I THE BARNACLE GOOSE FEEDING TRIAL

THE BARNACLE GOOSE FEEDING TRIAL TOBCE,LIA (TOBSEDA)

Do BARNACLE GEESE BRANTA LEUCOPSIS PREFER HIGH QUALITY FORAGE TO HIGH BIOMASS FOOD PATCHES?

A MSc research

project carried out at the University of Groningen (RUG), department

^of

Animal Ecology, by

Reinout Havinga.

Supervision by Drs.

Graaf

I )

)

I

)

)

)

I

I

1

Sandra and Dr.

van der

Julia Stahl.

REINOIJI F:tVI\cj,\ _TIlEBARNACLE GOOSE FEEDING TRIAL - TOI3SEDA

I

Ii

SUMMARY

According to the green wave concept, barnaclegeese migrate north in an attempt to benefit from an extended spring. The quality and palatability of the food plants is highest in this season. To investigate this idea we have collected information on preference for forage of different quality and biomass,_{along the}

flyway. This report is based on the results of this research in Tobseda, a breeding site in the Russian north._Earlierstudies have shown that the geese react strongly on plant quality. Our data suggest that they

I

^L

respond strongest to the amount of plant nitrogen availability per surface area, which is the product of vegetation quality and biomass. In natural systems vegetation quality usually decreases when biomass accumulates, but in our experiment we successfully broke this link. Hence the choice between quality and biomass was no longer a trade-off, and the results show that the geese are opportunistic feeders. Their choice is primarily based on the amount of nitrogen (proteins) that they can ingest, which in natural systems might be limited by increasing standing crop.

PE3IOME

CornacHo KOHULIL1HH 3CReHOl1 BOJ1H), 6eJIolle!U4e icaapxMHFPHI.,ytOT Ha cesep pajw Toro, To6bI flOJIJll1Th BCCHHH ce3oH-nepHoa, xora xaecso KOMOBMX pacTeHHll, HCflOJIb3CMMX HMHB flIUll, oco6emio scoxo. JJDi I1OBCKH 3Toi FHflOT3M MM iflBMH BbIdHHm,

JIll xaapic B

aecme icopa paceii c BMcoKol mulueBoll ueHHocmio lull 6o.mbmofl 6noMaccol B nepllou_Mnrpawrn.

3ra pa6om OCHOBH Ha pe3y.nblaTax, nonyeHHiIx Juisi rrriw, rHe3IsIuwxcsi Ha npll6pe)KHMx apmax BapeHueBa MOS1 (JTatna KonoKoJIxoBo Fy6M, oKpecTHocm noce.rIKa To6ceua).

flpeJuuyujHe Rcc.rle.uoBaHHsl nOKa3aJIH, 'ITO Lfli lyceil nmiesa UCHHOCTh pace MMCT 6oilbuioe 3Ha'IeHHe. Cor.nacHo HLIIHM 1LaHHMM, HaH6o.r!ee KO sipascea peawui Ha 0TH0CHTC.mbHOe coaepIcaHHe ycosieoro a3oTa Ha JIHHHU nuloIuauH. B npi1poae nwesasi LeHHocm pac'reHHl, o6bIqHo, CHHACaeTC$I npll HKOIJIHHH 6lloMaccbl. B 3KCHCHMHTJIbHMX YCJIOBMX HM yua.nocb flpeo.uOneTb 31)' 3BHHMOTh.

TaKHM o6pa3oM, BM6op xopa c 6o.nbiuol 6noMaccotl H BbI6op KOM c BMCOKOtI flHIUeBOll UHHOCTMO nepecraum 6iITb o6paTHo npOflOpuHOHaJ!LHMMH BeJIwmHaMH.

KopMslcb, ryc HCflOJIb3lOT onnoplyHllcTlNecKylo TaKTHKy. Pe3y.nbTamI noxa3a.rrn, 'ITO BbI6op xopia 6e.noLueiunf H K3KMH HBHMHO OCHOBH Ha xonwiecme asoia (To ecm, 6eJIKa) KoTopMl OHH cnoco6mi ycsonm. B npHpoae 3TOT napaeip MO)I(CT 6bIm orpawieu HKOHJ1CHHCM 6lloMaccb,.

SAMENVATFING

Volgens de 'groene golf' theorie trekken brandganzen naar het noorden om te kunnen profiteren van een verlengde lente. De vegetatie is in dat jaargetijde namelijk het best eetbaar voor de dieren. Om dit idee nader te beschouwen hebben we langs de gehele trekroute informatie verzameld over de voedselvoorkeur van de ganzen. Die voedselvoorkeur hangt af van twee factoren: kwaliteit en kwantiteit van de vegetatie. In dit versiag worden de resultaten van dit onderzoek uit Tobseda, een broedgebied in het Russische noorden, behandeld.

erdere studies hebben aangetoond dat de ganzen vooral kiezen voor de kwaliteit van de vegetatie, maar in onze studie vonden we een duidelijke keuze voor de combinatie van kwaliteit en kwantiteit: de hoeveelheid stikstof per oppervlakte-eeitheid. Dit heeft waarschijnlijk te maken met het feit dat in deze proef kwaliteit en kwantiteit niet van elkaar afhingen, wat in natuurlijke ecosystemen vaak we! zo is. Dan neemt de kwaliteit afbij een toenemende biomassa van de plant. Deze resultaten laten zien dat ganzen opportunistische forageerders zijn: hun keuze wordt primair bepaald door de hoeveelheid eetbare stikstof (eiwitten), die in natuurlijke systemen desalniettemin begrensd kan worden door de biomassavan de vegetatie.

