MARIE-LOU COULOMBE
EFFETS DE LA DENSITÉ DE POPULATION SUR LE COMPORTEMENT
D’APPROVISIONNEMENT ET LE BUDGET D’ACTIVITÉ DU CERF DE VIRGINIE
(ODOCOILEUS VIRGINIANUS) À L’ÎLE D’ANTICOSTI
Mémoire présenté
à la Faculté des études supérieures de l’Université Laval
dans le cadre du programme de maîtrise en biologie
pour l’obtention du grade de maître ès sciences (M.Sc.)
Département de biologie
FACULTÉ DES SCIENCES ET GÉNIE
UNIVERSITÉ LAVAL
QUÉBEC
2006
© Marie-Lou Coulombe, 2006
ii
Résumé
Nous avons étudié l’influence de la densité de population sur les déplacements, le budget
d’activité et l’utilisation de l’espace chez le cerf de Virginie en densités contrôlées
expérimentalement. Les déplacements et le budget d’activité étaient peu influencés par la
densité. Dans les densités contrôlées, les cerfs réduisaient leur activité avec l’augmentation
de biomasse de la végétation durant la saison ou selon le nombre d’années après coupe. En
densité naturelle, les cerfs passaient moins de temps en activité au début de l’été lorsque la
végétation était moins abondante. À haute densité, les cerfs ne recherchaient pas les zones
de couvert plus dense contrairement à ce qui se passait pour les cerfs à faible densité. Si la
quantité de végétation diminue avec l’augmentation de la densité de cerfs, nous prédisons
que les cerfs s’adapteront en augmentant leur temps d’alimentation ou, lorsque la
végétation sera fortement réduite, ils augmenteront leur temps de rumination et délaisseront
les milieux sous couvert.
iii
Abstract
We investigated the influence of population density on movements, activity budgets and
space utilization of white-tailed deer in a controlled-density experiment. Movements and
activity budgets were generally not greatly affected by density. Seasonal and annual
increases in vegetation abundance resulted in a reduction in the length of activity bouts
because the time required to gather food decreases when vegetation becomes more
abundant. In unenclosed areas, deer spent less time active at the beginning of the summer
and more time resting, likely to process less digestible forage. Deer at high density,
contrarily to deer at low density, did not select areas with dense cover. If population density
reduces forage availability, we predict that deer will adapt by feeding for longer periods,
particularly at the beginning of the summer when forage is more limited. Space utilization
in relation to food and cover is affected by population density.
iv
Avant-propos
Ce mémoire comprend trois articles écrits en anglais pour être publiés dans des revues
scientifiques ainsi qu’une introduction et une conclusion générale en français. Le premier
chapitre décrit une étude de validation d’une méthode que nous avons utilisée pour mesurer
les budgets d’activité décrits dans le chapitre 2. Steeve D. Côté a participé à l’élaboration
de l’étude ainsi qu’à la correction du manuscrit. Ariane Massé a participé au processus
complet, allant de la mise en place, en passant par la prise de données jusqu’à l’analyse et
la rédaction du manuscrit. Ce manuscrit a été soumis à la revue « Wildlife Society
Bulletin » et a été accepté en avril 2005. Le deuxième chapitre présente une étude qui visait
à quantifier les effets de la densité de population sur le comportement de déplacement et le
budget d’activité du cerf de Virginie (Odocoileus virginianus). Les coauteurs Jean Huot et
Steeve D. Côté ont participé à l’élaboration de l’étude ainsi qu’à la correction du manuscrit.
Le troisième chapitre expose la deuxième partie de l’étude qui s’intéressait aux effets de la
densité de population sur l’utilisation des sites en relation avec l’abondance de couvert et de
végétation. Les coauteurs Jean Huot et Steeve D. Côté ont participé à l’élaboration de
l’étude ainsi qu’à la correction du manuscrit.
Je tiens d’abord à remercier Jean Huot puisque c’est grâce à lui que j’ai pu participer au
merveilleux projet de la chaire de recherche industrielle CRSNG-Produits forestiers
Anticosti. Jean a su me faire confiance et ce même dans les moments les plus difficiles. Il a
aussi su démontrer une patiente incomparable et m’apporter des conseils judicieux dans la
prise et l’analyse des données ainsi que la correction des manuscrits. Ensuite, il m’est
essentiel de remercier Steeve Côté puisque c’est grâce à son soutien et à ses conseils si j’ai
pu terminer cette maîtrise avec autant de succès. Je dois le remercier aussi pour les
nombreuses fois qu’il est venu me visiter à Anticosti et pour tous les conseils qu’il a pu
apporter dans la mise en place, la prise de données, l’analyse et la rédaction.
Le laboratoire de Jean et Steeve est aussi reconnu pour son dynamisme incontestable! Je
remercie chacun de vous, membres du « love labo » pour l’aide que vous m’avez apporté à
travers ces années. Il m’est impossible de passer à côté de Jean-Pierre Tremblay puisque
v
sans lui, ce projet n’aurait pu aussi bien fonctionner. Merci pour ton aide, tes conseils et ta
bonne humeur. Un merci particulier à Sonia DeBellefeuille pour toute son aide autant au
niveau moral qu’au niveau de la correction ou de son aide sur le terrain. Tu fais vraiment un
travail exceptionnel! Aussi un merci spécial à Christian Dussault pour les conseils qu’il m’a
donné à partir du complexe G. Et puis, à François Fournier pour son aide dans l’analyse des
données et dans la correction du chapitre 1. Je voudrais remercier Sébastien Lefort et
Vanessa Viera puisque c’est grâce à eux si j’ai pu faire mon premier été de terrain à
Anticosti. Merci à Ariane Massé et Anouk Simard pour m’avoir donné tant de conseils,
d’aide sur le terrain et puis pour tant de discussions importantes en moment de détresse!
Merci, Joëlle pour tout le soleil que tu as su apporter dans mes journées et puis pour ton
aide intarissable sur le terrain. Merci aussi à Sandra Hamel pour sa gentillesse
incomparable et pour m’avoir accepté dans sa cabane sur la montagne. Aussi merci à toi
Daniel Sauvé pour toute l’aide que tu as donné sur le terrain et au bureau. Merci à vous mes
chères colocataires, Catherine Bajzak et Vanessa Viera qui ont su garder mon moral haut, et
puis aussi parce que « la journée la plus perdue est celle où on n’a pas ri ». Merci aussi à
Robert Weladji pour tous les commentaires sur les chapitres 2 et 3. Merci aussi à Martin
Barrette, Valérie Harvey, Julien Mainguy, Stéphanie Pellerin, Antoine St-Louis et Suzy
Tremblay pour toutes les discussions et votre soutien.
Ce projet détenait un terrain laborieux qui n’aurait pu être réalisé sans l’aide de dizaines de
personnes. Merci d’abord à Jean-Pierre Tremblay pour avoir établi le dispositif de densités
contrôlées. Les captures de cerfs ont nécessité l’aide d’une grande équipe. Alors, merci à
Laurier Breton et Bruno Rochette du Ministère des Ressources naturelles et de la Faune du
Québec ainsi qu’à Denis Duteau, François Fournier, Ariane Massé, Gérald Picard, Daniel
Sauvé, Anouk Simard, Jean-François Therrien et Jean-Pierre Tremblay. Merci à tous ceux
qui ont participé aux battues qui n’auraient certainement pas pu avoir lieu sans l’aide
incontournable de Gaétan Laprise et Danièle Morin. Merci aussi à Rémi Pouliot, Martin
Renière, Jean-François Therrien, Jescika Lavergne et Vanessa Viera pour l’aide
inépuisable qu’ils ont su m’apporter lors des suivis télémétriques. Merci aussi à Sonia
DeBellefeuille, Christian Dussault, Michel Duteau, Marie-Andrée Giroux, Léon L’Italien,
vi
Ariane Massé, Joëlle Taillon, Jean-Pierre Tremblay, Andréanne Tousignant et Vanessa
Viera pour leur contribution dans les inventaires de végétation.
Finalement, je dois souligner l’importance de l’aide des partenaires de la Chaire à l’île :
Produits forestiers Anticosti, la SÉPAQ, le MRNFQ, les résidents de l’île d’Anticosti. Ce
sont des partenaires incomparables pour la réussite d’un tel programme de recherche. Merci
aussi à Sophie Baillargeon pour sa contribution dans les analyses statistiques. Merci à
Christian Dussault, Daniel Fortin et Kim Lowell pour les conseils judicieux qu’ils ont pu
m’apporter dans les analyses du chapitre 2. Et merci à ma famille et amis qui m’ont soutenu
durant toutes ces années.
Un tel projet n’aurait pu avoir lieu sans l’appui financier et logistique de Produits forestiers
Anticosti inc., du conseil de recherches en sciences naturelles et en génie du Canada, du
Fonds québécois de la recherche sur la nature et les technologies, du Ministère des
Ressources naturelles et de la Faune du Québec ainsi que du Centre d’études nordiques.
vii
Table des matières
Table des matières ................................................................................................................ vii
Liste des tableaux .................................................................................................................. xi
Liste des figures .................................................................................................................... xii
Liste des annexes ................................................................................................................. xiv
Introduction générale .............................................................................................................. 1
Les populations de cervidés ................................................................................................ 1
L’île d’Anticosti ................................................................................................................. 3
Les effets de la densité de population sur le comportement ............................................... 4
Le comportement d’alimentation........................................................................................ 5
Le budget d’activité ........................................................................................................ 5
Effets des caractéristiques individuelles ......................................................................... 6
Influence des variables environnementales .................................................................... 7
La densité de population ................................................................................................. 8
La sélection d’un site d’alimentation ................................................................................. 8
Objectifs de l’étude ........................................................................................................... 10
Méthodologie .................................................................................................................... 11
Chapitre 1. Quantification and accuracy of activity data measured with VHF and GPS
telemetry ............................................................................................................................... 17
Résumé ............................................................................................................................. 18
Abstract ............................................................................................................................. 19
Introduction ...................................................................................................................... 20
Study area ......................................................................................................................... 23
Methods ............................................................................................................................ 24
Calibration of VHF and GPS motion sensors on captive white-tailed deer fawns ...... 24
Validation of GPS motion sensors on free-ranging deer .............................................. 25
Data analysis ................................................................................................................. 26
Validation of activity counts of GPS collars on free-ranging deer .............................. 29
Results .............................................................................................................................. 29
viii
Determination of specific behaviors ............................................................................. 31
Calibration of VHF and GPS motion sensors ............................................................... 31
Calculation of activity bouts ......................................................................................... 33
Activity of free-ranging deer ........................................................................................ 33
Discussion ......................................................................................................................... 37
Determination of specific behaviors ............................................................................. 37
VHF collars .................................................................................................................. 37
GPS collars ................................................................................................................... 39
Research and management implications........................................................................... 41
Acknowledgements .......................................................................................................... 41
Literature cited .................................................................................................................. 42
Chapitre 2. Influence of population density on white-tailed deer movements and activity
budgets .................................................................................................................................. 44
Résumé ............................................................................................................................. 45
Abstract ............................................................................................................................. 46
Introduction ...................................................................................................................... 47
Study area ......................................................................................................................... 49
Methods ............................................................................................................................ 49
Experimental design ..................................................................................................... 49
Deer captures ................................................................................................................ 50
Forage abundance ......................................................................................................... 52
Movements ................................................................................................................... 52
Activity budgets ............................................................................................................ 53
Analyses ........................................................................................................................... 55
Results .............................................................................................................................. 57
Forage abundance ......................................................................................................... 57
Movements ................................................................................................................... 57
Proportion of time spent active ..................................................................................... 57
Number of activity bouts .............................................................................................. 61
Length of active and inactive bouts .............................................................................. 66
ix
Discussion ......................................................................................................................... 67
Influence of population density .................................................................................... 67
Annual differences ........................................................................................................ 69
Seasonal differences ..................................................................................................... 70
Diel activity pattern ...................................................................................................... 72
Conclusion ........................................................................................................................ 72
Acknowledgements .......................................................................................................... 73
Literature cited .................................................................................................................. 73
Chapitre 3. Influence of forage abundance, cover and population density on white-tailed
deer space use ....................................................................................................................... 78
Résumé ............................................................................................................................. 79
Abstract ............................................................................................................................. 80
Introduction ...................................................................................................................... 81
Study area ......................................................................................................................... 83
Methods ............................................................................................................................ 83
Experimental design ..................................................................................................... 83
Deer captures ................................................................................................................ 84
Telemetry ...................................................................................................................... 84
Biomass and cover sampling ........................................................................................ 86
Analyses ....................................................................................................................... 87
Results .............................................................................................................................. 90
Spatial analysis ............................................................................................................. 90
Descriptive statistics ..................................................................................................... 97
Deer space use .............................................................................................................. 97
Discussion ....................................................................................................................... 101
Deer space use in relation to plant biomass and cover ............................................... 101
Limitations and strengths of the study ........................................................................ 103
Acknowledgments .......................................................................................................... 105
Literature cited ................................................................................................................ 106
Conclusion générale ........................................................................................................... 110
x
Validation des capteurs d’activité .................................................................................. 110
Les déplacements et le budget d’activité ........................................................................ 111
Le compromis couvert/nourriture ................................................................................... 114
Conclusions et recommandations ................................................................................... 117
Bibliographie générale ........................................................................................................ 119
xi
Liste des tableaux
Table 1–1. Individual and combined relative mean pulse rates (BPM) from variable-pulse
activity sensors that correctly classifieda the observed behaviors of captive white-tailed
deer fawns fitted with VHF collars during 21 November-23 December on Anticosti
Island, Québec. ............................................................................................................. 32
Table 1–2. Activity counts recorded during 4-minute intervals that correspondeda to
observed behaviors from captive white-tailed deer fawns fitted with GPS collars on
Anticosti Island, Québec............................................................................................... 34
Table 2–1. Characteristics of white-tailed deer used in an experiment on the effects of
population density on deer activity budgets on Anticosti Island, Québec.................... 51
Table 2–2. Comparisons of white-tailed deer summer movement rates in two controlled
densities according to age class, week and period of the day (Anticosti Island,
Québec). ........................................................................................................................ 59
Table 2–3. Proportion of time that white-tailed deer spent active in summer at two
controlled densities on Anticosti Island, Québec. ........................................................ 60
Table 2–4. Number of daily activity bouts (a) and length (min.) of active (b) and inactive
bouts (c) during summer of yearling and adult white-tailed deer at two controlled
densities on Anticosti Island, Québec. ......................................................................... 65
Table 3–1. Number of locations recorded in each diel period for radiocollared white-tailed
deer tracked in controlled-density enclosures on Anticosti Island, Québec. ................ 85
Table 3–2. Mean biomass (g/m²), lateral covera (/20 points) and canopy coverb (/4; ± SD)
according to deer density and stratum (forest stands and clear-cuts). Deer were kept in
3 sets of enclosures with 2 densities each on Anticosti Island, Québec. ...................... 99
Table 3–3. White-tailed deer relative space usea according to biomass, canopy cover and
lateral cover at 2 different densities (7.5 and 15 deer/km²) in a controlled-density
experiment on Anticosti Island, Québec for each diel period.b .................................. 100
xii
Liste des figures
Figure 1. Le bloc A et la disposition des 2 densités contrôlées dans les enclos construits sur
l’île d’Anticosti (Québec, Canada). Nous avons introduits 3 cerfs de Virginie dans 2
enclos de différentes grandeurs pour obtenir 2 densités différentes
……………………………………….....................................……….........................13
Figure 2. Localisation des sites où nous avons mesuré le budget d’activité, les déplacements
et l’utilisation de l’espace dans deux densités contrôlées dans des enclos situés sur
l’île d’Anticosti (Québec, Canada). La carte présente aussi les grands groupes
forestiers présents sur l’île en 1999…………………………………………………..14
Figure 1–1. Box plot representations of relative mean pulse rates (BPM) for VHF variable-
pulse sensors (a) and of activity counts for vertical (b), and horizontal (c) GPS activity
sensors for the 4 different behaviors observed. ............................................................ 27
Figure 1–2. Determination of a criterion to separate relative mean pulse rates (BPM) for
VHF collars equipped with variable-pulse motion sensors (a), and activity counts for
GPS collars equipped with double-axis motion sensors (b) into active and inactive
behaviors of white-tailed deer on Anticosti Island, Québec. ........................................ 28
Figure 1–3. Observed behavior and relative mean pulse rate (BPM) obtained during one
day for a white-tailed deer fawn fitted with a VHF collar equipped with a variable-
pulse sensor on Anticosti Island, Québec. .................................................................... 30
Figure 1–4. Relationship between observed and estimated proportion of daily active time
obtained with variable-pulse activity sensors of VHF collars fitted to 4 white-tailed
deer fawns each observed for 6 days ( x = 4 hr of observation per day) on Anticosti
Island, Québec. ............................................................................................................. 35
Figure 1–5. Mean activity counts (± 1 SE) recorded by horizontal and vertical sensors of
GPS collars fitted on free-ranging deer on Anticosti Island in summer and autumn
2001 (n = 8 deer) and 2002 (n = 8 deer). ...................................................................... 36
Figure 2–1. Mean plant biomass available to white-tailed deer in a controlled-density
experiment on Anticosti Island, Québec containing known densities (7.5 deer/km²:
black bars, 15 deer/km²: grey bars) of deer. ................................................................. 58
xiii
Figure 2–2. Activity budgets of white-tailed deer from Anticosti Island (Québec) according
to the number of years since the onset of a controlled-density experiment. ................ 62
Figure 2–3. Proportion of daily time spent active (a), number of daily activity bouts (b),
length of active (c), and inactive bouts (d) during summer for adult white-tailed deer
on Anticosti Island (Québec), pooled across years....................................................... 63
Figure 2–4. Proportion of daily time spent active (a), number of daily activity bouts (b),
length of active (c), and inactive bouts (d) during summer for yearling white-tailed
deer on Anticosti Island (Québec), pooled across years. .............................................. 64
Figure 3–1. Plant biomass (g/m²) available to white-tailed deer and interpolated by kriging
in 2003 for Block A on Anticosti Island, Québec. ................................................. 91−92
Figure 3–2. Lateral cover, or mean concealment (attributed to 4 classes 1: 0-25; 2: 26-50; 3:
51-75; 4: 76-100%) of the first 2 sections of a concealment board (2.5 m×0.3 m
divided in 0.5 m sections) in 2 opposite directions, available to white-tailed deer and
interpolated by kriging in 2003 for Block A on Anticosti Island, Québec. ............ 93−94
Figure 3–3. Canopy cover, or proportion of 20 points set at every 3 m from the center of
each sampling unit in 4 directions (east, southeast, southwest and west) where foliage
of >4 m trees was present, available to white-tailed deer and interpolated by kriging in
2003 for Block A on Anticosti Island, Québec. ..................................................... 95−96
Figure 3–4. Relationships between white-tailed deer relative space use (number of
overlapping buffers for a deer divided by the total number of positions for that deer in
every diel-period and at each random point) and plant biomass, lateral (mean
concealment value attributed by 4 classes of 25%) and canopy cover (proportion of 20
points where foliage of >4 m trees was present).. ........................................................ 98
xiv
Liste des annexes
Appendix 3−1. Spatial statistical data of biomass (a) lateral cover (b) and canopy cover (c)
in cuts and forests of enclosures containing different densities of white-tailed deer on
Anticosti Island, Québec............................................................................................. 126
1
Introduction générale
Les populations de cervidés
Depuis les dernières décennies, les populations de cervidés ont augmenté considérablement
dans plusieurs régions de l’Amérique du Nord et de l’Europe (Garrott et al. 1993). Dans
plusieurs régions, le déclin des populations de cerf de Virginie (Odocoileus virginianus)
dans les années 1950 avait pourtant incité les gestionnaires à diminuer la récolte par la
chasse (Rooney et Waller 2003, Côté et al. 2004). La diminution des prédateurs tels que le
loup (Canis lupus) a certes profité à l’augmentation des populations de cervidés mais la
diminution des prédateurs et la baisse de la chasse ne sont pas les seules responsables des
augmentations. En effet, la population humaine grandissante a fragmenté le territoire en le
défrichant pour en faire des champs agricoles ou simplement pour en récolter le bois (Côté
et al. 2004). Ces champs procurent un milieu favorable pour le cerf de Virginie en
augmentant la quantité de végétation disponible (Porter et Underwood 1999). Somme toute,
le cerf de Virginie est un animal qui s’adapte rapidement à son environnement et c’est
sûrement un facteur déterminant pour lequel les populations ont augmenté aussi
rapidement. Le cerf a trouvé avantage aux perturbations anthropiques.
Les cerfs à haute densité peuvent à leur tour modifier la structure et la composition des
communautés végétales (Rooney et Waller 2003, Côté et al. 2004). Le cerf de Virginie est
considéré comme une espèce généraliste qui sélectionne les plantes ou parties de plantes
dont il se nourrit afin d’optimiser l’acquisition d’énergie (Hofmann 1989). Les cerfs
peuvent consommer certaines espèces préférées à un point tel qu’à haute densité,
l’abondance de ces espèces diminue et certaines peuvent même disparaître (Healy 1997).
En effet, plusieurs études ont établi une relation directe entre l’intensité du broutement et
l’abondance des espèces préférées par le cerf (Balgooyen et Waller 1995, Rooney et Dress
1997).
Des études en enclos et des comparaisons insulaires ont démontré que les ongulés peuvent
même modifier la composition de la strate arborescente (Tilghman 1989, Healy 1997). En
effet, le broutement peut contribuer à l’échec de régénération et à la création d’ouvertures
2
qui contribuent indirectement à l’augmentation des graminées, des fougères et des
cypéracées (Anderson et al. 2001, Cooke et Farrell 2001, Kirby 2001, Rooney 2001,
Bradshaw et al. 2003) qui à leur tour, contribuent à l’échec de régénération (Stromayer et
Warren 1997).
Pour se défendre du broutement, certaines plantes produisent des métabolites secondaires
qui diminuent la digestibilité de leurs tissus ou encore développent des structures de
protection comme des épines (Schultz 1988, Hobbs 1996, Augustine et McNaughton 1998)
qui diminuent l’attrait des plantes pour les cerfs (Palo 1985, Bryant et al. 1991, Palo et
Robbins 1991, Bryant et al. 1992). Avec le temps, les espèces tolérantes ou résistantes au
broutement (sensu Boege et Marquis 2005) deviennent plus abondantes que les espèces
vulnérables (Hobbs 1996, Augustine et McNaughton 1998).
Ainsi, il semble qu’à haute densité les cerfs peuvent se retrouver dans des milieux où
l’abondance des plantes préférées est limitée. Les impacts du cerf sur la végétation peuvent
même être irréversibles sans une intervention anthropique (Stromayer et Warren 1997,
Augustine et al. 1998). Pour subsister, les cerfs doivent donc s’acclimater à leur milieu en
modifiant, notamment, leur comportement. En absence d’adaptation physiologique
nouvelle, s’ils ne peuvent combler leurs besoins énergétiques en modifiant leur
comportement, on peut prédire une diminution de la masse, de la reproduction et
éventuellement de la survie en fonction de la densité de la population (Clutton-Brock et al.
1987).
Dans un contexte où plusieurs populations de cervidés sont en croissance en Amérique du
Nord et en Europe et que ces populations ont des impacts importants sur leur
environnement, nous devons nous demander si et comment le comportement des cerfs est
modifié en fonction de l’augmentation de la compétition intra-spécifique et de la
diminution de la disponibilité de la végétation.
3
L’île d’Anticosti
L’île d’Anticosti est un cas particulier, où les fortes densités de cerfs ont causé des impacts
considérables sur la végétation. Depuis l’introduction d’environ 200 cerfs de Virginie sur
l’île d’Anticosti à la fin du 19e siècle, la population a connu une croissance rapide de sorte
que dès le milieu des années 1940, l’île était déjà identifiée à l’échelle nord-américaine
comme un endroit de surpopulation de cerf (Leopold et al. 1947). Le dernier inventaire a
estimé la densité à 16 cerfs/km2, soit un total de 127 000 têtes (Rochette et al. 2003). Cette
forte croissance serait due à la conjugaison de facteurs favorables à l’établissement du cerf
de Virginie tels que l’absence de prédateurs, le climat favorable, et les perturbations
naturelles et anthropiques qui prévalent sur Anticosti qui ont favorisé la création de bons
habitats pour le cerf en offrant une grande abondance de nourriture.
Déjà, les premiers impacts du broutement sur la végétation ont été notés dans les années
1930 (Marie-Victorin et Rolland-Germain 1969). Aujourd’hui, l’ampleur des dommages
causés par les cerfs est inquiétante pour l’avenir des forêts d’Anticosti (Potvin et al. 2003).
En effet, la composition spécifique des strates arbustives et herbacées a largement été
modifiée (Huot 1982, Potvin et al. 2003). Plusieurs espèces arbustives et herbacées sont
maintenant rares ou ont disparues (Potvin et al. 2000). Des plantes aussi résistantes que le
framboisier (Rubus idaeus) et l’épilobe à feuilles étroites (Epilobium angustifolium) ne se
retrouvent plus dans les parterres de coupe de l’île (Potvin et al. 2000). Les cerfs ont même
transformé la composition de la strate arborescente à l’échelle de l’île (Potvin et al. 2003).
Des études ont montré que les semis de sapin (Abies balsamea) sont fortement broutés et ce
même en été et jusqu’au centre de grandes coupes situées à plus de 800 m de la bordure de
la forêt (Potvin et Laprise 2002). Depuis les années 1930, les sapinières qui occupaient
initialement environ 40% de la superficie de l’île sont graduellement remplacées par des
peuplements d’épinette blanche (Picea glauca), une espèce qui est peu broutée (Potvin et
al. 2003). Les cerfs en sont donc arrivés à modifier leur environnement de façon globale. La
situation pourrait cependant changer de façon majeure au cours des prochaines décennies
puisque les sapinières, qui procurent aux cerfs la principale source d’alimentation hivernale
4
(environ 70% du régime alimentaire; Huot 1982, Lefort 2002) disparaissent rapidement
(Potvin et al. 2003).
Plusieurs indicateurs montrent déjà la vulnérabilité des cerfs à Anticosti; ils sont parmi les
plus petits en Amérique du Nord (Boucher et al. 2004) et les femelles ne se reproduisent en
général qu’à l’âge de deux ans et demi, soit une année plus tard qu’ailleurs (Potvin 1985).
De plus, elles ne se reproduisent pas chaque année et ont très peu de jumeaux. Étant donné
l’ampleur des effets du broutement du cerf sur la végétation et la diminution de la qualité de
l’habitat d’hiver, il est important de connaître si et comment la densité de population affecte
le comportement d’alimentation des cerfs afin de pouvoir adopter des méthodes de gestion
convenables au maintien des écosystèmes de l’île.
L’île d’Anticosti représente donc un milieu idéal pour poursuivre une telle étude puisque
les communautés végétales de l’île ont été profondément modifiées par le broutement du
cerf (Potvin et al. 2003). De plus, l’île se trouve à la limite nord de l’aire de répartition du
cerf de Virginie, l’été représente donc une saison critique pour les cerfs puisqu’ils doivent
profiter de la brève saison de croissance de la végétation pour rétablir leur condition
physique et amasser des réserves corporelles qui seront essentielles pendant la période
hivernale (Putman et al. 1996, Lesage et al. 2001). Ils ont donc avantage à optimiser leur
temps d’alimentation et leur sélectivité durant cette période.
Les effets de la densité de population sur le comportement
Quelques études en milieu naturel ont démontré que le comportement d’approvisionnement
et le budget d’activité différaient chez l’orignal (Alces alces) dans deux populations vivant
à différentes densités (Cederlund et al. 1989), chez le cerf Sika (Cervus nippon) au cours de
deux années pendant lesquelles la densité de population avait diminué (Borkowski 2000) ou
chez des cerfs de Virginie vivant dans des habitats et des densités différents (Rouleau et al.
2002). Les effets de la densité de population sur le comportement d’approvisionnement ont
été indirectement mesurés en contrôlant expérimentalement la quantité et/ou la qualité de la
biomasse végétale (Trudell et White 1981, Vivås et Sæther 1987) ou en mesurant le temps
passé à s’alimenter en fonction de la quantité de biomasse disponible (Gillingham et al.
5
1997). Malgré tout, en milieu naturel, les effets de la densité de population sur le
comportement d’alimentation et le budget d’activité des cervidés sont encore mal connus
étant donné la difficulté d’y observer des ongulés comme le cerf de Virginie en milieu
naturel et de contrôler la densité de population (Stenseth 1981, Borkowski 2000). Pourtant,
le comportement d’approvisionnement est déterminant dans l’expression des stratégies
d’histoire de vie adoptées par les animaux et son étude plus approfondie permettrait de
mieux comprendre l’impact des cervidés sur la végétation forestière (Miller 1997, Miller et
Ozoga 1997).
Depuis plusieurs années, les expériences de contrôle de la densité ont été utilisées pour
mieux comprendre les effets de la densité de population sur le comportement
d’alimentation des espèces domestiques (Hester et Kirby 1996). Les recherches manipulant
la densité de population sont maintenant fortement suggérées pour l’étude du
comportement des espèces sauvages puisqu’elles permettent de contrôler et de répliquer
directement différentes densités de cerfs (Hester et al. 2000, Gordon et al. 2004). Pour les
cervidés sauvages, ces études s’intéressent généralement à l’influence de la densité
d’herbivores sur l’abondance et la diversité des espèces végétales (Tilghman 1989, Hester
et al. 2000); cependant ces expériences sont aussi des outils exemplaires pour mesurer
l’influence de la densité sur le comportement d’approvisionnement et le budget d’activité.
Le comportement d’alimentation
Le comportement d’alimentation des herbivores peut être examiné selon deux angles
distincts et complémentaires qui nous permettront d’estimer les effets de la densité de
population sur le comportement : soit le budget d’activité ou la répartition du temps
octroyé à différentes activités et le choix de sites d’alimentation permettant de maximiser le
gain d’énergie par unité de temps.
Le budget d’activité
Pendant l’été, les cerfs consacrent en général 90–95% de leur temps passé en activité à
s’alimenter (Beier et McCullough 1990, Gillingham et al. 1997). Le temps passé en activité
représente donc majoritairement le temps passé en alimentation. Par ailleurs, le temps passé
6
inactif donne un indice du temps passé en rumination. En effet, une diminution de la qualité
de la végétation est reliée à une augmentation du temps nécessaire à la rumination et à une
augmentation du temps passé en inactivité (Mysterud 1998, Pérez-Barbería et Gordon
1999). Le temps passé actif peut être fonction des caractéristiques individuelles mais aussi
des conditions de l’environnement. Ainsi le rôle de la densité de population sur le
comportement des cerfs doit être considéré en fonction de ces deux groupes de facteurs.
Effets des caractéristiques individuelles
Le temps optimal qu’un individu passe à s’alimenter est fonction de son âge, sexe et statut
reproducteur puisque les demandes énergétiques reliées au métabolisme, à la croissance et à
la reproduction peuvent varier selon ces paramètres (Clutton-Brock et al. 1982). Puisque les
grands herbivores ont une demande énergétique absolue plus grande que les petits, il a été
suggéré que le temps passé en alimentation augmente avec la taille corporelle au niveau
inter et intra-spécifique (Bell 1971). Cependant, chez différentes espèces d’ongulés des
zones tempérées, Mysterud (1998) a trouvé que la proportion du temps passé en activité
diminuait de façon allométrique avec la taille corporelle. En effet, le taux métabolique est
lié allométriquement à la masse corporelle alors que la taille du rumen est reliée de façon
isométrique à la masse corporelle (Illius et Gordon 1987) de telle sorte que lorsque la masse
corporelle augmente, la taille du rumen devient proportionnellement plus grande par
rapport aux coûts métaboliques. Puisque le temps de passage de la végétation est
proportionnel à la taille du rumen, il a été suggéré que les plus gros ruminants pouvaient
consommer de la végétation de moins bonne qualité et ainsi, qu’ils passaient moins de
temps actif à la rechercher (Mysterud 1998, Pérez-Barbería et Gordon 1999). Par exemple,
chez les espèces où le dimorphisme sexuel est important, il a été démontré que les mâles
passent moins de temps en activité que les femelles mais ces différences diminuent lorsque
le dimorphisme est réduit (Zhang 2000, Shi et al. 2003) et augmentent avec l’importance du
dimorphisme (Moncorps et al. 1997, Ruckstuhl 1997, Ruckstuhl et Neuhaus 2002).
