ten Harkel and van der Meulen 1996
!"#$%&'()'*+$,-./'$.0'1&"(2#34+-%'5-&+(/4.'64#(2-&-(.'(.'&34'74/4&$&-(.'()'6+8'9($2&$:
6;.4'*+$22:$.02
1;&3(+<2=>'?$&&3-@2'AB'&4.'C$+D4:'$.0'E+$.D'F$.'04+'?4;:4.
G(;+%4>'A(;+.$:'()'74/4&$&-(.'G%-4.%4H'7(:B'IH'5(B'J'<A;.BH'KLLM=H'##B'NNOPNOQ
R;S:-2340'S8>'T#;:;2'R+422
G&$S:4'UVW>'http://www.jstor.org/stable/3236288
1%%42240>'QMXKYXQYYZ'QJ>ON
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=opulus.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
Opulus Press is collaborating with JSTOR to digitize, preserve and extend access to Journal of Vegetation
Science.
http://www.jstor.org
Journal of VegetationScience 6. 445-452, 1995
? IAVS;OpulusPress Uppsala. Printedin Sweden
445
Impact of grazing and atmospheric nitrogen deposition
on the vegetation of dry coastal dune grasslands
ten Harkel, Matthijs J.* & van der Meulen, Frank
Landscape and Environmenltal Research Group, University of Amsterdam, Nieuwe Prinsengr7acht 130, 1018 VZ
Amsterdam, The Netherlands; *Correspondingauthor. Tel. + 31 20 5257442; Fax + 31 20 5257431;
E-mail mth@fgb.frw.uva.nl
Abstract. A five-year experimentalstudy was carriedout to 1988, 1990; Bobbink 1991; Aerts & Heil 1993). So far,
examine the combined effects of grazing and atmospheric such systematic studies have not been initiated in the
nitrogen deposition on the vegetation of three dry dune dunes. The present study was initiated in 1988, in order
grasslands: shortspecies-rich,one shortspecies-poor,and
one
to understand the relations between atmospheric deposi-
one predominated tall graminoids.
by
Additional fertilizationwith nitrogen had no significant tion, grazing and vegetation dynamics. The present pa-
effect, neitherin grazednor in non-grazedplots. Exclusion of per deals with changes in species composition.
Since the 1970s some tall perennial graminoids,
grazingby rabbitsresultedin an increasein the frequencyof
perennialgraminoidsand a decrease in the frequencyof an- notably Ammophila arenaria, Avenulapubescens, Cala-
nual graminoids and herbs. Nevertheless, species diversity magrostis epigejos, Elymus athericus and Carex are-
remainedthe same in the species-richgrassland.During the naria, have become dominant in many originally short
experiment,the above-groundbiomass increasedin all non- dry dune grasslands of the Netherlands (C.T.M. Vertegaal
grazed plots and the amount of bare soil and mosses de- et al. int. report 1991). As a consequence, these grass-
creased.
lands show now a relatively low species diversity. From
The vegetation changes occurredmainly within one year
the point of view of nature conservation this is consid-
after the exclusion of grazing. An exception is the grass-
dominatedsite where the amountof Calamagrostisepigejos ered an unfavourable situation (Westhoff 1985). Nowa-
increased graduallyfrom ca. 20 % in the first two years to days, nature managers try to counteract this develop-
about50 % in the fourthand fifth year. ment and try to restore the original vegetation composi-
Grazingby rabbitsseems essential to preventgraminoids tion. This requires understanding of the processes respon-
to become predominantin the dry dunes. If graminoidsare sible for the predominance of graminoids.
dominant,grazingby horses can be an appropriate methodto In the past, tall graminoids were usually not domi-
restorethe originalgrasslandvegetation.After six months of nant because of the low nutrient status of sandy dune
grazingby horsesthe grass-dominated showed a decrease
site
soils (Willis & Yemm 1961; Boorman & Fuller 1982).
of the frequencyof perennialgraminoids,from 95 %to 80 %,
The present dominance of graminoids might be a result
and an increase of the frequency of perennial herbs, from
2.5 % to between 13 and 20 %. of the recent increase of atmospheric nitrogen deposi-
tion. From 1930 to 1980 the atmospheric nitrogen depo-
sition (NH4+ + NO3-) on coastal areas in the Netherlands
Keywords: Horse;Management; Meijendel;the Netherlands; increased from 4 kg N.ha-l.yr- to 14 kg N.ha-l.yr-1
Rabbit;Restorationecology; Species diversity. (Stuyfzand 1993). The predominance of graminoids due
to high nitrogen loads is also known from heathlands
Nomenclature: van der Meijden (1990). (Heil & Diemont 1983; Roelofs 1986; Bobbink et al.
1990; Aerts & Heil 1993) and chalk grasslands (Bobbink
et al. 1988; Bobbink 1991; Willems et al. 1993). Many
Introduction graminoids have a relative higher growth rate than other
plant species and are therefore able to benefit more from
Recently, semi-natural vegetation in the Netherlands increased nitrogen input. This is shown by fertilizing
has been under increasing influence of atmospheric experiments in grasslands (Willis & Yemm 1961; Willis
deposition of acid and eutrophic components. This has 1963; Van Hecke et al. 1981; Davy & Bishop 1984;
led to a decrease in nature conservation values. Bobbink 1991; Mountfordet al. 1993). When grasses
In chalk grasslands and heathlands the influence of aretall, light will not penetrate the soil andgermina-
into
atmospheric deposition has been studied to support man- tion of annualspecies is prevented(Bobbink& Willems
agement programs aiming at restoration of nature values 1991; Olff et al. 1993).
(Heil & Diemont 1983; Roelofs 1986; Bobbink et al. It seems thatgraminoidswill not dominatewhen the
446 ten Harkel,M.J. & van der Meulen, F.
vegetation is frequentlygrazedby animalssuch as rab- with a pH(CaCl2)of 4.2 to 5.4 on similar calcareous
bits (Williamset al. 1974;Bhadresa1977;Zeevalking& dunesandas the firstsite. The 10 cm thicktop soil has an
Fresco 1977). If therabbitpopulationdecreases,as in the organic matter content of 3 %. The site representsa
1950s (dueto myxomatosis),graminoids ableto take
are species-poor variant of the Violo-Corynephoretum as-
over (Ranwell 1960; Thomas 1963; Westhoff 1985). sociation(Westhoff& den Held 1969). A certaindomi-
Theeffect of rabbitgrazingis also experimentally shown nance of graminoidsoccurredhere from the beginning
in grazedgrasslands whereafterthe exclusionof grazing (Table 1). This site has been monitoredsince the sum-
tall graminoidsbecame dominant(Farrow1917; Hill et mer of 1988.
al. 1992). Even afterrestoration the rabbitpopulation,
of The third,grass-dominated is close to the second
site
graminoidsare able to remaindominantfor a long time site, butthe vegetationhas becometotallypredominated
as rabbitsdo not eat or enter such relatively tall grass by Carex arenaria, Festuca ovina + F. rubra and
swards.Non-grazedvegetationhas a relativelylow pro- Calamagrostis epigejos (Table 1). This site has been
tein content in comparison to grazed vegetation and monitoredsince the summerof 1988.
thereforeanimalsavoid the area (Bakkeret al. 1983).
The dominanceof graminoidscan be counteracted
by introducinglarge herbivores(Williams et al. 1974).
If heavy grazing is needed to permitseedling establish-
ment of annual species, it can also prohibitsuccessful
Table1. Totalnumber speciesandfrequency
of (averagefor
completion of the life cycle of these annual species allyears) plant
of speciesof thecontrol
plotsin thethree
sites
(Farrow1917; Watt 1962; Silvertownet al. 1992). Re- (grass-dominated before introduction horses);
site the of Rich
cently, horses and cows have been introduced Dutch
in = species-rich site; Poor = species-poor site; Grass = grass-
naturereservesto reducethe amountof biomassandthe dominated
site.
vegetation height and thus restore the original species Rich Poor Grass
composition.