SAMMENFATNING

Ifølge teorien om den 'grønne bølge' er det bramgAsens mening at drage norpA, fordi den kan udnytte en slags udstrakt vAr. PA denne Arstid er grsset af højeste nringsvrdi forgssene. For at fmde ud af om der er nogen mening i denne teori, har vi, langs hele bramgAsens trkrute, samlet viden om gssenes fødevaner:

spiser de hellere nAr der er høj kvalitet eller høj plantemasse? I denne rapport er beskrevet resultateme fra Tobseda, en yngleplads pA Barentshavskysten.

Indtil nu troede man at gssene vlger isr gnes af høj kvalitet, men i denne undersøgelse fandt vi at de faktisk er meget glade for grs i en kombination af høj kvalitet og høj masse: kvlstofmengde (proteiner) pro kvadratmeter. I naturen aftager kvaliteten nAr plantemassen bygger op, men i vores experiment brød vi denne sammenheng. Derfor var valget mellem masse og kvalitet ikke lngere en afvejning. Det viser at bramgAsen er opportunistisk i

sit fødevalg, at den i første omgang vlger den bedste føde i de

_største mengder. I naturen sker det dog, at det kan ikke kan lade siggøre pA en gang.

Ii.',

JrSun,verg,te,t Groningen BIBLIOTHEEK RU GRONINGEN

Bibliotheek Biologjsch Centrum

Kerklaan 30

— Postbus 14

^{2228 9017}

Msc research project. april 2004

9750 AA HAREN

₂

I

REIN0UT HAvINI\ TiILI3ARNAC'I L GOOSE FEEDING TRIAL. - TOI3SEDA

Ccatents

Page

Summary

2

Contents

3

Introduction

4

Study Site and Methods

6

Results

8

Discussion

12

References

14

Appendices

16

\lsc research project. april 2004 3

RIINOuI H\vIN1,\ Tiit I3ARNACLLGOOSE FI;FDINGTRIAL —TOBSEDA

INTRODUCTION

ft

M

any goose species are migratory, rearing their offspring in the arctic and spending

the winter in temperate regions. This

research deals with the continental barnacle goose (Branta leucopsis) population. Its flyway connects the wintering area in the Wadden Sea via the Baltic and Barents Sea to the breeding sites on Novaya Zemlya and Vaygach.

Until 1971 this population has bred exclusively on Novaya Zemlya and Vaygach. In the following years, the species founded breeding colonies in the Baltic, an area which before was used as a spring and autumn staging site only. From 1981 onwards the barnacle goose began to breed on the lowland shores and islets of the southern Barents Sea, too.

Almost the entire population winters around the Wadden Sea (Ganter et a!. 1999).

In the 1950's the population was at the lowest level with only 10.000 birds. Since then the number

of geese increased to 267.000 birds

in 1997, simultaneously with the extension of the breeding range (Ganter et al. 1999; Loonen et al. 1999).

However, ninety percent of the population still breeds in the core breeding grounds of Novaya Zemlya and Vaygach.

The barnacle goose is a herbivore that traditionally feeds on short swards of saltmarsh grasses. Although the world is full of green plants,

potential food sources for herbivores as it seems, only little of this apparent plenty is of interest to plant eaters. Herbivores are confronted with

constant change of their plant food that have

developed herbivore defence mechanisms like chemicals and fibrous tissues, which make the plant too sturdy to digest (White 1993). In this 'green desert' a herbivore has to look for opportunities to forage successfully.

The barnacle goose has a

relatively small digestive tract, which results in a relatively short food retention time (Demment & van Soest 1985).

For this reason the goose has to forage on easily digestible

plants. Vegetation with a low fibre

(cellulose) content is preferred (Prop & Vulink 1992). l'his has already been shown in the related brent goose Branta bernicla where experimental trials resulted in a choice for quality (Bos et al.

2002; Riddington et al. 1997). This is in correspondence with the general rule, visualised in appendix 1, that biomass and quality are negatively correlated as quality generally decreases when biomass is accumulated with the phenological age of a plant (Crawley 1983).

A theoretical graph with optimal food conditions on three staging sites is shown in Figure 1. In natural temperate and arctic ecosystems, the highest relative plant protein content is reached in Other locations Other locations

Figure 1: Theoretical relation between biomass and quality over time and along the migration route (Klimkowska 2003).

Nlsc research project, april 2004 4

I

I

I

Schiermonnikoog

I

4

Biomass

Gotland

A

Co

U) U) Co

E0

z

C,,

Tobseda

Quality

/

I

I

I

I

I

I

I

Is

I

March April

lllluI

May

Biomass and quality sufficient for geese Quality not sufficient for geese

June July

^Time

1

RIINoIJT HAVINGA THE BARNACLE GOOSE FEEDING TRIAL -TOBSEDA

early spring. In the course of the season, the food grass species will increase in biomass, but there will be a decrease in nitrogen content (Klimkowska 2003). This results in a crop that gets more difficult to digest as the season proceeds.