Des études ont démontré que les juvéniles passaient davantage de temps en activité que les
adultes puisqu’ils ont un taux métabolique plus élevé relativement à la taille de leur
système digestif (Bunnell et Gillingham 1985). L’influence de la densité de population
7
pourrait donc être différente selon l’âge des individus puisque les besoins énergétiques en
termes de croissance et de métabolisme varient entre les juvéniles et les adultes (Bunnell et
Gillingham 1985). Puisqu’ils utilisent les ressources différemment, l’augmentation de la
densité de population pourrait aussi influencer différemment ces deux groupes d’âge. Par
exemple, la survie des juvéniles serait davantage affectée par la densité que la survie des
adultes (Jorgenson et al. 1997).
Influence des variables environnementales
L’abondance et la qualité des plantes peuvent varier au cours de l’année. À Anticosti, la
saison de croissance débute à la fonte des neiges entre le début et la mi-mai (Ressources
naturelles Canada 2005). On observe alors une croissance rapide des plantes dans les
milieux ouverts et puis, graduellement, une croissance des plantes herbacées dans les sous-
bois. Les nouvelles pousses sont riches en protéines mais au cours de l’été, la végétation
augmente en abondance et sa teneur en fibres augmente, ce qui a pour effet de diminuer sa
digestibilité (Tremblay 1981, Robbins 1983, Van Soest 1994). Ces changements dans la
qualité et l’abondance de la végétation pourraient avoir un impact sur le comportement
d’alimentation en modifiant le temps nécessaire à l’acquisition et à l’assimilation de la
végétation et, donc, de l’énergie (Van Soest 1982). Par exemple, étant donné la diminution
de qualité de la végétation pendant l’été, Beier et McCullough (1990) ont trouvé que le
temps passé en activité augmentait pendant l’été et qu’au contraire, pendant l’hiver, pour
conserver leur énergie, les cerfs diminuaient leur activité.
Le budget d’activité est aussi influencé par des variables abiotiques telles que l’heure de la
journée et les conditions météorologiques. En effet, les cerfs ont tendance à être plus actifs
au lever et au coucher du soleil (Zagata et Haugen 1974, Kammermeyer et Marchinton
1977). De plus, plusieurs études ont démontré qu’il existe des relations entre les conditions
météorologiques et l’activité des herbivores. Par exemple, les cerfs sont moins actifs et
utilisent davantage les milieux fermés lorsque les conditions météorologiques sont
défavorables (p. ex. vents forts et précipitations abondantes pendant l’hiver; Miller 1970,
Zagata et Haugen 1974, Drolet 1976, Beier et McCullough 1990).
8
La densité de population
La biomasse de plantes préférées par individu diminue généralement en fonction de la
densité de population (Boucher et al. 2004) mais le taux de consommation ne diminue pas
nécessairement en fonction de la densité d’herbivores (Fortin et al. 2004). Néanmoins, à
forte densité d’herbivores, le temps passé en alimentation pourrait augmenter en réponse à
une diminution de la biomasse (Wickstrom et al. 1984, Renecker et Hudson 1986).
L’augmentation de la population peut donc contraindre les individus à rester actifs plus
longtemps et à se déplacer davantage pour acquérir une même quantité de nourriture
(Herbers 1981, Trudell et White 1981, Clutton-Brock et al. 1982, Gates et Hudson 1983,
Moncorps et al. 1997). Les cervidés peuvent aussi répondre à la diminution de l’abondance
de végétation en se nourrissant de manière moins sélective et en diminuant leurs
déplacements (Gates et Hudson 1983). Cependant, puisque le temps de rumination
augmente avec la quantité de fibres des espèces tolérantes au broutement (Baker et Hobbs
1986, Spalinger et al. 1986), la consommation de plantes de moindre qualité nécessite un
temps de rumination plus long, ce qui contribue à la diminution du temps passé en activité.
La sélection d’un site d’alimentation
En second lieu, les individus doivent sélectionner des endroits pour s’alimenter. Pour les
cervidés, un bon site d’alimentation représente habituellement un compromis entre la
proximité d’un abri (couvert) et l’abondance de nourriture (Tufto et al. 1996).
Le couvert peut être séparé en deux parties : 1) le couvert vertical représente l’abri formé
par la projection des cimes de la canopée jusqu’au sol, il est généralement formé de
végétation, 2) le couvert latéral représente l'obstruction latérale et peut être formé de
végétation ou de topographie (Mysterud et Østbye 1999). Le couvert latéral diminue
habituellement le risque de prédation donc le temps que les animaux doivent consacrer aux
comportements de vigilance (Mysterud et Østbye 1995) et en absence de prédateurs, il est
considéré comme jouant un rôle « psychologique » dans la sélection d’habitat en relation
aux prédateurs fantômes du passé (Byers 1997, Mysterud et Østbye 1999). De plus, les
animaux sont exposés à des conditions météorologiques moins stressantes dans les milieux
9
fermés (couvert latéral ou vertical élevés) que dans les milieux ouverts (Ozoga 1968, Huot
1974, Beier et McCullough 1990, Schimtz 1991, Mysterud et Østbye 1999). Cependant, les
milieux ouverts offrent généralement une plus grande abondance de nourriture estivale pour
les herbivores (Hanley 1984). On peut donc prédire que les cervidés se nourriront
principalement près des bordures puisqu’ils minimisent ainsi le compromis entre trouver un
abri contre les prédateurs et les conditions météorologiques plus difficiles et la disponibilité
de la nourriture qui est plus grande dans les milieux ouverts (Keay et Peek 1980, Tufto et
al. 1996).
Le budget d’activité, les conditions environnementales et le risque de prédation peuvent
varier selon la période du jour. Ainsi, l’importance du couvert peut aussi varier selon ces
facteurs. Dans plusieurs études, on a observé que les cerfs préfèrent utiliser les milieux
ouverts pendant la nuit soit parce que la prédation (Altendorf et al. 2001), le harcèlement
par les insectes (Mysterud et Østbye 1999) ou les activités anthropiques, telles la chasse
(Kilgo et al. 1998) et les activités agricoles (Rouleau et al. 2002) sont réduits pendant les
périodes de noirceur. Beier et McCullough (1990) ont trouvé que les cerfs utilisent aussi
des environnements ouverts pendant les périodes de noirceur, excepté en été où les cerfs
utilisent aussi des milieux ouverts pendant le jour. Ils proposèrent que les cerfs pourraient
utiliser les milieux ouverts pendant le jour puisque les graminées et leur couvert végétal
grand et dense offrent un couvert suffisant pour se cacher des prédateurs et s’abriter des
fortes températures. Le couvert de la canopée et le couvert latéral, couplés à l’abondance de
végétation peuvent donc jouer des rôles importants dans la sélection d’un site
d’alimentation.
Les coupes forestières procurent de tels habitats ouverts, où l’on rencontre une grande
quantité de végétation, entremêlées à des îlots forestiers qui présentent des milieux plus
fermés mais fournissant peu de nourriture (Masters et al. 1993). Les cerfs sélectionnent
habituellement ces coupes par rapport aux îlots forestiers lorsqu’elles offrent davantage de
nourriture et qu’elles ont suffisamment de couvert latéral (Lyon et Jensen 1980). Tierson et
al. (1985) avaient trouvé que les cerfs arrêtaient leur migration aux sites d’hivernage pour
se nourrir dans les coupes.
10
La densité de population peut aussi jouer un rôle important quant au compromis
couvert/nourriture en diminuant la quantité de végétation disponible (Healy et al. 1997) et
en augmentant le nombre d’individus dans un endroit donné. Le compromis
couvert/nourriture pourrait donc aussi être modifié selon la densité. En effet, on a trouvé
que les cerfs évitent de se nourrir dans les milieux ouverts à moins qu’ils ne se trouvent à
haute densité parce qu’ils y consacraient trop de temps aux comportements de vigilance
(Lesage et al. 2000). Rouleau et al. (2002) avaient trouvé que les cerfs vivant à haute
densité dans les milieux agricoles utilisent ces milieux aussi pendant la nuit contrairement
aux cerfs à faible densité. Pendant l’été, dans les milieux agricoles, l’augmentation de
l’utilisation des habitats ouverts pendant le jour et la nuit pourrait donc refléter les impacts
de fortes densités et l’abondance plus faible des plantes. En effet, la sélection des sites
d’alimentation peut être modifiée à haute densité puisque la nourriture est limitée, les cerfs
doivent donc quitter le couvert pour se nourrir dans des endroits où la nourriture est plus
abondante (Mysterud et Østbye 1999). En milieu forestier, il reste encore à savoir si la
densité a un impact sur l’utilisation des coupes pendant l’été.
Objectifs de l’étude
Les effets de la densité de population sur le comportement d’approvisionnement et le
budget d’activité des cerfs sont encore mal connus. Les expériences en densités contrôlées
sont des outils efficaces pour mesurer directement son influence. Au cours de cette étude,
notre objectif était donc de mieux comprendre le rôle de la densité de population sur les
déplacements et le budget d’activité de cerfs de Virginie (Chapitre 2) ainsi que sur le
compromis couvert/nourriture (Chapitre 3) montré par les cerfs lorsqu’ils sélectionnent leur
habitat. Nous avons utilisé une approche expérimentale en manipulant la densité de cerfs
afin de tester les hypothèses suivantes:
Hypothèse 1: Le comportement d’approvisionnement (déplacements et utilisation des sites
d’alimentation ou de couvert) des cerfs est déterminé par l’abondance de la nourriture qui
diminue avec l’augmentation de la densité de population.
11
Hypothèse 2: Les cerfs répondent à la diminution de l’abondance de la nourriture à haute
densité en augmentant la proportion de temps passé en activité et la durée des périodes
d’activité possiblement de façon à sélectionner la végétation de meilleure qualité même si
celle-ci est moins abondante.
Hypothèse 2 alternative: Les cerfs répondent à la diminution de l’abondance de la
nourriture à haute densité en modifiant leur budget d’activité possiblement de façon à
maximiser l’acquisition d’énergie à partir d’une végétation de faible qualité.
Afin d’étudier l’influence de la densité de population sur le comportement d’alimentation
du cerf, nous présentons 3 chapitres qui permettront de tester les hypothèses. Nous
présentons dans le chapitre 1 une étude qui justifie l’utilisation de capteurs d’activité afin
de mesurer le budget d’activité des cerfs. Ensuite, dans le chapitre 2 nous montrons les
effets de la densité de population sur le comportement de déplacement et le budget
d’activité du cerf de Virginie. Finalement, le chapitre 3 considère des effets de la densité de
population sur l’utilisation des sites en relation avec l’abondance de couvert et de
végétation.
Méthodologie
Dans le premier chapitre, nous avons déterminé la précision des capteurs d’activité à
impulsions variables des colliers VHF et des capteurs à deux axes des colliers GPS à
mesurer l’activité des cerfs. À cette fin, 4 cerfs de Virginie ont été munis de colliers
émetteurs VHF et 4 cerfs ont été munis de colliers GPS dans des enclos de 50×80 m. Nous
avons directement observé l’activité des individus dont les signaux étaient simultanément
enregistrés soit, pour les colliers VHF, dans un récepteur automatisé qui mesurait
l’intervalle moyen entre deux impulsions pendant 65 impulsions ou pour les colliers GPS,
dans l’enregistreur de données des colliers GPS à toutes les 5 minutes. Ensuite, nous avons
comparé les données observées et obtenues pour évaluer la précision des capteurs à
discerner différents comportements actifs (p.ex. alimentation vs. déplacements) et inactifs
(p.ex. repos vs. debout) et développé une méthode pour quantifier les périodes d’activité
des individus munis de colliers VHF. La proportion du temps actif, la durée des périodes
12
d’activité, d’inactivité et le nombre de périodes par jour estimés à partir des données des
colliers ont alors été comparés avec les données observées. Nous avons enfin évalué si les
données d’activité des colliers GPS pouvaient décrire des patrons d’activité journaliers de
16 cerfs en liberté sur l’île d’Anticosti (Québec, Canada).
Afin d’étudier le rôle de la densité sur le comportement d’approvisionnement et le budget
d’activité, nous avons mis en place un dispositif de densité contrôlée formé de 3 blocs.
Dans chaque bloc, deux densités contrôlées ont été mises en place en disposant 3 individus
dans un enclos de 40 ha (7.5 cerfs/km²) et 3 individus dans un enclos de 20 ha (15
cerfs/km²; Figure 1). Les enclos ont été disposés dans des coupes forestières effectuées en
2001 pour lesquelles 30% de la surface forestière était maintenue pour servir d’abris. Les
cerfs étaient munis de colliers émetteurs VHF équipés de capteurs d’activité. Nous avons
mesuré le budget d’activité des individus dans un des blocs la première année (A) et la
troisième année d’application du traitement de densité contrôlée. La deuxième année, le
budget d’activité des cerfs a été étudié dans 2 blocs (A, C). Les déplacements et l’utilisation
de l’espace disponible ont été étudiés dans un de ces blocs (A) la première et dans les 3
blocs (A, B, C) la deuxième année après l’application du traitement des densités contrôlées
(Figure 2). Nous avons équipé de colliers émetteurs et quantifié le budget d’activité de 4
femelles adultes dans des coupes non clôturées pendant une année, soit en 2003 ou la
deuxième année après le début des traitements contrôlés (T; Figure 2).
Dans le deuxième chapitre, nous avons mesuré l’influence de la densité de population sur
les déplacements et le budget d’activité des cerfs. D’abord, à chaque année, afin de
connaître la disponibilité de la végétation selon les densités, 20 parcelles ont été placées
dans la coupe et 20 autres parcelles ont été placées dans les îlots forestiers pour chaque
enclos ce qui donnait un total de 80 parcelles par bloc. À toutes les parcelles, la biomasse
herbacée a été estimée pour les espèces les plus importantes pour le cerf en évaluant
visuellement le pourcentage de recouvrement. À l’aide de régressions établies entre le
pourcentage de recouvrement et de la biomasse, nous avons estimé la biomasse disponible
dans chaque parcelle (Bonham 1989). Le nombre d’échantillons requis pour établir les
13
Figure 1. Le bloc A et la disposition des 2 densités contrôlées dans les enclos construits sur
l’île d’Anticosti (Québec, Canada). Nous avons introduit 3 cerfs de Virginie dans 2 enclos
de différentes grandeurs pour obtenir 2 densités différentes.
Coupe
Forêt résiduelle
200 m
3 cerfs dans 20 ha ou
15 cerfs/km²
3 cerfs dans 40 ha ou
7.5 cerfs/km²
14
Figure 2. Localisation des sites où nous avons mesuré le budget d’activité, les déplacements
et l’utilisation de l’espace de cerfs de Virginie munis de colliers VHF dans deux densités
contrôlées dans des enclos situés sur l’île d’Anticosti (Québec, Canada). La carte présente
aussi les grands groupes forestiers présents sur l’île en 1999.
A
BC
T
Épinette blanche
Sapin
Épinette noire
Tourbières
60 km
Pessières blanches
Sapinières
Pessières noires
Tourbières
15
régressions a été établi de façon empirique en reportant graphiquement les coefficients de
régression en fonction du nombre d’échantillons jusqu’à l’obtention d’une asymptote
(Frontier 1983). Le pourcentage de recouvrement a été estimé visuellement à l’intérieur de
deux quadrats de 1 m2 choisis aléatoirement à l’intérieur d’un cadre de 10×10 m centré sur
le milieu de la parcelle. La biomasse totale a été comparée entre les blocs et les années à
l’aide d’une analyse de variance en tenant compte des blocs comme facteurs aléatoires. Les
déplacements ont été estimés par la distance entre deux localisations consécutives séparées
de moins de 3 heures. Les positions étaient obtenues par triangulation à l’aide de stations au
sol localisées avec un GPS. Le budget d’activité a été quantifié à partir d’un récepteur
automatisé. Nous avons mesuré l’influence de la densité sur les déplacements, la proportion
du temps passé en activité, le nombre de périodes d’activité, et la durée des périodes
d’activité et d’inactivité entre les années, les semaines et les périodes du jour pour tous les
adultes et les juvéniles à l’aide d’analyses de variance en tenant compte des blocs et des
années comme facteurs aléatoires.
Dans le troisième chapitre, nous avons mesuré l’influence de la densité de population sur
l’utilisation de l’espace dans les enclos. Nous avons utilisé des parcelles pour estimer la
quantité de biomasse et la quantité de couvert latéral et vertical disponible. La végétation
dans les unités expérimentales a été évaluée selon un plan d’échantillonnage par degrés.
Afin de caractériser uniformément l’enclos, une grille a été générée dans ArcView GIS
avec des carreaux de 2 ha. Chaque carreau comprenait 5 parcelles qui ont été disposées
aléatoirement à l’aide de l’extension « Generate-randomly distributed points » de ArcView
GIS. Les points générés par cette méthode ont été transférés dans un GPS et retrouvés sur le
site. À chaque parcelle, la biomasse herbacée a été estimée par espèce en évaluant le
pourcentage de recouvrement de la même façon que dans le Chapitre 2. La fermeture du
couvert arborescent (arbres >4 m) a été déterminée par la projection verticale des cimes au-
dessus de 20 points équidistants de 3 m distribués sur 4 axes (est, sud-est, sud-ouest et
ouest) couvrant un demi-cercle et partant du centre de la parcelle. Le couvert latéral a été
mesuré à l’aide d’une planche à profil (2.5×0.3 m divisée en sections de 50 cm) située à 15
m du centre de la parcelle dans deux directions différentes (Nudds 1977).
16
Pour tous les enclos, nous avons ensuite caractérisé la quantité de biomasse et de couvert
vertical et latéral disponibles à l’aide d’une méthode géostatistique. Le krigeage est une
méthode qui permet d’interpoler les valeurs entre deux points en mesurant la relation
spatiale qui existe entre cette variable et l’espace (Cressie 1993). L’erreur des localisations
par triangulation des cerfs étant grande (107 m), nous avons placé une pastille d’erreur de
100 m autour de chaque localisation. Puisque les pastilles étaient grandes par rapport à la
taille des enclos, nous ne pouvions pas les considérer indépendantes les unes des autres.
Nous avons donc placé une grille avec des carreaux de 150×150 m sur chaque enclos et tiré
aléatoirement un point dans chaque carreau. À chaque point, le nombre de pastilles qui se
superposaient à chaque période du jour et la quantité de biomasse et de couvert latéral et
vertical ont été obtenus à partir des cartes préalablement établies. Ensuite, pour mesurer la
relation entre l’utilisation d’un certain point par les cerfs et des variables d’habitat, nous
avons simplement utilisé une analyse de régression en tenant compte des blocs et des
années comme des facteurs aléatoires.
17
Chapitre 1. Quantification and accuracy of activity data
measured with VHF and GPS telemetry
Marie-Lou Coulombe,
Ariane Massé et Steeve D. Côté
Ce chapitre a été accepté dans la revue « The Wildlife Society Bulletin » en avril 2005 et il
est maintenant sous presse. « The Wildlife Society » nous a accordé la permission de le
reproduire dans ce mémoire.
18
Résumé
Afin de valider l’utilisation de la télémétrie pour quantifier l’activité des ongulés, nous
avons équipé 8 cerfs (Odocoileus virginianus) en captivité avec des colliers pour
déterminer la précision des capteurs d’activité des colliers VHF et GPS ainsi que la
performance de colliers VHF pour mesurer les budgets d’activité. Chez 16 cerfs en milieu
naturel munis de colliers GPS, nous avons évalué si les capteurs pouvaient mesurer des
patrons d’activité journaliers. Les données VHF correspondaient aux observations dans
74% des cas et en considérant 3 échantillons successifs, nous avons augmenté la précision à
84% et déterminé avec succès 87% des périodes d’activité. Les valeurs obtenues à partir du
capteur vertical des colliers GPS étaient plus précises (92%) que les données obtenues à
partir du capteur horizontal (83%) et décrivaient correctement des pics d’activité à l’aube et
au crépuscule. Nous concluons donc que les colliers GPS et VHF, en utilisant 3
échantillons successifs, peuvent être utilisés pour quantifier l’activité des grands herbivores.
19
Abstract
Quantifying activity budgets and determining the accuracy of behavioral data obtained by
telemetry is essential to understand the behavior of animals that are difficult to observe. We
fitted 8 captive white-tailed deer (Odocoileus virginianus) with VHF or GPS collars to
determine the accuracy of VHF variable-pulse sensors and GPS dual-axis sensors and
validate the performance of VHF telemetry for the measurement of activity budgets. We
also evaluated whether instantaneous activity counts could measure daily activity patterns
of 16 free-ranging deer fitted with GPS collars on Anticosti Island (Québec, Canada).
Comparison of VHF telemetry data and visual observations of active (feeding, moving and
standing) and inactive (resting) deer behaviors were correct in 74% of the scans. By using
the activity values of 3 successive VHF scans, we increased accuracy to 84% of the
observed behaviors and detected 87% of observed activity bouts. The accuracy of GPS
activity data varied with orientation of the sensor: activity counts of vertical sensors (92%
agreement) were better able to predict observed behaviors than activity counts from
horizontal sensors (83% agreement). GPS activity sensors detected peaks of activity after
dawn and at dusk in free-ranging deer. We conclude that dual-axis GPS motion sensors can
be used to reliably record activity data and successive scans from VHF sensors can
precisely detect activity bouts in large herbivores.
20
Introduction
Conventional Very High Frequency (VHF) telemetry and animal tracking with Global
Positioning System (GPS) collars allow animal ecologists to quantify the activity of
wildlife and have been used to measure time budgets of species that are difficult to observe.
Initially, signal strength (Singer et al. 1981, Cederlund et al. 1989, Hölzenbein and
Schwede 1989) and linear distance between relocation points determined by radiotelemetry
(Sparrowe and Springer 1970, Kammermeyer and Marchinton 1977), and later by GPS
collars (Merrill and Mech 2003), were used to measure activity budgets of many species.
The interpretation of signal evenness, however, has been found to be subjective and
influenced by particular animals and the environment between the transmitter and the
antenna (Garshelis et al. 1982, Gillingham and Bunnell 1985, Rouys et al. 2001). The use
of relocation distances has been criticized because estimates of radiolocations have large
errors (for VHF telemetry) and distance traveled may misclassify stationary, but active,
animals as inactive (Craighead et al. 1973, White and Garrott 1990, Rouys et al. 2001).
VHF and GPS activity sensors made it possible for biologists to quantify remotely
continuous or instantaneous activity data.
Three types of VHF activity sensors have been used. Reset sensors are equipped with a
timer and a mercury switch that initiate a pulse rate change when the switch is not triggered
within a specified time lapse. Tip-switch sensors transmit different pulse rates depending
on orientation of the sensor. The number of pulse rate changes in a specific time period
may be used to index activity. Reset sensors and tip-switch sensors were found to be easily
triggered by head and comfort movements made by resting animals (Garshelis et al. 1982,
Gillingham and Bunnell 1985). Nonetheless, a strong correlation (r = 0.9) was found
between distance moved by black bears (Ursus americanus) and activity measured by reset
sensors (Garshelis et al. 1982). The proportion of time spent active measured with
tip-switch sensors could be estimated from telemetry data with 90% accuracy in a study of
black-tailed deer (Odocoileus hemionus columbianus; Gillingham and Bunnell 1985). To
refine the use of tip-switch sensors, Beier and McCullough (1988) increased sampling
interval from 1 to 5.25 minutes and this led to the correct classification of 98% of
21
individual samples. In a validation study of tip-switch sensors used to measure time budget
of Dall’s sheep (Ovis dalli), examination of scan pattern changes rather than fixed time
samples also increased accuracy of activity detection (Hansen et al. 1992). Variable-pulse
sensors were developed because it was thought that by adding extra pulses to every switch
movement, specific behaviors such as moving versus feeding could be identified from
different pulse patterns. As for tip-switch sensors, variable-pulse sensors are triggered by
individual movements and changes in pulse rates are not dependent on a specific time-delay
as for reset sensors. The first versions of variable-pulse sensors assessed movement
periodically (every 0.25 sec), but an increase in animal movement did not necessarily result
in higher pulse rates because instantaneous samples of movements missed concentrations of
rapid pulses (Gillingham and Bunnell 1985). Rather than sampling movement
instantaneously, later versions of variable-pulse sensors, such as the ones we used,
integrated the amount of movements by adding pulses to a base pulse rate. Depending on
scan duration (1–5 min), Relyea et al. (1994) found that 74 to 81% of the scans
discriminated resting deer from non-resting deer, but that variable-pulse sensors could not
discriminate amongst different active or inactive behaviors. Errors with variable-pulse
sensors are still associated with head and comfort movements in resting periods or with
sensors that fail to detect movements while animals are active, but keep their head still for
extended periods.
Methods for effectively gathering continuous VHF sensor data on numerous animals have
also been developed. In the beginning, researchers listened directly to signal changes
(Garshelis et al. 1982). Other systems registered signal variations on strip charts
(Gillingham and Bunnell 1985, Beier and McCullough 1988, Hansen et al. 1992).
Gathering data was still time consuming because every signal change had to be manually
recorded. In the 1990s, new automated systems that recorded time and pulse rates
electronically in an immediately usable form were developed (Relyea et al. 1994). We used
a version of this automated datalogging system to gather and analyze data on deer activity
budgets.
22
In the past decade, GPS collars have been equipped with different motion sensors: tilt-
switch activity sensors that tally head-down occurrence (Rumble et al. 2001) and activity
counters composed of dual-axis motion sensors sensitive to vertical and horizontal head and
neck movements (Moen et al. 1996, Turner et al. 2000, Adrados et al. 2003). The first
validation tests of activity counters were conducted on GPS 1000 collars (Lotek
Engineering, Newmarket, Ontario, Canada; Moen et al. 1996, Turner et al. 2000, Adrados
et al. 2003). Both the vertical and the horizontal sensors of GPS 1000 collars consist of a
cylinder that contains a small sphere. An integrated datalogger registers the number of
times that the sphere hits the extremities of the cylinders in a specific time interval. The
activity counts of GPS 1000 collars are combined values of the vertical and the horizontal
sensors. When the GPS fix interval is longer than the activity observation window, the
activity value recorded is averaged over the GPS fix interval. For example, if the GPS fixes
are taken every 2 hours and the activity counts are recorded at 5-minute intervals, then the
reported activity counts would be the average of 24 observations for every 2-hour period.
Moen et al. (1996), Turner et al. (2000) and Adrados et al. (2003) validated these activity
counters with captive animals and were able to classify correctly 91%, 95% and 69% of the
active samples and 75%, 91% and 89% of the inactive samples, respectively. Moen et al.
(1996) also validated activity sensors on free-ranging moose (Alces alces) and found that
the amount of time that moose were active as estimated from activity sensors was
comparable to daily time active reported in other studies of moose for the same region.
Moen et al. (1996) suggested that activity counts should be recorded during a time interval
≤10 minutes and that they should not be averaged over the entire GPS fix interval.
Recently, new models of GPS collars have allowed us to record 2 activity counts, one for
the vertical and one for the horizontal sensor. Furthermore, each value is the actual activity
count recorded during the observation window directly preceding the reported GPS
position, and not an average like for the GPS 1000 collars. These new sensors, however,
have never been validated in the field but may provide a considerable improvement over
older models for fine-scale analysis of foraging behavior and habitat use because they allow
the correlation of actual activity counts to reported GPS positions and corresponding
habitat. A comparison of activity counts measured on captive deer fitted with GPS collars
23
and the corresponding observed behavior would allow the verification of sensor accuracy.
Another approach to validate activity sensor data would be to look at circadian activity
patterns of free-ranging animals fitted with GPS collars. Circadian activity peaks
synchronized with dawn and dusk are widely observed in white-tailed deer (Odocoileus
virginianus; Montgomery 1963, Kammermeyer and Marchinton 1977, Beier and
McCullough 1990, Rouleau et al. 2002). If the sensors can track activity peaks as daylight
changes throughout seasons, then this would indicate that the sensors are reliable.
Our main objective was to validate the use of motion sensors to estimate the activity of
free-ranging large herbivores. We wanted to 1) determine the accuracy of VHF variable-
pulse sensors and GPS dual-axis sensors by comparing sensor data with observed behavior
2) determine the performance of VHF variable-pulse sensors to estimate activity budgets
and 3) verify the ability of instantaneous activity counts generated by GPS motion sensors
to estimate daily activity patterns of free-ranging deer.
Study area
We conducted this study on Anticosti, a 7,943-km2 island located in the Gulf of St.
Lawrence, Québec, Canada (49° 28’ N, 63° 00’ W). The climate was maritime and
characterized by cool summers and by mild and long winters. Mean daily temperature is
15°C in July and –14°C in January (Environment Canada 1993). The boreal forest that
prevailed on Anticosti was dominated by balsam fir (Abies balsamea), white spruce (Picea
glauca), and black spruce (Picea mariana) (Rowe 1972). White-tailed deer were introduced
on Anticosti in 1896 and in the absence of predators, their numbers increased rapidly to
>100,000. Currently, deer density is about 20/km2 and severe impacts of browsing on the
vegetation have occurred across the whole island (Potvin et al. 2003).
24
Methods
Calibration of VHF and GPS motion sensors on captive white-tailed deer
fawns
Deer captures
We captured 8 white-tailed deer fawns between 21 November and 12 December 2003 with
dartguns, Stephenson box-traps or cannon nets. Deer were released in 2 50×80 m semi-
natural enclosures that contained cover, forage and daily supplemental food placed in a
feeder. Low tree branches and shrubs were absent in the enclosures. The Animal Care and
Use Committee of Université Laval, Québec, Canada approved all capture methods
(reference number 2003-014).
VHF collars
LMRT-3 VHF collars equipped with STO-2a variable-pulse sensors (Lotek Engineering,
Newmarket, Ontario, Canada) were fitted on 4 deer. Each collar had a board fitted parallel
to the ground on the bottom of the transmitter case to which a tilt-switch oriented
perpendicular to the spine of the animal (horizontal sensor) was attached. Pulses were
automatically added each time the switch was triggered. Transmitter signals from the
collars were received and recorded in a SRX-400 Version W9 receiver-datalogger (Lotek
Engineering, Newmarket, Ontario, Canada) connected to a multidirectional antenna and a
12 V battery to ensure a constant electrical input.
The receiver was programmed to measure duration between 2 pulses for 65 consecutive
pulses, record mean pulse rate and then automatically switch to scan another transmitter.
The time needed to record 65 pulses was thus dependent on pulse rate. As the SRX receiver
scanned one transmitter at a time and because 4 individuals were followed each day, a
measure of pulse rate for each deer was obtained approximately every 4 minutes. At the end
of the day, data were downloaded on a laptop computer with Winhost software 1.0.0.1
(Lotek Engineering, Newmarket, Ontario, Canada).
25
GPS collars
The GPS 2200R collars (Lotek Engineering, Newmarket, Ontario, Canada) that we used
were equipped with dual-axis motion sensors (vertical and horizontal) that recorded the
number of times a switch was triggered during the 4 minutes immediately preceding a GPS
fix. Both sensors are fixed on a board parallel to the ground in the transmitter case. The
vertical sensor is oriented parallel to the spine of the animal and the horizontal sensor is
oriented perpendicular to the spine of the animal. We obtained GPS locations every 5
minutes. The maximum number of events that could be recorded during each 4-minute
interval was 255.
Behavioral observations
We waited 48 hours after deer had been introduced into the enclosures before beginning
visual observations so that deer could habituate to wearing a VHF or a GPS collar. Deer
were observed during mornings and afternoons from a 4-m high observation tower between
21 November and 23 December 2003. When needed, we used 8×42 binoculars or 20–
25×60 spotting scopes. The time and type of each behavior were recorded on tape
recorders. Watches were set to match the time on the receiver and the GPS collars. We
recorded 4 behaviors during observations: feeding, moving, standing, or resting.