This paper reports on an experiment in a Dutch Number of species
coastal dune area where fertilizationwith nitrogenand Annual herbs 4 3 1
exclusion of grazingwere combinedas a treatment. The Perennial herbs 9 3 3
Annual graminoids 2 1
purpose was to define the main reason for the domi- Perennial graminoids 6 6 6
nance of tall graminoids in large parts of comparable Annual herbs
dune areas. This dominance may be caused by high Cerastiluml semidecandrllm 1-5 <1 0
atmosphericdeposition of nitrogen, or the absence of El-odiliuti c(icultariiu 5- 10 <1 0
grazingby rabbits,or a combinationof both.Inaddition, Se'I'cio) svlvaticus 0 1-5
the effects of the introduction horsesandthe implica-
of Teeshiali(a illcicaulis 0 <1 0
Veronica arvcnsis <1 0 0
tions for the managementof dune grasslandswill be 1-5
Violla (cr1'tisii 0
discussed. Perennial herbs 0
Galilum verum 1-5 <1
1-5
0
Hiera t'i71I pilosellc <1 0
Material and Methods Holcusl /l(anatlus 0 <1
Leontodlonl saxatilis 1-5 0 0
Lotus corni(iculcatus 1-5 0
Study sites Myosotis arvenlsis 1-5 0
Ohonis repens 10- 20 0
The study was carriedout in threegrasslandsites in Rumnex acetosella 0 10- 20 0
Sedumi acre <1 0 0
Meijendel,a coastal dune areanorthof The Hague, the Senecio jacobaea 5-10 < 1
1-5
Netherlands.This area is managedby the Dune Water 1-5
Tarax acum sect. Erythrospe rma 0 0
Companyof Zuid-Holland(D.Z.H.). Annual graminoids
The first, species-richsite is located at a distanceof Aila praecox 1-5 5- 10 1-5
0.5 km fromthe coast. The sandysoil is calcareous,with Phleuni arenar-iuni <1 0 0
2- 3 % CaCO3, and with a very low organic matter Perennial graminoids
Agrostis stolonifera + A. capillaris 0 5- 10 0
content(2 %in the 10cm top soil). Phytosociologically,
Calamagriostis epigejos 1-5 10 - 20 10 - 20
the vegetationcan be assignedto the Festuco-Galietum Carex arenaaia 5- 10 5-10 >20
ma-ritimi association(Westhoff& den Held 1969). This Festuca ovina + F. orubra 1-5 >20 10 - 20
site has been monitoredsince the summerof 1989. Koeleria macrantha <1 0 1-5
The second, species-poor site is situated 2.5 km Luzula campestris 10 - 20 10 - 20 1-5
Poa pratensis 5- 10 < 1 <1
from the coast. The soil is decalcified down to 50 cm
- Impactof grazingand atmosphericnitrogendeposition- 447
Treatments Statisticalanalysis
At all threesites five homogeneousquadrats 2 m x
of The resultswere analyzedstatisticallywith the com-
2 m were selected. Each quadratwas divided into four puter package SAS (Anon. 1987). The counts of the
experimentalplots of I m x 1m: (1) control; (2) ferti- plants were standardizedto frequencies and will be
lized; (3) excluded from grazing by rabbits and (4) called frequenciesfurtheron.
combinationof fertilizationand exclusion of grazing. The effect of grazingandfertilizationon the follow-
The treatmentswere allocated randomly to the plots ing three frequencieswere tested with the SAS-proce-
(randomizedquadrat design). dure ANOVA with repeated measurements:species
In the autumnof 1990, horses were introducedin a groups (annualor perennialherbs and graminoids),in-
partof Meijendel to reduce severe grass encroachment dividualspecies, surfacewithoutphanerogamic species
andto possibly restorethe originalspecies composition. (baresoil and mosses). The effects on species diversity
The horsesgrazedfor six monthsat the grass-dominated (see below) and biomass were also tested. A signifi-
site;thereafter horseswereexcludedfromthequadrats
the cance leve of 0.05 was chosen. The effects of the treat-
because they demolishedthe exclosures. ments were analysed separatelyfor each site. The ef-
The fertilizerwas appliedto simulatethe effects of fects of treatmentson the grass-dominatedsite were
high loads of atmosphericnitrogen deposition. It was analysedtwofold: before the introduction the horses
of
assumedthatadditionof nitrogenwould speed up grass and after six months of grazingby horses.
encroachmentand cause the associatedloss of charac- Species diversity was measured as the Shannon
teristicherbs. Wiener index:
From October 1988 until April 1992 the fertilizer
appliedconsistedof pellets of NH4NO3 (addedin spring SD = P ln i (1)
and autumn)equalto an extradepositionof 25 kg N/ha/
yr. As weekly unsaturatedsoil water analysis by ce- where SD = species diversity andPi = the proportionof
ramiccups showed thatlarge amountsleached through species i (on a frequencybasis).
the surfacesoil within a few weeks, a differentmethod The moss species were regardedas one group,and were
of fertilizationwas applied. Since April 1992 quadrats not includedin this calculation.
were fertilized with a 0.0069 M (NH4)2SO4 solution
once every fortnight,which was equal to an extranitro-
gen input of 50 kg N/ha/yr. For osmotic reasons NaCl Results
(sea salt)was addedin this solutionup to the same NaCl
concentration the rainwater.
in The non-fertilizedplots The species presentin the threesites were dividedin
obtained an equal amount of NaCl solution without four groups: annual herbs, perennial herbs, annual
(NH4)2SO4. The yearly amount of solution given was graminoids and perennial graminoids (Table 1). The
equivalentto an extraprecipitation 26 mm/yr,which
of grazedquadrats the species-richsite containthe larg-
of
is only 3 % of the averageyearly precipitation. est number of herb species, 13. The numbers of the
species-poorand the grass-dominated were six and
site
Measurements four respectively. The numberof graminoidspecies is
equal in the sites, i.e. seven or eight. However, in the
Once a year in July, the species composition of the grazedplots the frequencyof perennialgraminoids(Ta-
vegetation was recordedin each plot by counting the ble 2) is much higher in the grass-dominated (92 %
site
presence of all separatephanerogamsand of the rest priorto grazingby horses) thanin the species-poorsite
(bare soil and mosses) with a point frequencymethod (63 %) and the species-rich site (44 %). However, the
(Mueller-Dombois& Ellenberg1974), using 100 cross- amountsof perennialgraminoidspecies in the species-
points of a grid. rich and grass-dominatedsystems are the same (Table
The standingcropbiomass (dead+living)was meas- 1).Thespeciesresponsible thedominance gramino-
for of
ured each year in July by cutting,drying and weighing ids in grass encroachedstages are alreadypresentin the
threerandomlyselected subsamplesof 20 cm x 20 cm, species-richstage.
each at the edge of all plots. The centralpartof each plot
(60 cm x 60 cm) has been left untouchedfor long-term Effects offertilization and grazing on species groups
monitoring.
The impactof nitrogenfertilizationon the change of
species groups and the three other variables was not
statisticallysignificant at all sites! Thereforefertilized
448 ten Harkel, M.J. & van der Meulen, F.
Table 2. Average frequency over all years of the species groups in the grazed plots and exclosure plots of the three sites.
Species group Annual Perennial Annual Perennial
herbs herbs graminoids graminoids
[%] [%] [%] [%]
Treatment G E G E G E G E
Site
Species-rich 7.5 ** 1.7 ** 41.5* 33.6 * 4.9 2.6 43.8 ** 56.0 **
Species-poor 1.2 0.1 24.4* 12.2* 10.5 4.3 62.7** 82.2**
Grass-dominated1 1.1 0.3 2.5 2.5 3.9 3.5 92.4 93.7
Grass-dominated2 0.1 0.2 12.8 * 4.1 * 6.9 ** 0.4 ** 80.0 ** 95.3 **
G = grazedplots; I = before the introduction horses;
of
E = exclosure plots; 2 = afterthe introduction horses.
of
Significance of differencesbetween grazedand exclosureplots: * = p < 0.05; **= p < 0.01.
and non-fertilized plots were taken together during the between grazed and non-grazed plots (Table 2). In con-
statistical analysis of the effects of grazing. trast, the frequency of annual graminoids and perennial
The effects of the exclusion of grazing by rabbits on herbs increased (respectively 0.4% to 7%, and 4% to
the species groups are shown in Table 2. The exclusion 13%). The horse-grazed plots showed an increase of the
resulted in an increase of perennial graminoids in both amount of bare soil and mosses, and of the species
the species-rich and species-poor sites. No increase was diversity. In the same time the biomass showed a de-
observed in the grass-dominated site, because the fre- crease.
quency of perennial graminoids was high already from
the beginning (92 %). The frequency of perennial herbs Eff#cts of fertilization and grazing on indilidual
decreased in both the species-rich and species-poor is
sites. In the species-rich site the frequency of annual
herbs also showed a decrease. Fertilization with nitrogen did not affect individual
Exclusion of grazing at the species-rich and the species. Only Luzula cnampestris decreased in the
species-poor site reduced the amount of bare soil and exclosures of the grass-dominated site. To test the effect
mosses, while the biomass of the stand increased (Table of grazing on individual species, fertilized and non-
3). Moreover, a decrease in species diversity was seen in fertilized plots were also taken together.
the species-poor site. The significant changes of individual species after
Before the introduction of horses on the grass-domi- the exclusion from grazing are shown in Table 4.
nated site, the amount of bare soil and mosses became Some species showed a decrease at the species-poor
lower in the exclosure plots. Six months of grazing by site. No increases were detected. However, in the last
horses resulted in a difference of perennial graminoids paragraph an increase in perennial graminoids was
Table 3. Average percentage of all years of bare soil and mosses, species diversity and biomass of the three sites.