Food requirements are at its highest during

spring migration, when the breeding season is

approaching. It has been suggested that during this migration the barnacle geese take advantage of the

Figure 2: Grazing barnacle geese on a saltmarsh during spring staging.

spring flush of fresh grass at each staging site

(Figure 2). The migrating flocks of barnacle geese

are 'surfing' north upon a green wave of easily

digestible grass (Owen 1980)

To fmd

out whether this 'green wave' hypothesis holds ground, there is need to collect data on forage quality (Klimkowska 2003) together with information on goose forage preferences along the flyway (Veen 2004; Lubbe 2003)

In this experiment we will try to reveal forage preference, comparing vegetation enhanced in quality and biomass with unmanipulated swards, in a full-factorial set-up. These manipulated vegetation plots are presented to wild geese, as their grazing pressure on the different treatments provides us with insights in forage preferences.

HYPOTHESIS

We hypothesise that barnacle geese prefer high quality to high biomass food sources.

We will test the reaction of geese to vegetation with different levels of biomass and quality, where the negative correlation between these two factors is disrupted experimentally. This will tell us more about the factors on which the geese base their choice. We will repeat the experiment on a second moment in the season. This way we can find out whether choice changes in time, perhaps urged by different biomass levels.

I7LI Mscresearch project, april 2004 5

RilNol rr HWINGA THEBARNACLE (IOOSE FEEDING TRIAL - TOBSEDA

STUDY SITE AND METhODS

T he experiment was carried out on the

subarctic saltmarsh (approximately 6 km2)

near Tobseda (68°35'N, 52°20'E) on the

Kolokolkova Bay, Timancoast, Russia (Appendix 3). The experiment was carried out in spring and early summer, between the 1 0th of June and the 3l of July 2003. A calendar of the experiment together with temperature data can be found in Appendix 4.

The first Tobseda breeding colony was established probably around 1990 (vd Jeugd et al. 2003). In 2003 the area was visited by non-breeding barnacle geese all summer. The experimental plots were close to a barnacle goose breeding colony. In th area, about 1650 birds were present on the ^12th of June. On the ^28thofJuly 5796 moulting adults and 4209 juveniles were present. Mean nest initiation was in June 12 and peakhatch on July 12. A total number of 2110 nest were found on the peninsula (Litvin et a!. 2004).

The saltmarsh was completely i'e

from June 23 onwards. Mean daily temperat.rc " June

ranged between -2 and +14 °C. In July temperatures ranged between +4 and + 17°C (Appendix 4).

Two dominant saltmarsh plant species (Figure 3) form the major part of the goose diet and were studied in his experiment. These are Carex subspathacea and Puccinellia phiyganodes. Both species grow in a closely mixed sward throughout the marsh, and form the staple diet of local barnacle geese.

Figure 3: Vegetation composition of the saltmarsh (point-quadrate data).

A

full factorial design was used, combining exclosed and fertilised vegetation in order to generate vegetation patches with higher

biomass and/or higher quality (Figure 4). This approach includes a control treatment for both

factors (Bos et al. 2002; Riddington et al. 1997):

:

: :: :

— . I—

: : : ₁

Figure 4: Schematic illustration four plots under manipulation.

Not exciosed + not fertilised (C);

Exclosed + not fertilised (B);

Not exciosed + fertilised (Q);

Exciosed+ fertilised (BQ).

This set-up was replicated six times with 20 to 150 mees between blocks, and approximately 1 metre between square plots of 1 6m2 within a block.

The outermost blocks are about 600m away from each other. The whole experiment was carried out twice in the season, producing an overall of 2x6 replicates. The plots were selected for homogeneity of the swards.

MANIPuLATIONS

Biomass was enhanced by excluding grazers from the vegetation over a three-week period, to allow biomass accumulation. Exclosures were made of strawberrynetting, bamboosticks and string. A string-cross was stretched out over the exclosure to prevent airborne geese from entering.

Quality of the vegetation was enhanced by

adding dissolved artificial fertiliser to the plot with

a watering can. Brackish water from adjacent

gullies was used as a solvent. For each plot, 1 kg fertiliser was used containing an N:P:K ratio of 16:16:16 (Appendix 6). This amounts to 10 g of nitrogen, lOg of phosphorus and lOg of potassium per square meter. This manipulation was carried out

\isc researchproject, april 2004 ₆

I

4m

C Q

of a block of

Carex .5

E

V

I

'I

t

I I

3 July 22 July

I.

'I

I

on the same day as the exclosures for the biomass- treatment were erected.

Three weeks after the manipulations were initiated, the biomass-exclosures were removed and wild geese could enter the experiment. The date of opening of the first set was July 3, of the second set July 22.

GOOSE VISIT

To assess goose visitation (jreference) we used dropping counts. Counts were done every 2-3 days.

For the analysis we used the cumulative dropping number from opening until all plots were visited at least once.

VEGETATION MEASUREMENTS

Prior to the start of the choice experiment, the following vegetation measures were taken:

Carex and Puccinellia tillers were counted in 5.5x5.5cm frames (20 repeats per plot). On each plot, biomass was sampled by collecting 50 tillers of each species. These tillers were dried for 48h at 60°C and then weighed. Final standing crop dry

bioniass values [gm2] are calculated using the

following formula: [#tillersinean tiller weightm2].

For data on plant quality, we collected leaf tips of both plant species (0.5g dry weight per sample),

of which the C:N ratio was assessed with the

element analyser (Buurman et al. 1996).

Msc research project. april 2004

For each plot vegetation cover was recorded using the point-quadrate method (Grant 1981).