Validation of GPS motion sensors on free-ranging deer
We monitored 16 free-ranging white-tailed deer does equipped with GPS 2200R collars
between July and November 2001 (n = 8) and 2002 (n = 8). Does were captured in late June
or early July in peat bogs with a net-gun fired from a helicopter. Handling time was less
than 5 minutes and deer were released at the capture site. We set GPS fix interval to 2 hours
and recorded activity counts during the 4 minutes immediately preceding every GPS
location. We predicted that if the motion sensors record activity accurately, then we could
track activity peaks at dawn and dusk as daylight changes throughout summer and fall.
26
Data analysis
Calibration of VHF and GPS motion sensors
All VHF mean pulse rates (BPM) were divided by the transmitter’s respective base pulse
rate (from 59–64 pulses per minute). A sensor that had not moved would thus give a
relative BPM of 1. Due to receiver or collar variations, pulse rate could be smaller but close
to base pulse rate (i.e. >0.95 of the base pulse rate). We rounded these data to 1 because we
considered them as equivalent to base pulse rate. When movements occurred, scans should
then give a relative BPM higher than 1 (up to about 2.5). The dual-axis motion sensors of
GPS 2200R collars provided 2 distinct activity counts, one for vertical and the other for
lateral head and neck movements. Number of events ranged from 0–255.
To test whether we could relate ranges of relative BPM and GPS activity counts to different
behaviors, we selected the periods when deer were observed performing only 1 of the 4
behaviors identified for the whole sampling period. We then compared means and ranges of
relative BPM and GPS activity counts (Figure 1-1). We determined the percentage of
correctly classified scans for VHF and GPS sensors when deer were active (i.e. feeding,
moving or standing) or inactive (resting) for the entire sampling period using different
mean relative BPM for VHF collars and different activity counts for GPS collars as
separation thresholds. Samples with relative BPM and activity counts under the threshold
value were considered inactive and samples over the threshold value were considered
active. The value of the best threshold was determined by plotting the percentage of
correctly classified samples when deer were active and inactive against all possible relative
BPM or activity count thresholds (Figure 1-2).
Calculation of activity bouts with VHF collars
To compare the results obtained with VHF collars to observed activity budgets and
calculate the accuracy of scans we combined the information from 1 scan with that of the
following 2 successive scans, i.e. a time interval of 10–15 minutes. We depicted activity
bouts using the following criterions: an inactive bout began when at least 3 consecutive
scans considered as inactive were recorded, and conversely an animal was classified as
27
Figure 1–1. Box plot representations of relative mean pulse rates (BPM) for VHF variable-
pulse sensors (a) and of activity counts for vertical (b), and horizontal (c) GPS activity
sensors for the 4 different behaviors observed. Collars were fitted to white-tailed deer
fawns that were simultaneously observed on Anticosti Island, Québec during 21 November-
23 December 2003. Solid middle bars correspond to the median for each behavior observed
and dashed bars symbolize the mean value. Error bars represent 90% quantiles and dots are
outlier values (<10% of data). Numbers in parentheses represent the number of telemetry
samples for each behavioral category. Distribution of VHF relative BPM and GPS activity
counts show clearly how values overlap between active (feeding, moving and standing) and
inactive (resting) behaviors.
Behaviors
Feeding Moving Standing Resting
Ho
rizo
nta
l acti
vit
y c
ou
nts
(G
PS
co
llars
)
0
25
50
75
100
125
150
175
200
225
250
Mean
rela
tiv
e B
PM
(V
HF
co
llars
)1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.4
2.6(128) (44) (63) (662)
Vert
ical
acti
vit
y c
ou
nts
(G
PS
co
llars
)
0
25
50
75
100
125
150
175
200
225
250
(234) (10) (18) (472)
a)
b)
c) (234) (10) (18) (472)
Figure 1. Box plot representations of relative mean pulse rates (BPM) for VHF variable-
pulse sensors (a) and of activity counts for vertical (b), and horizontal (c) GPS activity
sensors for the 4 different behaviors observed. Collars were fitted to white-tailed deer
fawns that were simultaneously observed on Anticosti Island, Québec. Solid middle bars
correspond to the median for each behavior observed and dashed bars symbolize the mean
value. Error bars represent 90% quantiles and dots are outlier values (<10% of data).
Numbers in parentheses represent the number of telemetry samples for each behavioral
category. Distribution of VHF relative BPM and GPS activity counts show clearly how
values overlap between active (feeding and moving) and inactive (standing and lying)
behaviors.
28
Figure 1–2. Determination of a criterion to separate relative mean pulse rates (BPM) for
VHF collars equipped with variable-pulse motion sensors (a), and activity counts for GPS
collars equipped with double-axis motion sensors (b) into active and inactive behaviors of
white-tailed deer on Anticosti Island, Québec. The percentages of correctly classified
samples in relation to all possible separation thresholds of relative BPM or activity counts
are illustrated. The best separation threshold is found where the 2 curves intersect or where
the classification error of active and inactive behaviors is the smallest.
Mean relative BPM threshold
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
Pro
po
rtio
n o
f co
rrectl
y c
lass
ifie
d b
eh
av
iors
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Inactive
Active
Activity count threshold
0 25 50 75 100 125 150 175 200 225 250
Pro
po
rtio
n o
f co
rrectl
y c
lass
ifie
d b
eh
av
iors
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
a)
b)
VHF collars
GPS collarsObserved accuracy
Selected separation threshold
29
active when at least 3 consecutive scans considered as active were recorded (Figure 1-3).
To compare our method of estimating active and inactive bout durations from telemetry-
derived data to observed data, we simply correlated the observed and estimated proportions
of time active and quantified the number of correctly classified activity bouts for each day
of observation. We also calculated the mean daily duration of active and inactive bouts and
compared the estimated and observed results with Student's t-tests. When needed, we either
square-root or log transformed bout length to achieve data normality (Zar 1999).
Validation of activity counts of GPS collars on free-ranging deer
GPS collars recorded 12 4-minute activity counts per day (every 2 hr) for each free-ranging
deer for both vertical and horizontal motion sensors. Data were classified into 4 periods of
the day and the numbers of 4-minute activity counts varied for each period: dawn (from
half an hour before sunrise to an hour after, n = 1), dusk (from an hour before sunset to half
an hour after, n = 1), day (n = 4–7), and night (n = 3–6). Mean activity counts were thus
used for day and night. For each sensor, we used repeated-measures ANOVAs (proc GLM;
SAS 1989) to evaluate the influence of the period of the day on activity counts from July to
November and we compared the results with the LSMEANS statements of SAS (1989). We
ensured that the residuals were normally distributed (Shapiro-Wilk test) and were
homogeneous by visual examination of the plots. Horizontal activity counts were square-
root transformed to normalize the residuals (Zar 1999). Unless specified, all results are
presented as x ± 1 SE.
Results
We collected 102 hours of observation on 4 captive deer wearing 4 different VHF collars
and 105 hours on 4 other captive deer wearing 3 different GPS collars. We observed each
deer with a VHF collar for a total of 6 days and each deer with a GPS collar for 3 to 6 days
(approx 4 hr of observation per day). Deer spent 30 ± 2% of their time feeding, 13 ± 1%
standing, 13 ± 2% moving, and 43 ± 4% resting (n = 42 deer-days) during the daily
observations. We obtained simultaneous data of relative BPM and behavioral observations
30
Figure 1–3. Observed behavior and relative mean pulse rate (BPM) obtained during one
day for a white-tailed deer fawn fitted with a VHF collar equipped with a variable-pulse
sensor on Anticosti Island, Québec. An inactive bout (white rectangles) began when at least
3 inactive scans (relative BPM of 1) were observed. To return to an active bout (black
rectangles), at least 3 active scans (relative BPM higher than 1) had to be observed. Note
the presence of single inactive scans surrounded by active scans (a), and single active scans
(b) surrounded by inactive scans that did not change the calculation of the duration of the
bout.
Time
08:00 10:00 12:00 14:00 16:00
Mean
rela
tiv
e B
PM
0.5
1.0
1.5
2.0
2.5
3.0
3.5
a b
Estimated activity bouts
Observed activity bouts
31
for 1,357 relative BPM samples from VHF collars and 1,126 4-minute samples from GPS
collars.
Determination of specific behaviors
For VHF collars, the ranges of 90% quantiles for relative BPM of resting (1–1.02), standing
(1–1.16), feeding (1–2.02) and moving (1–1.69) overlapped considerably (Figure 1-1a)
because they all included 1. Ranges of active behaviors, however, were much larger than
for inactive behaviors (Figure 1-1a). For GPS collars, 90% quantiles for vertical activity
counts ranged from 0 to 4 for resting, from 2 to 148 for standing, from 4 to 166 for feeding
and from 2 to 237 for moving and thus also overlapped considerably (Figure 1-1b). Overlap
ranges were even greater for horizontal activity counts (Figure 1-1c). Even if relative BPM
and activity counts of standing deer resembled those of resting deer, standing deer were
considered active because they were observed during active bouts and were standing only
for short periods (VHF: x = 33 ± 1 seconds; GPS: x = 19 ± 2 seconds).
Calibration of VHF and GPS motion sensors
Deer were either 100% active (n = 662) or 100% inactive (n = 662) for 1,324 scans of the
VHF collars. We identified the best separation threshold of relative BPM as 1 (i.e. the
smallest classification error, Figure 1-2a). The best separation criterion was found where
the 2 curves intersected or where the classification error of active and inactive behaviors
was the smallest (Relyea et al. 1994). For example, if we had used a separation threshold of
1.3 mean relative BPM for VHF collars, we would have correctly classified 97% of the
inactive samples (Figure 1-2a, top curve), but only 20% of the active samples (Figure 1-2a,
bottom curve). We correctly discriminated 87% and 61% of inactive and active samples,
respectively, for a total of 74% of correctly classified samples (Table 1-1). Most errors
occurred when deer were observed feeding or moving, but VHF telemetry signals indicated
that they were inactive. Using all 4-minute intervals when deer were completely inactive
(n = 472) or active (n = 614) (Figure 1-2b), we found that for GPS collars the best cut-off
activity count to separate active from inactive samples was a value of 10 for both vertical
and horizontal sensors. This separation criterion allowed us to correctly classify 92% and
32
Table 1–1. Individual and combined relative mean pulse rates (BPM) from variable-pulse
activity sensors that correctly classifieda the observed behaviors of captive white-tailed deer
fawns fitted with VHF collars during 21 November-23 December on Anticosti Island,
Québec.
Percentage of relative mean BPM correctly classified
Individual scans Combination of 3 scansc
Deer Id Nb Inactive Active Total Inactive Active Total
30 297 90 61 74 93 86 89
35 341 88 50 70 95 72 84
39 310 84 99 92 82 99 91
40 376 86 32 60 95 53 75
Total 1,324 87 61 74 92 77 84
a The threshold value for relative mean BPM was determined graphically (Figure 1-2a).
b N refers to the number of scans when deer were observed either 100% active or 100%
inactive.
c To determine whether each individual scan should be considered active or not, we used
the information of the following 2 scans. Samples were considered active until at least 3
successive scans changed to inactive relative mean BPM values, and vice-versa.
33
83% of the samples for the vertical and the horizontal activity sensor, respectively
(Table 1-2).
Calculation of activity bouts
Using the values of 3 successive VHF scans, we correctly estimated 92% (n = 662) of the
scans related to inactive behaviors and 77% (n = 662) of the scans related to active
behaviors and thus correctly classified 84% of the scans (Table 1-1). A strong correlation
existed between the proportion of time active observed and estimated by VHF activity
sensors (r = 0.81; n = 24 deer-days; P ≤ 0.001; Figure 1-4). Mean time observed active per
day was slightly higher ( x = 56 ± 5%, n = 24 deer-days), but not significantly different,
from estimated time spent active ( x = 50 ± 6%, n = 24 deer-days; t23 = 1.86; P = 0.08). In
addition, we correctly estimated 89% of inactive bouts and 82% of active bouts. We found
no difference between the mean duration of active bouts estimated by variable-pulse
sensors ( x = 62 ± 9 minutes, n = 23 deer-days) and those observed ( x = 64 ± 8 minutes,
n = 23 deer-days; t22 = -0.45; P = 0.65). The duration of inactive bouts estimated
( x = 75 ± 9 minutes, n = 23 deer-days) and observed ( x = 71 ± 8 minutes, n = 23
deer-days; t22 = -0.90; P = 0.44) were also comparable. We also did the comparisons for
each deer separately to account for individual deer-collar effects, and found no differences.
Activity of free-ranging deer
During the summer and fall of 2001 and 2002, we found an influence of the period of the
day (vertical: F3,54 = 7.69, P ≤ 0.001; horizontal: F3,54 = 6.89, P ≤ 0.001) and month
(vertical: F4,216 = 46.93, P ≤ 0.001; horizontal: F4,216 = 4.67, P ≤ 0.001) on mean activity
counts of free-ranging does. Mean activity counts decreased from July to November
(Figure 1-5). The vertical and horizontal activity sensors revealed 2 daily activity peaks that
were synchronized just after dawn and during dusk from July to November (Figure 1-5).
The activity peaks closely tracked the changes of daytime duration. The interaction
between period of the day and month was highly significant for horizontal sensors
(F12,216 = 6.80, P ≤ 0.001) but, although the direction of the results was similar, the
interaction was not significant for vertical sensors (F12,216 = 1.43, P = 0.15).
34
Table 1–2. Activity counts recorded during 4-minute intervals that correspondeda to
observed behaviors from captive white-tailed deer fawns fitted with GPS collars on
Anticosti Island, Québec.
Percentage of 4-minute intervals correctly classified
Vertical sensor Horizontal sensor
Deer Id Nb Inactive Active Inactive Active
29 451 90 96 87 63
32 219 95 96 64 100
33 296 95 79 94 97
41 120 88 90 59 99
Total 1,086 92 83
a The threshold value for activity counts was determined graphically for each sensor
(Figure 1-2b).
b N refers to the number of periods observed when each deer was either 100% active or
100% inactive.
35
Figure 1–4. Relationship between observed and estimated proportion of daily active time
obtained with variable-pulse activity sensors of VHF collars fitted to 4 white-tailed deer
fawns each observed for 6 days ( x = 4 hr of observation per day) on Anticosti Island,
Québec.
Observed proportion of daily time active
0.0 0.2 0.4 0.6 0.8 1.0
Est
imate
d p
rop
ort
ion
of
dail
y t
ime a
cti
ve
0.0
0.2
0.4
0.6
0.8
1.0
001.0,81.0 Pr
36
Figure 1–5. Mean activity counts (± 1 SE) recorded by horizontal and vertical sensors of
GPS collars fitted on free-ranging deer on Anticosti Island in summer and autumn 2001
(n = 8 deer) and 2002 (n = 8 deer). Grey sections indicate dawn and dusk, white sections
daytime and black sections nighttime. Identical letters identify mean activity counts that did
not differ statistically between periods of the day: the top row show the results for the
horizontal sensor and the bottom row for the vertical sensor. Analyses were performed
separately every month for each sensor.
Time
July
0 2 4 6 8 10 12 14 16 18 20 22
Act
ivit
y co
un
ts
0
25
50
75
100
125
150
175
200
225
September
Time
0 2 4 6 8 10 12 14 16 18 20 22
Act
ivit
y co
un
ts
0
25
50
75
100
125
150
175
200
225
October
0 2 4 6 8 10 12 14 16 18 20 22
Act
ivit
y co
un
ts
0
25
50
75
100
125
150
175
200
225
November
Time
0 2 4 6 8 10 12 14 16 18 20 22
Act
ivit
y co
un
ts0
25
50
75
100
125
150
175
200
225
0 2 4 6 8 10 12 14 16 18 20 22
Act
ivit
y co
un
ts
0
25
50
75
100
125
150
175
200
225
August
a ab b
a aab b
a aa b
a ab b
a abab b
a ab b
a acbc b
a ab c
a abb
a aa b
Time
Time
Vertical sensor
Horizontal sensor
37
Discussion
Our study revealed that activity sensors of both VHF and GPS collars can provide reliable
information on activity budgets and patterns of large herbivores. Data from VHF and GPS
activity sensors allowed us to determine accurately 74% and 88% of observed behaviors,
respectively, by using a unique separation criterion for each type of collar. In addition,
activity sensors of VHF collars accurately detected 87% of activity bouts.
Determination of specific behaviors
Particular behaviors could not be identified from mean relative BPM for VHF collars or
activity counts for GPS collars. Validation studies have never been successful in assigning
distinct patterns or ranges of activity sensor data to more specific behaviors than active and
inactive behaviors (Gillingham and Bunnell 1985, Beier and McCullough 1988, Hansen et
al. 1992, Relyea et al. 1994). A deer may engage in several different activities during a 1-
minute interval. Feeding and walking deer often move their heads similarly and thus trigger
switches equally. Standing deer do not trigger the switch often and pulse patterns resemble
those of resting deer. This resulted in strong overlap of relative BPM and activity counts
between behaviors, and observations were thus classified by the proportion of time that
deer were observed active or inactive. Deer feeding, moving or standing were considered
active and resting deer were considered inactive. Distinguishing between feeding and
walking is not essential in most studies interested in foraging behavior because deer spend
most (90–95%) of their active time foraging during the plant-growing season (Beier and
McCullough 1988, Gillingham et al. 1997).
VHF collars
We developed a criterion that correctly distinguished 74% of relative BPM samples. The
accuracy of our collars was comparable to other validation studies of variable-pulse sensors
(73–81%; Gillingham and Bunnell 1985, Relyea et al. 1994), but lower than for tip-switch
activity sensors (>90%; Gillingham and Bunnell 1985, Beier and McCullough 1988). Tip-
switch sensors have possibly given more accurate results than variable-pulse sensors
38
because switches are oriented parallel to the spine of the animal (vertical sensors) and are
thus more sensitive to foraging and walking movements.
Inactive samples were correctly classified in 87% of the cases and accuracy was high and
constant among deer (84–90%). Sensors transmitted an active signal only 13% of the time
when deer were resting compared to 27% for mule deer (Odocoileus hemionus; Relyea et
al. 1994) and 14% for black-tailed deer (Gillingham and Bunnell 1985). A large proportion
of active behaviors (39%), however, were misclassified into inactive behaviors. By
analyzing the information of 3 successive scans, we reduced this percentage to 23% and
correctly classified 87% of active bouts. Misclassification of active behaviors was also
rather high (10–63% of inactive pulses depending on the type of active behavior) for
variable-pulse sensors fitted on captive black-tailed deer (Gillingham and Bunnell 1985).
Relyea et al. (1994) found a higher and more constant accuracy (74–82%) for active
samples of free-ranging deer that had greater movement rates than captive deer. Differences
in study results might be due to the observation of captive deer that move slowly
(Gillingham and Bunnell 1985, Beier and McCullough 1988; this study). Many active
sequences did not induce a change in pulse rates because deer walked slowly with their
head in a horizontal position or kept their head down to feed for several minutes, with
imperceptible sensor movement.
Inaccuracy also varied between individuals (1–68%). Variability could be attributed to
individual differences in movement such as more head tipping while walking. The fit of the
collar around the neck could also influence how easily the switch is triggered by
movements. For example, a tighter fit could help to trigger the sensor more easily and
provide higher rate values. On the contrary, a looser fit could allow the collar to slide on the
neck rather than tilt the mercury switch during certain behaviors. We consider this
possibility as a minor problem in our study however, because all collars were adjusted the
same way. Head movements and tightening of the collar can vary depending on the age-sex
class of an individual and on the season. Researchers should, therefore, be careful in
analyzing activity data collected through long periods of time over many animals.
39
Taking into account 3 successive scans in our quantification of the length of activity bouts
allowed us to accurately determine 87% of active and inactive bouts. Active periods were
characterized by fast pulses interspersed by slow pulses. Using 3 successive scans
considerably decreased the effect of misclassified individual scans. Inactive bouts were
depicted by lasting and constant signals equal or very similar to base pulse rates and could
also be differentiated from active bouts, even if some individual scans were misclassified.
We conclude that variable-pulse sensors are a reliable method to quantify activity budgets
for behavioral studies when the information of successive scans is used, but individual
samples are less precise.
GPS collars The horizontal sensor of GPS collars was more sensitive than the vertical one and detected
head movements even when captive deer were resting or standing. These results are
consistent with the study of Relyea et al. (1994) who also found that horizontal sensors of
VHF collars were activated when deer were resting or standing. Due to its greater
sensitivity, the horizontal sensor lost some information because it reached the maximum
activity count (255) more often than the vertical sensor. Over-sensitivity to slight
movements is not necessarily recommended for activity sensors because it then becomes
more difficult to discriminate active from inactive behaviors and greater variations of
activity counts are recorded between individuals. Better classifications of active and
inactive samples with VHF motion sensitive tip-switches were also found using vertical
sensors (Beier and McCullough 1988, Hansen et al. 1992).
Our method provided a valid classification of active and inactive samples. Using a cut-off
value of 10 for both sensors allowed us to classify correctly 92% and 83% of the samples
for vertical and horizontal sensor, respectively. Other studies (Moen et al. 1996, Turner et
al. 2000) on GPS activity sensors used higher cut-off values to discriminate active from
inactive behaviors, but these sensors recorded a combined value of vertical and horizontal
sensor instead of two separated values. Adrados et al. (2003) developed an individually-
based method to discriminate active from inactive behaviors using the mean daily activity
count of each collar as a reference. This method is useful because it avoids bias due to
40
potential variations in collar tightening among animals and seasons. The horizontal sensor
of our GPS collars was more sensitive and variable than the vertical sensor and thus may
need to be calibrated for each individual or used with an individually-based method. An
individually-based method, however, is not necessary when using a vertical sensor that
records actual activity counts. The vertical sensor of our collars was very accurate and
using a cut-off value of 0 instead of 10, would still have correctly identified 86% and 97%
of active and inactive samples, respectively. The calibration of GPS motion sensors on
captive deer allowed us to validate the use of these sensors to quantify activity of free-
ranging animals. We also compared the magnitude and distribution of activity counts from
captive deer to those from free-ranging deer and found that they were similar. Therefore,
we are confident that the magnitude of the activity of free-ranging animals can be reliably
captured with the sensors.
GPS collars are widely used to study habitat use of large herbivores across seasons or years.
GPS positions are usually taken at intervals varying from 1 to 4 hours. Moen et al. (1996)
suggested that activity counts should be recorded during a time interval ≤10 minutes and
that activity counts should not be averaged over the whole GPS fix interval. The GPS
2200R collars that we used recorded actual activity counts over 4-minute intervals
immediately preceding every GPS fix and thus allowed the correlation of actual activity
counts to specific periods of time. The use of GPS activity sensors on free-ranging deer on
Anticosti Island allowed us to detect 2 activity peaks that were synchronized just after dawn
and during dusk from July to November, as well as a daytime decrease in activity for July
and August. Mean activity counts during daytime possibly increased from September to
November because of the approaching rut. The horizontal activity counts are higher and
more variable than the vertical activity counts and this may explain why the interaction
between period of the day and month was significant for the horizontal sensors, but not for
the vertical sensors. Circadian activity peaks are widely observed in white-tailed deer
(Montgomery 1963, Ozoga and Verme 1970, Kammermeyer and Marchinton 1977,
Rouleau et al. 2002). Similarly to our study, Beier and McCullough (1990) observed
morning activity peaks after sunrise and evening activity peaks during dusk. Even if we
41
used a coarse measure of activity taken every 2 hours, activity sensors could reliably track
the daily activity peaks of deer as daylight changed from summer to fall. Our results
indicate that GPS activity sensors can be used to estimate activity and are therefore an
advantageous tool to monitor the daily activity patterns of large herbivores.
Research and management implications
Activity sensor technology allows the possibility to quantify activity of animals, especially
for species difficult to observe in nature. We found that vertical sensors were more accurate
than horizontal sensors for both GPS and VHF collars (Gillingham and Bunnell 1985, Beier
and McCullough 1988). We encourage scientists and managers to use activity sensors to
record activity of their study animals, but we suggest using VHF vertical variable-pulse
sensors because they will possibly give more reliable results than horizontal sensors.
VHF and GPS activity sensors allow the quantification of continuous activity information
(active vs. inactive behaviors) to study the foraging behavior and assess fine-scale habitat
use and temporal activity patterns of wild mammals. For example, activity sensors can be
used to examine the influence of population density and resource abundance on activity
budgets. In addition, each GPS fix obtained from free-ranging individuals is accompanied
by an activity value that can be related to the environmental characteristics of each position.
This information can be used, for example, to analyze the effects of habitat quality on the
time budget of free-ranging deer at a fine scale. Activity data are necessary to improve our
understanding of foraging behavior and, more generally, of plant-herbivore relationships. In
the context of high deer density on Anticosti Island and elsewhere, activity data can
contribute to generate predictive models that will help wildlife managers and land use
planners to integrate plant-herbivore relationships into forest and wildlife management.
Acknowledgements
We thank L. Breton and B. Rochette from the Ministère des Ressources Naturelles, de la
Faune et des Parcs du Québec, as well as D. Duteau, G. Picard, F. Fournier, D. Sauvé, A.
Simard and J. Taillon, for capturing deer. J. Taillon and A. Tousignant helped with the
behavioral observations and R. Pouliot during the preliminary steps of the study. S.
42
DeBellefeuille, F. Fournier, J. Huot and many graduate colleagues reviewed an earlier draft
of the manuscript. We are also thankful to G. C.White, R. Moen and M. P. Gillingham for
their helpful comments on the manuscript. This project was funded by Produits forestiers
Anticosti inc., the Natural Sciences and Engineering Research Council of Canada and the
Fonds québécois de la recherche sur la nature et les technologies.
Literature cited
Adrados, C., H. Verheyden-Tixier, B. Cargnelutti, D. Pépin, and G. Janeau. 2003. GPS
approach to study fine-scale site use by wild red deer during active and inactive
behaviors. Wildlife Society Bulletin 31:544–552.
Beier, P., and D. R. McCullough. 1988. Motion-sensitive radio collars for estimating white-
tailed deer activity. Journal of Wildlife Management 52:11–13.
Beier, P., and D. R. McCullough. 1990. Factors influencing white-tailed deer activity
patterns and habitat use. Wildlife Monographs 109:1–51.
Cederlund, G., R. Bergström, and F. Sandegren. 1989. Winter activity patterns of females
in two moose populations. Canadian Journal of Zoology 67:1516–1522.
Craighead, J. J., F. C. Craighead, R. L. Ruff, and B. W. O'Gara. 1973. Home ranges and
activity patterns of non-migratory elk of the Madison drainage herd as determined
by radio-telemetry. Wildlife Monographs 33:1–50.
Environment Canada. 1993. Canadian climate normals, 1961-1990 / Normales climatiques
au Canada, 1961-1990. Environnement Canada, Service de l'environnement
atmosphérique, Programme climatologique canadien, Downsview, Ontario, Canada.
Garshelis, D. L., H. B. Quigley, C. R. Villarrubia, and M. R. Pelton. 1982. Assessment of
telemetric motion sensors for studies of activity. Canadian Journal of Zoology
60:1800–1805.
Gillingham, M. P., and F. L. Bunnell. 1985. Reliability of motion-sensitive radio collars for
estimating activity of black-tailed deer. Journal of Wildlife Management 49:951–
958.
Gillingham, M. P., K. L. Parker, and T. A. Hanley. 1997. Forage intake by black-tailed deer
in a natural environment: bout dynamics. Canadian Journal of Zoology 75:1118–
1128.
Hansen, M. C., G. W. Garner, and S. G. Fancy. 1992. Comparison of 3 methods for
evaluating activity of Dall's sheep. Journal of Wildlife Management 56:661–668.
Hölzenbein, S., and G. Schwede. 1989. Activity and movements of female white-tailed deer
during the rut. Journal of Wildlife Management 53:219–223.
Kammermeyer, K. E., and R. L. Marchinton. 1977. Seasonal change in circadian activity of
radio-monitored deer. Journal of Wildlife Management 41:315–317.
Merrill, S. B., and L. D. Mech. 2003. The usefulness of GPS telemetry to study wolf
circadian and social activity. Wildlife Society Bulletin 31:947–960.
Moen, R., J. Pastor, and Y. Cohen. 1996. Interpreting behavior from activity counters in
GPS collars on moose. Alces 32:101–108.
43
Montgomery, G. G. 1963. Nocturnal movements and activity rhythms of white-tailed deer.
Journal of Wildlife Management 27:422–427.
Ozoga, J. J., and L. J. Verme. 1970. Winter feeding patterns of penned white-tailed deer.
Journal of Wildlife Management 34:431–439.
Potvin, F., P. Beaupré, and G. Laprise. 2003. The eradication of balsam fir stands by white-
tailed deer on Anticosti Island, Québec: A 150-year process. Écoscience 10:487–
495.
Relyea, R. A., I. M. Ortega, and S. Demarais. 1994. Activity monitoring in mule deer:
Assessing telemetry accuracy. Wildlife Society Bulletin 22:656–661.
Rouleau, I., M. Crête, and J.-P. Ouellet. 2002. Contrasting the summer ecology of white-
tailed deer inhabiting a forested and an agricultural landscape. Écoscience 9:459–
469.
Rouys, S., J. Theuerkauf, and M. Krasińska. 2001. Accuracy of radio-tracking to estimate
activity and distances walked by European bison in the Bialowiża forest, Poland.
Acta Theriologica 46:319–326.
Rowe, J. S. 1972. Forest regions of Canada. Department of environment, Canadian Forest
Service Publication, Ottawa, Ontario, Canada.
Rumble, M. A., L. Benkobi, F. Lindzey, and R. S. Gamo. 2001. Evaluating elk habitat
interactions with GPS collars. Pages 11–17 in Tracking animals with GPS:
Proceedings of an international conference. The Macaulay Land Use Research
Institute, 12–13 March 2001, Aberdeen, United Kingdom.
SAS Institute. 1989. SAS/STAT User's guide, Version 5, SAS Institute Inc., Cary, North
Carolina, USA.
Singer, F. J., D. K. Otto, A. R. Tipton, and C. P. Hable. 1981. Home ranges, movements,
and habitat use of European wild boar in Tennessee. Journal of Wildlife
Management 45:343–353.
Sparrowe, R. D., and P. F. Springer. 1970. Seasonal activity patterns of white-tailed deer in
eastern South Dakota. Journal of Wildlife Management 34:420–431.
Turner, L. W., M. C. Udal, B. T. Larson, and S. A. Shearer. 2000. Monitoring cattle
behavior and pasture use with GPS and GIS. Canadian Journal of Animal Science
80:405–413.
White, G. C., and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data, Academic
Press, San Diego, California, USA.
Zar, J. H. 1999. Biostatistical analysis, Prentice Hall, Upper Saddle River, New Jersey,
USA.
44
Chapitre 2. Influence of population density on
white-tailed deer movements and activity budgets
45
Résumé
Nous avons mesuré l’influence de la densité de population sur les déplacements et le budget
d’activité de cerfs de Virginie se trouvant à différentes densités en milieux contrôlés et
naturels. Le budget d’activité, les déplacements et la biomasse de plantes disponibles ne
variaient généralement pas selon les densités contrôlées (7.5 et 15 cerfs/km²). Cependant,
nous avons trouvé des différences interannuelles reliées à l’augmentation de l’abondance de
végétation après coupe. En effet, suite à l’augmentation de l’abondance de végétation dans
les enclos à densités contrôlées, la durée des périodes passées en activité diminuait et les
cerfs augmentaient le nombre de périodes passées en activité par jour. Au cours de l’été,
lorsque la végétation a augmenté en abondance, les cerfs adultes à 7.5 cerfs/km²
diminuaient la proportion du temps passé en activité par jour mais pas à 15 cerfs/km². Au
début de l’été en milieu naturel (>20 cerfs/km²), la végétation est peu abondante et les cerfs
ont diminué le temps passé en activité possiblement afin d’augmenter le temps de
rumination d’une végétation de moins bonne qualité.
46
Abstract
Foraging decisions in herbivores may be affected by population density as this factor is
frequently related to changing availability of preferred plant species, net abundance of
biomass and variation in intra-specific competition. We studied the effect of population
density on white-tailed deer movements and activity budgets using a controlled-density
experiment. The controlled densities (7.5 deer/km² and 15 deer/km²) were obtained by
placing 3 deer in 2 enclosures of different sizes (20 ha and 40 ha) where forest was partially
harvested in 2001. We repeated the experiment with the same two densities in 3 different
locations. Summer activity budgets and movements of yearlings and adults were quantified
by VHF telemetry the first, second and third year after the onset of the controlled-density
experiment. During one year, we also measured the activity budget of 4 adults in an
unfenced area at a density of >20 deer/km². In each enclosure, biomass of 13 of the most
abundant and preferred plant species was measured in 40 plots every year.