Variable Bare + mosses Species diversity Biomass
[%] (see formula1) [g/m-2]
Treatment G E G E G E
Site
Species-rich 43.9 ** 8.5 ** 1.945 1.983 42 ** 264 **
Species-poor 57.1 ** 8.1** 1.635 * 1.356* 75 ** 467 **
Grass-dominated' 17.8** 5.6** 1.134 1.213 335 536
Grass-dominated2 45.2 ** 3.7 ** 1.664 1.465 * 256 ** 613 **
G = grazedplot; 1 = before the introduction horses;
of
E = exclosure plot; 2 = afterthe introduction horses.
of
Significance of differencesbetween grazedand exclosureplots: * =p < 0.05; **=p < 0.01.
- Impact of grazing and atmospheric nitrogen deposition - 449
shown.This contradiction mightbe due to an increaseof Vegetation change after the exclusion of grazing
differentperennialgraminoidsin each plot.
In the grass-dominatedsite no changes due to the Vegetation change was defined as the differencein
exclusion from grazingwere foundbefore the introduc- frequencybetweentwo yearsof the species thatchanged
tion of grazingby horses. The whole site is dominated in the non-grazed plots relative to the grazed plots
by perennialgraminoids. (Table 4). The difference remainedconstant after the
Species thatincreasedafterthe exclusion fromgraz- first year, the changes measured occurred during the
ing or decreased after the introductionof grazing by first year. The vegetation change was calculated sepa-
horses are all perennialgraminoids.One exception is ratelyfor frequencieswhich increasedor decreaseddue
the increase of Phleum arenariumat the species-rich to grazing.
site undernon-grazingconditions.This is unexpectedas At the species-rich site the vegetation change of
is
it is a small annualgrassandits germination normally species which benefited from the exclusion of grazing
preventedin a vegetationpredominated tall grasses
by was small, viz. from 11 % to 16 %. The vegetation
such as Festuca rubra + F. ovina and Poa pratensis changeof species which decreasedafterthe exclusion of
(Table 4). Species that showed a decrease after exclu- grazingall occurredduringthe first year.
sion or an increaseafterthe introductionof grazing by At the species-poorsite all changes occurredduring
horses,areherbsor smallperennialgraminoids,such as the first year.
Luzula campestris. At the grass-dominatedsite the vegetation change
for species which increasedaftergrazingby horses was
the same for species which increased after the exclu-
sion, i.e. from 3 % to 25 %. Calamagrostis epigejos is
the only species which increasedin the exclosures com-
paredto the controlplots. In spite of the introductionof
horses the frequency of Calamagrostis epigejos remained
the same (ca. 20 %) in the grazed plots, therefore it
increasedfrom 20 % to ca. 50 % in the exclosures. The
exclosures developed from a dominance of several
graminoids (Table 1) into a dominance by only one
species. The frequencyof the species favouredby graz-
Table4. Plant
species after
development excluding or
grazing ing showed an increasein the grazedplots from 7 % to
after of at site.
introduction horses thegrass-dominated 30%, while it remainedlow in the exclosures(3 %).The
increaseof these species occurredafterthe introduction
Species-rich site
Festuca ovina +F. rubra + of the horses. The vegetation level of the grazed plots
Poa pratensis + was lowered and this gave the plants an opportunityto
Phleumarenarium +
Cerastiumsemidecandrum grow.
Erodiumcicutarium
Luzulacampestris
Myosotisarvensis Discussion
Seneciojacobaea
Taraxacumsect. Erythrosperma
Viola curtisii This study showed thatmaintenanceof species-rich
Species-poor site dry dune grasslandsstronglydependson the activity of
Rumexacetosella grazinganimals,especiallyunderrelativelyhigh atmos-
Seneciojacobaea
pheric nitrogendeposition levels. When grazing is ex-
Grass-dominated site cluded, perennialgraminoidsbecome dominantwithin
Before introduction horses
of no change one ortwo years.These graminoids were alreadypresent
After introduction horses
of
in the three sites before the treatment(Table 1). As in
Aira praecox + other grazing experiments (Belsky 1986; Hill et al.
Luzulacampestris +
1992) no new plant species invadedthe plots.
Rumexacetosella ++
Seneciojacobaea +
Underrelatively high atmosphericnitrogendeposi-
Calamagrostisepigejos tion levels, fertilizationby nitrogenhad no effect on the
three sites during the 5 yr of this study. This is not
+ = increase (p < 0.05);
++ = increase (p < 0.01);
consistent with fertilizer experiments in other grass-
=decrease (p < 0.05); lands (Willis 1963; Van Hecke et al. 1981; Davy &
- =decrease (p < 0.01).
Bishop 1984; Bobbink et al. 1988; Bobbink 1991;
450 ten Harkel,M.J. & van der Meulen, F.
Mountfordet al. 1993). Effects of additionalnutrient These studies show that when no other changes
supply were always registeredwithin 5 yr. In our case occur, the vegetationis probablynot seriously affected
the amountof nitrogen fertilizerof 50 kg N/ha/yr and by an increasingavailabilityof nutrients.Whengrazing
background depositionof nitrogenof 10 kg N/ha/yr(ten is continued,changes will probablybe very slow and
Harkel et al. 1991) should be sufficient to accelerate cannotbe measuredwithin 5 yr.
changes in species composition if other resources are The main conclusion of this study is that with high
not limited.Bobbinket al. (1992) gave a criticalload of nitrogenlevels a dunegrasslanddoes not become grass-
nitrogenfor dune grasslandsof 20-30 kg N/ha/yr. dominatedas long as the vegetationis grazed,not even
Some explanations for these deviating results are when fertilized.When grazing is absent, the relatively
given below. high availabilityof nitrogen gives graminoidsthe op-
portunityto become dominant.Effects of a decreaseof
Phosphorusas a limitingfictor grazing intensity were reportedby Ranwell (1960): an
increasein the growthand floweringof graminoidsand
Fertilization experiments with different combina- a decreaseof herbswas foundafterthe disappearance of
tions of nutrients(nitrogen,phosphorus, potassium,cal- rabbits myxomatosisin a sand-dunesystem.Willis &
by
cium) showed that the nitrogencontent of the soil was Yemm (1961) andWillis (1963) showed, however,that
the most limiting factor (Milton 1940, 1947; Willis even without grazing a very nutrient-poor sandy dune
1963; Boorman& Fuller 1982; Kachi & Hirose 1983; system can remain species-rich. However, the system
Doughertyet al. 1990; Olff et al. 1993). However, also became dominatedby grasses after addition of nutri-
phosphoruscan be an importantlimitingfactor(Milton ents. Boorman& Fuller (1982) found that the addition
1940,1947; Willis 1963;Boorman& Fuller1982;Kachi of nutrientsto grazed dune swardscaused a decline of
& Hirose 1983; Dougherty et al. 1990). Phosphorus annualspecies, but,due to the grazing,perennialgrasses
limitationcan be an explanationwhy nitrogenfertiliz- did not dominatethe sward.When grazing animalsare
ation did not have any effect in this study. At the study presentin nutrient-poor systems like chalk grasslands
sites atmosphericphosphatedeposition of 0.3 kg P/ha/ (Bobbink 1991) and heathland(Farrow 1917; Heil &
yr (ten Harkel et al. 1991) and the water soluble soil Diemont 1983), no dominanceby graminoidsoccurs.In
phosphatecontentof ca. 0.3 kg P/ha in the 8 cm surface general,it is seen thatmoderategrazing increasesplant
soil (ten Harkel 1992) are very low. However, in this species diversity (Puertoet al. 1990; Gibson & Brown
study we focused on the role of nitrogenin the vegeta- 1991; Belsky 1992).
tion development.The availabilityof phosphorus the to The introductionof horses at the grass-dominated
vegetationin the Dutchdunes is currently being studied. site showed that the dominanceof graminoidscan be
If phosphatewould have been the limitingfactorin counteractedby introducinggrazing. Light penetrates
the past,the presentdominanceof graminoidsin partsof for
into the soil surfaceand providesopportunities ger-
the Dutch dunes cannotbe explained. Therefore,phos- mination of seedlings (Willems 1983; Bobbink &
phate can be a limiting factor only at high nitrogen Willems 1991). Herbs increase and species richness
levels. This was also shown by Doughertyet al. (1990). may be restored.When atmospheric depositionof nitro-
gen remainshigh, grazing will always be necessaryto
Changes in ecosystemsother than nutrlient
availability preservespecies-richdune grasslandsfor the future.
Research on heathland (Heil & Diemont 1983)
showed that a relatively old stand of Calluna vulgaris Acknowledgements. thank DuneWater
We the of
Company
(ca. 15 yr) was not significantly affected by nitrogen Zuid-Holland,notably its Departmentof Nature and Land-
application. The closed canopy of Calluna vulgaris scape Managementfor theirsubstantial financialsupportand
the Department Implementation Managementfor exten-
of of
preventedFestuca ovina to take advantageof the nitro- sive technicalfacilities to carryout the field experiments.