STATISTICS (Zar 1999)

A Multivariate Analysis of Variance (GLM) is used to defme the effect that each treatment had on the dependant variables (j)reference by geese, biomass and quality of the vegetation as well as vegetation cover). Fixed factors are treatment, replicate (nested within set) and set. A Tukey post hoc test is

performed to defme subsets. The

manipulation effect is tested defining fertilisation and exclusion plus the interaction as fixed factors.

To assess the effect of biomass and nitrogen content on goose preference, a Univariate Analysis of Variance (GLM) is used. Dependent variable is preference; covariates are biomass and percentage

of nitrogen of the sward (Pucc and

^Carex

combined). In addition, the interaction between the two covariates is taken into account. Fixed factors are set, replicate (nested within set) and treatment.

Results are significant at or below the level of p=0.05

When preference is tested, the biomass and quality characteristics of the two plant species are combined to values of the total sward.

RLIN()U1 1IAVIN(iA Ti IL BARNACLE GOOSE FELDING TRIAL - TOBSEDA

I

I

1

1

.1

I

I

I

I I

1

I I I

I

I

I I

7

RIJNO(JIHAVINGA

RESULTS

TIlE BARNACLE(( )OSLIN[L)ING TRIAL — TOBSEDA

WHAT WAS ThE EFFECT OF THE TREATMENTS ON THE VEGETATION?

Biomass of Puccinellia phryganodes was significantly higher on BQ plots in comparison to the natural vegetation (C). B and Q

plots

showed no significant increase in biomass of P.

phryganodes. For Carex subspathacea biomass was equal on all plots (Figure 5). The fertiliser treatment however did split the four treatments up into two subsets of low and high quality vegetation plots of both species (Figure 6). Table 1 gives a schematic overview of the hioms and quality differences between r'nents.

I

0

I

40

30

20

10

0

100

80

60

40

20

0

(ii) Puccineilia phryganodes

b

I

a a

a

C B 0 B0

treatment

0

C

•0

43

0

7

0

6

=

C

VC

.003

C Ca

Cl

e

4

(ii) Puccinellia phiyganodes T

sl

_s2

I

b

a a

I

¹

B 0 BO

treatment

I I

it

II

Figure 5. Biomass of Carex(I) and Puccinellia

(ii)per trstment.

Biomass is expressed as grams dryweight per square metre. Letters above the bars

indicate subsets.

Note the different scales for (i) and (ii).

Table 1: Overview of effect of treatment on Treatment not-fertilised+

not-exclosed(C) Quality Pucc. & Carex low

Biomass Puccinel!ia low

Biomass Carex normal

Figure 6. Quality of Carex (i) and Puccine!lla (ii) per treatment.

Quality is expressed as percentage of nitrogen of the plant tissue. Letters above the bars indicate subsets.

vegetation parameters biomass and quality.

not-fertilised+ fertilised+ fertilised+

exclosed(B) not-exciosed (Q) exciosed (BQ)

low high high

low low high

normal normal normal

I I

I

C B 0 80

C B Q BO

I

I

17D1 Msc research project, april 2004 ₈

.1

R1iNot.JT HAVINGA TilE BARNACLE GOOSE FEEDING TRIAL - TOBSEDA

WjiiT WAS ThE EFFECT OF ThE TREATMENTS ON PREFERENCE?

The geese showed no preference to B plots over the control (a=O.87). Q

^plots

were preferred above C & B plots (p<O.Ol) and BQ plots were preferred over C, B & Q

^plots

(p<O.Ol) (Figure 7). Three different preference levels can thus be detected:

Level a: B & C Levelb: Q

Level

c: BQ

The preference

to

fertilised

plots matches with the higher nitrogen content (Table

^{1 ' The}

high preference to the BQ plot within the fertilised plots matches

with the high biomass of P.

phryganodes in this plot.

Apparently the combination of a quality and biomass enhanced

sward has an amplifying effect on

^treatment

'It Mscresearchproject, april 2004 9

Figure 7:

Preference to the different

treatments. Black bars represent the first set and grey the second.

Letters above the bars indicate subsets.

C 8 0 80

THE EFFECTS OF ThE MANIPULATIONS ON BOTh VEGETATION AND PREFERENCE

The fertilisation did significantly alter all variables, except biomass. Exclusion of geese had a significant effect on all variables, except plant tissue quality. The combination of both manipulations did also make a difference in all variables except quality (Table 3).

Table3: Manova. Effect of vegetation manipulations on vegetation and 6ose response. Carex and Puccinellia data are combined (sward). Non-significar.-results are highlighted. Set and replicate

(nested within set) are corrected for.

Source Dependent Variable Type Ill Sum of Squares df Mean Square ^F

exclosed

fertilised

excifert

Sig.

Rinm

^A')7)

'

4?7.5O9 37.665 .000

Quality .L(.. .272 1.008 .323

BxQ ^{7.846 1 7.846 15.992 .000}

Dropping 22.175 ¹ 22.175 18.227 .000

Biomass 73.375 ¹ 73.375 .647 .427

Quality 65.813 ¹ 65.813 244.064 .000

BxQ 15.425 1 15.425 31.442 .000

Dropping 77.998 ¹ 77.998 64.113 .000

Biomass 908.234 1 908.234 8.005 .008

Quality .116 1 .116 .430 .516

BxQ 3.201 ¹ 3.201 6.525 .015

Dropping 17.521 1 17.521 14.402 .001

RLINOUT H\vINGA TilEBARNACI.I GOOSE FEH)ING TRIAL -TOBSEDA

WHAT EFFECT DO THE VEGETATION PARAMETERS BIOMASS AND QUALITY HAVE ON PREFERENCE?

