.

Knevel, I.C., Bekker, R.M., Bakker, J.P. & Kleyer, M.

2003. Life-history traits of the Northwest European

Flora: the LEDA database. Journal of Vegetation

Science 14: 611–614.

Kuhner, A. & Kleyer, M. 2008. A parsimonious

combination of functional traits predicting plant

response to disturbance and soil fertility. Journal of

Vegetation Science 19: 681–692.

Louault, F.P., Pillar, V.D., Aufrere, J., Garnier, E. &

Soussana, J.-F. 2005. Plant traits and functional types

in response to reduced disturbance in a semi-natural

grassland. Journal of Vegetation Science 16: 151–160.

McIntyre, S., et al. 1995. Plant life-history attributes –

their relationship to disturbance responses in

herbaceous vegetation. Journal of Ecology 83: 31–44.

Moog, D., Poschlod, P., Kahmen, S. & Schreiber, K.-F.

2002. Comparison of species composition between

different grassland management treatments after 25

years. Applied Vegetation Science 5: 99–106.

Oksanen, J., Kindt, R., Legendre, P. & O’Hara, R.B.

2007. vegan: Community Ecology Package version

1.8-5. Available at http://cran.r-project.org/. Accessed

February 2007.

Ozinga, W.A., Schaminee, J.H.J., Bekker, R.M., Bonn, S.,

Poschlod, P., Tackenberg, O., Bakker, J. &

Groenendael, J.M.v. 2005. Predictability of plant

species composition from environmental conditions is

constrained by dispersal limitation. Oikos 108: 555–561.

Ozinga, W., Hennekens, S.M., Schaminee, J.H.J., Smits,

N.A.C., Bekker, R.M., Romermann, C., Klimes, L.,

Bakker, J.P. & Van Groenendal, J.M. 2007. Local

above-ground persistence of vascular plants: life-

history trade-offs and environmental constraints.

Journal of Vegetation Science 18: 489–497.

Poschlod, P. & WallisDeVries, M.F. 2002. The historical

and socioeconomic perspective of calcareous

grasslands – lessons from the distant and recent past.

Biological Conservation 104: 361–376.

Poschlod, P., Bakker, J.P. & Kahmen, S. 2005. Changing

land use and its impact on biodiversity. Basic and

Applied Ecology 6: 93–98.

Poschlod, P., Karlık, P., Baumann, A. & Wiedmann, B.

2008. The history of dry calcareous grasslands near

Kallmunz (Bavaria) reconstructed by the application

of palaeoecological, historical and recent-ecological

methods. In: Szabo, P. & Hedl, R. (eds.) Human

nature: studies in historical ecology and environmental

history. Institute of Botany of the Czech Academy of

Sciences, Brno, CZ, pp. 130–143.

Poschlod, P., Bernhardt-Romermann, M., Donaubauer,

T., Hoffmann, J. & Sittig, E. 2009. Die Altersstruktur

von Pflanzenpopulationen – Ein naturschutzfachliches

- Substitutes for grazing in semi-natural grasslands - 1097

Bewertungskriterium. In: Schreiber, K.-F., Brauck-

mann, H.-J., Broll, G., Krebs, S. & Poschlod, P.

Artenreiches Grunland in der Kulturlandschaft. 30 Jahre

Offenhaltungsversuche Baden-Wurttemberg. Naturschutz-

Spectrum 97, LUBW, Verlag regionalkultur-ur,

Karlsruhe, DL, pp. 289–299.

Pywell, R.F., Bullock, J.M., Roy, D.B., Warman, L.,

Walker, K.J. & Rothery, P. 2003. Plant traits as

predictors of performance in ecological restoration.

Journal of Applied Ecology 40: 65–77.

Quetier, F., Thebault, A. & Lavorel, S. 2007. Plant traits

in a state and transition framework as markers of

ecosystem response to land-use change. Ecological

Monographs 77: 33–53.

Ribera, I., Doledec, S., Downie, I.S. & Foster, G.N. 2001.

Effect of land disturbance and stress on species traits

of ground beetle assemblages. Ecology 82: 1112–1129.

Romermann, C., Tackenberg, O., Jackel, A.-K. &

Poschlod, P. 2008. Eutrophication and fragmentation

are related to species’ rate of decline but not to

species rarity! Results from a functional approach.

Biodiversity and Conservation 17: 591–604.

Schaffers, A.P. & Sykora, K.V. (2000). Reliability of

Ellenberg indicator values for moisture, nitrogen and

soil reaction: a comparison with field measurements.

Journal of Vegetation Science 11: 225–244.

Schmidt, W. 1974. Bericht uber die Arbeitsgruppe

fur Sukzessionsforschung auf Dauer-flachen der

Internationalen Vereinigung fur Vegetationskunde.

Vegetatio 29: 69–73.

Schreiber, K.-F. 1997. Sukzessionen – Eine Bilanz der

Grunlandbracheversuche in Baden-Wurttemberg.

Veroffentlichungen PAO 23.

Thuiller, W., Richardson, D.M., Rouget, M., Proches, S.

& Wilson, J.R.U. 2006. Interactions between

environment, species traits and human uses describe

patterns of plant invasions. Ecology 87: 1755–1769.

WallisdeVries, M.F., Poschlod, P. & Willems, J.H. 2002.

Challenges for the conservation of calcareous

grasslands in northwestern Europe: integrating the

requirements of flora and fauna. Biological

Conservation 104: 265–273.

Wamelink, G.W.W., Joosten, V., van Dobben, H.F. &

Berendse, F. 2002. Validity of Ellenberg indicator values

judged from physico-chemical field measurements.

Journal of Vegetation Science 13: 269–278.

Watkinson, A.R. & Ormerod, S.J. 2001. Grasslands,

grazing and biodiversity: editors’ introduction. Jour-

nal of Applied Ecology 38: 223–237.

Westoby, M., Falster, D.S., Moles, A.T., Vesk, P.A. &

Wright, I.J. 2002. Plant ecological strategies: some

leading dimensions of variation between species.

Annual Review of Ecology and Systematics 33: 125–159.

Wisskirchen, R. & Haeupler, H. 1998. Standardliste der

Farn- und Blutenpflanzen Deutschlands. Ulmer,

Stuttgart (Hohenheim).

Supporting Information

Additional supporting information may befound in the online version of this article:

Appendix S1. Overview on species occurrencesat the differently managed study sites over the 30years of vegetation monitoring. Only species in-cluded in the analyses are shown. GR5 3 grazing,MO5mowing, 2M5mulching twice per year,1M5mulching once per year, M25 4 mulching ev-ery second year, SU5 succession. Nomenclaturefollows Wisskirchen & Haeupler (1998).

Please note: Wiley-Blackwell is not responsiblefor the content or functionality of any supportingmaterials supplied by the authors. Any queries(other than missing material) should be directed tothe corresponding author for the article.

Received 18 May 2008;

Accepted 26 June 2009.

Co-ordinating Editor: H. Bruelheide.

1098 Romermann, C. et al.