J.H.
gen application.However, when Calluna vulgaris died van Boxel, G.W. Heil and A. Kooijman assisted with the
aftera heatherbeetle infestationand the vegetationwas researchandgave valuablehelp duringthe compilationof this
opened up, Festuca ovina became dominantafter re- article. We thank P.D. Jungerius and J.M. Verstratenfor
peated applicationof nitrogen. criticalreadingof the text. D. Warmerdam acknowledged
is
Chalkgrasslandscan sustainhigh loads of nutrients for help with the correctionof the English text.
when the managementis appropriate. Mowing andhar-
vesting of vegetation at a site dominatedby Brachy-
podiumpinnatumwas adequateto restorea characteris-
tic species-richchalk grasslandwithin 5 yr (Bobbink&
Willems 1991).
- Impact of grazing and atmospheric nitrogen deposition - 451
References J. Veg. Sci. 2: 291-300.
Heil, G.W. & Diemont, W.H. 1983. Raised nutrientlevels
Anon. 1987. StatisticalAnalyzingSystemRelease 6.03. SAS change heathlandinto grassland.Vegetatio53: 113-120.
InstituteInc., Cary,NC. Hill, M.O., Evans, D.F. & Bell S.A. 1992. Long-termeffects
Aerts, R. & Heil G.W. (eds.) 1993. Heathlands:patterns and of excluding sheep from hill pasturesin North Wales. J.
processes in a changing environment. Kluwer Academic Ecol. 80: 1-13.
Publishers,Dordrecht. Kachi, N. & Hirose, T. 1983. Limiting nutrientsfor plant
Bakker,J.P., de Leeuw, J. & van Wieren, S.E. 1983. Micro- growthin coastal sand dune soils. J. Ecol. 71: 937-944.
patternsin grasslandvegetationcreatedand sustainedby Milton, W.E.J. 1940. The effect of manuring,grazing and
sheep-grazing.Vegetatio55: 153-161. cutting on the yield, botanicaland chemical composition
Belsky, A.J. 1986. Revegetation of artificialdisturbancesin of naturalhill pastures.I. Yield and botanicalsection. J.
grasslands of the Serengeti National Park, Tanzania. I. Ecol. 28: 326-356.
Colonizationof grazed and ungrazedplots. J. Ecol. 74: Milton, W.E.J. 1947. The yield, botanicalandchemical com-
419-437. position of naturalhill herbageundermanuring,control-
Belsky, A.J. 1992. Effects of grazing, competition, distur- led grazing and hay conditions. I. Yield and botanical
bance and fire on species composition and diversity in section.J. Ecol. 35: 65-95.
grasslandcommunities.J. Veg. Sci. 3: 187-200. Mountford,J.O., Lakhani,K.H. & Kirkham,F.W. 1993. Ex-
Bhadresa,R. 1977. Food preferenceof rabbits Oryctolagus perimentalassessment of the effects of nitrogenaddition
cuniculus L. at Holkham sand dunes, Norfolk. J. Appl. underhay-cuttingandaftermath grazingon the vegetation
Ecol. 14: 287-291. of meadows on a Somersetpeat moor. J. Appl. Ecol. 30:
Bobbink, R. 1991. Effects of nutrientenrichmentin Dutch 321-332.
chalk grassland.J. Appl.Ecol. 28: 28-41. Mueller-Dombois,D. & Ellenberg,H. 1974. Aims and meth-
Bobbink,R. & Willems, J.H. 1991. Impactof differentcutting ods of vegetation ecology. Wiley, New York, NY.
regimes on the performanceof Brachypodium pinnatum Olff, H., Huisman, J. & van Tooren, B.F. 1993. Species
in Dutch chalk grassland.Biol. Conserv.56: 1-21. dynamicsandnutrientaccumulation duringearlyprimary
Bobbink,R., Bik, L. & Willems,J.H. 1988. Effects of nitrogen succession in coastal sand dunes.J. Ecol. 81: 693-706.
fertilization on vegetation structureand dominance of Puerto, A., Rico, M., Matfas, M.D. & Garcia, J.A. 1990.
Brachypodium pinnatum (L.) Beauv. in chalk grassland. Variation structure diversityin Mediterranean
in and grass-
Acta Bot. Neerl. 37: 231-242. landsrelatedto trophicstatusandgrazingintensity.J. Veg.
Bobbink,R., Boxman,D., Fremstedt, Heil, G., Houdijk,A.
E., Sci. 1: 445-452.
& Roelofs, J. 1992. Critical loads for nitrogen eutro- Ranwell, D.S. 1960. Newborough Warren, Anglesey. III.
phication of terrestrialand wetland ecosystems based Changesin the vegetationon partsof the dunesystem after
upon changesin vegetationandfauna.In: Grennfelt,P. & the loss of rabbitsby myxomatosis.J. Ecol. 48: 385-397.
Th6rnelof,E. (eds.) Critical loadsfor nitrogen, pp. 111- Roelofs, J.G.M. 1986. The effect of airborne sulphur and
159. Nordic Council of Ministers,Copenhagen. nitrogen deposition on aquatic and terrestrialheathland
Bobbink,R., den Dubbelden,K. & Willems, J.H. 1989. Sea- vegetation.Experientia42: 372-377.
sonal dynamicsof phytomassandnutrientsin chalkgrass- Silvertown J., Watt, T.A., Smith, B. & Treweek, J.R. 1992.
land. Oikos 66: 216-224. Complex effects of grazing treatmenton an annual in a
Bobbink,R., Heil, G.W. & Raessen, M.B.A.G. 1990. Atmos- species-poorgrasslandcommunity.J. Veg. Sci. 3: 35-40.
pheric depositionand canopyexchangein heathlandeco- Stuyfzand,P.J. 1993. Hydrochemistryand hydrology of the
systems.DutchPriorityProgramme Acidification,Re-
on coastal dune area of the Western Netherlands. Ph.D.
port 119. Departmentof PlantEcology and Evolutionary Thesis, Free Universityof Amsterdam.
Biology, Universityof Utrecht. Thomas, A.S. 1963. Furtherchanges in vegetation since the
Boorman,L.A. & Fuller, R.M. 1982. Effects of addednutri- adventof myxomatosis.J. Ecol. 51: 151-183.
ents on dune swards grazed by rabbits.J. Ecol. 70: 345- ten Harkel,M.J. 1992. Abiotiek-biotiekinteractiesin de droge
355. duinenvan Meijendel.MeijendelMeded.(Den Haag) 24:
Davy, A.J. & Bishop, G.F. 1984. Response of Hieracium 65-89. (In Dutch.)
pilosella in Brecklandgrass heath to inorganicnutrients. ten Harkel, M.J., Heil, G.W., van der Hagen, H.G.J.M. &
J. Ecol. 72: 319-330. Witter,J.V. 1991. Abiotic-bioticinteractionsin dry dune
Dougherty, K.M., Mendelssohn, I.A. & Monteferrante, F.J. ecosystems. Mitt.Hydrograph.Dienst. Osterreich65/66:
1990. Effect of nitrogen,phosphorus potassiumaddi-
and 58-62.
tions on plantbiomass and soil nutrientcontentof a swale van der Meijden, R. 1990. Flora van Nederland. 21th ed.
barrier strandcommunityin Louisiana.Ann.Bot. 66: 265- Wolters-Noordhoff, Groningen.
271. van Hecke, P., Impens, I. & Behaeghe, T.J. 1981. Temporal
Farrow, E.P. 1917. On the ecology of the vegetation of variationof species composition and species diversity in
breckland.III. Generaleffects of rabbitson the vegeta- permanentgrasslandplots with different fertilizer treat-
tion. J. Ecol. 5: 1-18. ments. Vegetatio47: 221-232.
Gibson, C.W.D. & Brown, V.K. 1991. The effects of grazing Watt, A.S. 1962. The effect of excluding rabbitsfrom grass-
on local colonisationandextinctionduringearlysuccesion. land A (Xerobrometum) breckland,1936-60. J. Ecol.
in
452 ten Harkel, M.J. & van der Meulen, F.
50: 181-198. Williams, O.B., Wells, T.C.E. & Wells, D.A. 1974. Grazing
Westhoff, V 1985. Nature managementin coastal areas of managementof Woodwaltonfen: seasonal changesin the
WesternEurope.Vegetatio62: 523-532. diet of cattle and rabbits.J. Appl. Ecol. 11: 499-516.
Westhoff,V. & den Held, A.J. 1969. Plantengemeenschappen Willis, A.J. 1963. BrauntonBurrows:the effects on the veg-
in Nederland.Thieme, Zutphen. etation of the addition of mineral nutrientsto the dune
Willems, J.H. 1983. Species composition and above-ground soils. J. Ecol. 51: 353-374.
phytomassin chalk grasslandwith differentmanagement. Willis, A.J. & Yemm, E.W. 1961. Braunton Burrows;mineral
Vegetatio52: 171-180. nutrientstatusof the dune soils. J. Ecol. 49: 377-390.
Willems, J.H., Peet, R.K. & Bik, L. 1993. Changesin chalk- Zeevalking,H.J. & Fresco, L.F.M. 1977. Rabbitgrazingand
grassland structureand species richness resulting from species diversityin a dune area.Vegetatio 35: 193-196.
selective nutrientadditions.J. Veg. Sci. 4: 203-212.