Neither biomass per square metre nor the percentage of nitrogen separately as a covariate parameter has a significant effect on preference (respectively p=O.77; p=O.6O), but the

interaction of both characteristics does help to explain preference (p<0.01)(Table 2). From the fixed factors that were incorporated into this test, only treatment has an additional effect on

preference (p<O.OS).

Table2: Unianova. The effects of different factors and vegetation (sward) variables on preference.

FFixed factor, CCovariate Significant results are highlighted

Source Type Ill Sum of Squares df Mean Square F Sig.

FSet .215 1 .215 .293 .592

FReplicate(Set) 14.310 10 1.431 1.952 .077

FTreatment _{9.761 3 3.254 4.439 .011}

CBiomass _{.063 1 .063 .086} .771

CQuality .203 ¹ .203 .278 .602

CBiomass. Quality _7.656 1 7.656 10.444 .003

STANDING NITROGEN

The interaction of biomass and quality has a significant effect on preference as is shown in Table 2. Since this interaction term is defined as the product of the interacting factors, this

interaction can be interpreted as the supply of nitrogenm2 or 'standing nitrogen' (figure 8).

This 'standing nitrogen' can, as a single factor, be plotted against preference in a regression diagram. It is found that this factor explains 39% of the variation in preference in the first set

and even 65%

in

the second.

3.5

3.0

.9

2.5

C 2.0

0.0.

2 ^1.5 V

a)C)

Ca) 0.5 a)

0.0 1.0

nitrogen supply (gNm')

Figure 8: Explaining preference by nitrogen supply. The different shadings of the dots represent treatment origin. Note the different scales of the two graphs.

!II'\lsc research project, april 2004 10

I

.1

'I

I

11

)

I

I

I

0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 1 2 3 4 5

I

RFINo iTHAvINGA TIE E3ARNACI.F GoOSE FEEDINGTRIAL - TOE3SE[)A

1

I

No

EFFECT OF COVER

Vegetation cover differences did not have a significant effect on preference in this experiment, when tested together with nitrogen supply and

treatment

(Table 4).

I

Table 4 Anova Theeffectofvegetationcoveron preference FFIxed factor, CCovarlate Significant results are highlighted

I

Source Type Ill Sum of Squares df Mean Square ^{F Sig.}

FSet 7.425E-04 I 7.425E-04 .001 .970

FReplicate(Set) 10.669 10 1.067 2.024 .068

FTreatment _{8.462 3 2.821 5.350 .005}

I

CNitrogen supply 8.851 1 8.851 16.788 .000

CFoodvegetationcover .880 ¹ .880 1.670 .207

CCarexcover 2.173 ¹ 2.173 4.122 .052

Cpucc cover .871 1 .871 1.653 .209

1

I

I I

I

I I

I

I I I

I

7I ^\lsc

^researchproject, april 2004 ¹¹

I

RIouT HAVINGA TuE BARNACLE GOOSE FEEDING TRIAL -TOBSEDA

DISCUSSION

T

^he results of this experiment clearly show that the interaction of biomass and quality,

which could be expressed

as 'standing nitrogen' (grams of nitrogen per square metre), explains patch choice by foraging barnacle geese (Table 2, Figure 8). This seems to contrast with the hypothesis that quality is preferred over biomass.

CORRELATION BIOMASS/QUALITY

The hypothesised preference of quality over biomass is based on the general rule that these two parameters are correlated negatively (Appen!ix 1), thus producing ^a fairly straightforward choice mechanism: it's one or the other. A higher quality is usually preferred under these conditions (Lubbe 2003). In this experiment however we managed to annihilate this correlation (Appendix 2). This resulted in the fact that choice was not a trade-off between biomass and quality anymore:

if the

amount of food does not affect quality, which might still be the first selection criterion, why take little when there is a lot? Comparable results have been found in staging barnacle geese on Gotland (Veen

2004), and, with different methods, in staging

white-fronted geese on Iceland (Kristiansen et al.

2000).

WHAT DETERMINES CHOICE?

The main parameters in this experiment were the two vegetation characteristics that constitute to Optimal Foraging Theory (Belovsky 1997):

biomass and quality. These are the two main factors that determine the nutritional value of a food source to a herbivore. With help of a chemical analysis we are able to determine (small) variation in biomass and quality values of plants, and correlate this to food choice of the herbivore.

Barnacle geese show a very distinct reaction to protein availability, resulting in seasonal habitat shifts (Prins & Ydenberg 1985). It is interesting to see how these food characteristics cohere with vegetation characteristics that are more likely to be perceptible to the geese. Perhaps these characteristics, like grass colour, taste, texture, smell (Wallraff 2003), or a combination of these show to be highly correlated to an optimal forage choice.

Standing nitrogen is the main explanation for the variation in preference, but not the only one. An additional effect of treatment was found (Table 2).

There must thus have been another factor of

influence, strongly correlated

to treatment and

distributed unevenly with the nitrogen supply.

Perhaps vegetation colour (Appendix 5) was this factor. Differences in vegetation cover are not of great importance here (Table 4).

TWO FOOD SPECIES

In this report the proportion at which the two food species are ingested is not incorporated.