Adults were less active than yearlings at 7.5 but not at 15 deer/km². Otherwise, movements,
activity budgets and available biomass were similar between 7.5 deer/km² and 15 deer/km².
However, the available biomass increased through years and activity budgets varied
accordingly. As biomass increased, deer increased the number of daily activity bouts,
which also became shorter. With increasing biomass, it seemingly took less time for deer to
obtain sufficient forage to enter a rumination bout. Adult deer decreased time spent active
throughout summer, but only at 7.5 deer/km². However, adults in unfenced cutblocks were
less active at the beginning of the summer than deer at 7.5 deer/km² and increased time
spent active through summer. When vegetation was less abundant, such as in early summer,
deer at 7.5 deer/km² seemed to spend more time active gathering vegetation. In unfenced
areas at high density, forage was even less available, free-ranging deer in early summer had
to increase processing time to extract available energy, and thus decrease the proportion of
time spent active. This study demonstrates that relationships between density and foraging
behavior are complex and that controlled-density experiments may help to understand the
behavior of herbivores in relation to available resources.
47
Introduction
Summer is a critical season for herbivores in temperate and boreal regions to restore body
condition and build up body reserves for over-winter survival (Putman et al. 1996, Lesage
et al. 2001). Maximising energy intake requires that herbivores utilize the most profitable
plants in terms of energy content and foraging time necessary to crop them (Bunnell and
Gillingham 1985). Herbivores have access to less forage as population density increases
(Healy et al. 1997, Côté et al. 2004), hence, foraging strategies may become even more
critical at high density. Deer have become the dominant herbivores in most ecosystems of
North America and Europe and they have recently reached historic high densities over large
areas (Côté et al. 2004). The impacts of deer on ecosystem functioning are far reaching
(Côté et al. 2004) and the influence of population density on deer foraging behavior
therefore needs to be assessed.
Deer may change their behavior in response to changing availability of preferred plant
species, net abundance of forage biomass or social competition in relation to population
density. Many studies have obtained conflicting results regarding the influence of forage
quality and availability on herbivore behavior (Henriksen et al. 2003). Usually, herbivore
density is negatively related to forage availability and ungulates at high density may be
constrained to remain active for longer periods in order to ingest enough forage (Trudell
and White 1981, Moncorps et al. 1997). Deer may also increase search movements and
foraging time with increasing density to consume the most rewarding plants and plant parts
(Wickstrom et al. 1984, Bartmann et al. 1992). However, as forage is distributed in
spatially separated patches, movements and increased foraging time may also entail higher
time and energy costs (Murray 1991). Alternatively, deer may also respond to forage
depletion by foraging less selectively to reduce movement costs (Gates and Hudson 1983),
and their diel active time would then remain unaffected (Kohlmann and Risenhoover 1994).
By feeding on lower quality vegetation, i.e. plants with a high content of structural
compounds that reduce plant digestibility (Bryant and Kuropat 1980, Palo 1985, Côté
1998), the rate of passage of forage from the rumen to the lower digestive tract should slow
down and rumination time may then increase (Van Soest 1982, Spalinger et al. 1986).
48
Studies have shown that yearlings spend more time active than adults do as is expected
given their smaller mass and relatively smaller digestive system, higher metabolic rate and
growth energetic demands (Bunnell and Gillingham 1985, Côté et al. 1997, Shi et al. 2003).
As energetic demands increase allometrically (W0.75) with body size, larger individuals
need less energy per unit of mass than smaller individuals (Illius and Gordon 1987), and
because gut size increases linearly with body size and turnover time declines, this allows
larger individuals to extract more energy from lower quality forage than smaller individuals
(Demment and Van Soest 1985). Active time thus tends to decrease with increasing body
size (Moncorps et al. 1997, Ruckstuhl 1997, Mysterud 1998, Pérez-Barbería and Gordon
1999, Jeschke and Tollrian 2005). Given their different use of resources, an increase in
population density should thus affect differently juveniles and adults. For example, fawn
survival is more affected by density than adult survival (Jorgenson et al. 1997).
As vegetation increases in abundance during the growing season, it also becomes more
lignified and its protein content decreases (Hanley 1984). Time spent active could thus
differ between periods of plant growth and periods of plant senescence (Gates and Hudson
1983). Diel-patterns of activities have also been largely documented and studies have
shown that deer usually synchronize activity bouts with dawn and dusk (Kammermeyer and
Marchinton 1977, Beier and McCullough 1990) and shift their activity periods to times
when weather conditions are most favourable for thermoregulation (Beier and McCullough
1990).
Controlled-density experiments have been used as a research tool for many years to study
the foraging behavior of domestic animals and its use is now strongly encouraged for wild
ungulates (Hester et al. 2000, Gordon et al. 2004). Most studies have investigated the
effects of browser densities on vegetation abundance and diversity (Tilghman 1989, Hester
et al. 2000); however, controlled-density experiments may also be very useful to understand
how browsers modify their behavior at different population densities. Our general objective
was to assess how the daily and summer activity patterns of yearling and adult white-tailed
deer vary in relation to population density. We predicted that yearling and adult deer at high
density would be more active and have higher movement rates than those at lower densities
49
because of the increased time necessary to gather forage at high density. Alternatively,
yearling and adult deer could increase time spent inactive at high density to process
vegetation that is more fibrous. Total active and inactive times would then remain similar in
all densities but the length of inactive bouts would increase.
Study area
Anticosti Island (49° 28’ N, 63° 00’ W) is located in the Gulf of St. Lawrence, Québec,
Canada and covers 7,943 km2. Forests are naturally dominated by balsam fir (Abies
balsamea), white spruce (Picea glauca) and black spruce (P. mariana). White birch (Betula
papyrifera) and trembling aspen (Populus tremoloides) are irregularly found on the island.
Around 200 deer were introduced on the island at the turn of the 19th century. The
population spread and grew rapidly because of the absence of predators and the presence of
natural and human disturbances that created openings favourable to deer. Today, deer
densities of >20 deer/km² are found in many areas on the island (Potvin and Breton 2005).
Forest composition has been strongly modified by selective browsing, deciduous browse
species have almost disappeared and balsam fir stands are now being replaced by white
spruce stands (Potvin et al. 2003). The climate on Anticosti is typically maritime and
characterized by long and milder winters than on the continent (Huot 1982). Mean
temperatures are -12°C in January and 15°C in July and an average of 406 cm of snow and
630 mm of rain falls every year on the island (Environment Canada 1993).
Methods
Experimental design
Our experimental design is made of 4 blocks, three of which (A, B, C) consist in two
enclosures of different sizes (20 ha and 40 ha) in which we introduced deer. The last block
(T) is an unfenced area where density was estimated at >20 deer/km². We located all blocks
in balsam fir-dominated forests that were partially cut in the summer of 2001. Two of the
blocks (B, C) were localised in the center of the island (Jupiter River area); the other two
were located 130 km away in the western part of the island (A, T). Between 30 and 40% of
residual forest patches of different sizes (0.19–21.6 ha) were left uncut in the enclosures.
50
Water was easily accessible to deer in streams or artificial water holes in every enclosure.
To assess the effects of population density on activity and movement of deer, 2 controlled
densities were established in blocks A, B and C. Controlled densities were 7.5 deer/km2
(LDE; 40 ha enclosures with 3 deer) and 15 deer/km2 (HDE; 20 ha enclosures with 3 deer).
We chose these deer densities to cover a gradient that would include white-tailed deer
density levels proposed for sustainable tree regeneration (7 deer/km²; Tilghman 1989,
deCalesta and Stout 1997), the estimated density on Anticosti Island at the beginning of the
experiment (15.6 deer/km²; Rochette et al. 2003) and the local estimated in situ density
level in management areas adjacent to experimental blocks (>20 deer/km²). Our set up is
part of a larger study trying to determine which deer densities are compatible with forest
regeneration (Tremblay et al. in prep.).
Deer captures
We used different methods to capture deer: dart guns (Pneu-dart Inc, Williamsport,
Pennsylvania, USA), netguns (Coda Enterprises Inc., Mesa, Arizona, USA) shot from a
helicopter, Stephenson box traps and cannon nets baited with cattle feed and balsam fir
twigs. In June or early July of the first, second and third year after the onset of the
controlled-density experiment, deer were released in the study enclosures. On the second
year after the onset of the controlled-density experiment, we captured and released 4 adult
females fitted with VHF collars in the unfenced cutblocks (Table 2-1). The Animal Care
and Use Committee of Université Laval, Québec, Canada (Reference number 2005–008)
approved all capture methods.
All deer were fitted with VHF collars equipped with sto-2a variable-pulse activity sensors
(LMRT series, Lotek Engineering, Newmarket, Ontario, Canada). One adult male and 1
yearling male lost their collars and 1 yearling male had a malfunctioning collar. We
verified reproductive status of adult females by direct observation at capture and at the end
of summer. As only 2 reproductive females were monitored, we did not include
reproductive status as a variable in the analyses but verified if activity and movements were
comparable to the other females.
51
Table 2–1. Characteristics of white-tailed deer used in an experiment on the effects of
population density on deer activity budgets on Anticosti Island, Québec.
a Number of years since the onset of the controlled-density experiment.
b The number of days for which activity budgets of radio-collared deer were monitored
during each month and each year of the study.
c These females were observed with a fawn.
Deer no. Yeara Block Density (deer/km²) Age class Sex Number of days monitoredb
July August September
1 1 A 7.5 Adult Male 21 19
2 Yearling Male 21 19
3 Yearling Female 21 9
4 15 Yearling Female 21 19
5 Yearling Female 21 19
6 Yearling Female 20 19
7 2 A 7.5 Adult Femalec 15 22 13
8 Yearling Female 14 22 13
9 Adult Female 15 22 13
10 15 Adult Female 14 22 13
11 Adult Male 15 22 13
12 Yearling Male
13 2 B 7.5 Yearling Female
14 Yearling Male
15 Adult Female
16 15 Adult Female
17 Yearling Female
18 Adult Femalec
19 2 C 7.5 Yearling Female 15 7 5
20 Adult Female 14 8 5
21 Yearling Male 14 7 5
22 15 Adult Male 15 8 5
23 Yearling Female 13 8 5
24 Adult Male
25 2 T >20 Adult Female 0 2 5
26 Adult Female 9 8 8
27 Adult Female 9 6 4
28 Adult Female 9 9 8
29 3 A 7.5 Yearling Male
30 Adult Male
31 Yearling Male
32 15 Yearling Female
33 Yearling Male
34 Adult Male
35 3 C 7.5 Yearling Female 8 13 1
36 Yearling Female 0 12 1
37 Adult Male 7 10
38 15 Adult Male 13 29 1
39 Yearling Male 13 31 2
40 Yearling Male 13 30 2
52
Forage abundance
To assess plant biomass available in the enclosures, we randomly placed 20 sampling
points in cuts and 20 points under forest cover in each experimental unit. At each sampling
point, percent of plant cover was estimated in 2 1-m2 plots randomly located in a 10×10 m
quadrat centered on the sampling point. We quantified biomass of the following species:
Abies balsamea, Betula papyrifera, Cirsium spp., Coptis groenlandica, Cornus canadensis,
Epilobium angustifolium, Grass sp., Hieracium sp., Maianthemum canadense, Picea
glauca, Rubus idaeus, Rubus pubescens and Trientalis borealis. Plant biomass was assessed
using regressions between percent plant cover and mass of dried plants (Bonham 1989).
Number of samples needed for regressions was estimated empirically by plotting regression
coefficients with number of samples until an asymptote was reached (Frontier 1983). We
did not assess forage abundance in the experimental unit located outside the enclosures but
considered it comparable to the forage abundance of the experimental units during the first
year of the experiment.
Movements
In July and August 2002, the first year of the controlled-density experiment, we
radiotracked 6 deer in block A and in the second year after the onset of the experiment, 18
deer were tracked in 3 blocks (A, B, C). Deer were located with receivers (SRX-400
version W9, Lotek Engineering, Newmarket, Ontario, Canada and TR-2 scanner/receiver,
Telonics, Meza, Arizona, USA), a unidirectional yagi antenna and a compass. Telemetry
stations were positioned with a GPS Garmin (Garmin international, Olathe, Kansas, USA;
precision of <5 m) on forest roads adjacent to the enclosures. To limit human disturbance,
stations were generally located more than 100 m away from the enclosures. At least 3
azimuths differing by a minimum of 30º were obtained by moving between stations with a
vehicle (White and Garrott 1990). To reduce location error, positioning had to be
completed within 15 minutes (White and Garrott 1990). Periods of the day were evenly
sampled by separating them into 3 periods of 8 hours (8h00–16h00, 16h00–0h00 and
0h00–8h00). These 8-hour periods were rotated between 2 observers and between groups of
53
enclosures every 3 days to evenly sample the complete 24 hours of a day and all the
enclosures.
LOAS software (Location of a Signal; Version 2.07, Ecological Software Solutions,
Schwägalpstrasse, Urnäsch, Switzerland) was used to estimate positions and error
polygons. Error polygons were calculated with “Andrews” estimator. All locations were
plotted with LOAS on maps and were assigned Universal Transverse Mercator (UTM)
coordinates. The average error from plotted to actual locations was determined by placing
VHF collars at known locations throughout the enclosures and was estimated at 107 m
(SE = 88 m; n = 88 trials). We deleted telemetry locations with error polygons greater than
0.1 ha. After processing, 2,916 usable locations amongst the 3,251 recorded were obtained.
The minimum movement rate was estimated as the linear distance between two successive
deer-locations separated by less than 3 hours divided by the time elapsed between these 2
locations. One deer (deer #3 in Table 2-1) was also fitted with a GPS collar to verify the
influence of positioning error on movement rate estimation. The distance moved per hour
was equivalent between VHF and GPS collars (VHF: x = 241.1 ± 12.6 m/hour;
n = 196 movements; GPS: x = 247.4 ± 7.8 m/hour; n = 944 movements; F1,1138 = 0.12;
P = 0.73). The distance moved was slightly related to the time interval between two
locations (r = 0.11; P < 0.01; n = 2,219 movements) and the mean time interval between
positions was 1h38 (SE = 27 min., n = 2,219 movements).
Activity budgets
Variable-pulse activity sensors of VHF collars use mercury switches that add pulses to the
base pulse rate of the collar each time the switch is triggered. The number of pulses above
the base pulse rate indicates the degree of animal activity during the period when the pulses
were counted (Type STO-2A, Lotek Engineering, Newmarket, Ontario, Canada).
Transmitter signals of activity sensors were received and recorded in a SRX-400 version
W9 receiver-datalogger (Lotek Engineering, Newmarket, Ontario, Canada) connected to a
multidirectional antenna, a 12 V battery, and a solar panel.
54
The receiver was programmed to measure the duration between 2 pulses for 65 consecutive
pulses, record mean pulse rate and then automatically switch to scan another transmitter.
The time needed to record 65 pulses was thus dependent on pulse rate. As the SRX receiver
scanned one transmitter at a time and because 6 individuals were monitored each day, a
measure of pulse rate for each deer was obtained approximately every 6 minutes. Data were
downloaded in a portable computer with the help of Winhost software (version 1.0.0.1,
Lotek Engineering, Newmarket, Ontario, Canada).
Validation studies of activity sensors have obtained mixed results in overall reliability (74
to 98% accuracy) and have demonstrated that it is necessary to validate methods used to
measure activity budgets with direct animal observations (Gillingham and Bunnell 1985,
Beier and McCullough 1988, Relyea et al. 1994). We conducted our own validation study
with direct observations of deer in small enclosures (Coulombe et al. 2006). By combining
the information of 3 successive scans, we correctly assessed 87% of all activity bouts
(Coulombe et al. 2006). An inactive bout began when at least 3 inactive scans were
observed. To return to an active bout, at least 3 active scans had to be observed. Activity
data did not allow us to differentiate amongst different active or inactive behaviors, e.g.
resting could not be differentiated from ruminating or eating from moving. However, as it
was shown for the Odocoileus genus, activity periods not corresponding to feeding
activities (e.g. vigilance, social interactions) represent only 5 to 15% of time spent active
(Beier and McCullough 1990, Gillingham et al. 1997) and thus are not an important part of
the activity budget. Additionally, a decrease in plant quality is generally related to an
increase in rumination time, and simultaneously to an increase in time spent inactive
(Mysterud 1998, Pérez-Barbería and Gordon 1999). Time spent inactive is thus an
indication of rumination time.
At the onset of the controlled-density experiment, in July and August 2002, one block (A)
with 2 densities was available for the study of activity budgets (Table 2-1). The second
year, in July, August and September 2003, 2 blocks (A, C) were studied. The third year, in
July, August and September 2004, we monitored activity budgets of deer in 1 block (C).
Another unit (T) was located in an unfenced area where density was estimated by an aerial
55
survey to >20 deer/km2. We monitored the activity budget of these free-ranging deer from
July to September 2003. To analyse time budgets, we used the proportion of time spent
active, the length of active and inactive bouts and the number of activity bouts per day.
Analyses
We first tested if plant biomass available to deer differed between densities (7.5 deer/km²,
15 deer/km²), strata (clear-cuts, forests) and years since the onset of the controlled-density
experiment, using an analysis of variance with block as a random factor. We then
contrasted the mean distance moved and the proportion of time spent active between
densities, periods of the day (dawn: 1h30 before sunrise to 1h30 after sunrise, day: 1h30
after sunrise to 1h30 before sunset, dusk: 1h30 before sunset to 1h30 after sunset and night:
1h30 after sunset to 1h30 before sunrise) and week with block and year as random factors.
We also used week2 in certain analyses (because plant quality first increases but then
decreases throughout the summer) but if the quadratic term did not significantly explain
extra variability, it was removed. Periods of the day and weeks were repeated for each deer,
we thus used a repeated measures analysis with density as the treatment and block as
replicates (Proc Mixed, SAS Version 9.1, SAS Institute Inc., Cary, North Carolina, USA).
We considered the enclosure as the experimental unit to take into account the potential
problem of autocorrelation between individuals in a same enclosure. Mean length of active
and inactive bouts and number of daily activity bouts were compared between densities and
week (or week’s quadratic term) for yearlings and adults with block and year as random
factors and with weeks (or week’s quadratic term) as repeated measures. We did not have
enough individuals for each year and density to account for sex, age and density in a full
model. We therefore chose to compare the effects of density between different age groups
(yearlings and adults) of both sexes with an analysis of variance because the effects of
density may differ between growing individuals (i.e. yearlings) and adults as their activity
budgets vary (Bunnell and Gillingham 1985, Shi et al. 2003). To account for differences
between years in movements and time budgets, we used simple contrast comparisons of the
random factor “block×year” (McLean et al. 1991). After models were developed, we
compared with a Z-test the means for adults and yearlings.
56
Because weather conditions vary daily and weekly and can influence deer activity patterns
(Beier and McCullough 1990), we verified the influence of weather on deer movements and
on the proportion of time spent active. We obtained mean hourly temperature data from the
Environment Canada meteorological station located in the western part of the island.
Although mean hourly temperature is not a complete weather descriptor because weather
conditions may also depend on the combination of solar radiation, wind velocity, air
temperature, and other factors, it provides a practical approximation of weather conditions
(Beier and McCullough 1990). For every period of the day, we thus related the proportion
of time spent active for adults and yearlings to the mean temperature for the period and
tested if slopes were significantly different from zero and if they differed between densities.
As normality was violated because of extreme movement or proportion of time spent active
values, we used a robust regression (proc Robustreg, SAS Version 9.1, SAS Institute Inc.,
Cary, North Carolina, USA).
We could not integrate deer in the unfenced area in the same models as above because we
had only one unfenced block (that contained only adults). To test for differences between
the adult activity budget in controlled densities and in the unfenced area, we subtracted
each value of activity from controlled densities to the corresponding value from the
unfenced area, for each week and period of the day. Differences were tested in a model
containing density, period of the day and week (or its quadratic term) with block and year
as random factors. If the estimated parameter confidence intervals contained zero, we
concluded that there was no difference between activity budgets in controlled densities and
in natural densities.
After each analysis, residuals were examined for normality and variance homogeneity.
Significance was set at 0.05 and all results, unless specified, are presented as
means ± standard error. To simplify the presentation of results, we do not show interaction
terms, as they, unless specified, were not significant.
57
Results
Forage abundance
Total biomass of forage did not vary with deer density (F1,2 = 0.03; P = 0.87), but was
greater 2 and 3 years after the onset of the controlled-density experiment then during the
first year (F2,20 = 18.53; P < 0.01; Figure 2-1). Total biomass was greater in cutblocks than
under forest cover 2 and 3 years after the onset of the experiment but not during the first
year (F2,20 = 3.73; P = 0.04; Figure 2-1).
Movements
Mean temperature did not affect deer movements or activity budgets during any period of
the day (all P-values > 0.05), we therefore did not include temperature in any further
analyses. Movement rates were similar between controlled densities for both yearlings and
adults (Table 2-2). In addition, movement rates did not differ between adults and yearlings
(Z = 0.70, P = 0.50). Movement rates did not change according to the number of years after
onset of the controlled-density experiment for both yearlings (1 year: x = 109 ± 7 m/hour;
n = 38 daily periods; 2 years: x = 111 ± 6 m/hour; n = 92; F1,106 = 0.00, P = 1) and adults
(1 year: x = 107 ± 11 m/hour; n = 19; 2 years: x = 116 ± 6 m/hour; n = 114; F1,113 = 1.09,
P = 0.3) or according to periods of the day or number of weeks since the beginning of the
summer (Table 2-2).
Proportion of time spent active
Yearlings were active about 73-75% of the time, both in low (LDE) and in high (HDE)
density enclosures (Table 2-3). Adults in LDE were about 14% more active than those in
HDE (Table 2-3). Adults in unfenced areas were as active ( x = 0.72 ± 0.01; n = 21 daily
periods) as those in controlled density enclosures (Table 2-3; LDE: t2 = 1.23, P = 0.34;
HDE: t2 = -2.34, P = 0.14). Yearlings were more active than adults in HDE (Z = -1.98,
P = 0.05), but not in LDE (Z = -0.44, P = 0.66). For adults, the proportion of time spent
active
58
Year(s) after onset of the controlled-density experiment
1 2 3
Bio
mass
(g
/m²)
0
25
50
75
100
125
150
175
200
1 2 3
Bio
mass
(g
/m²)
0
5
10
15
20
25
30
35
a
b
c
b
c
a) Forest
b) Cutblock
a
Figure 2–1. Mean plant biomass available to white-tailed deer in a controlled-density
experiment on Anticosti Island, Québec containing known densities (7.5 deer/km²: black
bars, 15 deer/km²: grey bars) of deer. Biomass was estimated 1, 2, and 3 years after the
onset of a controlled-density experiment in residual forest stands (a) and cutblocks (b, note
the different scales on the Y axes). Biomass was compared between densities, years and
stratum (forest vs. cutblock) with block (3 sites) as a random factor. Bars with different
letters are statistically different.
59
Table 2–2. Comparisons of white-tailed deer summer movement rates in two controlled densities according to age class,
week and period of the day (Anticosti Island, Québec).
a Movements were compared between densities and weeks for yearlings (a) and adults (b) with block (3 sites) and year as
random factors and period of the day and weeks as repeated measures.
b Numbers in parentheses are the number of values used in the analysis, each corresponding to a specific period of the day,
week, density and block.
Factor Movement (m/hour)a DF F-value P-value
a) Yearlings Density 7.5 deer/km²: 120.0 ± 5.9 (77)b 15 deer/km²: 97.1 ± 6.9 (53) 1,5 2.84 0.15
n = 12 deer Week slope: -0.36 ± 0.41 (130) 1, 108 1.39 0.36
Period of the day dawn
118.4 ± 9.0 (31)
day
105.8 ± 6.8 (38)
dusk
120.6 ± 11.0 (25)
night
102.1 ± 10.3 (36)
3,15 1.42 0.28
b) Adults Density 7.5 deer/km²: 124.1 ± 7.6 (75) 15 deer/km²:102.6 ± 6.7 (58) 1,5 0.29 0.61
n = 9 deer Week slope: 0.28 ± 0.39 (133) 1,101 0.02 0.89
Period of the day dawn
128.6 ± 13.7 (34)
day
116.6 ± 7.9 (37)
dusk
120.6 ± 11.0 (25)
night
102.1 ± 10.3 (37)
3,15 1.46 0.26
60
Table 2–3. Proportion of time that white-tailed deer spent active in summer at two controlled densities on Anticosti
Island, Québec.
Factor Proportion of time spent active† DF F-value P-value
a) Yearlings Density 7.5 deer/km²: 0.75 ± 0.03 (102) †† 15 deer/km²: 0.73 ± 0.03 (72) 1,5 0.98 0.34
n = 13 deer Week slope: 0.0027 ± 0.014 (174) 1,153 0.24 0.62
Week × density difference between slopes at 7.5 and 15 deer/lm²: -0.0018 ± 0.018 (174) 1,153 2.37 0.13
Period of the day dawn
0.70a††† ± 0.03 (47)
day
0.75a ± 0.03 (44)
dusk
0.90b ± 0.03 (44)
night
0.62c ± 0.04 (39)
3,15 3.22 0.05
b) Adults Density 7.5 deer/km²: 0.73 ± 0.03 (98) 15 deer/km²: 0.63 ± 0.04 (60) 1,4 24.9 0.03
n = 9 deer Week slope: 0.0045 ± 0.015 (147) 1,138 0.51 0.48
Week × density difference between slopes at 7.5 and 15 deer/lm²: -0.0306 ± 0.019 (147) 1,138 11.12 0.03
Period of the day dawn
0.63 ± 0.04 (42)
day
0.69 ± 0.04 (41)
dusk
0.82 ± 0.04 (39)
night
0.59 ± 0.04 (36)
3,12 0.78 0.53
† The proportion of time spent active was compared between densities, weeks since the beginning of summer and periods
of the day for yearlings (a) and adults (b) with block (3 sites) and year as random factors and periods of the day and week
as repeated measures.
† † Numbers in parentheses are number of values used in the analysis, each corresponding to a specific period of the day,
week, density and block.
††† Values of proportion of time spent active between periods of the day with different letters are statistically different
(P < 0.05).
61
decreased from the first to the third year after the onset of the controlled-density experiment
(1 vs. 2 years: F1,142 = 1.85, P = 0.18; 1 vs. 3 years: F1,142 = 8.52, P < 0.01; 2 vs. 3 years:
F1,142 = 4.47, P = 0.04). For yearlings, time spent active was similar through years (1 vs. 2
years: F1,158 = 0.93, P = 0.33; 1 vs. 3 years: F1,158 = 0.12, P = 0.73; 2 vs. 3 years:
F1,158 = 1.54, P = 0.22; Figure 2-2a). For adults, the proportion of time spent active
decreased throughout summer in LDE (t138 = -2.14; P = 0.03) but not in HDE (t138 = 0.96;
P = 0.34; Table 2-3, Figure 2-3a). For yearlings, time spent active did not change through
summer in both densities (Table 2-3, Figure 2-4a). Adults in the unfenced area spent less
time active during July than adults in LDE (F1,36 = 4.14, P = 0.05), but no differences
occurred in other months or with deer in HDE. Yearlings were more active at dusk than at
any other period of the day and less active at night than during the day or at dawn
(Table 2-3). Although adults were also 19% more active at dusk than at night, the
difference was not significant (Table 2-3).
Number of activity bouts
The number of daily activity bouts did not differ with density or weeks since the beginning
of the summer for both yearlings and adults (Table 2-4). The number of daily activity bouts
did not differ between deer in the unfenced area ( x = 8.8 ± 2.2; n = 15 weeks) and those in
controlled-density enclosures (Table 2-4; LDE: t4 = -0.04; P = 0.97; HDE: t4 = 1.99;
P = 0.12). Yearlings and adults had a similar number of activity bouts (LDE: Z = 0.62,
P = 0.54; HDE: Z = -0.16, P = 0.88). Yearlings had more activity bouts 2 and 3 years after
the onset of the controlled-density experiment than after 1 year (1 vs. 2 years: F1,37 = 75.32,
P < 0.01; 1 vs. 3 years: F1,34 = 83.62, P < 0.01; Figure 2-2b) and adults had more activity
bouts 3 years after the onset of the controlled-density experiment than 2 years after it (1 vs.
3 years: F1,34 = 1.34, P = 0.26, 2 vs. 3 years: F1,34 = 6.28, P = 0.02; Figure 2-2b). Despite
small sample size, we detected a gradual decrease in the number of activity bouts for deer
in unfenced areas during the summer. Deer in unfenced areas had more activity bouts than
deer in LDE until the end of July, but less bouts at the end of August (week×week:
F1,8 = 12.95, P < 0.01; Table 2-4; Figure 2-3b).
62
Figure 2–2. Activity budgets of white-tailed deer from Anticosti Island (Québec) according
to the number of years since the onset of a controlled-density experiment. Data were pooled
between two controlled deer densities (7.5 and 15 deer/km²). Comparisons were made by
simple contrasts in a model accounting for density, period of the day and weeks since the
beginning of the summer with block and year as random factors. Bars with different letters
are statistically different.
Adults Yearlings
Nu
mb
er
of
dail
y a
cti
vit
y b
ou
ts
0
2
4
6
8
10
12
14
16
Adults Yearlings
Pro
po
rtio
n o
f ti
me a
cti
ve
0.5
0.6
0.7
0.8
0.9a) b)
Adults Yearlings
Len
gth
of
inacti
ve b
ou
ts (
min
.)
20
30
40
50
60
70
80
1
2
3
d)
Adults Yearlings
Len
gth
of
acti
ve b
ou
ts (
min
.)
20
40
60
80
100
120
140c)
a
a b
a a
a
aab
b
a a
b
a
b
c
a
ab
b
a
b
c
a a
b
year(s) after onset
of the controlled-
density experiment
63
Figure 2–3. Proportion of daily time spent active (a), number of daily activity bouts (b),
length of active (c), and inactive bouts (d) during summer for adult white-tailed deer on
Anticosti Island (Québec), pooled across years. Each circle corresponds to the mean for one
block-density during one week. Regression lines were drawn from model predicted values
(Tables 2–3, 2–4).
7.5 deer/km²
Date
07/02 07/16 07/30 08/13 08/27 09/10
Pro
po
rtio
n o
f ti
me s
pen
t acti
ve
0.2
0.4
0.6
0.8
1.0
Date
07/02 07/16 07/30 08/13 08/27 09/10
Len
gth
of
acti
ve b
ou
ts (
min
.)
0
20
40
60
80
100
120
140
160
Date
07/02 07/16 07/30 08/13 08/27 09/10
Len
gth
of
inacti
ve b
ou
ts (
min
.)
0
20
40
60
80
Date
07/02 07/16 07/30 08/13 08/27 09/10
Nu
mb
er
of
dail
y a
cti
vit
y b
ou
ts
4
6
8
10
12
14
15 deer/km²
>20 deer/km²
a)
c) d)
b)
64
Figure 2–4. Proportion of daily time spent active (a), number of daily activity bouts (b),
length of active (c), and inactive bouts (d) during summer for yearling white-tailed deer on
Anticosti Island (Québec), pooled across years. Each circle corresponds to the mean for one
block-density during one week. Regression lines were drawn from model predicted values
(Tables 2–3, 2–4).
Date
07/02 07/16 07/30 08/13 08/27 09/10
Nu
mb
er
of
dail
y a
cti
vit
y b
ou
ts
4
6
8
10
12
14
16
18
Date
07/02 07/16 07/30 08/13 08/27 09/10
Len
gth
of
inacti
ve b
ou
ts (
min
.)
0
20
40
60
80
Date
07/02 07/16 07/30 08/13 08/27 09/10
Len
gth
of
acti
ve b
ou
ts (
min
.)