Received 23 January1995;
Revision received 23 May 1995;
Accepted 5 September1995.
6;.4'*+$22:$.02
1;&3(+<2=>'?$&&3-@2'AB'&4.'C$+D4:'$.0'E+$.D'F$.'04+'?4;:4.
G(;+%4>'A(;+.$:'()'74/4&$&-(.'G%-4.%4H'7(:B'IH'5(B'J'<A;.BH'KLLM=H'##B'NNOPNOQ
R;S:-2340'S8>'T#;:;2'R+422
G&$S:4'UVW>'http://www.jstor.org/stable/3236288
1%%42240>'QMXKYXQYYZ'QJ>ON
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=opulus.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
Opulus Press is collaborating with JSTOR to digitize, preserve and extend access to Journal of Vegetation
Science.
http://www.jstor.org
Journal of VegetationScience 6. 445-452, 1995
? IAVS;OpulusPress Uppsala. Printedin Sweden
445
Impact of grazing and atmospheric nitrogen deposition
on the vegetation of dry coastal dune grasslands
ten Harkel, Matthijs J.* & van der Meulen, Frank
Landscape and Environmenltal Research Group, University of Amsterdam, Nieuwe Prinsengr7acht 130, 1018 VZ
Amsterdam, The Netherlands; *Correspondingauthor. Tel. + 31 20 5257442; Fax + 31 20 5257431;
E-mail mth@fgb.frw.uva.nl
Abstract. A five-year experimentalstudy was carriedout to 1988, 1990; Bobbink 1991; Aerts & Heil 1993). So far,
examine the combined effects of grazing and atmospheric such systematic studies have not been initiated in the
nitrogen deposition on the vegetation of three dry dune dunes. The present study was initiated in 1988, in order
grasslands: shortspecies-rich,one shortspecies-poor,and
one
to understand the relations between atmospheric deposi-
one predominated tall graminoids.
by
Additional fertilizationwith nitrogen had no significant tion, grazing and vegetation dynamics. The present pa-
effect, neitherin grazednor in non-grazedplots. Exclusion of per deals with changes in species composition.
Since the 1970s some tall perennial graminoids,
grazingby rabbitsresultedin an increasein the frequencyof
perennialgraminoidsand a decrease in the frequencyof an- notably Ammophila arenaria, Avenulapubescens, Cala-
nual graminoids and herbs. Nevertheless, species diversity magrostis epigejos, Elymus athericus and Carex are-
remainedthe same in the species-richgrassland.During the naria, have become dominant in many originally short
experiment,the above-groundbiomass increasedin all non- dry dune grasslands of the Netherlands (C.T.M. Vertegaal
grazed plots and the amount of bare soil and mosses de- et al. int. report 1991). As a consequence, these grass-
creased.
lands show now a relatively low species diversity. From
The vegetation changes occurredmainly within one year
the point of view of nature conservation this is consid-
after the exclusion of grazing. An exception is the grass-
dominatedsite where the amountof Calamagrostisepigejos ered an unfavourable situation (Westhoff 1985). Nowa-
increased graduallyfrom ca. 20 % in the first two years to days, nature managers try to counteract this develop-
about50 % in the fourthand fifth year. ment and try to restore the original vegetation composi-
Grazingby rabbitsseems essential to preventgraminoids tion. This requires understanding of the processes respon-
to become predominantin the dry dunes. If graminoidsare sible for the predominance of graminoids.
dominant,grazingby horses can be an appropriate methodto In the past, tall graminoids were usually not domi-
restorethe originalgrasslandvegetation.After six months of nant because of the low nutrient status of sandy dune
grazingby horsesthe grass-dominated showed a decrease
site
soils (Willis & Yemm 1961; Boorman & Fuller 1982).
of the frequencyof perennialgraminoids,from 95 %to 80 %,
The present dominance of graminoids might be a result
and an increase of the frequency of perennial herbs, from
2.5 % to between 13 and 20 %. of the recent increase of atmospheric nitrogen deposi-
tion. From 1930 to 1980 the atmospheric nitrogen depo-
sition (NH4+ + NO3-) on coastal areas in the Netherlands
Keywords: Horse;Management; Meijendel;the Netherlands; increased from 4 kg N.ha-l.yr- to 14 kg N.ha-l.yr-1
Rabbit;Restorationecology; Species diversity. (Stuyfzand 1993). The predominance of graminoids due
to high nitrogen loads is also known from heathlands
Nomenclature: van der Meijden (1990). (Heil & Diemont 1983; Roelofs 1986; Bobbink et al.
1990; Aerts & Heil 1993) and chalk grasslands (Bobbink
et al. 1988; Bobbink 1991; Willems et al. 1993). Many
Introduction graminoids have a relative higher growth rate than other
plant species and are therefore able to benefit more from
Recently, semi-natural vegetation in the Netherlands increased nitrogen input. This is shown by fertilizing
has been under increasing influence of atmospheric experiments in grasslands (Willis & Yemm 1961; Willis
deposition of acid and eutrophic components. This has 1963; Van Hecke et al. 1981; Davy & Bishop 1984;
led to a decrease in nature conservation values. Bobbink 1991; Mountfordet al. 1993). When grasses
In chalk grasslands and heathlands the influence of aretall, light will not penetrate the soil andgermina-
into
atmospheric deposition has been studied to support man- tion of annualspecies is prevented(Bobbink& Willems
agement programs aiming at restoration of nature values 1991; Olff et al. 1993).
(Heil & Diemont 1983; Roelofs 1986; Bobbink et al. It seems thatgraminoidswill not dominatewhen the
446 ten Harkel,M.J. & van der Meulen, F.
vegetation is frequentlygrazedby animalssuch as rab- with a pH(CaCl2)of 4.2 to 5.4 on similar calcareous
bits (Williamset al. 1974;Bhadresa1977;Zeevalking& dunesandas the firstsite. The 10 cm thicktop soil has an
Fresco 1977). If therabbitpopulationdecreases,as in the organic matter content of 3 %. The site representsa
1950s (dueto myxomatosis),graminoids ableto take
are species-poor variant of the Violo-Corynephoretum as-
over (Ranwell 1960; Thomas 1963; Westhoff 1985). sociation(Westhoff& den Held 1969). A certaindomi-
Theeffect of rabbitgrazingis also experimentally shown nance of graminoidsoccurredhere from the beginning
in grazedgrasslands whereafterthe exclusionof grazing (Table 1). This site has been monitoredsince the sum-
tall graminoidsbecame dominant(Farrow1917; Hill et mer of 1988.
al. 1992). Even afterrestoration the rabbitpopulation,
of The third,grass-dominated is close to the second
site
graminoidsare able to remaindominantfor a long time site, butthe vegetationhas becometotallypredominated
as rabbitsdo not eat or enter such relatively tall grass by Carex arenaria, Festuca ovina + F. rubra and
swards.Non-grazedvegetationhas a relativelylow pro- Calamagrostis epigejos (Table 1). This site has been
tein content in comparison to grazed vegetation and monitoredsince the summerof 1988.
thereforeanimalsavoid the area (Bakkeret al. 1983).
The dominanceof graminoidscan be counteracted
by introducinglarge herbivores(Williams et al. 1974).
If heavy grazing is needed to permitseedling establish-
ment of annual species, it can also prohibitsuccessful
Table1. Totalnumber speciesandfrequency
of (averagefor
completion of the life cycle of these annual species allyears) plant
of speciesof thecontrol
plotsin thethree
sites
(Farrow1917; Watt 1962; Silvertownet al. 1992). Re- (grass-dominated before introduction horses);
site the of Rich
cently, horses and cows have been introduced Dutch
in = species-rich site; Poor = species-poor site; Grass = grass-
naturereservesto reducethe amountof biomassandthe dominated
site.
vegetation height and thus restore the original species Rich Poor Grass
composition.