Instead, the sward is tested as a homogeneous food

source. Considering the diet in future research

however might increase our understanding of the feeding ecology of the goose.

During the writing of this report this ratio was assessed. It was found that the proportion of Carex subspathacea and Puccinellia phryganodes in the droppings is 1:1 (vd Graaf & Stahl 2004). These data suggest that the geese select for C.

subspathacea, since this species is less dominant than P. phi'yganodes in the saltmarsh vegetation cover and biomass.

TROUBLESHOOTING

Preference was measured by counting droppings after the geese had been present on all four different treatments. A disadvantage of this assessment could be that the geese might keep frequenting plots with a higher carrying capacity. This would mean that there is a bias, rather than a preference, towards high biomass. To avoid this problem it could be helpful to have observation data from the foraging geese (Bos et al. 2002). The behaviour of the geese on the moment of arrival to a block of treatments could be interesting.

The most promising aproach to the research question is a truly full-factorial set-up.

Unfortunately this was not achieved here (Table 1;

Figure 5 and 6). An improvement could be to

situate the experiment in a place with a somewhat higher natural grazing pressure, where differences in biomass are more likely to develop. Mowing the low-biomass plots is more labour intensive, but more accurate, too. It can also be problematic that in a nutrient limited system, artificial fertiliser not only enhances plant quality, but also biomass.

PASTURE SELECTION

The enhancement of vegetation quality by means of adding commercial fertiliser might seem artificial when trying to understand natural feeding ecology mechanisms. However in unnianipulated ecosystems on Schiermonnikoog and Spitsbergen

barnacle geese have been observed to actively move to sites that have been subject to faecal

discards of gulls (Bazely et al. 1991; Prop et al.

1984). These sites showed raised nitrogen levels in the vegetation. Also local terrain shifts from natural foraging habitats to agricultural grasslands (Prop et al. 1998; Black et al. 1991; Prins & Ydenberg

1985) show that the barnacle goose is able to

actively select a feeding ground. On a larger scale

Iei \lsc research project, april 2004 12

I' I'

If

It

I

II

II

this supports the theory that the geese follow the 'green

wave' of spring vegetation growth.

TIMINGOFRAISING YOUNG

In natural temperate and arctic ecosystems, the largest proportion of proteins per plant biomass is reached in early spring. Herbivores and other animals can seize the opportunity of this bounty by adjusting the reproductive season to peak forage quality. This is shown in e.g. the trophic life-cycle relationship between oak, caterpillars and fits. Here

fit peak hatch coincides with the caterpillar peak hatch on oak, which in its turn coincides with the budding of oakleaves in spring (Perrins 1991).

With reference to this example it can be argued that the strategy of the migrating geese might be to

have their peak hatch coincide with the most optimal foraging conditions (Prop & de Vries

1993), which might be found in the arctic summer.

This mechanism is different from the 'green wave theory' as mentioned earlier, but these two mechanisms could coexist.

I

I

\lsc research project, april 2004 13

I I

Rr1NouT HAVINGA TuE BARNACLEGOOSE FEEDING TRIAL. - TOBSEDA

I

I

I I

I I I I

I I

I I I I I

I

REI\OI ir HAVINGA

Tiii

I3ARNACLE GOOSE f1EDtN( TRIAL - TOBSEDA

REFERENCES

I

Bazely DR, Ewins PJ, McCleery RH (1991).

Possible effects of local enrichment by gulls on feeding-site selection by wintering Barnacle Geese Branta leucopsis. This 133, 111-114.

Belovsky GE (1997). Optimal foraging and community structure: the allometry of herbivore food selection and competition. Evolutionary Ecology 11,641-672.

Black JM, Deerenberg C, Owen M (1991).

Foraging behaviour and site selection of barnacle geese Brantaleucopsisin a traditional and newly colonised spring staging habitat. Ardea 79, 349- 358.

Bos D, Drent RH, Rubinigg M, Stahl J (2002).

The relative importance of food biomass and quality for patch choice in brent geese. In: Bos D (ed) Grazing in coastal grasslands; brent geese and facilitation by herbivory RUG, Gromngen, pp 23- 41.

Buurman P, Lagen BV, Velthorst EJ (1996).

Manual for soil and water analysis. Backhuys Publishers, Leiden.

Crawley MJ (1983). Herbivory. Blackwell Scientific Publications, Oxford.

Demment MW, van Soest PJ (1985). A nutritional explanation for body-size patterns of ruminant and nonruminant herbivores. American Naturalist 125, 641-672.

Ganter B, Larsson K, Syroecbkovsky EV, Litvin KE, Leito A, Madsen J (1999). Barnacle goose Branta leucopsis: Russia/Baltic. In: Madsen J, Cracknell G, Fox T (eds) Goose populations of the western palearctic; a review of status and

distribution pp 270-2 83.

Grant SA (1981). Sward components. In:

Hodgson J, Baker RD, Davies A, Laidlaw AS, Leaver JD (eds) Sward Measurement Handbook British Grassland Society, Hurley, pp 7 1-92.

Klimkowska A (2003). Geese on a green wave:

weather parameters and seasonal changes in food availability and food quality in relation to the barnacle goose (Branta leucopsis) migration along

the east Atlantic flyway. Lelystad, RIZA.

Kristiansen JN, Fox T, Nachman G (2000).