0
20
40
60
80
100
120
140
160
180
200
Date
07/02 07/16 07/30 08/13 08/27 09/10
Pro
po
rtio
n o
f ti
me s
pen
t acti
ve
0.5
0.6
0.7
0.8
0.9
1.0a)
c) d)
b)
7.5 deer/km²
15 deer/km²
Table 2–4. Number of daily activity bouts (a) and length (min.) of active (b) and inactive bouts (c) during summer of yearling
and adult white-tailed deer at two controlled densities on Anticosti Island, Québec.
a The number of activity bouts and length of active and inactive bouts were compared between densities and weeks since the
beginning of summer for yearlings and adults with block (3 sites) and year as random factors and weeks as repeated measures.
b Numbers in parentheses are sample sizes or number of values used in the analysis, each corresponding to a specific week,
density and block.
Factor Parameter estimatesa DF F-value P-value
a) Yearlings Density 7.5 deer/km²: 10.1 ± 1.5 (20)b 15 deer/km²: 9.9 ± 1.5 (18) 1,4 0.60 0.48
n = 13 deer Week slope: 0.06 ± 0.13 (38) 1,3 0.02 0.89
Adults Density 7.5 deer/km²: 8.9 ± 1.1 (8) 15 deer/km²: 10.2 ± 1.1 (8) 1,4 2.04 0.23
n = 9 deer Week slope: -0.05 ± 0.20 (16) 1,3 0.35 0.56
b) Yearlings Density 7.5 deer/km²: 77.3 ± 16.0 (29) 15 deer/km²: 104.6 ± 17.3 (24) 1,7 3.07 0.12
n = 13 deer Week slope:-4.52 ± 3.00 (53) 1,42 0.88 0.35
Adults Density 7.5 deer/km²: 104.6 ± 18.1 (29) 15 deer/km²: 68.3 ± 20.0 (23) 1,7 0.85 0.39
n = 9 deer Week slope:-0.60 ± 2.40 (52) 1,41 0.06 0.81
c) Yearlings Density 7.5 deer/km²: 43.7 ± 6.1 (29) 15 deer/km²: 48.9 ± 6.7 (20) 1,5 2.72 0.16
n = 13 deer Week slope: 1.93 ± 0.90 (49) 1,4 0.46 0.50
Week × density difference between slopes: -3.01 ± 1.14 (49) 1,4 7.24 0.01
Adults Density 7.5 deer/km²: 50.5 ± 6.1 (28) 15 deer/km²: 58.2 ± 6.3 (22) 1,5 3.51 0.12
n = 9 deer Week slope: -0.60 ± 0.99 (50) 1,41 0.08 0.78
65
66
Length of active and inactive bouts
The length of active bouts did not differ with density for yearlings (Table 2-4). For adults, it
was also similar for all densities, including the unfenced area ( x = 93.0 ± 3.8 min.; n = 15
weeks; LDE vs. HDE: Table 2-4; LDE vs. unfenced area: t5 = -0.12, P = 0.91; HDE vs.
unfenced area: t5 = -0.16, P = 0.30). Adults and yearlings also had active bouts of similar
lengths (Table 2-4; LDE: Z = 0.80, P = 0.42; HDE: Z = 1.03, P = 0.30). Active bouts were
longer 1 year compared to 2 (F1,39 = 21.79, P < 0.01) and 3 (F1,39 = 27.22, P < 0.01) years
after the onset of the controlled-density experiment for adults and greater the first year than
the third year after the onset of the controlled density experiment for yearlings (F1,49 = 4.03,
P = 0.05; Figure 2-2c). The length of active bouts did not change through summer in any
age-class or density (Table 2-4; Figures 2–3c and 2–4c).
The length of inactive bouts was similar for yearlings and adults in all controlled densities
(Table 2-4). In the unfenced area, the length of inactive bouts ( x = 50.3 ± 5.9 min.; n = 10
weeks) was similar to LDE (Table 2-4; t3 = 1.38, P = 0.26), but deer in HDE had slightly
longer inactive bouts than deer in the unfenced area (Table 2-4; t3 = 3.20, P = 0.05). The
length of inactive bouts was also similar for adults and yearlings (Table 2-4; LDE:
Z = 1.13, P = 0.26; HDE: Z = -1.37, P = 0.17). The length of inactive bouts, however,
gradually increased from the first to the third year after the onset of the controlled-density
experiment for adults (1 vs. 2 years: F1,46 = 9.72, P < 0.01; 2 vs. 3 years: F1,46 = 11.72,
P < 0.01; Figure 2-2d). For yearlings, inactive bouts were also longer the third year than
after the second (F1,45 = 3.85, P = 0.05) and the first year (F1,45 = 3.87, P = 0.05) after the
onset of the controlled-density experiment (Figure 2-2d). For adults, the length of inactive
bouts did not vary with time in LDE and HDE (Table 2-4; Figure 2-3d), but deer from the
unfenced area had longer inactive bouts compared to deer in LDE until the end of July but
shorter inactive bouts after (week×week: F1,19 = 6.10, P = 0.02; Figure 2-3d). The length of
inactive bouts tended to decrease throughout the summer for yearlings in LDE (t40 = -1.60;
P = 0.12), and increase in HDE (t40 = 2.15; P = 0.04; Figure 2-4d).
67
Discussion
Influence of population density
Cervids may adjust their behavior to lower quantity and quality of vegetation by either
increasing movement rates (Wickstrom et al. 1984, Bartmann 1992), spending more time
active (Trudell and White 1981, Moncorps et al. 1997) or by increasing the length of
inactive bouts to process vegetation of lower quality (VanSoest 1982, Spalinger et al.
1986). We predicted that population density would influence movement rates, proportion of
time spent active or length of inactive and active bouts because population density is
inversely related to the quantity and quality of available vegetation. To our knowledge, our
study is the first to control for cervid population density using experimental enclosures to
test these predictions.
Movement rates of deer did not differ between controlled densities. In agricultural and
forested landscapes differing in population density, Rouleau et al. (2002) also monitored
movements of white-tailed deer and found that variations in movement rates were related to
ecological differences between landscapes and not to population density. In our study, we
used adjacent and ecologically comparable enclosures and found that there was no
influence of population density on deer movement rates. Despite the error associated with
each telemetry position, movement rates for a deer fitted with both GPS (high precision)
and VHF (low precision) collars were comparable. We applied the same protocol to
measure movement rates at both densities and we are therefore confident that movements
between densities are comparable. Fences could have limited deer movements and biased
movement rates. In addition, deer fitted with GPS collars on Anticosti Island in unfenced
clear-cuts moved at a much slower rate ( x = 78.9 ± 32.8 m/hour; n = 4 deer; unpublished
data) than deer in both enclosures (Table 2-2). Consequently, it appears that the fence did
not refrain deer movements and that population density did not influence movement rates at
controlled densities. For mule deer (Odocoileus hemionus), it was found that movement
rate was lower when forage was more abundant (Bartmann 1992, Wickstrom et al. 1984).
We did not detect differences, however, in plant biomass between the two experimental
densities because the treatment was not applied for long enough or because the difference
68
in density was not sufficiently large. Observations from a companion study suggest that
differences in biomass between treatments arise when density differences are greater and
when the treatment is applied for one year longer (Tremblay 2005). Additionally, both
controlled densities (7.5 and 15 deer/km²) were lower than the average deer density found
under natural conditions on the island (i.e., >20 deer/km²) and this might have prevented us
from observing a difference in movement rates. However, although plant biomass increased
from the first to the second year of the treatment application, we did not observe any
additional difference between movement rates of deer in controlled densities.
Flexible time-activity budgets may allow animals to circumvent the effects of declining
food abundance at high population density (Cederlund et al. 1989, Beier and McCullough
1990, Borkowski 2000). In winter, for example, cervids usually reduce activity in response
to food scarcity (Moen 1978, Georgii 1981, Risenhoover 1986). Observational studies have
found that concentrate feeders such as white-tailed deer spend 20 to 68% of their time
active in summer (Bunnell and Herstad 1989). Depending on density and age, we found
that deer spent 63 to 75% of the time active in summer. Compared to observational studies,
these activity values are thus rather high. Extra time needed for foraging, because of the
high density prevailing on Anticosti Island, might cause such high activity rates. However,
high activity values were found even in low-density enclosures two and three years after the
onset of the experience, when biomass availability had significantly increased. For
yearlings, we found no difference in the proportion of time spent active, number of daily
activity bouts or length of active and inactive bouts between controlled densities. Adults,
however, were more active at low density than at high density. During the growing season,
high plant biomass and availability, especially at low density, may allow adult deer to be
more selective in their search of plant species or plant parts (Owen-Smith and Novellie
1982, Belovsky 1984). At low density, adult deer would increase time spent active because
high quality food is more abundant and is then more cost-effective to search for (Trudell
and White 1981, Moncorps et al. 1997), although this did not occur for yearling deer.
Yearlings spent a larger part of their diel time active in high-density enclosures than adult
deer and this may have prevented them from increasing time spent active in low-density
69
enclosures. Because of their smaller mass and digestive system, higher metabolic rate and
growth energetic demands, yearlings generally spend more time active than adults (Bunnell
and Gillingham 1985, Shi et al. 2003). Clutton-Brock et al. (1987) proposed that
differential use of resources among ungulates should increase at high density when
resources are low since larger individuals, who have higher energy requirements are
excluded from areas of low plant biomass used by smaller individuals. At high density,
yearlings spent more time active than adults probably because they continued to feed on
less abundant high quality vegetation that could not be used by adult deer that had to gather
more abundant vegetation given their higher mass.
Annual differences
Plant biomass was higher the second and third year after the initiation of the controlled-
density experiment than during the first year (Figure 2-1). Similarly to a study comparing
activity budgets in two moose (Alces alces) populations at different plant biomass densities
(Cederlund et al. 1989), we found that the repartition of time between active and inactive
bouts varied through years. Indeed, the length of active bouts decreased through years as
the length of inactive bouts and the number of daily activity periods increased (Figure 2-2).
The decrease in the length of active bouts was probably due to an increase in food
availability rather than to an improvement of diet quality. As biomass increased, deer could
fill their rumen faster before entering a rumination bout, which could also explain the
higher number of daily activity bouts 3 years after the initiation of the controlled-density
experiment compared to the first year. Longer inactive bouts may be due to the extra
available time for processing vegetation (Moncorps et al. 1997) or the possibility to reduce
exposition to adverse environmental conditions. Indeed, when forage quantity increases,
time spent inactive decreases as deer may select higher quality forage and spend less time
ruminating (Mysterud 1998, Pérez-Barbería and Gordon 1999), but this relationship does
not appear to hold when high quality vegetation is available and abundant. For example, in
mountain goats (Oreamnos americanus), the proportion of time spent ruminating relative to
the time spent inactive was negatively related to availability of high quality vegetation (S.
Hamel, unpublished data). As the abundance of high quality vegetation increased, goats
70
spent relatively less time ruminating and more time resting in protected areas because it
was faster for them to fill their digestive tract and ruminate vegetation. Correspondingly, in
our study under abundant forage conditions, deer increased the length of inactive bouts.
Alternatively, if the abundance and diversity of available plant species increase when deer
density is reduced, herbivores may select more digestible and profitable forage and thus the
extra time spent searching for higher quality food may be compensated by the higher net
energy gain (Cederlund et al. 1989). On a daily basis, as deer select more profitable forage,
they require less time to process vegetation and can spend more time searching for highly
digestible plants or plant parts (Bartmann et al. 1992, WallisDeVries 1996). Under this
scenario, we should have expected inactive bouts to decrease in length as the active bouts
increase in length, the opposite of what we found. Additionally, if deer increase selectivity
as biomass increases in adjacent enclosures, we should have expected movement rates to
increase (Gates and Hudson 1983, Kohlmann and Risenhoover 1994). However, movement
rate did not change between the first and the second year of the experiment. We therefore
conclude that deer responded to an increase in plant biomass the first and second year of the
experiment by reducing the time necessary to fill their rumen and increasing the number of
foraging bouts per day. This change in their foraging behavior allowed them to gain about
25% more mass during summer than deer under natural conditions on the island, i.e. at high
density (Simard et al., in prep.).
Seasonal differences
Forage quantity and quality normally vary during and between seasons. On Anticosti, the
vegetation-growing season begins when snow melts in early May (Natural resources
Canada 2005). New shoots are rich in protein and easily digestible (Van der Wall et al.
2000). Through the growing season, vegetation on the Island increases in structural
compounds and decreases in protein content (Tremblay 1981) and thus digestibility
decreases (Robbins 1983, Van Soest 1994). Constraints due to body size and energy
requirements suggest that adults and yearlings may have to allocate their activity budgets
differently (Bunnell and Gillingham 1985). We found that through summer, juveniles and
adults responded differently to an increase in forage abundance and to a decline in plant
71
digestibility. An increase in time spent feeding through the plant growing season is normal
for young individuals and has been reported for the mouflon (Ovis musimon; Moncorps et
al. 1997) and feral goats (Capra hircus; Shi et al. 2003). Yearlings grow during summer,
their energy requirements also increase and they need more forage and thus increase
feeding time (Bunnell and Gillingham 1985). For yearlings in our study, however, the
proportion of time spent active did not increase in either density but the length of inactive
bouts increased at high density during summer. This may have been caused by a decrease in
vegetation quality and thus an increase in processing time. The proportion of time spent
active was constantly high at both densities and thus yearlings might not have had time
available to increase time spent active through summer. In addition, controlled-densities
might not have been sufficiently different to lead to a different response between densities.
Activity gradually decreased during summer in adults at low density. Abundance of
vegetation gradually increases through summer and adult deer may need to spend less time
active, searching and eating vegetation as the season progresses. Forage biomass as of mid-
July may exceed deer metabolic demands, and they could thus meet their energy
requirements by foraging fewer hours per day (Beier and McCullough 1990). This is
consistent with our interpretation above that as biomass increases deer may be as selective,
but they decrease the amount of time required to fill their rumen.
Until the end of July, deer under natural density (>20 deer/km2) had access to less
digestible forage than those under controlled densities (biomass in unfenced areas was
similar or lower than year 1 in Figure 2-1), leading to less time spent active and longer
inactive bouts necessary to digest forage. As the abundance of forage increased through
summer, deer in natural areas increased the proportion of time spent active, but still had
fewer daily activity bouts than in controlled densities. An increase in foraging time could
be expected if deer increase selectivity as vegetation becomes more abundant (Wickstrom
et al. 1984). This is also supported by the observation that active bouts became gradually
longer and inactive bouts decreased in length as summer progressed.
72
Diel activity pattern
For yearlings at both densities, the proportion of time spent active was higher at dawn but
especially at dusk, and lower during the night than during daytime. Peaks of activity at
dawn but especially at dusk have been largely documented in deer and are part of their
diurnal foraging-ruminating cycle (Kammermeyer and Marchinton 1977, Beier and
McCullough 1990). Foraging bouts usually occur following and before periods of darkness,
so deer may gather food in periods of daylight, when foraging is more profitable. As in
many other studies, we found that peaks of activity at dawn were not as constant as those at
dusk (Skogland 1983, Beier and McCullough 1990), and they probably occurred just after
dawn (Beier and McCullough 1990). The interval of 1h30 before and after sunrise may also
be too large to detect differences between dawn and daytime. We did not detect statistically
higher activity rates at dawn and dusk for adults, but activity was nonetheless 16 to 28%
higher at dusk than during other periods of the day (Table 2-3).
Conclusion
We expected that population density would influence deer foraging behavior as increased
density is often related to a decrease in forage abundance. Forage abundance did not differ
between 7.5 and 15 deer/km2 during the course of our study, and deer density had a limited
impact on deer movements and time-budgets. However, as demonstrated by inter-annual
differences in activity and activity budgets of free-ranging deer, when forage abundance is
reduced at high population density, deer activity budgets may change to compensate for a
diminution in available energy. Considering the behavioral plasticity of white-tailed deer, it
is not surprising that relationships between population density, plant quantity and foraging
behavior are complex. On one end, when deer density is high, plant biomass is low and
individual energy acquisition is limited by low forage abundance (Fowler 1981, Sæther et
al. 1996). As for deer under natural conditions on Anticosti Island, the length of inactivity
bouts are also longer because deer are limited in their digestive capabilities (Van Soest
1982). Evidences that energy acquisition is low for deer on Anticosti Island exist since deer
on Anticosti are 40% smaller (Boucher et al. 2004) and age at primiparity occurs one year
later than deer from the source population and other populations on the continent
73
(Goudreault 1980). We found that when forage increased in abundance deer increased the
length of activity bouts probably to select more digestible forage. On the other end, when
deer population density was low and forage abundant, deer increased net energy acquisition
by decreasing the length of activity bouts and increasing the number of daily activity bouts
per day. Controlled-density experiments could help to better understand the behavior of
herbivores in relation to available resources as they directly manipulate population density.
Acknowledgements
We thank L. Breton and B. Rochette from the Ministère des Ressources naturelles, et de la
Faune du Québec, as well as D. Duteau, F. Fournier, G. Picard, D. Sauvé, A. Simard, J.
Taillon, and J.-P. Tremblay for capturing deer. R. Pouliot, M. Renière, V. Viera and
especially J.-F. Therrien thankfully assisted to radiotrack deer. S. DeBellefeuille, C.
Dussault, M. Duteau, M.-A. Giroux, L. L’Italien, A. Massé, J. Taillon, J.-P. Tremblay, A.
Tousignant, and V. Viera helped with vegetation sampling. We are also most grateful to J.-
P. Tremblay for the establishment of the enclosures and biomass data. R. Weladji, S.
DeBellefeuille and many graduate colleagues reviewed an earlier draft of the manuscript.
We are also thankful to S. Baillargeon for help with the statistical analyses and to the
Centre d’études nordiques for computer support. This project was funded by Produits
forestiers Anticosti inc., the Natural Sciences and Engineering Research Council of Canada
and the Fonds québécois de la recherche sur la nature et les technologies.
Literature cited
Bartmann, R. M., G. C. White, and L. H. Carpenter. 1992. Compensatory mortality in a
Colorado mule deer population. Wildlife Monographs 121:1–39.
Beier, P., and D. R. McCullough. 1988. Motion-sensitive radio collars for estimating white-
tailed deer activity. Journal of Wildlife Management 52:11–13.
Beier, P., and D. R. McCullough. 1990. Factors influencing white-tailed deer activity
patterns and habitat use. Wildlife Monographs 109:1–51.
Belovsky, G. E. 1984. Herbivore optimal foraging: a comparative test of three models.
American Naturalist 124:97–115.
Bonham, C. D. 1989. Measurements for terrestrial vegetation. John Wiley and Sons, New
York, New York, USA.
Borkowski, J. 2000. Influence of the density of a sika deer population on activity, habitat
use, and group size. Canadian Journal of Zoology 78:1369–1374.
74
Boucher, S., M. Crête, J.-P. Ouellet, C. Daigle, and L. Lesage. 2004. Large-scale trophic
interactions: White-tailed deer growth and forest understory. Écoscience 11:286–
295.
Bryant, J. P., and P. J. Kuropat. 1980. Selection of winter forage by subarctic browsing
vertebrates: The role of plant chemistry. Annual Review of Ecology and
Systematics 11:261–285.
Bunnell, F. L., and A. S. Harestad. 1989. Activity budgets and body mammals weight in
mammals: How sloppy can mammals be? Current Mammalogy 2:245-305.
Bunnell, F. L., and M. P. Gillingham. 1985. Foraging behavior: dynamics of dining out.
Pages 53–79 in R. J. Hudson and R. G. White, editors. Bioenergetics of wild
herbivores. CRC Press, Boca Raton, Florida, USA.
Cederlund, G., R. Bergström, and F. Sandegren. 1989. Winter activity pattern of females in
two moose populations. Canadian Journal of Zoology 67:1516–1522.
Clutton-Brock, T. H., G. R. Iason, and F. E. Guinness. 1987. Sexual segregation and
density-related changes in habitat use in male and female Red deer (Cervus
elaphus). Journal of Zoology (London) 211:275–289.
Côté, S. D. 1998. In vitro digestibilities of summer forages utilized by the Rivière George
Caribou herd. Arctic 51:48–54.
Côté, S. D., T. P. Rooney, J.-P. Tremblay, C. Dussault, and D. M. Waller. 2004. Ecological
impacts of deer overabundance. Annual Review of Ecology, Evolution and
Systematics 35:113–147.
Côté, S. D., J. A. Schaefer, and F. Messier. 1997. Time budgets and synchrony of activities
in muskoxen: the influence of sex, age and season. Canadian Journal of Zoology
75:1628–1635.
Coulombe, M.-L., A. Massé, and S. D. Côté. 2006. Quantification and accuracy of activity
data measured with VHF and GPS telemetry. Wildlife Society Bulletin. In press.
DeCalesta, D. S., and S. L. Stout. 1997. Relative deer density and sustainability: a
conceptual framework for integrating deer management with ecosystem
management. Wildlife Society Bulletin 25:252–258.
Demment, M. W., and P. J. Van Soest. 1985. A nutritional explanation for body-size
patterns of ruminant and nonruminant herbivores. American Naturalist 125:641–
672.
Environment Canada. 1993. Canadian climate normals, 1961-1990- Normales climatiques
au Canada 1961-1990, volume 6-Québec. Environnement Canada, Service de
l'environnement atmosphérique, Programme climatologique canadien, Downsview,
Ont.
Focardi, S., P. Marcellini, and P. Montanaro. 1996. Do ungulates exhibit a food density
threshold? A field study of optimal foraging and movement patterns. Journal of
Animal Ecology 65:606-620.
Fowler, C. W. 1981. Density dependence as related to life history strategy. Ecology
62:602–610.
Frontier, S. 1983. Stratégies d'échantillonnage en écologie. Presses de l'Université Laval,
Québec, Québec, Canada.
Gates, C. C., and R. J. Hudson. 1983. Foraging behaviour of wapiti in a boreal forest
enclosure. Canadian-Field Naturalist 110:197–206.
75
Georgii, B. 1981. Activity patterns of female red deer (Cervus elaphus L.) in the Alps.
Oecologia 49:127–136.
Gillingham, M. P., and F. L. Bunnell. 1985. Reliability of motion-sensitive radio collars for
estimating activity of black-tailed deer. Journal of Wildlife Management 49:951–
958.
Gillingham, M. P., K. L. Parker, and T. A. Hanley. 1997. Forage intake by black-tailed deer
in a natural environment: bout dynamics. Canadian Journal of Zoology 75:1118–
1128.
Gordon, I. J., A. J. Hester, and M. Festa-Bianchet. 2004. The management of wild large
herbivores to meet economic, conservation and environmental objectives. Journal of
Applied Ecology 41:1021–1031.
Goudreault, F. 1980. Reproduction et dynamique de la population de cerfs de l'île
d'Anticosti. M.Sc thesis. Université Laval, Québec, Québec, Canada.
Hanley, T. A. 1984. Habitat patches and their selection by wapiti and black-tailed deer in a
coastal montane coniferous forest. Journal of Applied Ecology 21:423–436.
Healy, W. M., D. S. deCalesta, and S. L. Stout. 1997. A research perspective on white-
tailed deer overabundance in the northeastern United States. Wildlife Society
Bulletin 25:259–263.
Henriksen, S., R. Aanes, B.-E. Sæther, T. H. Ringsby, and J. Tufto. 2003. Does availability
of resources influence grazing strategies in female Svalbard reindeer? Rangifer
23:25–37.
Hester, A. J., L. Edenius, R. M. Buttenschøn, and A. T. Kuiters. 2000. Interactions between
forests and herbivores: the role of controlled grazing experiments. Forestry 73:381–
391.
Huot, J. 1982. Body condition and food resources of white-tailed deer on Anticosti Island,
Québec. Ph.D. thesis. University of Alaska, Fairbanks, Alaska, USA.
Illius, A. W., and I. J. Gordon. 1987. The allometry of food intake in grazing ruminants.
Journal of Animal Ecology 56:989–999.
Jeschke, J., and R. Tollrian. 2005. Predicting herbivore feeding times. Ethology 111:187 –
206.
Jorgenson, J. T., M. Festa-Bianchet, J.-M. Gaillard, and W. D. Wishart. 1997. Effects of
age, sex, disease, and density on survival of bighorn sheep. Ecology 78:1019–1032.
Kammermeyer, K. E., and R. L. Marchinton. 1977. Seasonal change in circadian activity of
radio-monitored deer. Journal of Wildlife Management 41:315–317.
Kohlmann, S. G., and K. L. Risenhoover. 1994. Spatial and behavioral response of white-
tailed deer to forage depletion. Canadian Journal of Zoology 72:506–513.
Lesage, L., M. Crête, J. Huot, and J.-P. Ouellet. 2001. Evidence for a trade-off between
growth and body reserves in northern white-tailed deer. Oecologia 126:30–41.
McLean, R. A., W. L. Sanders, and W. W. Stroup. 1991. A unified approach to mixed
linear models. The American Stastistician 45:54–64.
Moen, A. N. 1978. Seasonal changes in heart rates, activity, metabolism, and forage intake
of white-tailed deer. Journal of Wildlife Management 42:715–738.
Moncorps, S., P. Boussès, D. Réale, and J.-L. Chapuis. 1997. Diurnal time budget of the
mouflon (Ovis musimon) on the Kerguelen archipelago: influence of food resources,
age, and sex. Canadian Journal of Zoology 75:1828–1834.
76
Murray, M. G. 1991. Maximizing energy retention in grazing ruminants. Journal of Animal
Ecology 60:1029–1045.
Mysterud, A. 1998. The relative roles of body size and feeding type on activity time of
temperate ruminants. Oecologia 113:442–446.
Natural resources Canada. 2005. The atlas of Canada. Available at
http://atlas.gc.ca/site/english/maps/environment/climate/snowcover/endsnowcover
[accessed on 20 august 2005].
Owen-Smith, N., and P. Novellie. 1982. What should a clever ungulate eat? American
Naturalist 119:151–178.
Palo, R. T. 1985. Chemical defence in birch: Inhibition of digestibility in ruminants by
phenolic extracts. Oecologia 68:10–14.
Pérez-Barbería, F. J., and I. J. Gordon. 1999. The relative roles of phylogeny, body size and
feeding style on the activity time of temperate ruminants: a reanalysis. Oecologia
120:193–197.
Potvin, F., P. Beaupré, and G. Laprise. 2003. The eradication of balsam fir stands by white-
tailed deer on Anticosti Island, Québec: A 150-year process. Écoscience 10:487–
495.
Potvin, F., and L. Breton. 2005. Testing 2 aerial survey techniques on deer in fenced
enclosures - visual double-counts and thermal infrared sensing. Wildlife Society
Bulletin 33:317–325.
Putman, R. J., J. Langbein, A. J. M. Hewison, and S. K. Sharma. 1996. Relative roles of
density-dependent and density-independent factors in population dynamics of
British deer. Mammal Review 26:81–101.
Relyea, R. A., I. M. Ortega, and S. Demarais. 1994. Activity monitoring in mule deer:
Assessing telemetry accuracy. Wildlife Society Bulletin 22:656–661.
Risenhoover, K. L. 1986. Winter activity patterns of moose in interior Alaska. Journal of
Wildlife Management 50:727–734.
Robbins, C. T. 1983. Wildlife feeding and nutrition. Academic Press, New York, New
York, USA.
Rochette, B. A., A. Gingras, and F. Potvin. 2003. Inventaire aérien du cerf de Virginie de
l'île d'Anticosti- été 2001. Société de la faune et des parcs du Québec, Direction de
l'aménagement de la faune de la Côte-Nord et direction de la recherche sur la faune,
Québec, Québec, Canada.
Rouleau, I., M. Crête, and J.-P. Ouellet. 2002. Contrasting the summer ecology of white-
tailed deer inhabiting a forested and an agricultural landscape. Écoscience 9:459–
469.
Ruckstuhl, K. E. 1997. Foraging behaviour and sexual segregation in bighorn sheep.
Animal Behaviour 56:99–106.
Sæther, B.-E., R. Andersen, O. Hjeljord, and M. Heim. 1996. Ecological correlates of
regional variation in life history of the moose Alces Alces. Ecology 77:1493–1500.
Shi, J., R. I. M. Dunbar, D. Buckland, and D. Miller. 2003. Daytime activity budgets of
feral goats (Capra hircus) on the Isle of Rum: influence of season, age, and sex.
Canadian Journal of Zoology 81:803–815.
Skogland, T. 1983. The effects of density dependent resource limitation on size of wild
reindeer. Oecologia 60:156–168.
77
Spalinger, D. E., C. T. Robbins, and T. A. Hanley. 1986. The assessment of handling time
in ruminants: the effect of plant chemical and physical structure on the rate of
breakdown of plant particles in the rumen of mule deer and elk. Canadian Journal of
Zoology 64:312–321.
Tilghman, N. G. 1989. Impacts of white-tailed deer on forest regeneration in northwestern
Pennsylvania. Journal of Wildlife Management 53:524–532.
Tremblay, J. 1981. Qualité de la nourriture d'été du cerf de Virginie à l'île d'Anticosti.
M.Sc. thesis. Université Laval, Québec, Québec, Canada.
Tremblay, J.-P. 2005. Relations entre les perturbations induites par les cervidés et la
dynamique de régénération des écosystèmes forestiers boréaux. Ph.D. thesis.
Université Laval, Québec, Québec, Canada.
Trudell, J., and R. G. White. 1981. The effect of forage structure and availability on food
intake, biting rate, bite size and daily eating time of reindeer. Journal of Applied
Ecology 18:63–81.
Van der Wall, R., N. Madan, S. van Lieshout, C. Dormann, R. Langvatn, and S. D. Albon.
2000. Trading forage quality for quantity? Plant phenology and patch choice by
Svalbard reindeer. Oecologia 123:108–115.
Van Soest, P. J. 1982. Nutritional ecology of the ruminant: ruminant metabolism,
nutritional strategies, the cellulolytic fermentation, and the chemistry of forages and
plant fibers. O & B Books, Corvallis, Oregon, USA.
Van Soest, P. J. 1994. Nutritional ecology of the ruminant, 2nd edition. Cornell University
Press, Ithaca, New York, USA.
WallisDeVries, M. F. 1996. Effects of resource distribution patterns on ungulate foraging
behaviour: a modelling approach. Forest Ecology and Management 88:167–177.
White, G. C., and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic
press, San Diego, California, USA.
Wickstrom, M. L., C. T. Robbins, T. A. Hanley, D. E. Spalinger, and S. M. Parish. 1984.
Food intake and foraging energetics of elk and mule deer. Journal of Wildlife
Management 48:1285–1301.
78
Chapitre 3. Influence of forage abundance, cover and
population density on white-tailed deer space use
79
Résumé
L’influence de la densité de population sur la répartition du cerf de Virginie dans l’espace
en relation avec la biomasse et le couvert disponibles a été mesurée dans 3 blocs formés de
2 enclos contenant respectivement 7.5 (faible densité) et 15 cerfs/km2 (haute densité). La
biomasse et le couvert latéral et vertical ont été interpolés par krigeage et les localisations
télémétriques des cerfs ont été divisées en 3 périodes quotidiennes. Pendant l’aube et le
crépuscule, aux deux densités, l’utilisation de l’espace était positivement reliée à la
biomasse de la végétation. À faible densité, la répartition des cerfs était aussi positivement
reliée au couvert vertical à l’aube et au crépuscule. Pendant le jour, la distribution des cerfs
était positivement reliée à la biomasse mais seulement à haute densité où les cerfs
réduisaient aussi l’utilisation d’endroits avec un couvert latéral dense. Cette étude souligne
l’influence de la densité de population et de la période du jour sur la répartition des cervidés
dans l’espace ainsi que la valeur des expériences en densités contrôlées et de la
géostatistique pour mieux comprendre les facteurs qui influencent le comportement
d’approvisionnement des grands herbivores.
80
Abstract
Deer population densities are increasing in many areas of North America and Europe and
are strongly affecting species composition and structure of plant communities. The effects
of increased density on deer foraging behavior, however, have received little attention. We
assessed the influence of population density on deer space use in relation to vegetation
abundance and cover. We quantified space use in 3 blocks with 2 enclosures each
containing 3 radio-collared deer at densities of 7.5 (low density) and 15 deer/km2 (high
density), respectively. Vegetation biomass, canopy and lateral cover were interpolated by
kriging and deer observations were divided into 3 periods: dawn/dusk, day, and night. At
dawn and dusk, in both densities, space use was positively related to forage abundance.