This paper reports on an experiment in a Dutch Number of species
coastal dune area where fertilizationwith nitrogenand Annual herbs 4 3 1
exclusion of grazingwere combinedas a treatment. The Perennial herbs 9 3 3
Annual graminoids 2 1
purpose was to define the main reason for the domi- Perennial graminoids 6 6 6
nance of tall graminoids in large parts of comparable Annual herbs
dune areas. This dominance may be caused by high Cerastiluml semidecandrllm 1-5 <1 0
atmosphericdeposition of nitrogen, or the absence of El-odiliuti c(icultariiu 5- 10 <1 0
grazingby rabbits,or a combinationof both.Inaddition, Se'I'cio) svlvaticus 0 1-5
the effects of the introduction horsesandthe implica-
of Teeshiali(a illcicaulis 0 <1 0
Veronica arvcnsis <1 0 0
tions for the managementof dune grasslandswill be 1-5
Violla (cr1'tisii 0
discussed. Perennial herbs 0
Galilum verum 1-5 <1
1-5
0
Hiera t'i71I pilosellc <1 0
Material and Methods Holcusl /l(anatlus 0 <1
Leontodlonl saxatilis 1-5 0 0
Lotus corni(iculcatus 1-5 0
Study sites Myosotis arvenlsis 1-5 0
Ohonis repens 10- 20 0
The study was carriedout in threegrasslandsites in Rumnex acetosella 0 10- 20 0
Sedumi acre <1 0 0
Meijendel,a coastal dune areanorthof The Hague, the Senecio jacobaea 5-10 < 1
1-5
Netherlands.This area is managedby the Dune Water 1-5
Tarax acum sect. Erythrospe rma 0 0
Companyof Zuid-Holland(D.Z.H.). Annual graminoids
The first, species-richsite is located at a distanceof Aila praecox 1-5 5- 10 1-5
0.5 km fromthe coast. The sandysoil is calcareous,with Phleuni arenar-iuni <1 0 0
2- 3 % CaCO3, and with a very low organic matter Perennial graminoids
Agrostis stolonifera + A. capillaris 0 5- 10 0
content(2 %in the 10cm top soil). Phytosociologically,
Calamagriostis epigejos 1-5 10 - 20 10 - 20
the vegetationcan be assignedto the Festuco-Galietum Carex arenaaia 5- 10 5-10 >20
ma-ritimi association(Westhoff& den Held 1969). This Festuca ovina + F. orubra 1-5 >20 10 - 20
site has been monitoredsince the summerof 1989. Koeleria macrantha <1 0 1-5
The second, species-poor site is situated 2.5 km Luzula campestris 10 - 20 10 - 20 1-5
Poa pratensis 5- 10 < 1 <1
from the coast. The soil is decalcified down to 50 cm
- Impactof grazingand atmosphericnitrogendeposition- 447
Treatments Statisticalanalysis
At all threesites five homogeneousquadrats 2 m x
of The resultswere analyzedstatisticallywith the com-
2 m were selected. Each quadratwas divided into four puter package SAS (Anon. 1987). The counts of the
experimentalplots of I m x 1m: (1) control; (2) ferti- plants were standardizedto frequencies and will be
lized; (3) excluded from grazing by rabbits and (4) called frequenciesfurtheron.
combinationof fertilizationand exclusion of grazing. The effect of grazingandfertilizationon the follow-
The treatmentswere allocated randomly to the plots ing three frequencieswere tested with the SAS-proce-
(randomizedquadrat design). dure ANOVA with repeated measurements:species
In the autumnof 1990, horses were introducedin a groups (annualor perennialherbs and graminoids),in-
partof Meijendel to reduce severe grass encroachment dividualspecies, surfacewithoutphanerogamic species
andto possibly restorethe originalspecies composition. (baresoil and mosses). The effects on species diversity
The horsesgrazedfor six monthsat the grass-dominated (see below) and biomass were also tested. A signifi-
site;thereafter horseswereexcludedfromthequadrats
the cance leve of 0.05 was chosen. The effects of the treat-
because they demolishedthe exclosures. ments were analysed separatelyfor each site. The ef-
The fertilizerwas appliedto simulatethe effects of fects of treatmentson the grass-dominatedsite were
high loads of atmosphericnitrogen deposition. It was analysedtwofold: before the introduction the horses
of
assumedthatadditionof nitrogenwould speed up grass and after six months of grazingby horses.
encroachmentand cause the associatedloss of charac- Species diversity was measured as the Shannon
teristicherbs. Wiener index:
From October 1988 until April 1992 the fertilizer
appliedconsistedof pellets of NH4NO3 (addedin spring SD = P ln i (1)
and autumn)equalto an extradepositionof 25 kg N/ha/
yr. As weekly unsaturatedsoil water analysis by ce- where SD = species diversity andPi = the proportionof
ramiccups showed thatlarge amountsleached through species i (on a frequencybasis).
the surfacesoil within a few weeks, a differentmethod The moss species were regardedas one group,and were
of fertilizationwas applied. Since April 1992 quadrats not includedin this calculation.
were fertilized with a 0.0069 M (NH4)2SO4 solution
once every fortnight,which was equal to an extranitro-
gen input of 50 kg N/ha/yr. For osmotic reasons NaCl Results
(sea salt)was addedin this solutionup to the same NaCl
concentration the rainwater.
in The non-fertilizedplots The species presentin the threesites were dividedin
obtained an equal amount of NaCl solution without four groups: annual herbs, perennial herbs, annual
(NH4)2SO4. The yearly amount of solution given was graminoids and perennial graminoids (Table 1). The
equivalentto an extraprecipitation 26 mm/yr,which
of grazedquadrats the species-richsite containthe larg-
of
is only 3 % of the averageyearly precipitation. est number of herb species, 13. The numbers of the
species-poorand the grass-dominated were six and
site
Measurements four respectively. The numberof graminoidspecies is
equal in the sites, i.e. seven or eight. However, in the
Once a year in July, the species composition of the grazedplots the frequencyof perennialgraminoids(Ta-
vegetation was recordedin each plot by counting the ble 2) is much higher in the grass-dominated (92 %
site
presence of all separatephanerogamsand of the rest priorto grazingby horses) thanin the species-poorsite
(bare soil and mosses) with a point frequencymethod (63 %) and the species-rich site (44 %). However, the
(Mueller-Dombois& Ellenberg1974), using 100 cross- amountsof perennialgraminoidspecies in the species-
points of a grid. rich and grass-dominatedsystems are the same (Table
The standingcropbiomass (dead+living)was meas- 1).Thespeciesresponsible thedominance gramino-
for of
ured each year in July by cutting,drying and weighing ids in grass encroachedstages are alreadypresentin the
threerandomlyselected subsamplesof 20 cm x 20 cm, species-richstage.
each at the edge of all plots. The centralpartof each plot
(60 cm x 60 cm) has been left untouchedfor long-term Effects offertilization and grazing on species groups
monitoring.
The impactof nitrogenfertilizationon the change of
species groups and the three other variables was not
statisticallysignificant at all sites! Thereforefertilized
448 ten Harkel, M.J. & van der Meulen, F.
Table 2. Average frequency over all years of the species groups in the grazed plots and exclosure plots of the three sites.
Species group Annual Perennial Annual Perennial
herbs herbs graminoids graminoids
[%] [%] [%] [%]
Treatment G E G E G E G E
Site
Species-rich 7.5 ** 1.7 ** 41.5* 33.6 * 4.9 2.6 43.8 ** 56.0 **
Species-poor 1.2 0.1 24.4* 12.2* 10.5 4.3 62.7** 82.2**
Grass-dominated1 1.1 0.3 2.5 2.5 3.9 3.5 92.4 93.7
Grass-dominated2 0.1 0.2 12.8 * 4.1 * 6.9 ** 0.4 ** 80.0 ** 95.3 **
G = grazedplots; I = before the introduction horses;
of
E = exclosure plots; 2 = afterthe introduction horses.
of
Significance of differencesbetween grazedand exclosureplots: * = p < 0.05; **= p < 0.01.
and non-fertilized plots were taken together during the between grazed and non-grazed plots (Table 2). In con-
statistical analysis of the effects of grazing. trast, the frequency of annual graminoids and perennial
The effects of the exclusion of grazing by rabbits on herbs increased (respectively 0.4% to 7%, and 4% to
the species groups are shown in Table 2. The exclusion 13%). The horse-grazed plots showed an increase of the
resulted in an increase of perennial graminoids in both amount of bare soil and mosses, and of the species
the species-rich and species-poor sites. No increase was diversity. In the same time the biomass showed a de-
observed in the grass-dominated site, because the fre- crease.
quency of perennial graminoids was high already from
the beginning (92 %). The frequency of perennial herbs Eff#cts of fertilization and grazing on indilidual
decreased in both the species-rich and species-poor is
sites. In the species-rich site the frequency of annual
herbs also showed a decrease. Fertilization with nitrogen did not affect individual
Exclusion of grazing at the species-rich and the species. Only Luzula cnampestris decreased in the
species-poor site reduced the amount of bare soil and exclosures of the grass-dominated site. To test the effect
mosses, while the biomass of the stand increased (Table of grazing on individual species, fertilized and non-
3). Moreover, a decrease in species diversity was seen in fertilized plots were also taken together.
the species-poor site. The significant changes of individual species after
Before the introduction of horses on the grass-domi- the exclusion from grazing are shown in Table 4.
nated site, the amount of bare soil and mosses became Some species showed a decrease at the species-poor
lower in the exclosure plots. Six months of grazing by site. No increases were detected. However, in the last
horses resulted in a difference of perennial graminoids paragraph an increase in perennial graminoids was
Table 3. Average percentage of all years of bare soil and mosses, species diversity and biomass of the three sites.
Variable Bare + mosses Species diversity Biomass
[%] (see formula1) [g/m-2]
Treatment G E G E G E
Site
Species-rich 43.9 ** 8.5 ** 1.945 1.983 42 ** 264 **
Species-poor 57.1 ** 8.1** 1.635 * 1.356* 75 ** 467 **
Grass-dominated' 17.8** 5.6** 1.134 1.213 335 536
Grass-dominated2 45.2 ** 3.7 ** 1.664 1.465 * 256 ** 613 **
G = grazedplot; 1 = before the introduction horses;
of
E = exclosure plot; 2 = afterthe introduction horses.
of
Significance of differencesbetween grazedand exclosureplots: * =p < 0.05; **=p < 0.01.