Does size matter? Maximising nutrient and biomass

intake by shoot size selection amongst herbivorous geese. Ardea 88[2], 119-125.

Litvin KE, Eichhorn G, Drent RH, Gurtovaya E, Stahl J, vd GraafAJ, vd Jeugd HP, Gregersen J, Karagicheva J, Havinga RM (2004). Results Kolokolkova Bay. In: Noordhuis R, van Eerden MR (eds) PRISM / NWO Progress Report 2003, 1

-PechoraDelta Expedition Lelystad, pp 95-129.

Loonen MJJJ, Bruinzeel LW, Black JM, Drent RH (1999). The benefits of large broods in barnacle geese: a study using natural and experimental manipulations. Journal of Animal Ecology 68, 753-768.

Lubbe G (2003). Competitie voor Festuca tussen brandganzen en hazen; de rol van biomassa en kwaliteit. Van Hall Instituut.

Owen M (1980). Movements and migration. In:

I

Wildgeese of the world Batsford, London, pp 97- 99.

Perrins CM (1991). Tits and their caterpillar food supply. This 133[supplement 1], 49-54.

Prins HHTh, Ydenberg RC (1985). Vegetation growth and seasonal habitat shift of the barnacle goose (Branta leucopsis). Oecologica 66, 122-125.

Prop J, Black JM, Shimmings P, Owen M (1998). The spring range of barnacle geese Branta leucopsis in relation to changes in land

management and climate. Biological Conservation 86, 339-346.

Prop J, de Vries J (1993). Impact of snow and food conditions on the reproductive performance of Barnacle Geese Branta leucopsis. Ornis

Scandinavica 24, 110-12 1.

Prop J, van Eerden MR, Drent RH (1984).

Reproductive success of the Barnacle Goose Branta leucopsis in relation to food exploitation on the breeding grounds, western Spitsbergen. Norsk Polarinstitutets Skrifter 181, 87-117.

Prop J, Vulink T (1992). Digestion by barnacle geese in the annual cycle: the interplay between retention time and food quality. Functional Ecology 6, 180-189.

I

Riddington R, Hassal M, Lane SJ (1997). The selection of grass swards by brent geese Branta b,

bernicla: interactions between food quality and quantity. Biological Conservation 81, 153-160.

'!II Msc research ploject, april 2004

14

I

1

I

I

1

I

1

I

I

REINOUTHAvINGA THE BARNACLE GOOSE FEEDING TRIAL - TOBSEDA

vd GraafAJ, Stahl J (2004). Ratio of Carex experiment on Gotland. Rijksuniversiteit ,I

subspathacea: Puccinellia phryganodes in barnacle Groningen.

goose droppings from Kolokolkova Bay is 1:1.

Waliraff HG (2003). Zur olfaktorischen

,I

vdJeugd HP, Gurtovaya E, Eichhorn G, Litvin Navigation der Vogel [Olfactory navigation by KE, Mineev OY, van Eerden MR (2003). Breeding birds]. Journal fur Ornithologie 144, 1-32.

barnacle geese in Kolokolkova Bay, Russia:

number of breeding pairs, reproductive success and White TCR (1993). The Inadequate

'

morphology,Polar Biology 26, 700-706. Environment. Springer Verlag, Berlin.

Veen GF (2004). Effects of biomass and quality Zar JH (1999). Biostatistical Analysis. Prentice of forage by breeding and spring staging barnacle Hall, Upper Saddle River.

1 ^geese (Branta leucopsis). Field observations and

I

1

I I I

I

I I

I

1

I I

I I

_

7I'\lsc research project, april 2004 ¹

I

________________ _____________________________

0) U)U)

z

Co

a

0'

Appendix 2: Correlation between vegetation quality and biomass of the manipulated vegetation plots. Note the different scales of the x-axes.

two species in the

7[ \lsc research project, april 2004

¹⁶

1

V

It

I I

RIJ\ouI HAVINGA Tiic BARNACI1GOOSE ILIDIN(; TRIAL-TOBSEDA

APPENDIXES

Biomass and quality

6

Loliumperenne

/ ^Festuca

^rubra

0,

A ^{50 100 150 200 250 300 350 400 450 500 550 600}

/

biomass gIm2

Puccinelliamaritima

I

I

I

I

I I I

I I I

I

Carex subspathacea Puccinelila phiyganodes

Biomass Appendix 1: Schematical relationship between quality and biomass I season (right); to the left an example of this relation in three species from the temperate saltmarsh (Klimkowska 2003).

6

10 20 30 40 50 60 0 20 40 60 80 100

Biomass (gUm')

RE1'OIJT HAVINGA THE BARNACLEGOOSE FEEDING TRIAL - TOBSEDA

Appendix 3a: Geographical position of study site within the migration pattern of the continental barnacle goose (other subpopulations migrate between the British isles and Greenland or Svalbard).

Sand/ Dunes

Appendix 3b: Study site in a close-up.

1!I Msc research

proJect, april 2004 17

Barents Sea

— . ,

. •

Saltmarsh Bog/ Tundra

Kolokolkova Bay

S 1km

RIl\o n' HAVINGA TuE BARNACLE GoOSE FH;DING TRIAL -TOBSEDA

C)0 U, CoI-

CoI..

Co

EC,

I-

Co

-10

I II

Date

• max mm - - - - 'Mean Appnthx 4a: Weatherdata from ToDsed0 rween 25 may and 17 august 2003.