During the day, deer space use was also positively related to forage abundance but only at
high density. At low density, habitat use was positively related to canopy cover during
dawn/dusk. Deer decreased the use of areas with dense lateral cover during the day at high
density, but no relationship was found at low density. Contrarily to our prediction, deer did
not use open habitats more frequently at night than during the day. This study underlines
the effect of population density and diel periods on the utilization of available resources and
the value of controlled density experiments and geostatistics to disentangle the factors
affecting foraging behavior of large herbivores.
81
Introduction
In summer, northern ungulates typically devote most of their time to finding and consuming
food (Beier and McCullough 1990). Foraging decisions are thus important to an herbivore’s
energy and time budget. When food acquisition is the primary determinant of patch
selection, use of habitat patches should be positively related to forage availability
(MacArthur and Pianka 1966, McNamara et al. 1993). There are, however, several other
constraints such as heat or wind exposure (Belovsky 1981) and predation risk (Berger
1991) that can affect foraging behavior independently of forage characteristics. For cervids,
a good foraging site is usually represented by a trade-off between proximity to protective
cover and abundance of vegetation (Kotler et al. 1994).
Cover may be divided into two components: (1) the canopy cover corresponds to the
projection of the tree crowns to the ground and (2) the lateral cover which is made up of
concealing understory vegetation or topography. Lateral cover reduces predation risk and
thus time devoted to vigilance (Altendorf et al. 2001) and, in the absence of predators, has
been considered to play a “psychological” role in habitat selection related to the ghosts of
predators past (Byers 1997, Mysterud and Østbye 1999). Animals are also typically
exposed to milder weather conditions (e.g. temperature, wind, precipitations) in closed
habitats than in open habitats (Mysterud and Østbye 1999). Open sites generally offer more
abundant forage in summer (Hanley 1984), but present a higher risk of predation (Tufto et
al. 1996) and higher thermoregulatory costs due to an increased heat load (Beier and
McCullough 1990). Cervids thus often prefer feeding near edges of forests and open
habitats because edges minimize the trade-off between exposition to predators and/or harsh
weather, and forage abundance (Keay and Peek 1980, Tufto et al. 1996).
Clear-cuts provide open habitats that offer abundant food resources to deer and they are
usually interspersed with forest stands that present dense canopy cover but low forage
availability (Masters et al. 1993). Deer select cutblocks when clearings produce higher food
resources than forest stands and if they provide sufficient hiding cover (Lyon and Jensen
1980). Clearings have been used many times to enhance forage production and thus
improve deer habitat conditions (Masters et al. 1993). Tierson et al. (1985) found that deer
82
stopped migrating to traditional winter ranges to feed in recently logged areas but summer
home ranges were not modified by the presence of cutblocks.
It has been proposed that population density modulates the trade-off between using habitats
rich in forage and habitats rich in cover (Mysterud and Østbye 1999) because population
density is generally negatively related to forage abundance (Healy et al. 1997) and
increases intraspecific competition (Clutton-Brock et al. 1982). Indeed, Lesage et al. (2000)
found that white-tailed deer generally increased the use of forests over agricultural fields
when competition for forage in forest stands was low but, as deer density increased, the use
of open areas increased, likely because forage was more abundant in open habitats. Another
study showed that because of the low availability of forage in the forest due to high
population density, deer adapted to feeding in agricultural crops at night (Rouleau et al.
2002). In agricultural landscapes, population density and landscape composition may thus
affect the degree to which deer feed on crops. An increase in the use of open habitats, such
as agricultural fields, in summer may indicate the low abundance of forage in areas with
adequate cover and the impacts of high density on space use (Mysterud and Østbye 1999).
Population densities of many cervid populations in North America are rapidly increasing
(Côté et al. 2004). Although of high interest, it is unknown whether increasing density is
modifying deer behavior, especially in relation to trade-offs between selection for forage or
cover in the context of landscapes composed of forests and clear-cuts.
Controlled-browsing experiments have been used for many years to study the foraging
behavior of domestic animals and their use is now advocated for wild ungulates (Hester et
al. 2000). Controlled-browsing studies generally investigate the effects of different
herbivore densities on vegetation abundance and diversity (Tilghman 1989, Hester et al.
2000); however, they can also be very useful tools to understand the effects of population
density on the foraging behavior of wild ungulates.
Deer use open habitats more often during the night than during the day (Beier and
McCullough 1990, Rouleau et al. 2002) and since their activity peaks at dawn and dusk
(Beier and McCullough 1990), the trade-off between using forage or cover patches may
83
also depend on diel periods (Mysterud et al. 1999). Therefore, our objective was to address
the influence of population density on deer space use in relation to vegetation abundance
and cover by experimentally controlling population density in large enclosures in which we
obtained 2 replicated densities. We also examined how daylight and activity peaks
influence the forage/cover trade-off by separating our radiolocations into 3 daily periods
(dawn/dusk, day, night). We predicted that deer at high density would use open areas of the
enclosures where forage is more abundant independently of cover characteristics. However,
deer at low density should use areas of the enclosures in relation to available cover because
competition for forage is less influential at low density.
Study area
Anticosti Island (Québec, Canada, 49° 28’ N, 63° 00’ W) is located at the northern fringe of
the white-tailed deer range in North America and covers 7,943 km2. Forests are naturally
dominated by balsam fir (Abies balsamea), white spruce (Picea glauca) and black spruce
(P. mariana). White birch (Betula papyrifera) and trembling aspen (Populus tremoloides)
are irregularly found on the island. About 220 deer were introduced on the island at the turn
of the 19th century. In the absence of predation, the population spread and grew rapidly.
Today, deer densities of >20 deer/km² are found in most areas on the island (Potvin and
Breton 2005). Deer have modified the original forest and greatly reduced the abundance of
deciduous woody vegetation on the island (Potvin et al. 2003, Tremblay et al. 2005). The
climate of Anticosti is maritime and characterized by longer and milder winters compared
to the white-tailed deer range on the continent (Huot 1982). Mean temperatures are -12°C
in January and 15°C in July, snow precipitation averages 406 cm annually and rainfall 630
mm (Environment Canada 1993).
Methods
Experimental design
Our experimental design consists in 3 sets of enclosures (A, B, C) in which we introduced
24 deer during 2 different years. Enclosures were located in balsam fir dominated forests
that were partially cut in the early summer of 2001. Water was easily accessible to deer at
84
many streams and artificial water holes in every enclosure. One block (A) was erected in
the western part of the island near Simonne Lake and two blocks (B and C) were erected
130 km to the east in the central part of the island near the Jupiter River. Between 30 and
40% of residual forest stands of different sizes (0.19 ha - 21.6 ha) were left in the
enclosures. In 2002, one site (A) was studied and in 2003, the 3 sites were studied. To test
the influence of deer density, the blocks were divided in two enclosures to obtain densities
of 7.5 deer/km2 (40 ha enclosure with 3 deer; LDE) and 15 deer/km2 (20 ha enclosure with
3 deer; HDE). We used different animals in 2002 and 2003 (Table 3-1).
Deer captures
In late June, we fitted 6 deer in 2002 and 18 deer in 2003 with VHF collars (LMRT series)
equipped with sto-2a variable pulse activity sensors (Table 3-1). We used different methods
to capture deer: dart guns, Stephenson box traps and cannon nets baited with cattle feed and
balsam fir twigs. Deer were released in the study enclosures shortly after capture. The
Animal Care and Use Committee of Université Laval, Québec, Canada (2005-008)
approved all capture methods. We verified reproductive status of adult females by direct
observation at capture and at the end of summer. Since only 2 monitored females had a
fawn, we did not include reproductive status in the analyses.
Telemetry
In July and August 2002, we radiotracked 6 deer in block A and in July and August 2003,
we radiotracked 16 deer in the 3 blocks (Table 3-1). One adult male lost its collar and 1
yearling male had a malfunctioning collar. Deer were located with telemetric receivers
(SRX-400 version W9, Lotek Engineering, Newmarket, Ontario, Canada and a TR-2
scanner/receiver, Telonics, Meza, Arizona, USA), unidirectional antennas and compasses.
Telemetry stations were positioned with a GPS Garmin (Garmin international, Olathe,
Kansas, USA; precision of <5 m) on trails adjacent to the enclosures. To limit human
disturbance, trails were generally located more than 100 m away from the enclosures. At
least 3 azimuths differing by a minimum of 30º were obtained by moving between stations
with a vehicle (White and Garrott 1990). To reduce location error, positioning had to be
completed within 15 minutes (White and Garrott 1990). 24-h days were evenly divided
85
Table 3–1. Number of locations recorded in each diel period for radiocollared white-tailed
deer tracked in controlled-density enclosures on Anticosti Island, Québec.
Number of locations
ID Year Block Density Age Sex Dawn/dusk Day Night
(deer/km²)
1 2002 A 7.5 Adult Male 71 124 55
2 Yearling Male 68 125 56
3 Yearling Female 52 108 37
4 15 Yearling Female 66 117 62
5 Yearling Female 69 123 58
6 Yearling Female 73 125 56
7 2003 A 7.5 Adult Female 24 46 15
8 Yearling Female 20 43 21
9 Adult Female 22 50 18
10 15 Adult Female 27 47 19
11 Adult Male 25 38 18
12 Yearling Male 0 0 0
13 2003 B 7.5 Yearling Female 21 44 16
14 Yearling Male 19 48 17
15 Adult Female 20 42 17
16 15 Adult Female 19 46 21
17 Yearling Female 17 48 16
18 Adult Female 22 46 16
19 2003 C 7.5 Yearling Female 24 41 22
20 Adult Female 28 43 21
21 Yearling Male 22 34 26
22 15 Adult Male 24 43 21
23 Yearling Female 21 46 21
24 Adult Male 0 0 0
86
into 3 periods of 8 hours (8h00–16h00, 16h00-0h00 and 0h00–8h00). These 8-hour periods
were evenly sampled and rotated between 2 observers and between groups of enclosures
every 3 days. During sampling periods of 8 hours, deer were positioned about every 2
hours.
LOAS software (Location Of A Signal Version 2.07, Ecological Software Solutions,
Schwägalpstrasse, Urnäsch, Switzerland) was used to estimate positions and error
polygons. Error polygons were calculated with “Andrews” estimator. All locations were
plotted with LOAS software on maps and were assigned Universal Transverse Mercator
(UTM) coordinates. The average error from plotted to actual locations was determined by
using control transmitters set at known locations throughout the enclosures and was
estimated at 107 m (SE = 88 m; n = 88 trials). We removed locations with error polygons
greater than 0.01 ha. After processing, we kept 2,916 locations from the 3,251 original
locations. Positions were assigned to 3 diel-periods: dawn and dusk (1h30 before sunrise to
1h30 after sunrise, and 1h30 before sunset to 1h30 after sunset), day (1h30 after sunrise to
1h30 before sunset), and night (1h30 after sunset to 1h30 before sunrise).
Biomass and cover sampling
To characterize uniformly the vegetation of the enclosures, 5 sampling points were drawn
with the « Generate-randomly distributed points » extension of ArcView GIS (ArcView
GIS Version 3.1, Environmental systems research institute, Redlands, California, USA) in
every 2 ha squares of a grid superposed to each enclosure. We found the sampling stations
in the field with a GPS Garmin. At each sampling point, percent of plant cover was
estimated in two 1-m2 quadrats randomly chosen in a 10×10 m quadrat centered at the
sample point. In block A, the same sampling points were used in 2002 and 2003.
Plant biomass was estimated for every major plant component of deer diet and the most
abundant species on Anticosti (Huot 1982) using regressions between percent of plant
cover and mass of the corresponding dried plant biomass (Bonham 1989). We analysed the
following species: Abies balsamea, Betula papyrifera, Cirsium spp., Coptis groenlandica,
Cornus canadensis, Epilobium angustifolium, Grass sp., Hieracium sp., Maianthemum
87
canadense, Picea glauca, Rubus idaeus, Rubus pubescens and, Trientalis borealis. The
number of samples needed for regressions was estimated empirically by plotting regression
coefficients with number of samples until an asymptote was reached (Frontier 1983). We
summed the biomass values of all plants for each quadrat and used the mean value of the
two quadrats for each sample point in the analyses.
At each sampling point, canopy cover was estimated by vertically projecting foliage (>4 m
trees) to 20 points distributed equally on the ground every 3 m in four directions (east,
southeast, southwest and west) from the center of the sampling unit. Each point was judged
as with cover or not and canopy cover corresponded to the sum of all sampled directions
(value of 1 for each point with cover). Lateral cover was measured with a cover board (2.5
m×0.3 m divided in 0.5 m sections) in 2 opposite directions by attributing board
concealment to 4 classes (1: 0-25; 2: 26-50; 3: 51-75; 4: 76-100%; Nudds 1977). We used
the mean value from the first two sections of the board (0-1 m) and values from both
directions were averaged.
Analyses
Data points could not account for spatial relationships and spatial correlation between
biomass abundance and cover. We thus estimated biomass and cover abundances with a
geostastistical software (Geostatistical analyst; ArcMap 9.0, Environmental systems
research institute, Redlands, California, USA). Geostatistics is a branch of applied statistics
that focuses on the detection, modeling, and estimation of spatial patterns in spatially
correlated data (Rossi et al. 1992). Spatial autocorrelation occurs because samples collected
closer to each other are more similar than samples collected farther apart. This particularly
occurs when the variable sampled is spatially structured (e.g. in patches). In their simplest
form, geostatistics involve 2 steps: 1) characterizing the spatial structure of the variable
with a variogram, thus defining the degree of autocorrelation between the data points; and
2) predicting values between measured points based on the estimated degree of
autocorrelation (Robertson 1987). Semivariance is the average measure of the variance
associated to any two sampled points in a given distance class. For example, with a lag size
of 50 m, a mean variance value would be obtained for each distance class (0-50 m, 51-100
88
m, 101-150 m, etc.). In spatially structured data, changes of semivariance represented in a
variogram usually augment with increasingly separated points (Cressie 1993). Variograms
may also include a directional value to weight for directional changes in the factor (e.g.
slope direction; Cressie 1993). The semivariogram models provide the following
parameters: 1) the nugget effect which indicates the residual spatial variability below the
lag size that cannot be modeled with the current sampling resolution, 2) the sill, which
defines the asymptotic value of semivariance; and 3) the range, defined as the distance over
which autocorrelation is present (Cressie 1993).
Semivariograms were prepared individually for forest stands and cuts of each enclosure
because vegetation and cover drastically change between these two habitats (Masters et al.
1993). Biomass values were log-transformed to normalize data. By trial and error, we
determined the best fit of spherical variograms until a maximum lag distance of 125 m (i.e.,
half of the minimum enclosure dimension) in 5-m increments and in all directions (Jurado-
Expòsito et al. 2004). The best-fitted values, determined by cross-validation results, of
nugget, sill and range were calculated and recorded for further analysis in ordinary kriging
(Cressie 1993). Ordinary kriging is an interpolation technique that uses observed values
associated with X and Y coordinates and estimates values for all locations within the
sampled coordinates with the help of a variogram describing how values change through
space (Cressie 1993). We used kriging values to validate the fitted variogram through cross-
validation. This procedure is based on the systematic removal of observations, one by one,
from the raw data set, which is then estimated by kriging (Isaaks and Svriastava 1989).
Kriging provides an error term for each estimated value, thus giving a measure of reliability
for the interpolations. Biases in estimation errors were evaluated using the standardised root
mean squared error (Appendix 3−1; RMSE; Isaaks and Svriastava 1989). The nugget value
divided by the total variance (sill) gives an estimation of the spatial dependence (Appendix
3−1; Jurado-Expòsito et al. 2004).
Rettie and McLoughlin (1999) recommended the use of buffers to account for telemetry
error in habitat selection studies. In our study, the error of locations obtained by telemetry
was 107 m on average and thus larger than most forest stands present in the enclosures. In
89
70% of the trials with hidden collars at known locations, collars were located less than 100
m away from the real position. Furthermore, up to 100 m, the probability of obtaining a
position closer to the real position was not a function of distance (F1,9 = 0.11; P = 0.75).
We thus took into account the mean location error by placing a 100 m buffer around each
location. Since the buffers were quite large compared to the size of the enclosures, they
overlapped considerably and were thus dependent on one another. To take this into account,
we randomly placed 1 point in every 150×150-m square of a grid placed over each
enclosure with the extension “Simple random sample” (ArcView GIS, Version 3.1,
Environmental systems research institute, Redlands, California, USA). We used 150×150-
m squares because they were sufficiently large compared to the buffer size and allowed for
a reasonable number of sampling points for the regression (20-30 points). For each of the 3
diel periods, we counted the total number of overlapping buffers at each point for each deer
with the extension “dissect overlaps” (ArcView GIS, Version 3.1, Environmental systems
research institute, Redlands, California, USA). We then divided the total number of
overlapping buffers for each random point by the total number of locations for this deer
during each diel period to describe a relative use of the enclosures that was independent of
the number of positions taken on an individual.
Resources and relative use (mean relative number of overlapping buffers per diel-period)
were evaluated at random points distributed across the total surface of the enclosures. We
estimated relationships between relative use and resource abundance by means of
regression analyses. We did not consider resource availability in the models because a
random and independent sample of points were already considered in the analysis. For each
diel period, the relationship between relative use for all deer in a particular enclosure and
biomass, lateral and canopy cover was quantified using a linear mixed model with block
and year as random factors (Proc Mixed, SAS Version 9.1, SAS Institute Inc., Cary, North
Carolina, USA). Correlation coefficients between factors and absolute values were all <0.4,
we thus included all variables in the model. We performed regressions and then compared
slopes between densities and determined if slopes differed from 0 for each density. As deer
may often use sites with intermediate biomass because plants of the early phenological
90
stages have more nutrients and offer a high rate of energy intake (Stewart et al. 2000), we
also tested if the relationship between biomass abundance and mean relative use was
parabolic and if it varied between densities. We did not have enough animals in every
block, density and year to include sex in the models, but we contend that sex may be taken
into consideration in further studies because space use in relation to food or cover may vary
between sexes (Beier and McCullough 1990, DePerno et al. 2003).
To compare resource abundance between densities and stratum (residual forest stands or
clear-cuts), we compared the mean values of total biomass, vertical cover and lateral cover
available at the random points located on the sampling grid with an analysis of variance
with block and year as random factors. As only one block was studied for two years, year
could not be treated as a fixed effect. We verified the normality of the residuals and the
homogeneity of variance by visual examinations of the residual plots. Significance levels
were 0.05.
Results
Spatial analysis
The relationships between space and biomass, lateral cover and canopy cover varied
between forests and cuts, and between enclosures (Appendix 3−1). RMSE values ranged
from 0.27 to 2.38 but most values were close to 1 ( x = 1.01; SE = 0.30; n = 48 models;
Appendix 3−1), indicating that variability was generally not biased towards higher or lower
values than those measured. The nugget effect was greater than zero in many cases,
indicating that observations separated by small distances were highly variable (Isaaks and
Srivastava 1989; Appendix 3−1). A general feature of the variograms was the relatively
close nugget and C values (i.e. the variance explained by the spatial variation in the data;
Appendix 3−1), indicating that spatial autocorrelation was imperceptible in some cases.
Spatial relationships, however, provided a good representation of how biomass
(Figure 3-1), lateral (Figure 3-2) and canopy cover (Figure 3-3) varied within clear-cuts and
forest stands and allowed us to examine how
91
Forage abundance (g/m²)
0 – 20
20 – 60
60 – 140
140 – 230
230 – 380
a) Dawn/dusk relative use
0.00 – 0.02
0.02 – 0.06
0.06 – 0.10
0.10 – 0.18
0.18 – 0.30
0 100 20050 Meters
b) Daylight relative use
0.01 – 0.03
0.03 – 0.05
0.05 – 0.08
0.08 – 0.13
0.13 – 0.18
92
Figure 3–1. Plant biomass (g/m²) available to white-tailed deer and interpolated by kriging
in 2003 for Block A on Anticosti Island, Québec. Semivariogram values are in
Appendix 3−1. Relative use was measured in every 150×150 m square of a grid superposed
on the Block by drawing a random sampling point and calculating its relative use (i.e. the
number of overlapping buffers for a deer divided by the total number of positions for that
deer in every diel-period) at this location. Use was averaged for the deer at 7.5 deer/km²
and at 15 deer/km² during dawn and dusk (a), during the day (b) and at night (c).
c) Nighttime relative use
0.00 – 0.02
0.02 – 0.04
0.04 – 0.07
0.07 – 0.11
0.11 – 0.18
93
b) Daylight relative use
0.01 – 0.03
0.03 – 0.05
0.05 – 0.08
0.08 – 0.13
0.13 – 0.18
a) Dawn/dusk relative use
0.00 – 0.02
0.02 – 0.06
0.06 – 0.10
0.10 – 0.18
0.18 – 0.30
Lateral cover (/4)
0.40
0.4 – 0.8
0.8 – 1.2
1.2 – 1.6
1.6 – 2.0
2.0 – 2.4
0 100 20050 Meters
94
Figure 3–2. Lateral cover, or mean concealment (attributed to 4 classes 1: 0-25; 2: 26-50; 3:
51-75; 4: 76-100%) of the first 2 sections of a concealment board (2.5 m×0.3 m divided in
0.5 m sections) in 2 opposite directions, available to white-tailed deer and interpolated by
kriging in 2003 for Block A on Anticosti Island, Québec. Semivariogram values are in
Annex 3−1. Relative use was measured in every 150×150 m square of a grid superposed on
the block by drawing a random sampling point and calculating its relative use (i.e. the
number of overlapping buffers for a deer divided by the total number of positions for that
deer in every diel-period) at this location. Use was averaged for the deer at 7.5 deer/km²
and at 15 deer/km² during dawn and dusk (a), during the day (b) and at night (c).
c) Nighttime relative use
0.00 – 0.02
0.02 – 0.04
0.04 – 0.07
0.07 – 0.11
0.11 – 0.18
95
b) Daylight relative use
0.01 – 0.03
0.03 – 0.05
0.05 – 0.08
0.08 – 0.13
0.13 – 0.18
a) Dawn/dusk relative use
0.00 – 0.02
0.02 – 0.06
0.06 – 0.10
0.10 – 0.18
0.18 – 0.30
0 100 20050 Meters
Canopy cover (/20)
0.0 – 0.7
0.7 – 2.4
2.4 – 4.9
4.9 – 8.5
8.5 – 13.7
96
Figure 3–3. Canopy cover, or proportion of 20 points set at every 3 m from the center of
each sampling unit in 4 directions (east, southeast, southwest and west) where foliage of
>4 m trees was present, available to white-tailed deer and interpolated by kriging in 2003
for Block A on Anticosti Island, Québec. Semivariogram values are in Annex 3−1. Relative
use was measured in every 150×150 m square of a grid superposed on the block by drawing
a random sampling point and calculating its relative use (i.e. the number of overlapping
buffers for a deer divided by the total number of positions for that deer in every diel-period)
at this location. Use was averaged for the deer at 7.5 deer/km² and at 15 deer/km² during
dawn and dusk (a), during the day (b) and at night (c).
c) Nighttime relative use
0.00 – 0.02
0.02 – 0.04
0.04 – 0.07
0.07 – 0.11
0.11 – 0.18
97
these variables could be related to the utilization of the enclosures by deer.
Descriptive statistics
Vegetation abundance was negatively related to canopy cover at both densities (LDE:
r = -36, P < 0.01, n = 98 random points; HDE: r = -0.21, P = 0.02, n = 60 random points).
Lateral cover, however, was not significantly related to biomass abundance at both
densities (LDE: r = 0.16, P = 0.11, n = 98 random points; HDE: r = 0.15, P = 0.24, n = 60
random points). Biomass in the enclosures did not vary with density but, as expected, was
greater in clear-cuts than under forest cover (Table 3-2). Lateral and canopy cover were
positively correlated in the low-density enclosures (r = 0.36, P < 0.01, n = 98 random
points) but not in the high-density enclosures (r = 0.14, P = 0.27, n = 60 random points).
Lateral and canopy cover did not differ among the enclosures but obviously, canopy cover
was more important in forest stands than in cuts. Relative use was higher in the high-
density enclosures than in the low-density enclosures (Figure 3-4), likely because the
enclosures at high density were twice smaller than at low density. Relative use between all
3 diel periods were correlated at both densities (LDE; r’s > 0.47, P’s < 0.01; HDE;
r’s > 0.46, P’s < 0.01).
Deer space use
At dawn and dusk, relative use increased with biomass at both densities and the relationship
between relative use and canopy cover differed between densities (Table 3-3). The slope
parameter was positive for deer in low-density enclosures (t144 = 2.66; P < 0.01) but was
not different from 0 for deer in high-density enclosures (t144 = -1.25; P = 0.21; Figure 3-4c),
indicating that deer at low density used sites with dense canopy cover more often than deer
at high density. In no diel periods or densities did the quadratic term of biomass was related
to mean relative use (F’s1, 142 < 2.89; P’s > 0.09). At dawn and dusk, deer did not select
areas with denser lateral cover at any density (Table 3-3). During the day, the slope
parameter relating deer space use to biomass abundance for deer in high-density enclosures
was positive (t144 = -1.70; P = 0.09), but the slope for deer in low-density enclosures did not
differ from zero (t144 = -1.13; P = 0.25; Figure 3-4d), indicating that deer at high density
used more often sites with high biomass during the day than sites with low
98
Figure 3–4. Relationships between white-tailed deer relative space use (number of
overlapping buffers for a deer divided by the total number of positions for that deer in every
diel-period and at each random point) and plant biomass, lateral (mean concealment value
attributed by 4 classes of 25%) and canopy cover (proportion of 20 points where foliage of
>4 m trees was present). Results are shown for dawn and dusk (a to c), day (d to f) and
night (g to i). Data for biomass and cover were obtained by kriging and taken at one
random location in each 150×150 m square of a grid superposed on enclosures containing
7.5 deer/km² (solid circles and line) or 15 deer/km² (empty circles and dashed line) on
Anticosti Island, Québec. For each diel period, regression lines were obtained from mixed
models relating relative use to biomass, canopy cover and lateral cover for each density,
with block and year as random factors (statistics are in Table 3-3).
99
Table 3–2. Mean biomass (g/m²), lateral covera (/20 points) and canopy coverb (/4; ± SD) according to deer density and stratum
(forest stands and clear-cuts). Deer were kept in 3 sets of enclosures with 2 densities each on Anticosti Island, Québec.
ANOVAs
Density (deer /km²) Clear-cut Forest Factor DF F Value P-value
a) Biomass 7.5 81.3 ± 38.6 29.6 ± 12.1 Density 1,3 0.17 0.71
15 68.9 ± 20.2 34.1 ± 4.4 Stratum 1,6 20.10 <0.01
Density × Stratum 1,6 0.76 0.42
b) Lateral cover 7.5 2.7 ± 0.9 3.0 ± 0.8 Density 1,3 7.30 0.07
15 3.1 ± 0.8 3.3 ± 0.4 Stratum 1,6 2.13 0.19
Density × Stratum 1,6 0.04 0.84
c) Canopy cover 7.5 1.0 ± 0.4 9.8 ± 3 Density 1,3 1.00 0.39
15 1.8 ± 3.0 6.8 ± 2 Stratum 1,6 38.06 <0.01
Density × Stratum 1,6 2.97 0.14
a Mean concealment value (attributed to 4 classes 1: 0-25; 2: 26-50; 3: 51-75; 4: 76-100%) of the first 2 sections of a
concealment board (2.5 m×0.3 m divided in 0.5 m sections) in 2 opposite directions.
b Proportion of 20 points set at every 3 m from the center of each sampling unit in 4 directions (east, southeast, southwest and
west) where foliage of >4 m trees was present.
100
Table 3–3. White-tailed deer relative space usea according to biomass, canopy cover and
lateral cover at 2 different densities (7.5 and 15 deer/km²) in a controlled-density
experiment on Anticosti Island, Québec for each diel period.b
Source of variation Parameter Std DF F-value P-value
a) Dawn/dusk Density -0.1350 0.0827 1,3 2.66 0.201
Biomass 0.0006 0.0004 1,14 3.82 0.053 *
Biomass × densityc -0.0002 0.0045 1,14 0.16 0.691
Lateral cover -0.0207 0.0222 1,14 2.36 0.127
Lateral cover × density 0.0007 0.0265 1,14 0.000 0.980
Canopy cover -0.0005 0.0039 1,14 0.31 0.580
Canopy cover × density 0.0124 0.0039 1,14 6.66 0.011 **
b) Day Density -0.1727 0.0685 1,3 6.36 0.086
Biomass 0.0006 0.0004 1,14 0.97 0.326
Biomass × density -0.0008 0.0004 1,14 4.10 0.045 **
Lateral cover -0.0355 0.0184 1,14 0.77 0.382
Lateral cover × density 0.05171 0.0219 1,14 5.56 0.020 **
Canopy cover -0.0017 0.0032 1,14 0.37 0.544
Canopy cover × density 0.0009 0.0040 1,14 0.06 0.810
c) Night Density -0.0466 0.1044 1,3 0.20 0.699
Biomass 0.0003 0.0006 1,14 0.81 0.370
Biomass × density -0.0001 0.0006 1,14 0.03 0.853
Lateral cover -0.0049 0.0279 1,14 1.03 0.311
Lateral cover × density -0.0234 0.0334 1,14 0.52 0.471
Canopy cover -0.0050 0.0049 1,14 0.02 0.902
Canopy cover × density 0.0092 0.0061 1,14 2.29 0.132
* P < 0.10, ** P < 0.05
aRelative space use represents the total number of overlapping buffers divided by the total
number of positions taken on a particular deer for each of the 3 diel periods at one point in
every 150×150 m square of a grid superposed on each enclosure.
bThe relationships between use and biomass, lateral and canopy cover were quantified
using linear mixed models with block and year as random factors.
cInteraction parameters represent the differences between the slopes at 7.5 deer/km² and
15 deer/km².
101
but not deer at low density. Space use was not related to canopy cover at any density during
the day. The slope for deer space use in relation to lateral cover during the day did not
differ from 0 at low density (t144 = 1.35; P = 0.18), but was negative for deer at high density
(t144 = -1.93; P = 0.05; Table 3-3, Figure 3-4e). This revealed that deer at high density,
compared to deer at low density, decreased the use of sites with dense lateral cover during
the day. At night, relative space use did not vary with deer density, biomass, canopy or
lateral cover (Table 3-3, Figure 3-4g to i).
Discussion
An increase in population density generally leads to a decrease in the abundance of plant
species preferred by deer (Healy et al. 1997) and an increase in intraspecific competition
(Clutton-Brock et al. 1982). To our knowledge, our study is the first to manipulate
population density to test its influence on deer use of forage and cover resources. As
observed in agricultural landscapes (Lesage et al. 2000, Rouleau et al. 2002), we predicted
that deer would use more frequently open habitats than closed habitats at high density
because more forage was available in open areas and thus competition was likely lower
than in closed habitats. At dawn and dusk, deer used space in relation to forage abundance
at high density. Deer at low density used space in function of biomass, but also used sites
with high canopy cover during dawn and dusk. During the day, deer space use was also
positively related to forage abundance at high density but decreased as a function of lateral
cover. It thus seems that the use of cover by deer differed with density. We predicted that
deer would use open habitats more often at night than during the rest of the day because
foraging in open habitats is then less costly in terms of thermoregulation costs (Parker and
Gillingham 1990) and because deer are more secure in covered areas during daylight
(Kufeld et al. 1988, Naugle et al. 1997). Contrarily to our prediction, however, deer did not
use areas of higher biomass and lower canopy cover at night.
Deer space use in relation to plant biomass and cover
Resource distribution, such as forage and cover, is known to affect the spacing patterns of
ungulates (e.g. Bowyer et al. 1998, DePerno et al. 2003, Palmer et al. 2003). The
102
abundance of available biomass may determine the time a herbivore spends in a particular
patch (MacArthur and Pianka 1966, McNamara et al. 1993) and when energy acquisition is
the predominant activity, more time should be devoted to patches where biomass is high
(Wickstrom et al. 1984). In cutblocks, other studies reported that white-tailed deer space
use was related to forage abundance (Stewart et al. 2000, Rothley 2002). Similarly, Cimino
and Lovari (2003) found that removing food in clearings incited roe deer to use woodlots
instead of open habitats.