- Impact of grazing and atmospheric nitrogen deposition - 449
shown.This contradiction mightbe due to an increaseof Vegetation change after the exclusion of grazing
differentperennialgraminoidsin each plot.
In the grass-dominatedsite no changes due to the Vegetation change was defined as the differencein
exclusion from grazingwere foundbefore the introduc- frequencybetweentwo yearsof the species thatchanged
tion of grazingby horses. The whole site is dominated in the non-grazed plots relative to the grazed plots
by perennialgraminoids. (Table 4). The difference remainedconstant after the
Species thatincreasedafterthe exclusion fromgraz- first year, the changes measured occurred during the
ing or decreased after the introductionof grazing by first year. The vegetation change was calculated sepa-
horses are all perennialgraminoids.One exception is ratelyfor frequencieswhich increasedor decreaseddue
the increase of Phleum arenariumat the species-rich to grazing.
site undernon-grazingconditions.This is unexpectedas At the species-rich site the vegetation change of
is
it is a small annualgrassandits germination normally species which benefited from the exclusion of grazing
preventedin a vegetationpredominated tall grasses
by was small, viz. from 11 % to 16 %. The vegetation
such as Festuca rubra + F. ovina and Poa pratensis changeof species which decreasedafterthe exclusion of
(Table 4). Species that showed a decrease after exclu- grazingall occurredduringthe first year.
sion or an increaseafterthe introductionof grazing by At the species-poorsite all changes occurredduring
horses,areherbsor smallperennialgraminoids,such as the first year.
Luzula campestris. At the grass-dominatedsite the vegetation change
for species which increasedaftergrazingby horses was
the same for species which increased after the exclu-
sion, i.e. from 3 % to 25 %. Calamagrostis epigejos is
the only species which increasedin the exclosures com-
paredto the controlplots. In spite of the introductionof
horses the frequency of Calamagrostis epigejos remained
the same (ca. 20 %) in the grazed plots, therefore it
increasedfrom 20 % to ca. 50 % in the exclosures. The
exclosures developed from a dominance of several
graminoids (Table 1) into a dominance by only one
species. The frequencyof the species favouredby graz-
Table4. Plant
species after
development excluding or
grazing ing showed an increasein the grazedplots from 7 % to
after of at site.
introduction horses thegrass-dominated 30%, while it remainedlow in the exclosures(3 %).The
increaseof these species occurredafterthe introduction
Species-rich site
Festuca ovina +F. rubra + of the horses. The vegetation level of the grazed plots
Poa pratensis + was lowered and this gave the plants an opportunityto
Phleumarenarium +
Cerastiumsemidecandrum grow.
Erodiumcicutarium
Luzulacampestris
Myosotisarvensis Discussion
Seneciojacobaea
Taraxacumsect. Erythrosperma
Viola curtisii This study showed thatmaintenanceof species-rich
Species-poor site dry dune grasslandsstronglydependson the activity of
Rumexacetosella grazinganimals,especiallyunderrelativelyhigh atmos-
Seneciojacobaea
pheric nitrogendeposition levels. When grazing is ex-
Grass-dominated site cluded, perennialgraminoidsbecome dominantwithin
Before introduction horses
of no change one ortwo years.These graminoids were alreadypresent
After introduction horses
of
in the three sites before the treatment(Table 1). As in
Aira praecox + other grazing experiments (Belsky 1986; Hill et al.
Luzulacampestris +
1992) no new plant species invadedthe plots.
Rumexacetosella ++
Seneciojacobaea +
Underrelatively high atmosphericnitrogendeposi-
Calamagrostisepigejos tion levels, fertilizationby nitrogenhad no effect on the
three sites during the 5 yr of this study. This is not
+ = increase (p < 0.05);
++ = increase (p < 0.01);
consistent with fertilizer experiments in other grass-
=decrease (p < 0.05); lands (Willis 1963; Van Hecke et al. 1981; Davy &
- =decrease (p < 0.01).
Bishop 1984; Bobbink et al. 1988; Bobbink 1991;
450 ten Harkel,M.J. & van der Meulen, F.
Mountfordet al. 1993). Effects of additionalnutrient These studies show that when no other changes
supply were always registeredwithin 5 yr. In our case occur, the vegetationis probablynot seriously affected
the amountof nitrogen fertilizerof 50 kg N/ha/yr and by an increasingavailabilityof nutrients.Whengrazing
background depositionof nitrogenof 10 kg N/ha/yr(ten is continued,changes will probablybe very slow and
Harkel et al. 1991) should be sufficient to accelerate cannotbe measuredwithin 5 yr.
changes in species composition if other resources are The main conclusion of this study is that with high
not limited.Bobbinket al. (1992) gave a criticalload of nitrogenlevels a dunegrasslanddoes not become grass-
nitrogenfor dune grasslandsof 20-30 kg N/ha/yr. dominatedas long as the vegetationis grazed,not even
Some explanations for these deviating results are when fertilized.When grazing is absent, the relatively
given below. high availabilityof nitrogen gives graminoidsthe op-
portunityto become dominant.Effects of a decreaseof
Phosphorusas a limitingfictor grazing intensity were reportedby Ranwell (1960): an
increasein the growthand floweringof graminoidsand
Fertilization experiments with different combina- a decreaseof herbswas foundafterthe disappearance of
tions of nutrients(nitrogen,phosphorus, potassium,cal- rabbits myxomatosisin a sand-dunesystem.Willis &
by
cium) showed that the nitrogencontent of the soil was Yemm (1961) andWillis (1963) showed, however,that
the most limiting factor (Milton 1940, 1947; Willis even without grazing a very nutrient-poor sandy dune
1963; Boorman& Fuller 1982; Kachi & Hirose 1983; system can remain species-rich. However, the system
Doughertyet al. 1990; Olff et al. 1993). However, also became dominatedby grasses after addition of nutri-
phosphoruscan be an importantlimitingfactor(Milton ents. Boorman& Fuller (1982) found that the addition
1940,1947; Willis 1963;Boorman& Fuller1982;Kachi of nutrientsto grazed dune swardscaused a decline of
& Hirose 1983; Dougherty et al. 1990). Phosphorus annualspecies, but,due to the grazing,perennialgrasses
limitationcan be an explanationwhy nitrogenfertiliz- did not dominatethe sward.When grazing animalsare
ation did not have any effect in this study. At the study presentin nutrient-poor systems like chalk grasslands
sites atmosphericphosphatedeposition of 0.3 kg P/ha/ (Bobbink 1991) and heathland(Farrow 1917; Heil &
yr (ten Harkel et al. 1991) and the water soluble soil Diemont 1983), no dominanceby graminoidsoccurs.In
phosphatecontentof ca. 0.3 kg P/ha in the 8 cm surface general,it is seen thatmoderategrazing increasesplant
soil (ten Harkel 1992) are very low. However, in this species diversity (Puertoet al. 1990; Gibson & Brown
study we focused on the role of nitrogenin the vegeta- 1991; Belsky 1992).
tion development.The availabilityof phosphorus the to The introductionof horses at the grass-dominated
vegetationin the Dutchdunes is currently being studied. site showed that the dominanceof graminoidscan be
If phosphatewould have been the limitingfactorin counteractedby introducinggrazing. Light penetrates
the past,the presentdominanceof graminoidsin partsof for
into the soil surfaceand providesopportunities ger-
the Dutch dunes cannotbe explained. Therefore,phos- mination of seedlings (Willems 1983; Bobbink &
phate can be a limiting factor only at high nitrogen Willems 1991). Herbs increase and species richness
levels. This was also shown by Doughertyet al. (1990). may be restored.When atmospheric depositionof nitro-
gen remainshigh, grazing will always be necessaryto
Changes in ecosystemsother than nutrlient
availability preservespecies-richdune grasslandsfor the future.
Research on heathland (Heil & Diemont 1983)
showed that a relatively old stand of Calluna vulgaris Acknowledgements. thank DuneWater
We the of
Company
(ca. 15 yr) was not significantly affected by nitrogen Zuid-Holland,notably its Departmentof Nature and Land-
application. The closed canopy of Calluna vulgaris scape Managementfor theirsubstantial financialsupportand
the Department Implementation Managementfor exten-
of of
preventedFestuca ovina to take advantageof the nitro- sive technicalfacilities to carryout the field experiments.
J.H.
gen application.However, when Calluna vulgaris died van Boxel, G.W. Heil and A. Kooijman assisted with the
aftera heatherbeetle infestationand the vegetationwas researchandgave valuablehelp duringthe compilationof this
opened up, Festuca ovina became dominantafter re- article. We thank P.D. Jungerius and J.M. Verstratenfor
peated applicationof nitrogen. criticalreadingof the text. D. Warmerdam acknowledged
is
Chalkgrasslandscan sustainhigh loads of nutrients for help with the correctionof the English text.
when the managementis appropriate. Mowing andhar-
vesting of vegetation at a site dominatedby Brachy-
podiumpinnatumwas adequateto restorea characteris-
tic species-richchalk grasslandwithin 5 yr (Bobbink&
Willems 1991).