30

25

20

15 10 5 0 -5

I I

I I

1.2 -

I

1.0-

0.6

caO.4•

a2

0.0•

Peakhatch l2JuIy

I

16/6 Set I

30/6

I I

14/7 28/7

Treated

Set 2

._..________

^Ocened

Appendix 4b: Timing of the experiment and natural grazing pressure. Day/month are expressed on the x-axis.

11/8

I I

Msc research project, april 2004 18

I

REINOIJ1' HAVINGA TilE BARNACLE GOOSE FEE[)ING TRIAL -TOBSEDA

Hp.

Appendix 5: An exciosed as well as fertilised plot. Note the different colour of the,1rass inside the exciosure. This was due to the fertilisation treatment, rather than to the exclusion of barnacle geese.

Appendix 6: Fertiliser bag.

MI$CTPY1IUMS no tiiM4jj

HLlTPOAMMOoc

(uov. 18%. 4ooøop•_{W4.e,ue ;}

oe—

1e%,,.,2.

_C.

$1IWCIInKvaTeIlbHwA $11eMSHTOS

aoT, OQoçI'SlHS) npeje-

HSleHo WIS io6ps.* s pe

u1sae.ii KYfl&rYP KS

S

no. -

- fl*I

lSfle

onoo H SOCUSSflQ

$L4KS'U 1) O90r#

I1 cI 1b1W$4I fllV5X H 100123

.'

H5 serb1yIs4a.

fl

^11UlR4IlN

PANIIIIHY .acsoa 1110051 30.40!M.

flps fl1IISnrus J1UPOSbSSN KYCISPHUNOS NIIOCRT 200-300 ia n1ZHyIo Kily, noonseI*owq

n

faa 'uaoeo-IaaNw. oyw,ryo

^{rpnoai ne.}

1001011 so aOe,ss cp 6ora.npn- CTIOflb$W NP?O N

6o- It"

rk.,p,.a I ic 1.wquv'.

2-3 .II 055011ru f r.w-. c '1)- W'C'.WiNtk!11WIM.('OIl 9 K05 y$SHUS I55'VIK pearrnn 110r10eUSluicCn* Ikl-C1W CIacz '10U$4 350s11c1 (n,qiay'

UflIl pbIxJ*i4Ne).0 sns*iiwl cooN

flO.964Wi rM I'5065t501111001$ no-

co-a 9551110 110510 lulae1 lOç

QKHH COU4SMIILI11 r,po6oo- 28r.

HIIOTO.n,No:

R ^OGOW.

141400. r. XMM10I l]t1M c6c'..

L

fleacpojcoa, 20

V'I Pqnisa. 8 ,nn.• (006) 67-14-72.

mn4c; p95) 572e6-68 flposloo.y,'p65ISS3y5Ki 010C 111005yPOl 9NMN. (eSCt9 4W

)KEJIAEMDAM

&CIATWX YPOKAEBt

Mepw6s300.alocnl flpi. ps8oieco8iscie l,501a rai.aoN nirIs.a, noayS1ec n19

'4aTlIpPmN.flocJI. pi6omi cr'aayer aialTb pyus s sma C swnow

flsps10 poM no

flpii nosa, S nlais - lcMKTb

0coujcs IEKS11SC1SOM11M No6pS 11111455 SNy W1150wcy1Y550u.

r11 0565 TS*4$0 aTKKSTIy) O'OMCJ1#S5H0W r10rnarN.aU,.H - ISTb •51111Th HSflbdU CIsIlINOl 100b1 P 155101N S0T'. fls1190650.

'MWOC?11 91115*11. uçtsia KIM 0C15- 1151nocrpsq..ujero S ae11suMI,- 'CS ylpSAnsHHe (nps0s6o wuov 4Kyb0 )fluuSoTvy Mfl$ N$CTpykI,NtO s'

AI4.1T6 B CYxO6I 3AXPbITOM

IuMt:uHsIi, OTZIEJ1I40 or

I1EKAPCTB V IECAX. HE$OCfl6ThbX IETRIA

1 )KCTtW

05L5l^WR1

JTL'410071 0WK5K1S 10L*K

as

COI* I('l3100'Os)'005 H

•P11SP5*t

0IIOCflIP 45 IV' WSpie,i-j,±15,

19o91-e

Hasp Iqa1pC.Ms455P*1kl..a9 4*ZS.0512-l

•lcIINw0.5*I5 'asp as55 21C45S.D40I$4,js

•1JI1NJ

IS iP.

.

⁰⁴

I7L Msc research project, april 2004 ¹⁹

1

REEN0UTHAVINGA

Tn

BARNACLE GOOSE FEEDING TRIAL - TOBSEDA

THANKS CPACHBO BEDANKT DANKE TAK

I

Sandra and Julia for their support in the field and during the writing of the report.

I

Allthe fine company during the expedition.

Sandra, Julia, Kostja, Lena, Rudi, Ashley, Götz, Yura, Jens, Yulia, Henk.

Also Vasja and Alexander Moskvien from Tobseda.

Yulia Karagicheva for producing the Russian summary and Jens for correcting the one in Danish.

All the funding and coordinating support from Gromngen university and RIZA Lelystad (PRISM project).

The Syktyvkar biology institute and Moscow bird ringing centre for their cooperation in getting the permission to travel to the study site.

1

I I I

I I I

I I I

I I

L[3

\lsc research project, april 2004 20