In most studies, deer spent proportionally more time feeding during the hours surrounding
sunrise and sunset than during the day or at night (Beier and McCullough 1990).
Accordingly, at dawn and dusk, we found that deer space use was positively associated to
forage abundance at both densities. In a given environment, foraging in areas of higher
biomass benefits deer by allowing a higher rate of energy intake and, because these patches
are usually more profitable, deer can stay longer in areas where forage is more abundant
(Wickstrom et al. 1984). When forage is limited, cervids increase the use of habitats that
have a greater abundance of forage to increase the rate of energy acquisition (Van der Wall
2000, Dussault et al. 2005).
Compared to open habitats, forested habitats are cooler in summer and daily variations in
temperature are lower than in open habitats, providing adequate thermal and radiation cover
that allow animals to minimize thermoregulatory costs (Parker and Gillingham 1990). At
dawn and dusk, deer space use was therefore also positively related to canopy cover, at
least at low density. In our study, forage biomass was negatively related to the amount of
canopy cover at both densities (Table 3-2). Differences among slopes of relative use at low
and high densities suggest that deer used space according to forage abundance and canopy
cover at low density, but deer at high density did not use areas of dense canopy cover.
Studies have shown that white-tailed deer are capable of simultaneously considering
multiple habitat factors and circumstantially adjusting their behavior (Schmitz 1992,
Naugle et al. 1997, Rothley 2002). Possibly, this compromise could not be achieved at high
density because increased competition forced deer to stay in open areas where forage was
readily available. In agreement with findings in agricultural-forest landscapes (Lesage et al.
103
2000, Rouleau et al. 2002), it seems that population density affects the use of resources in
forested areas.
During daytime, deer space use was positively related to forage biomass at high density but
not at low density. Using areas with high biomass of forage thus seemed as important
during the day as during dawn and dusk for deer at high density, but deer at low density
decreased their use of open areas with high forage biomass during daytime, possibly to
reduce the costs of exposition to high radiation. Beier and McCullough (1990) found that
deer often used open areas with low canopy cover in summer because of tall grasses and
lateral cover that provided deer with a protection from harsh weather or, in their study,
predation. Deer at high density in our study, however, used areas in relation to forage
abundance and reduced utilization with increasing lateral cover, confirming our prediction
that deer habitat use is based more strongly on forage biomass than on cover as population
density increases.
Many authors reported that deer use of open habitats is higher at night than during daytime
because deer are safer in darkness (Kufeld et al. 1988, Naugle et al. 1997). Deer in our
study, however, did not use open areas more often at night at any density. Consequently,
lateral and canopy cover on Anticosti Island seem to be more important for
thermoregulation in daytime than at night. Cover removal had a strong effect on diurnal
locations, but not on the night locations of female roe deer because of the lower impact of
an increase in visibility during the night than during the day (Cimino and Lovari 2003).
Rothley (2002) also observed that interspersion of cover and forage was important for
habitat utilization of white-tailed deer in presence of hunters, but that in the absence of
disturbance, only forage abundance was an important factor. Surprisingly, however, space
use in our study was not related to forage abundance at night at any density.
Limitations and strengths of the study
Sampling of habitat characteristics in deer habitat use studies is generally restricted by time
and labour and, consequently, only a small portion of the study areas can be described in
details. For that reason, researchers often estimate mean biomass and available cover as one
104
value each for all different habitat types (Mysterud et al. 1999) and assume that vegetation
and cover are distributed homogeneously over each habitat (Sweeney et al. 1984). In fine-
scale habitat studies, however, means cannot account for heterogeneity present in the
environment. A better representation of the environment can be achieved by taking
surrounding sampling points into account (Turner and Gardner 1995). For example, as in
our study, patches of high forage abundance may occur in small forest openings or patches
of denser canopy (Figure 3-1) or lateral cover may remain in clear-cuts (Figure 3-2). The
high variability in the distribution of biomass and cover data (Figure 3-4) reveals the
heterogeneous distribution of resources and how a mean value for clear-cuts and for forest
stands would bias the results. The use of geospatial GIS software to account for spatial
relationships is of major interest for fine-scale habitat studies (Turner and Gardner 1995),
but still has received little attention. Even if a drawback of such a method in large study
areas is that the estimated interpolation errors increase as the distance between sampled
points augments, for sample points adequately distributed spatially, this method may
provide additional insights into ungulate foraging behavior and fine-scale habitat use
studies.
There is also an inherent error associated with radiotelemetry and if radiolocations are
taken as exact locations, this may introduce a bias in the data (White and Garrott 1990,
Rettie and McLoughlin 1999). It has been recommended to use visual observations to study
fine-scale space use (Rettie and McLoughlin 1999); however, for white-tailed deer and
other hard to observe species, this is generally not possible without disturbing their natural
behavior. As suggested by Rettie and McLoughlin (1999), the use of buffers allowed us to
consider telemetry error and to develop a spatial representation of how deer utilized the
enclosures. Our approach gave conservative results as the relationships were based on
spatial relationships and not on individual data points distributed through space.
The use of controlled density experiments has proven useful in describing how
communities respond to known herbivore densities (Tilghman 1989, deCalesta 1994,
Hester et al. 2000, McShea and Rappole 2000, Horsley et al. 2003) and has lead to a better
understanding of critical herbivore densities thresholds allowing vegetation regeneration
105
(Hester et al. 2000). No study has used, however, these controlled browsing experiments to
investigate the role of population density on the behavior of large herbivores. Although
based on a small number of deer, the use and replication of enclosures with controlled
densities allowed us to test directly the influence of known population density on deer
foraging behavior.
Considering diel-periods was also essential to assess space use in relation to forage
abundance and cover because space use varied during the 24-h period. During the day, deer
space use was positively related to forage abundance but only at high density. Deer at high
density restrained their use of areas with greater lateral cover. It thus appears that
population density influences the use of available habitat constraining deer at higher
densities to feed in areas with lower canopy or lateral cover but higher food resources than
deer at lower densities.
In the long-term, forest exploitation likely benefits most cervid populations by increasing
forage availability and quality (Ford et al. 1994). Deer at high-densities may use clear-cuts
as crop fields for foraging, irrespectively of available cover. This, however, may be costly
for the forest industry, especially if deer eat seedlings of commercial species (Côté et al.
2004). The use of large clear-cuts as a management strategy to dissuade deer from
browsing in open areas may not be a sufficient method to limit browsing (Potvin and
Laprise 2002), as lateral cover may also be used for protection.
Acknowledgments
We thank L. Breton and B. Rochette from the Ministère des Ressources naturelles et de la
Faune du Québec, as well as D. Duteau, F. Fournier, G. Picard, D. Sauvé, A. Simard, J.
Taillon, J.-F. Therrien and J.-P. Tremblay for help capturing deer. R. Pouliot, M. Renière,
J.-F. Therrien and V. Viera thankfully assisted to radiotrack deer and S. Debellefeuille, C.
Dussault, M.-A. Giroux, A. Massé, J. Taillon, J.-P. Tremblay, A. Tousignant, and V. Viera
helped for vegetation sampling. We are also indebted to J.-P. Tremblay for the
establishment of the enclosures. R.B. Weladji, S. DeBellefeuille and many graduate
colleagues reviewed an earlier draft of the manuscript. We are also thankful to S.
106
Baillargeon, D. Fortin and K. Lowell for help with statistical analyses. This project was
funded by Produits forestiers Anticosti inc. and the Natural Sciences and Engineering
Research Council of Canada.
Literature cited
Altendorf, K. B., J. W. Laundré, C. A. López González, and J. S. Brown. 2001. Assessing
effects of predation risk on foraging behavior of mule deer. Journal of Mammalogy
82:430–439.
Beier, P., and D. R. McCullough. 1990. Factors influencing white-tailed deer activity
patterns and habitat use. Wildlife Monographs 109:1–51.
Belovsky, G. E. 1981. Optimal activity times and habitat choice of moose. Oecologia
48:22–30.
Berger, J. 1991. Greater Yellowstone's native ungulates: myths and realities. Conservation
Biology 5:353–363.
Bonham, C. D. 1989. Measurements for terrestrial vegetation. John Wiley and Sons, New
York, New York, USA.
Bowyer, R. T., J. G. Kie, and V. Van Ballenberghe. 1998. Habitat selection by neonatal
black-tailed deer: Climate, forage, or risk of predation? Journal of Mammalogy
79:415–425.
Byers, J. A. 1997. American pronghorn: social adaptations and the ghosts of predators past.
University of Chicago Press, Chicago, Illinois, USA.
Cimino, L., and S. Lovari. 2003. The effects of food or cover removal on spacing patterns
and habitat use in roe deer (Capreolus capreolus). Journal of Zoology (London)
261:299–305.
Clutton-Brock, T. H., F. E. Guinness, and S. D. Albon. 1982. Red deer: Behaviour and
ecology of two sexes. The University of Chicago Press, Chicago, Illinois, USA.
Côté, S. D., T. P. Rooney, J.-P. Tremblay, C. Dussault, and D. M. Waller. 2004. Ecological
impacts of deer overabundance. Annual Review of Ecology, Evolution and
Systematics 35:113–147.
Cressie, N. A. C. 1993. Statistics for spatial data. John Wiley & Sons, New York, New
York, USA.
deCalesta, D. S. 1994. Effect of white-tailed deer on songbirds within managed forests in
Pennsylvania. Journal of Wildlife Management 58:711–718.
DePerno, C. S., J. A. Jenks, and S. L. Griffin. 2003. Multidimensional cover characteristics:
Is variation in habitat selection related to white-tailed deer sexual segregation?
Journal of Mammalogy 84:1316–1329.
Dussault, C., R. Courtois, J.-P. Ouellet, and I. Girard. 2005. Space use of moose in relation
to food availability. Canadian Journal of Zoology 83:1431-1437.
Environment Canada. 1993. Canadian climate normals, 1961-1990- Normales climatiques
au Canada 1961-1990, volume 6-Québec. Environnement Canada, Service de
l'environnement atmosphérique, Programme climatologique canadien, Downsview,
Ontario.
107
Ford, W. M., A. S. Johnson, and P. H. Hale. 1994. Nutritional quality of deer browse in
southern Appalachian clearcuts and mature forests. Forest Ecology and
Management 67:149–157.
Frontier, S. 1983. Stratégies d'échantillonnage en écologie. Presses de l'Université Laval,
Québec, Québec, Canada.
Hanley, T. A. 1984. Habitat patches and their selection by wapiti and black-tailed deer in a
coastal montane coniferous forest. Journal of Applied Ecology 21:423–436.
Healy, W. M., D. S. deCalesta, and S. L. Stout. 1997. A research perspective on white-
tailed deer overabundance in the northeastern United States. Wildlife Society
Bulletin 25:259–263.
Hester, A. J., L. Edenius, R. M. Buttenschøn, and A. T. Kuiters. 2000. Interactions between
forests and herbivores: the role of controlled grazing experiments. Forestry 73:381–
391.
Horsley, S., S. Stout, and D. deCalesta. 2003. White-tailed deer impact on the vegetation
dynamics of a northern hardwood forest. Ecological Applications 13:98–118.
Huot, J. 1982. Body condition and food resources of white-tailed deer on Anticosti Island,
Québec. Ph.D. thesis. University of Alaska, Fairbanks, Alaska, USA.
Isaaks, E. H., and R. M. Srivastava. 1989. Applied geostatistics. Oxford University Press,
New York, New York, USA.
Jurado-Expòsito, M., F. López-Granados, J. L. González-Andújar, and L. García-Torres.
2004. Spatial and temporal analysis of Convolvulus arvensis L. populations over
four growing seasons. European Journal of Agronomy 21:287–296.
Keay, J. A., and J. M. Peek. 1980. Relationships between fires and winter habitat of deer in
Idaho. Journal of Wildlife Management 44:372–380.
Kotler, B. P., J. E. Gross, and W. A. Mitchell. 1994. Applying patch use to assess aspects of
foraging behavior in Nubian Ibex. Journal of Wildlife Management 58:299–307.
Kufeld, R. C., D. C. Bowden, and D. L. Schrupp. 1988. Habitat selection and activity
patterns of female mule deer in the Front Range, Colorado. Journal of Range
Management 41:515–522.
Lesage, L., M. Crête, J. Huot, A. Dumont, and J.-P. Ouellet. 2000. Seasonal home range
size and philopatry in two northern white-tailed deer populations. Canadian Journal
of Zoology 78:1930–1940.
Lyon, L. J., and C. E. Jensen. 1980. Management implications of elk and deer use of clear-
cuts in Montana. Journal of Wildlife Management 44:352–362.
MacArthur, R. H., and E. R. Pianka. 1966. On optimal use of a patchy environment.
American Naturalist 100:603–609.
Masters, R. E., R. L. Lochmiller, and D. M. Engle. 1993. Effects of timber harvest and
prescribed fire on white-tailed deer forage production. Wildlife Society Bulletin
21:401–411.
McNamara, J. M., A. I. Houston, and W. W. Weisser. 1993. Combining prey choice and
patch use—what does rate-maximizing predict? Journal of Theoretical Biology
164:219–238.
McShea, W. J., and J. H. Rappole. 2000. Managing the abundance and diversity of breeding
bird populations through manipulation of deer populations. Conservation Biology
14:1161–1170.
108
Mysterud, A., P. K. Larsen, R. A. Ims, and E. Østbye. 1999. Habitat selection by roe deer
and sheep: does habitat ranking reflect resource availability? Canadian Journal of
Zoology 77:776–783.
Mysterud, A., and E. Østbye. 1999. Cover as a habitat element for temperate ungulates:
effects in habitat selection and demography. Wildlife Society Bulletin 27:385–394.
Naugle, D. E., J. A. Jenks, B. J. Kernohan, and R. R. Johnson. 1997. Effects of hunting and
loss of escape cover on movements and activity of female white-tailed deer,
Odocoileus virginianus. Canadian Field-Naturalist 111:595–600.
Nudds, T. D. 1977. Quantifying the vegetation structure of wildlife cover. Wildlife Society
Bulletin 5:113–117.
Palmer, S. C. F., A. J. Hester, D. A. Elston, I. J. Gordon, and S. E. Hartley. 2003. The perils
of having tasty neighbors: grazing impacts of large herbivores at vegetation
boundaries. Ecology 84:2877–2890.
Parker, K. L., and M. P. Gillingham. 1990. Estimates of critical thermal environments for
mule deer. Journal of Range Management 43:73–80.
Potvin, F., P. Beaupré, and G. Laprise. 2003. The eradication of balsam fir stands by white-
tailed deer on Anticosti Island, Québec: A 150-year process. Écoscience 10:487–
495.
Potvin, F., and L. Breton. 2005. Testing 2 aerial survey techniques on deer in fenced
enclosures - visual double-counts and thermal infrared sensing. Wildlife Society
Bulletin 33:317–325.
Potvin, F., and G. Laprise. 2002. Suivi de la banque de semis de sapin baumier sur l'île
d'Anticosti en relation avec le broutement du cerf. Société de la faune et des parcs
du Québec, Québec, Canada.
Rettie, W. J., and P. D. McLoughlin. 1999. Overcoming radiotelemetry bias in habitat-
selection studies. Canadian Journal of Zoology 77:1175–1184.
Robertson, G. P. 1987. Geostatistics in ecology: interpolating with known variance.
Ecology 68:744–748.
Rossi, R., D. J. Mulla, A. G. Journel, and E. H. Franz. 1992. Geostatistical tools for
modeling and interpreting ecological spatial dependence. Ecological Monographs
62:277–314.
Rothley, K. D. 2002. Use of multiobjective optimization models to examine behavioural
trade-offs of white-tailed deer habitat use in forest harvesting experiments.
Canadian Journal of Forest Research 32:1275–1284.
Rouleau, I., M. Crête, and J.-P. Ouellet. 2002. Contrasting the summer ecology of white-
tailed deer inhabiting a forested and an agricultural landscape. Écoscience 9:459–
469.
Schmitz, O. J. 1992. Optimal diet selection by white-tailed deer: balancing reproduction
with starvation risk. Evolutionary Ecology 6:125–141.
Stewart, K. M., T. E. Fulbright, and D. L. Drawe. 2000. White-tailed deer use of clearings
relative to forage availability. Journal of Wildlife Management 64:733–741.
Sweeney, J. M., M. E. Garner, and R. P. Burkert. 1984. Analysis of white-tailed deer use of
forest clear-cuts. Journal of Wildlife Management 48:652–655.
109
Tierson, W. C., G. F. Mattfeld, R. W. Sage Jr., and D. F. Behrend. 1985. Seasonal
movements and home ranges of white-tailed deer in the Adirondacks. Journal of
Wildlife Management 49:760–769.
Tilghman, N. G. 1989. Impacts of white-tailed deer on forest regeneration in northwestern
Pennsylvania. Journal of Wildlife Management 53:524–532.
Tremblay, J.-P., I. Thibault, C. Dussault, J. Huot, et S. D. Côté. 2005. Long-term decline in
white-tailed deer browse supply: can lichens and litterfall act as alternative food
sources that preclude density-dependent feedbacks? Canadian Journal of Zoology
83:1087-1096.
Tufto, J., R. Andersen, and J. Linnell. 1996. Habitat use and ecological correlates of home
range size in a small cervid: the roe deer. Journal of Animal Ecology 65:715–724.
Turner, M. G., and R. H. Gardner. 1995. Ecological dynamics at broad scales. Bioscience
45:S29–S35.
Van der Wall, R., N. Madan, S. Van Lieshout, C. Dormann, R. Langvatn, and S. D. Albon.
2000. Trading forage quality for quantity? Plant phenology and patch choice by
Svalbard reindeer. Oecologia 123:108-115.
White, G. C., and R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic
press, San Diego, California, USA.
Wickstrom, M. L., C. T. Robbins, T. A. Hanley, D. E. Spalinger, and S. M. Parish. 1984.
Food intake and foraging energetics of elk and mule deer. Journal of Wildlife
Management 48:1285–1301.
110
Conclusion générale
Cette étude a permis d’évaluer les effets de la densité de population sur le comportement
d’approvisionnement et le budget d’activité d’un grand herbivore, le cerf de Virginie.
D’abord, nous avons validé l’utilisation des capteurs d’activité pour mesurer le budget
d’activité des cerfs munis de colliers VHF (Coulombe et al. 2006). Nous nous sommes
ensuite intéressés aux déplacements ainsi qu’au budget d’activité des cerfs adultes et
juvéniles et nous avons étudié leur variabilité selon la densité, le nombre d’années après le
début de l’expérience de densité contrôlée, ainsi qu’au cours de l’été et de la journée. Enfin,
nous avons mesuré l’utilisation des enclos par les cerfs afin de mieux comprendre comment
la densité de population influence l’utilisation de l’espace en fonction du couvert et de la
nourriture disponibles.
Validation des capteurs d’activité
Afin d’étudier l’influence de la densité de population sur le budget d’activité, nous devions
d’abord valider l’utilisation des capteurs d’activité horizontaux à impulsions variables des
colliers VHF à l’aide d’observations directes. Au début, nous avons tenté de discriminer
différents comportements actifs (p.ex. cerfs en alimentation vs. en déplacement) et inactifs
(p.ex. cerfs en comportement de repos vs. en vigilance) mais tel que démontré dans d’autres
études (Gillingham et Bunnell 1985, Beier et McCullough 1988, Hansen et al. 1992, Relyea
et al. 1994), nous n’avons pu différencier ces comportements en se basant sur les données
obtenues par télémétrie. Les comportements ont donc simplement été séparés en
comportements actifs ou inactifs. Les données télémétriques individuelles étaient
correctement classifiées en comportements actifs et inactifs dans 74% des cas et
introduisaient donc des erreurs dans la quantification des périodes d’activité. En intégrant
l’information de 3 données d’activité successives, nous avons correctement identifié 84%
des comportements actifs et inactifs et 87% des périodes d’activité. Nous avons donc
conclu que les capteurs d’activité horizontaux utilisés pourraient décrire avec suffisamment
de précision les budgets d’activité des cerfs à l’étude. Cependant, à l’aide des capteurs
d’activité à deux axes des colliers GPS qui ont été validés simultanément, il nous a aussi été
111
possible de déterminer que des capteurs d’activité orientés verticalement seraient plus
précis dans l’évaluation des budgets d’activité.
Les déplacements et le budget d’activité
Dans un milieu donné, en absence de prédateurs, l’abondance des ressources contrôle
généralement le comportement d’approvisionnement. La flexibilité des comportements
permet aux individus de faire face à une diminution de l’abondance des ressources causée
par une augmentation de la densité de population (Cederlund et al. 1989, Beier et
McCullough 1990, Borkowski 2000).
Tel que mentionné dans l’introduction générale, nous voulions vérifier les hypothèses
suivantes : 1) le comportement d’approvisionnement (déplacements et budget d’activité)
des cerfs est déterminé par l’abondance de la nourriture qui diminue avec l’augmentation
de la population. Les cerfs répondent à la diminution de l’abondance de la nourriture à
haute densité en modifiant leur budget d’activité possiblement de façon (a) à sélectionner la
végétation de meilleure qualité même si celle-ci est moins abondante en augmentant le
temps de recherche de la végétation et le taux de déplacement ou (b) les cerfs répondent à
la diminution de l’abondance de la nourriture à haute densité en augmentant le temps passé
à la rumination de la végétation de moindre qualité, ce qui accroit la durée des périodes
d’inactivité. Cette étude est la première à tester ces hypothèses en contrôlant pour la densité
de population dans des enclos de façon expérimentale.
Les taux de déplacement des cerfs ne différaient pas entre les densités contrôlées. Rouleau
et al. (2002) avaient trouvé que les taux de déplacement différaient entre des individus se
trouvant en milieu forestier et ceux en milieu agricole à différentes densités mais les
différences étaient reliées à des différences écologiques entre les milieux. Dans notre étude,
les cerfs étaient placés dans des milieux adjacents et écologiquement comparables et malgré
les densités différentes, nous n’avons pas trouvé de différences dans le taux de
déplacement. Cependant, notre étude s’est déroulée dans un milieu récemment perturbé par
une coupe forestière. Les différences de biomasse entre les années étaient donc plus
grandes qu’entre les densités. Nous n’avons pu détecter de différences en ce qui concerne la
112
biomasse disponible entre les densités contrôlées puisque les densités n’étaient pas
suffisamment différentes et parce que l’expérience n’avait pas été appliquée pendant assez
longtemps. Les conclusions d’une étude connexe suggèrent que des différences de
biomasse entre les densités se manifestent lorsque l’écart entre les densités est plus grand
(Tremblay 2005). Néanmoins, le taux de déplacement n’était pas différent entre la première
et la deuxième année d’application du traitement de densité contrôlée même si l’abondance
de végétation avait augmentée. De plus, les changements dans la quantité de biomasse entre
les densités n’étaient pas assez grands dans notre étude pour affecter directement la plupart
des caractéristiques du budget d’activité des cerfs. Cependant, les densités contrôlées (7.5
et 15 cerfs/km²) étaient plus faibles que la densité naturelle habituellement présente sur l’île
(>20 cerfs/km²).
En effet, pour les juvéniles, aucune différence n’a été trouvée dans la proportion du temps
passé en activité, la durée des périodes d’activité ou d’inactivité et le nombre de périodes
d’activité par jour entre les densités. Par ailleurs, nous avons trouvé que les adultes dans les
enclos à 7.5 cerfs/km² passaient une plus grande proportion de temps en activité par jour
que ceux à 15 cerfs/km². Nous avons également observé que les juvéniles passaient plus de
temps en activité que les adultes, mais seulement à haute densité. Cependant, nous
n’excluons pas la possibilité que nos conclusions soient influencées par un rapport
mâle/femelle différent entre les traitements.
La végétation disponible dans les coupes était plus abondante la deuxième et la troisième
année que la première année après le début de l’expérience de densité contrôlée. Bien que
les cerfs passaient une proportion totale de temps en activité équivalente durant les trois
années, nous avons trouvé qu’avec une augmentation de l’abondance de la végétation, la
répartition du temps actif a changé. En effet, la durée des périodes d’activité a diminué au
cours des années tandis que la durée des périodes d’inactivité et le nombre de périodes
d’activité par jour ont augmenté. Lorsque la biomasse a augmenté, les cerfs pouvaient donc
remplir leur rumen plus rapidement avant de débuter une période de rumination, ce qui
pourrait expliquer l’augmentation du nombre de périodes d’activité 3 ans après le début de
l’application du traitement (Moncorps et al. 1997). Lorsque la biomasse a augmenté au
113
cours des années, les cerfs pouvaient remplir leur rumen en moins de temps et donc
commencer plus rapidement une période de rumination. Ce changement dans le budget
d’activité avec l’augmentation de la biomasse permettrait aux cerfs de gagner 25% de plus
en masse corporelle pendant l’été par rapport aux cerfs en densité naturelle (Simard et al.,
en prép). Nous avons aussi trouvé que la durée des périodes d’inactivité a diminué au cours
des années. Bien que le temps passé à ruminer augmente généralement en fonction du
temps passé inactif lorsque la qualité de la végétation diminue (Mysterud 1998, Pérez-
Barbería and Gordon 1999), ceci ne s’applique peut-être pas lorsque la végétation de bonne
qualité est abondante (S. Hamel, comm. pers.). Les herbivores augmenteraient alors la
durée des périodes inactives en restant au repos sous couvert sans nécessairement
augmenter le temps de rumination parce qu’ils auraient ainsi davantage de temps disponible
grâce à l’abondance de végétation pour diminuer l’exposition aux conditions
environnementales plus stressantes des milieux ouverts.
À Anticosti, la saison de croissance de la végétation débute à la fonte des neiges entre le
début et la mi-mai (Ressources naturelles Canada 2005). Les nouvelles pousses sont riches
en protéines et facilement digestibles (Van der Wall et al. 2000). Au cours de l’été, la
végétation augmente en lignine et sa concentration en protéines décroît (Tremblay 1981) et,
en conséquence, la digestibilité des plantes diminue (Robbins 1983, Van Soest 1994). Pour
tous les cerfs, les taux de déplacements étaient semblables au cours de l’été. Cependant, les
adultes et les juvéniles ont répondu différemment aux changements saisonniers dans la
qualité et l’abondance de la végétation. En effet, la durée des périodes d’inactivité a
augmenté chez les juvéniles à haute densité. Les adultes à faible densité ont
progressivement diminué la proportion du temps passé en activité au cours de l’été. Puisque
l’abondance de la végétation augmente durant l’été, les cerfs ont eu besoin de moins de
temps pour chercher et ingérer la végétation. De plus, il est possible que l’abondance de la
nourriture à partir du mois de juillet excédait la biomasse nécessaire pour les besoins
énergétiques des cerfs et ceux-ci auraient donc pu rencontrer leurs besoins énergétiques en
moins de temps (Beier et McCullough 1990).
114
Jusqu’au début du mois d’août, les cerfs en densité naturelle ont passé davantage de temps
à ruminer et en conséquence, ils ont passé moins de temps en activité par jour et avaient des
périodes d’inactivité plus longues que les cerfs à faible densité. En densité naturelle, les
cerfs sont nombreux et les jeunes pousses disparaissent rapidement. Les cerfs en densité
naturelle auraient donc eu accès à de la nourriture moins abondante et probablement de
moins bonne qualité que ceux à des densités contrôlées. Lorsque la végétation est devenue
plus abondante au cours de l’été, les cerfs en densité naturelle ont augmenté la proportion
du temps passé en activité. L’augmentation de la proportion du temps passé en activité
pourrait être reliée à une augmentation de leur sélectivité. Ceci est aussi appuyé par le fait
qu’au cours de l’été, la durée des périodes d’activité a augmenté et la durée des périodes
d’inactivité a diminué.
Pour les juvéniles dans les deux densités contrôlées, la proportion du temps passé en
activité était plus élevée à l’aube et au crépuscule et moins élevée durant la nuit que
pendant le jour et l’aube. Des pics d’activité au cours de la journée ont été rapportés dans
plusieurs études et font partie du cycle d’activité des cerfs (Kammermeyer et Marchinton
1977, Beier et McCullough 1990). Ils représentent des périodes d’alimentation intenses en
prévision ou suivant une période de noirceur pendant laquelle la recherche de nourriture est
probablement moins efficace. Comme dans plusieurs études, les pics au crépuscule étaient
plus constants que ceux à l’aube (Skogland 1983, Beier et McCullough 1990). Comme
nous avons considéré un intervalle équivalent de 1h30 après et avant le lever du soleil,
n’avons pas détecté de pic d’activité après l’aube comme d’autres études (Beier et
McCullough 1990). Aucune différence statistique de l’activité en fonction de la période de
la journée n’a été trouvée pour les adultes. Cependant, comme pour les juvéniles, la
proportion du temps d’activité était de 16 à 28% plus élevée au crépuscule que pendant les
autres périodes de la journée.
Le compromis couvert/nourriture
L’augmentation de la densité de population dans un endroit est généralement reliée à une
diminution de l’abondance des espèces préférées par le cerf (Healy et al. 1997) et à une
augmentation de la compétition intraspécifique (Clutton-Brock et al. 1982). Nous pensions
115
que les cerfs utiliseraient davantage les milieux ouverts à haute densité puisque, bien que
les coûts reliés à la thermorégulation y soient plus élevés, la végétation y est plus abondante
et donc la compétition moins intense qu’en milieu fermé. De plus, puisque l’utilisation de
l’espace peut varier selon la période du jour (Cimino et Lovari 2003), nous avons pris en
compte ce facteur.
En général, les cerfs passent proportionnellement plus de temps à s’alimenter à l’aube et au
crépuscule (Beier et McCullough 1990). Il est donc énergétiquement plus avantageux pour
eux de passer davantage de temps à ces moments dans des endroits où la nourriture est
abondante (Wickstrom et al. 1984). À l’aube et au crépuscule, dans les deux densités, les
cerfs utilisaient l’espace en fonction de l’abondance de la végétation. Les cerfs à faible
densité utilisaient aussi, pendant ces périodes, des endroits où le couvert arborescent était
abondant. Les habitats fermés, pendant l’été, sont plus frais que les milieux ouverts et
présentent un couvert qui minimise les coûts de thermorégulation (Parker et Gillingham
1990). Dans notre étude, la quantité de biomasse était négativement reliée à la densité du
couvert vertical dans les deux densités. Ainsi, il semble que les cerfs à faible densité
conciliaient l’utilisation de nourriture et de couvert mais les cerfs à haute densité utilisaient
l’espace uniquement en fonction de la quantité de nourriture. Le compromis n’était
probablement pas possible à haute densité parce que la compétition pour la nourriture était
plus élevée et poussait les individus à utiliser davantage des milieux ouverts pour se nourrir
car la quantité de nourriture y était plus grande. De plus, contrairement aux cerfs à faible
densité, les cerfs à haute densité utilisaient l’espace pendant le jour en fonction de
l’abondance de nourriture. Aussi, les cerfs à haute densité pendant le jour évitaient les sites
où le couvert latéral était dense pour se nourrir dans des endroits où l’abondance de
nourriture était plus élevée.
Nous supposions que les cerfs utiliseraient davantage les endroits découverts la nuit
puisque le comportement d’approvisionnement y est moins coûteux en termes de
thermorégulation (Parker et Gillingham 1990) et parce que le risque d’être repéré par un
prédateur est plus faible (Kufeld et al. 1988, Naugle et al. 1997). Cependant, pendant la
nuit, les cerfs n’utilisaient pas l’espace en fonction de l’abondance de nourriture ou de
116
couvert. Le couvert latéral et vertical semblaient donc plus importants pendant le jour,
l’aube et le crépuscule lorsque les coûts de thermorégulation sont plus importants. Rothley
(2002) a aussi trouvé que l’abondance de nourriture entremêlée de milieux fermés était
importante lorsque les coûts reliés à l’utilisation des milieux ouverts étaient élevés, mais
qu’en absence de dérangements seule l’abondance de nourriture était importante.
Cependant, nous n’avons pas trouvé que l’abondance de nourriture influençait l’utilisation
de l’espace pendant la nuit.