- Impact of grazing and atmospheric nitrogen deposition - 451
References J. Veg. Sci. 2: 291-300.
Heil, G.W. & Diemont, W.H. 1983. Raised nutrientlevels
Anon. 1987. StatisticalAnalyzingSystemRelease 6.03. SAS change heathlandinto grassland.Vegetatio53: 113-120.
InstituteInc., Cary,NC. Hill, M.O., Evans, D.F. & Bell S.A. 1992. Long-termeffects
Aerts, R. & Heil G.W. (eds.) 1993. Heathlands:patterns and of excluding sheep from hill pasturesin North Wales. J.
processes in a changing environment. Kluwer Academic Ecol. 80: 1-13.
Publishers,Dordrecht. Kachi, N. & Hirose, T. 1983. Limiting nutrientsfor plant
Bakker,J.P., de Leeuw, J. & van Wieren, S.E. 1983. Micro- growthin coastal sand dune soils. J. Ecol. 71: 937-944.
patternsin grasslandvegetationcreatedand sustainedby Milton, W.E.J. 1940. The effect of manuring,grazing and
sheep-grazing.Vegetatio55: 153-161. cutting on the yield, botanicaland chemical composition
Belsky, A.J. 1986. Revegetation of artificialdisturbancesin of naturalhill pastures.I. Yield and botanicalsection. J.
grasslands of the Serengeti National Park, Tanzania. I. Ecol. 28: 326-356.
Colonizationof grazed and ungrazedplots. J. Ecol. 74: Milton, W.E.J. 1947. The yield, botanicalandchemical com-
419-437. position of naturalhill herbageundermanuring,control-
Belsky, A.J. 1992. Effects of grazing, competition, distur- led grazing and hay conditions. I. Yield and botanical
bance and fire on species composition and diversity in section.J. Ecol. 35: 65-95.
grasslandcommunities.J. Veg. Sci. 3: 187-200. Mountford,J.O., Lakhani,K.H. & Kirkham,F.W. 1993. Ex-
Bhadresa,R. 1977. Food preferenceof rabbits Oryctolagus perimentalassessment of the effects of nitrogenaddition
cuniculus L. at Holkham sand dunes, Norfolk. J. Appl. underhay-cuttingandaftermath grazingon the vegetation
Ecol. 14: 287-291. of meadows on a Somersetpeat moor. J. Appl. Ecol. 30:
Bobbink, R. 1991. Effects of nutrientenrichmentin Dutch 321-332.
chalk grassland.J. Appl.Ecol. 28: 28-41. Mueller-Dombois,D. & Ellenberg,H. 1974. Aims and meth-
Bobbink,R. & Willems, J.H. 1991. Impactof differentcutting ods of vegetation ecology. Wiley, New York, NY.
regimes on the performanceof Brachypodium pinnatum Olff, H., Huisman, J. & van Tooren, B.F. 1993. Species
in Dutch chalk grassland.Biol. Conserv.56: 1-21. dynamicsandnutrientaccumulation duringearlyprimary
Bobbink,R., Bik, L. & Willems,J.H. 1988. Effects of nitrogen succession in coastal sand dunes.J. Ecol. 81: 693-706.
fertilization on vegetation structureand dominance of Puerto, A., Rico, M., Matfas, M.D. & Garcia, J.A. 1990.
Brachypodium pinnatum (L.) Beauv. in chalk grassland. Variation structure diversityin Mediterranean
in and grass-
Acta Bot. Neerl. 37: 231-242. landsrelatedto trophicstatusandgrazingintensity.J. Veg.
Bobbink,R., Boxman,D., Fremstedt, Heil, G., Houdijk,A.
E., Sci. 1: 445-452.
& Roelofs, J. 1992. Critical loads for nitrogen eutro- Ranwell, D.S. 1960. Newborough Warren, Anglesey. III.
phication of terrestrialand wetland ecosystems based Changesin the vegetationon partsof the dunesystem after
upon changesin vegetationandfauna.In: Grennfelt,P. & the loss of rabbitsby myxomatosis.J. Ecol. 48: 385-397.
Th6rnelof,E. (eds.) Critical loadsfor nitrogen, pp. 111- Roelofs, J.G.M. 1986. The effect of airborne sulphur and
159. Nordic Council of Ministers,Copenhagen. nitrogen deposition on aquatic and terrestrialheathland
Bobbink,R., den Dubbelden,K. & Willems, J.H. 1989. Sea- vegetation.Experientia42: 372-377.
sonal dynamicsof phytomassandnutrientsin chalkgrass- Silvertown J., Watt, T.A., Smith, B. & Treweek, J.R. 1992.
land. Oikos 66: 216-224. Complex effects of grazing treatmenton an annual in a
Bobbink,R., Heil, G.W. & Raessen, M.B.A.G. 1990. Atmos- species-poorgrasslandcommunity.J. Veg. Sci. 3: 35-40.
pheric depositionand canopyexchangein heathlandeco- Stuyfzand,P.J. 1993. Hydrochemistryand hydrology of the
systems.DutchPriorityProgramme Acidification,Re-
on coastal dune area of the Western Netherlands. Ph.D.
port 119. Departmentof PlantEcology and Evolutionary Thesis, Free Universityof Amsterdam.
Biology, Universityof Utrecht. Thomas, A.S. 1963. Furtherchanges in vegetation since the
Boorman,L.A. & Fuller, R.M. 1982. Effects of addednutri- adventof myxomatosis.J. Ecol. 51: 151-183.
ents on dune swards grazed by rabbits.J. Ecol. 70: 345- ten Harkel,M.J. 1992. Abiotiek-biotiekinteractiesin de droge
355. duinenvan Meijendel.MeijendelMeded.(Den Haag) 24:
Davy, A.J. & Bishop, G.F. 1984. Response of Hieracium 65-89. (In Dutch.)
pilosella in Brecklandgrass heath to inorganicnutrients. ten Harkel, M.J., Heil, G.W., van der Hagen, H.G.J.M. &
J. Ecol. 72: 319-330. Witter,J.V. 1991. Abiotic-bioticinteractionsin dry dune
Dougherty, K.M., Mendelssohn, I.A. & Monteferrante, F.J. ecosystems. Mitt.Hydrograph.Dienst. Osterreich65/66:
1990. Effect of nitrogen,phosphorus potassiumaddi-
and 58-62.
tions on plantbiomass and soil nutrientcontentof a swale van der Meijden, R. 1990. Flora van Nederland. 21th ed.
barrier strandcommunityin Louisiana.Ann.Bot. 66: 265- Wolters-Noordhoff, Groningen.
271. van Hecke, P., Impens, I. & Behaeghe, T.J. 1981. Temporal
Farrow, E.P. 1917. On the ecology of the vegetation of variationof species composition and species diversity in
breckland.III. Generaleffects of rabbitson the vegeta- permanentgrasslandplots with different fertilizer treat-
tion. J. Ecol. 5: 1-18. ments. Vegetatio47: 221-232.
Gibson, C.W.D. & Brown, V.K. 1991. The effects of grazing Watt, A.S. 1962. The effect of excluding rabbitsfrom grass-
on local colonisationandextinctionduringearlysuccesion. land A (Xerobrometum) breckland,1936-60. J. Ecol.
in
452 ten Harkel, M.J. & van der Meulen, F.
50: 181-198. Williams, O.B., Wells, T.C.E. & Wells, D.A. 1974. Grazing
Westhoff, V 1985. Nature managementin coastal areas of managementof Woodwaltonfen: seasonal changesin the
WesternEurope.Vegetatio62: 523-532. diet of cattle and rabbits.J. Appl. Ecol. 11: 499-516.
Westhoff,V. & den Held, A.J. 1969. Plantengemeenschappen Willis, A.J. 1963. BrauntonBurrows:the effects on the veg-
in Nederland.Thieme, Zutphen. etation of the addition of mineral nutrientsto the dune
Willems, J.H. 1983. Species composition and above-ground soils. J. Ecol. 51: 353-374.
phytomassin chalk grasslandwith differentmanagement. Willis, A.J. & Yemm, E.W. 1961. Braunton Burrows;mineral
Vegetatio52: 171-180. nutrientstatusof the dune soils. J. Ecol. 49: 377-390.
Willems, J.H., Peet, R.K. & Bik, L. 1993. Changesin chalk- Zeevalking,H.J. & Fresco, L.F.M. 1977. Rabbitgrazingand
grassland structureand species richness resulting from species diversityin a dune area.Vegetatio 35: 193-196.
selective nutrientadditions.J. Veg. Sci. 4: 203-212.
Received 23 January1995;
Revision received 23 May 1995;
Accepted 5 September1995.