L’inventaire des habitats utilisés par le cerf est souvent limité par des contraintes de temps
et de main d’œuvre. Pour cette raison, les chercheurs estiment souvent une valeur moyenne
par catégorie d’habitat (Mysterud et al. 1999) et supposent que la biomasse et le couvert
disponibles sont homogènes à travers cet habitat (Sweeney et al. 1984). Cependant, pour les
études fines de sélection d’habitat, les moyennes ne peuvent représenter l’hétérogénéité
spatiale présente dans l’environnement et une meilleure représentation peut être obtenue en
tenant compte des relations spatiales existant entre plusieurs points d’échantillonnage
(Turner et Gardner 1995). Par exemple, nous avons trouvé une grande variabilité dans les
données à cause de la présence d’arbres laissés dans la coupe ou d’ouvertures dans les
forêts qui laissent entrer la lumière et favorisent la croissance des végétaux. Vu leurs petites
tailles, il est possible d’estimer les ressources disponibles dans l’ensemble des enclos à des
intervalles réguliers et ainsi de décrire des zones représentatives de la biomasse et du
couvert latéral et vertical disponible à l’aide des méthodes géostatistiques. Il existe aussi
une erreur associée aux localisations télémétriques et si l’on considère les localisations sans
en tenir compte, cela peut donner des résultats erronés (White et Garrott 1990, Rettie et
McLoughlin 1999). Il a été recommandé d’utiliser des observations directes pour faire des
études fines d’utilisation de l’espace (Rettie et McLoughlin 1999); cependant, pour le cerf
de Virginie et plusieurs espèces de cervidés, ceci n’est généralement pas possible sans
déranger leur comportement naturel. Tel que recommandé par Rettie et McLoughlin
(1999), nous avons donc utilisé des pastilles d’erreurs de taille correspondante à l’erreur de
télémétrie. L’utilisation de ces nouvelles techniques nous a permis d’obtenir des résultats
novateurs, mais conservateurs quant à l’utilisation des enclos à différentes densités.
117
Conclusions et recommandations
Nous avions prédit que la densité de population influencerait le comportement
d’approvisionnement puisqu’elle est reliée à une diminution de l’abondance de la
végétation. Cependant, les déplacements et le budget d’activité étaient semblables entre les
densités. À une densité de 15 cerfs/km², les cerfs avaient un budget d’activité et un taux de
déplacement semblable à ceux à 7.5 cerfs/km². Cependant, les adultes à 15 cerfs/km² n’ont
pas diminué leur activité pendant l’été comme ceux à 7.5 cerfs/km².
Selon les différences interannuelles et les budgets d’activité observés des cerfs en densité
naturelle, on peut supposer que si la densité de population affecte suffisamment
l’abondance et la qualité de la végétation, alors les cerfs modifieront leur comportement
afin de compenser pour la diminution ou l’augmentation de la quantité d’énergie
disponible. En effet, selon nos résultats, il existe une série de changements
comportementaux reliés à une diminution de la quantité de végétation due à une
augmentation de la densité de population. Dans un milieu où la végétation est peu
abondante, les cerfs réduisent d’abord le temps passé en activité pour ruminer plus
longtemps la végétation plus fibreuse. Ensuite, lorsque la biomasse augmente en réponse à
une diminution de la densité, les cerfs deviennent plus sélectifs et augmentent le temps
passé en activité. Finalement, la durée des périodes d’activité diminue à nouveau lorsque la
biomasse devient si abondante que le temps nécessaire pour remplir le rumen diminue. Les
cerfs à haute densité, contrairement aux cerfs à faible densité, n’utilisaient pas pendant
l’aube, le crépuscule et la journée des milieux où le couvert était plus abondant.
L’utilisation de l’espace serait donc aussi modifiée par la densité de population. En effet,
l’utilisation des milieux ouverts augmenterait à haute densité lorsque la compétition pour la
nourriture est plus élevée.
Maintenant, on peut se demander si, et à combien de cerfs/km2, la densité affectera les traits
d’histoire de vie du cerf de Virginie car malgré sa grande capacité d’adaptation, il est limité
par ses capacités de digestion puisqu’il doit ruminer (Van Soest 1982). Des études
s’intéressant au taux d’acquisition des réserves corporelles à différentes densités
permettraient de connaître si le taux d’acquisition des réserves est réduit à ≥15 cerfs/km².
118
Certains indicateurs démontrent que les densités élevées de cerfs retrouvées à l’île
d’Anticosti ont entraîné des impacts sur les traits d’histoire de vie de celui-ci. En effet, en
comparaison aux cerfs de la population source du continent, les cerfs d’Anticosti sont 40%
plus petits (Boucher et al. 2004) et les femelles se reproduisent pour la première fois un an
plus tard (Goudreault 1980).
Généralement, les coupes forestières de petite superficie avantagent les populations de
cervidés en leur procurant davantage de nourriture à proximité du couvert (Ford et al.
1994). Cependant, l’utilisation de grandes coupes pour réduire le broutement du cerf dans
des zones où le couvert vertical est éloigné ne semble pas être une solution satisfaisante
pour limiter l’utilisation des coupes par le cerf à Anticosti (Potvin et Laprise 2002). La nuit
et le couvert latéral fourni par les grandes graminées sont des moyens alternatifs au couvert
vertical que peuvent utiliser les cerfs pour s’alimenter en milieu ouvert. À Anticosti, cela
semble être le cas puisque des relevés ont montré que les semis de sapin (Abies balsamea)
sont fortement broutés et ce même en été et jusqu’au centre de grandes coupes situées à
plus de 800 m de la bordure de la forêt (Potvin et Laprise 2002).
Le cerf de Virginie est vraisemblablement une espèce plastique qui peut rapidement
s’adapter aux changements d’abondance et de qualité de la végétation. Il a trouvé le moyen
de subsister à l’île d’Anticosti, dans des conditions où beaucoup d’autres grands herbivores
ne pourraient subsister à de telles densités. En effet, malgré la faible végétation disponible,
il accumule suffisamment de réserves pour survivre aux hivers rigoureux qui sévissent à
l’île en modifiant son comportement d’approvisionnement.
119
Bibliographie générale
Altendorf, K. B., J. W. Laundré, C. A. López González, et J. S. Brown. 2001. Assessing
effects of predation risk on foraging behavior of mule deer. Journal of Mammalogy
82:430–439.
Anderson, R. C., E. A. Corbett, M. R. Anderson, G. A. Corbett, et T. M. Kelley. 2001. High
white-tailed deer density has negative impact on tallgrass prairie forbs. Journal of
the Torrey Botanical Society 128:381–392.
Augustine, D. J., L. E. Frelich, et P. A. Jordan. 1998. Evidence for two alternate stable
states in an ungulate grazing system. Ecological Applications 8:1260–1269.
Augustine, D. J., et S. J. McNaughton. 1998. Ungulate effects on the functional species
composition of plant communities: herbivore selectivity and plant tolerance. Journal
of Wildlife Management 62:1165–1183.
Baker, D. L., et N. T. Hobbs. 1986. Strategies of digestion: digestive efficiency and
retention time of forage diets in montane ungulates. Canadian Journal of Zoology
65:1978–1984.
Balgooyen, C. P., et D. M. Waller. 1995. The use of Clintonia borealis and other indicators
to gauge impacts of white-tailed deer on plant communities in northern Wisconsin,
USA. Natural Areas Journal 15:308–318.
Beier, P., et D. R. McCullough. 1988. Motion-sensitive radio collars for estimating white-
tailed deer activity. Journal of Wildlife Management 52:11–13.
Beier, P., et D. R. McCullough. 1990. Factors influencing white-tailed deer activity patterns
and habitat use. Wildlife Monographs 109:1–51.
Bell, R. H. V. 1971. A grazing ecosystem in the Serengenti. Scientific American 225:
86–93.
Boege, K., et R. J. Marquis. 2005. Facing herbivory as you grow up: the ontogeny of
resistance in plants. Trends in Ecology & Evolution 20:441–448.
Bonham, C. D. 1989. Measurements for terrestrial vegetation. John Wiley and Sons, New
York, New York, USA.
Borkowski, J. 2000. Influence of the density of a sika deer population on activity, habitat
use, and group size. Canadian Journal of Zoology 78:1369–1374.
Boucher, S., M. Crête, J.-P. Ouellet, C. Daigle, et L. Lesage. 2004. Large-scale trophic
interactions: White-tailed deer growth and forest understory. Écoscience 11:286–
295.
Bradshaw, R. H. W., G. E. Hannon, et A. M. Lister. 2003. A long-term perspective on
ungulate-vegetation interactions. Forest Ecology and Management 181:267–280.
Bryant, J. P., F. D. Provenza, J. Pastor, P. B. Reichardt, T. P. Clausen, et J. T. du Toit.
1991. Interactions between woody plants and browsing mammals mediated by
secondary metabolites. Annual Review of Ecology and Systematics 22:431–446.
Bryant, J. P., T. Reichardt, et T. P. Clausen. 1992. Chemically mediated interactions
between woody plants and browsing mammals. Journal of Range Management
45:18–24.
120
Bunnell, F. L., et M. P. Gillingham. 1985. Foraging behavior: dynamics of dining out.
Pages 53–79 dans R. J. Hudson et R. G. White, editors. Bioenergetics of wild
herbivores. CRC Press, Boca Raton, Florida, USA.
Byers, J. A. 1997. American pronghorn: social adaptations and the ghosts of predators past.
University of Chicago Press, Chicago, Illinois, USA.
Cederlund, G., R. Bergström, et F. Sandegren. 1989. Winter activity pattern of females in
two moose populations. Canadian Journal of Zoology 67:1516–1522.
Cimino, L., and S. Lovari. 2003. The effects of food or cover removal on spacing patterns
and habitat use in roe deer (Capreolus capreolus). Journal of Zoology (London)
261:299–305.
Clutton-Brock, T. H., F. E. Guinness, et S. D. Albon. 1982. Red deer: Behaviour and
ecology of two sexes. The University of Chicago Press, Chicago, Illinois, USA.
Clutton-Brock, T. H., G. R. Iason, et F. E. Guinness. 1987. Sexual segregation and density-
related changes in habitat use in male and female Red deer (Cervus elaphus).
Journal of Zoology (London) 211:275–289.
Cooke, A. S., et L. Farrell. 2001. Impact of muntjac deer (Muntiacus reevesi) at Monks
wood national reserve, Cambridgeshire, eastern England. Forestry 74:242–249.
Côté, S. D., T. P. Rooney, J.-P. Tremblay, C. Dussault, et D. M. Waller. 2004. Ecological
impacts of deer overabundance. Annual Review of Ecology, Evolution and
Systematics 35:113–147.
Coulombe, M.-L., A. Massé, et S. D. Côté. 2006. Quantification and accuracy of activity
data measured with VHF and GPS telemetry. Wildlife Society Bulletin. Sous presse.
Cressie, N. A. C. 1993. Statistics for spatial data. John Wiley & Sons, New York, New
York, USA.
Drolet, C. A. 1976. Distribution and movements of white-tailed deer in southern New
Brunswick in relation to environmental factors. Canadian Field-Naturalist 90:123–
136.
Ford, W. M., A. S. Johnson, et P. H. Hale. 1994. Nutritional quality of deer browse in
southern Appalachian clearcuts and mature forests. Forest Ecology and
Management 67:149–157.
Fortin, D., M. S. Boyce, et E. H. Merrill. 2004. Multi-tasking by mammalian herbivores:
overlapping processes during foraging. Ecology 85:2312-2322.
Frontier, S. 1983. Stratégies d'échantillonnage en écologie. Presses de l'Université Laval,
Québec, Québec, Canada.
Garrott, R. A., P. J. White, et C. A. V. White. 1993. Overabundance: An issue for
conservation biologists? Conservation Biology 7:946–949.
Gates, C. C., et R. J. Hudson. 1983. Foraging behaviour of wapiti in a boreal forest
enclosure. Canadian-Field Naturalist 110:197–206.
Gillingham, M. P., et F. L. Bunnell. 1985. Reliability of motion-sensitive radio collars for
estimating activity of black–tailed deer. Journal of Wildlife Management 49:951–
958.
Gillingham, M. P., K. L. Parker, et T. A. Hanley. 1997. Forage intake by black-tailed deer
in a natural environment: bout dynamics. Canadian Journal of Zoology 75:1118–
1128.
121
Gordon, I. J., A. J. Hester, et M. Festa-Bianchet. 2004. The management of wild large
herbivores to meet economic, conservation and environmental objectives. Journal of
Applied Ecology 41:1021–1031.
Goudreault, F. 1980. Reproduction et dynamique de la population de cerfs de l'île
d'Anticosti. Mémoire de M.Sc. Université Laval, Québec, Québec, Canada.
Hanley, T. A. 1984. Habitat patches and their selection by wapiti and black-tailed deer in a
coastal montane coniferous forest. Journal of Applied Ecology 21:423–426.
Hansen, M. C., G. W. Garner, et S. G. Fancy. 1992. Comparison of 3 methods for
evaluating activity of dall's sheep. Journal of Wildlife Management 56:661–668.
Healy, W. M. 1997. Influence of deer on the structure and composition of oak Forests in
central Massachusetts. Pages 249–266 dans W. J. McShea, H. B. Underwood, et J.
H. Rappole, editors. The science of overabundance: deer ecology and population
management. Smithsonian Institution press, Washington, Washington, USA.
Healy, W. M., D. S. deCalesta, et S. L. Stout. 1997. A research perspective on white-tailed
deer overabundance in the northeastern United States. Wildlife Society Bulletin
25:259–263.
Herbers, J. M. 1981. Time resources and laziness in animals. Oecologia 49:252–262.
Hester, A. J., L. Edenius, R. M. Buttenschøn, et A. T. Kuiters. 2000. Interactions between
forests and herbivores: the role of controlled grazing experiments. Forestry 73:381–
391.
Hester, A. J., et K. J. Kirby. 1996. Effects of season and intensity of sheep grazing on tree
regeneration in a British upland woodland. Forest Ecology and Management 88:99–
106.
Hobbs, N. T. 1996. Modification of ecosystems by ungulates. Journal of Wildlife
Management 60:695–713.
Hofmann, R. R. 1989. Evolutionary steps of ecophysiological adaptation and diversification
of ruminants: a comparative view of their digestive system. Oecologia 78:443–457.
Huot, J. 1974. Winter habitat of white-tailed deer at Thirty-one mile Lake, Québec.
Canadian Field-Naturalist 88:293–301.
Huot, J. 1982. Body condition and food resources of white-tailed deer on Anticosti Island,
Québec. Thèse de Ph.D. University of Alaska, Fairbanks, Alaska, USA.
Illius, A. W., et I. J. Gordon. 1987. The allometry of food intake in grazing ruminants.
Journal of Animal Ecology 56:989–999.
Jorgenson, J. T., M. Festa-Bianchet, J.-M. Gaillard, et W. D. Wishart. 1997. Effects of age,
sex, disease, and density on survival of bighorn sheep. Ecology 78:1019–1032.
Kammermeyer, K. E., et R. L. Marchinton. 1977. Seasonal change in circadian activity of
radio-monitored deer. Journal of Wildlife Management 41:315–317.
Keay, J. A., et J. M. Peek. 1980. Relationships between fires and winter habitat of deer in
Idaho. Journal of Wildlife Management 44:372–380.
Kilgo, J. C., R. F. Labisky, et D. E. Fritzen. 1998. Influences of hunting on the behavior of
white-tailed deer: Implications for conservation of the Florida panther. Conservation
Biology 12:1359–1364.
Kirby, K. J. 2001. The impact of deer on the ground flora of British broadleaved woodland.
Forestry 74:219–229.
122
Kufeld, R. C., D. C. Bowden, and D. L. Schrupp. 1988. Habitat selection and activity
patterns of female mule deer in the Front Range, Colorado. Journal of Range
Management 41:515-522.
Lefort, S. 2002. Habitat hivernal du cerf de Virginie (Odocoileus virginianus) à l'île
d'Anticosti. Mémoire de M.Sc. Université Laval, Québec, Québec, Canada.
Leopold, A., L. K. Sowls, et D. L. Spencer. 1947. A survey of over-populated deer ranges
in the United States. Journal of Wildlife Management 11:162–177.
Lesage, L., M. Crête, J. Huot, A. Dumont, et J.-P. Ouellet. 2000. Seasonal home range size
and philopatry in two northern white-tailed deer populations. Canadian Journal of
Zoology 78:1930–1940.
Lesage, L., M. Crête, J. Huot, et J.-P. Ouellet. 2001. Evidence for a trade-off between
growth and body reserves in northern white-tailed deer. Oecologia 126:30–41.
Lyon, L. J., et C. E. Jensen. 1980. Management implications of elk and deer use of clear-
cuts in Montana. Journal of Wildlife Management 44:352–362.
Marie-Victorin, F., et F. Rolland-Germain. 1969. Flore de l'Anticosti-Minganie. Presses de
l'université de Montréal, Montréal, Québec, Canada.
Masters, R. E., R. L. Lochmiller, et D. M. Engle. 1993. Effects of timber harvest and
prescribed fire on white-tailed deer forage production. Wildlife Society Bulletin
21:401–411.
Miller, F. L. 1970. Distribution patterns of black-tailed deer (Odocoileus hemionus
columbianus) in relation to environment. Journal of Mammalogy 51:248–260.
Miller, K. V. 1997. Considering social behavior in the management of overabundant white-
tailed deer populations. Wildlife Society Bulletin 25:279–281.
Miller, K. V., et J. J. Ozoga. 1997. Density effects on deer sociobiology. Pages 136–150
dans W. J. McShea, H. B. Underwood, et J. H. Rappole, editors. The science of
overabundance: deer ecology and population management. The Smithsonian
Institution, Washington, Washington, USA.
Moncorps, S., P. Boussès, D. Réale, et J.-L. Chapuis. 1997. Diurnal time budget of the
mouflon (Ovis musimon) on the Kerguelen archipelago: influence of food resources,
age, and sex. Canadian Journal of Zoology 75:1828–1834.
Mysterud, A. 1998. The relative roles of body size and feeding type on activity time of
temperate ruminants. Oecologia 113:442–446.
Mysterud, A., P. K. Larsen, R. A. Ims, et E. Østbye. 1999. Habitat selection by roe deer and
sheep: does habitat ranking reflect resource availability? Canadian Journal of
Zoology 77:776–783.
Mysterud, A., et E. Østbye. 1995. Bed-site selection by the European roe deer (Capreolus
capreolus) in southern Norway during winter. Canadian Journal of Zoology 73:924–
932.
Mysterud, A., et E. Østbye. 1999. Cover as a habitat element for temperate ungulates:
effects in habitat selection and demography. Wildlife Society Bulletin 27:385–394.
Naugle, D. E., J. A. Jenks, B. J. Kernohan, and R. R. Johnson. 1997. Effects of hunting and
loss of escape cover on movements and activity of female white-tailed deer,
Odocoileus virginianus. Canadian Field-Naturalist 111:595-600.
Nudds, T. D. 1977. Quantifying the vegetation structure of wildlife cover. Wildlife Society
Bulletin 5:113–117.
123
Ozoga, J. J. 1968. Variations in microclimate in a conifer swamp deeryard in northern
Michigan. Journal of Wildlife Management 32:574–584.
Palo, R., et C. T. Robbins. 1991. Plant defences against mammalian herbivory. CRC Press,
Boca Raton, Florida, USA.
Palo, R. T. 1985. Chemical defence in birch: Inhibition of digestibility in ruminants by
phenolic extracts. Oecologia 68:10–14.
Parker, K. L., et M. P. Gillingham. 1990. Estimates of critical thermal environments for
mule deer. Journal of Range Management 43:73–80.
Pérez-Barbería, F. J., et I. J. Gordon. 1999. The relative roles of phylogeny, body size and
feeding style on the activity time of temperate ruminants: a reanalysis. Oecologia
120:193–197.
Porter, W. F., et H. B. Underwood. 1999. Of elephants and blind men: Deer management in
the U.S. National Parks. Ecological Applications 9:3–9.
Potvin, F. 1985. Exploitation et dynamique de la population de cerfs d'Anticosti de 1975 à
1984. Ministère du Loisir de la Chasse et de la Pêche, Direction générale de la
faune, Québec, Québec, Canada.
Potvin, F., P. Beaupré, A. Gingras, et D. Pothier. 2000. Le cerf et les sapinières de l'île
d'Anticosti. Direction de la faune et des habitats. Société de la Faune et des Parcs du
Québec, Québec, Canada.
Potvin, F., P. Beaupré, et G. Laprise. 2003. The eradication of balsam fir stands by white-
tailed deer on Anticosti Island, Québec: A 150-year process. Écoscience 10:487–
495.
Potvin, F., et G. Laprise. 2002. Suivi de la banque de semis de sapin baumier sur l'île
d'Anticosti en relation avec le broutement du cerf. Société de la faune et des parcs
du Québec, Québec, Québec, Canada.
Putman, R. J., J. Langbein, A. J. M. Hewison, et S. K. Sharma. 1996. Relative roles of
density-dependent and density-independent factors in population dynamics of
British deer. Mammal Review 26:81–101.
Relyea, R. A., I. M. Ortega, et S. Demarais. 1994. Activity monitoring in mule deer:
Assessing telemetry accuracy. Wildlife Society Bulletin. 22:656–661.
Renecker, L. A., et R. J. Hudson. 1986. Seasonal foraging rates of free-ranging moose.
Journal of Wildlife Management 50:143–147.
Ressources naturelles Canada. 2005. L'atlas du Canada. Disponible à
http://atlas.gc.ca/site/francais/maps/environment/climate/snowcover/endsnowcover/.
[Visité le 20 août 2005]
Rettie, W. J., et P. D. McLoughlin. 1999. Overcoming radiotelemetry bias in habitat-
selection studies. Canadian Journal of Zoology 77:1175–1184.
Robbins, C. T. 1983. Wildlife feeding and nutrition. Academic Press, New York, New
York, USA.
Rochette, B. A., A. Gingras, et F. Potvin. 2003. Inventaire aérien du cerf de Virginie de l'île
d'Anticosti- été 2001. Société de la faune et des parcs du Québec, Direction de
l'aménagement de la faune de la Côte-Nord et direction de la recherche sur la faune,
Québec, Québec, Canada.
Rooney, T. P. 2001. Deer impacts on forest ecosystems: a North American perspective.
Forestry 74:201–208.
124
Rooney, T. P., et W. J. Dress. 1997. Species loss over sixty-six years in the ground-layer
vegetation of Heart’s content, an old-growth forest in Pennsylvania, USA. Natural
Areas Journal 17:297–305.
Rooney, T. P., et D. M. Waller. 2003. Direct and indirect effects of white-tailed deer in
forest ecosystems. Forest Ecology and Management 181:165–176.
Rothley, K. D. 2002. Use of multiobjective optimization models to examine behavioural
trade-offs of white-tailed deer habitat use in forest harvesting experiments.
Canadian Journal of Forest Research 32:1275–1284.
Rouleau, I., M. Crête, et J.-P. Ouellet. 2002. Contrasting the summer ecology of white-
tailed deer inhabiting a forested and an agricultural landscape. Écoscience 9:459–
469.
Ruckstuhl, K. E. 1997. Foraging behaviour and sexual segregation in bighorn sheep.
Animal Behaviour 56:99–106.
Ruckstuhl, K. E., et P. Neuhaus. 2002. Sexual segregation in ungulates: a comparative test
of three hypotheses. Biological Review 77:77–96.
Schimtz, O. J. 1991. Thermal constraints and optimization of winter feeding and habitat
choice in white-tailed deer. Holarctic Ecology 14:104–111.
Schultz, J. C. 1988. Plant responses induced by herbivory. Trends in Ecology & Evolution
3:45–49.
Shi, J., R. I. M. Dunbar, D. Buckland, et D. Miller. 2003. Daytime activity budgets of feral
goats (Capra hircus) on the Isle of Rum: influence of season, age, and sex.
Canadian Journal of Zoology 81:803–815.
Skogland, T. 1983. The effects of density dependent resource limitation on size of wild
reindeer. Oecologia 60:156–168.
Spalinger, D. E., C. T. Robbins, et T. A. Hanley. 1986. The assessment of handling time in
ruminants: the effect of plant chemical and physical structure on the rate of
breakdown of plant particles in the rumen of mule deer and elk. Canadian Journal of
Zoology 64:312–321.
Stenseth, N. C. 1981. Optimal food selection: Some further considerations with special
reference to the grazer-hunter distinction. American Naturalist 117:457–475.
Stromayer, K. A. K., et R. J. Warren. 1997. Are overabundant deer herds in the eastern
United States creating alternate stable states in forest plant communities? Wildlife
Society Bulletin 25:227–234.
Sweeney, J. M., M. E. Garner, et R. P. Burkert. 1984. Analysis of white-tailed deer use of
forest clear-cuts. Journal of Wildlife Management 48:652–655.
Tierson, W. C., G. F. Mattfeld, R. W. Sage Jr., et D. F. Behrend. 1985. Seasonal
movements and home ranges of white-tailed deer in the Adirondacks. Journal of
Wildlife Management 49:760–769.
Tilghman, N. G. 1989. Impacts of white-tailed deer on forest regeneration in northwestern
Pennsylvania. Journal of Wildlife Management 53:524–532.
Tremblay, J. 1981. Qualité de la nourriture d'été du cerf de Virginie à l'île d'Anticosti.
Mémoire de M.Sc. Université Laval, Québec, Québec, Canada.
Trudell, J., et R. G. White. 1981. The effect of forage structure and availability on food
intake, biting rate, bite size and daily eating time of reindeer. Journal of Applied
Ecology 18:63–81.
125
Tufto, J., R. Andersen, et J. Linnell. 1996. Habitat use and ecological correlates of home
range size in a small cervid: the roe deer. Journal of Animal Ecology 65:715–724.
Turner, M. G., et R. H. Gardner. 1995. Ecological dynamics at broad scales. Bioscience
45:S29–S35.
Van der Wall, R., N. Madan, S. van Lieshout, C. Dormann, R. Langvatn, and S. D. Albon.
2000. Trading forage quality for quantity? Plant phenology and patch choice by
Svalbard reindeer. Oecologia 123:108–115.
Van Soest, P. J. 1982. Nutritional ecology of the ruminant: ruminant metabolism,
nutritional strategies, the cellulolytic fermentation, and the chemistry of forages and
plant fibers. O & B Books, Corvallis, Oregon, USA.
Van Soest, P. J. 1994. Nutritional ecology of the ruminant, 2nd edition. Cornell University
Press, Ithaca, New York, USA.
Vivås, H. J., et B.-E. Sæther. 1987. Interactions between a generalist herbivore, the moose
Alces alces and its food resources: an experimental study of winter foraging
behaviour in relation to browse availability. Journal of Animal Ecology 56:509–
520.
White, G. C., et R. A. Garrott. 1990. Analysis of wildlife radio-tracking data. Academic
press, San Diego, California, USA.
Wickstrom, M. L., C. T. Robbins, T. A. Hanley, D. E. Spalinger, et S. M. Parish. 1984.
Food intake and foraging energetics of elk and mule deer. Journal of Wildlife
Management 48:1285–1301.
Zagata, M. D., et A. O. Haugen. 1974. Influence of light and weather on observability of
Iowa deer. Journal of Wildlife Management 38:220–228.
Zhang, E. 2000. Daytime activity budgets of the Chinese water deer. Mammalia 64:163–
172.
126
Appendix 3−1. Spatial statistical data of biomass (a) lateral cover (b) and canopy cover (c)
in cuts and forests of enclosures containing different densities of white-tailed deer on
Anticosti Island, Québec.
Year Block Density Stratum Na Rangeb C0c Cd RMSEe C-Ratiof
(deer/km²) (m)
a) biomass 2002 A 7.5 Cutg 87 245 1.07 1.08 0.96 0.50
Forest 22 240 1.19 1.18 0.85 0.50
15 Cut 30 280 0.00 1.46 1.92 0.00
Forest 21 250 1.07 1.13 0.82 0.49
2003 A 7.5 Cut 87 280 0.91 0.16 0.51 0.85
Forest 21 180 0.27 0.07 0.78 0.80
15 Cut 30 250 0.87 0.00 0.58 1.00
Forest 21 150 0.78 0.51 0.27 0.60
2003 B 7.5 Cut 53 120 1.01 1.16 0.97 0.46
Forest 46 420 0.07 0.06 1.10 0.57
15 Cut 26 300 0.76 0.21 0.77 0.79
Forest 34 300 0.38 0.16 1.00 0.71
2003 C 7.5 Cut 61 315 0.38 0.51 0.71 0.43
Forest 33 180 0.56 0.00 1.22 1.00
15 Cut 27 350 0.89 0.00 0.73 1.00
Forest 27 315 0.23 0.14 0.88 0.62
b) lateral cover 2002 A 7.5 Cut 87 420 1.06 0.71 1.07 0.60
Forest 22 280 10.1 2.66 1.12 0.79
15 Cut 30 250 0.45 0.01 1.03 0.98
Forest 21 180 6.34 17.8 0.92 0.26
2003 A 7.5 Cut 87 240 0.00 6.63 2.38 0.00
Forest 21 180 13.5 3.42 1.01 0.80
15 Cut 30 420 0.16 0.14 1.19 0.54
Forest 21 180 6.04 17.9 0.92 0.25
2003 B 7.5 Cut 53 350 1.64 3.45 1.21 0.32
Forest 46 500 19.2 7.42 0.97 0.72
15 Cut 26 420 0.20 0.14 1.12 0.59
Forest 34 300 11.5 11.8 1.03 0.49
2003 C 7.5 Cut 61 350 2.04 1.41 1.07 0.59
Forest 33 600 12.6 4.37 1.00 0.74
15 Cut 27 500 0.32 0.76 1.26 0.29
Forest 27 660 5.89 28.7 1.10 0.17
c) canopy cover 2002 A 7.5 Cut 86 360 1.03 0.40 0.99 0.72
Forest 22 180 0.47 0.66 1.02 0.41
15 Cut 30 200 0.64 0.00 1.01 1.00
Forest 21 276 0.82 0.77 0.92 0.52
127
Year Block Density Stratum Na Rangeb C0c Cd RMSEe C-Ratiof
(deer/km²) (m)
2003 A 7.5 Cut 87 350 0.61 0.28 1.02 0.68
Forest 21 150 0.00 1.50 0.86 0.00
15 Cut 30 150 0.85 0.00 0.99 1.00
Forest 21 120 1.00 0.52 0.90 0.66
2003 B 7.5 Cut 53 450 0.22 0.49 0.98 0.31
Forestg 46 300 0.28 0.32 0.98 0.47
15 Cut 26 300 0.43 0.00 1.07 1.00
Forest 34 300 0.27 0.30 1.00 0.47
2003 C 7.5 Cut 61 420 0.31 0.27 1.01 0.54
Forest 33 350 0.32 0.35 1.09 0.48
15 Cut 27 300 0.10 0.09 0.93 0.53
Forest 27 420 0.08 0.25 1.08 0.25 a Number of sample points where biomass, canopy and lateral cover were measured and
used to quantify the semivariogram and cross-validations.
b Range (m) is the distance over which spatial autocorrelation was detected.
c The nugget effect (C0) or the inherent data variability below which the minimum lag
distance cannot be modelled with the current sampling resolution.
d The variance (C) associated to the spatial variability in the data.
e Biases in estimation errors evaluated by the standardised root mean squared error (RMSE)
from the cross-validation analysis. The RSME should be close to 1 if the predicted standard
errors are valid. If the RSME is greater than 1, the variability is underestimated in the
predictions. If the RSME is less than 1 the variability is overestimated.
f The C-ratio, given by the nugget value divided by the total variance (sill or (C0+C)),
defines the asymptotic value of semivariance and gives an estimation of the dependence
between estimated values of biomass or cover and their position in space. A ratio lower
than 0.25 usually represents values that are highly spatially correlated, a ratio between 0.25
and 0.75 usually corresponds to values that are moderately correlated and a ratio near 1
represents values that are not spatially correlated (Jurado-Expòsito et al. 2004).
g These models were developed with directional semivariograms (direction of 350º) as they
increased the fit of the estimated values.
Top Related