sÉlection d'habitat du liÈvre d'amÉrique en forÊt …le chapitre 2 a été publié...
TRANSCRIPT
JAMES HODSON
SÉLECTION D'HABITAT DU LIÈVRE D'AMÉRIQUE
EN FORÊT BORÉALE IRRÉGULIÈRE AMÉNAGÉE
Thèse présentée
à la Faculté des études supérieures de l‘Université Laval
dans le cadre du programme de doctorat en biologie
pour l‘obtention du grade de Philosophiae Doctor (Ph.D.)
DÉPARTEMENT DE BIOLOGIE
FACULTÉ DES SCIENCES ET GÉNIE
UNIVERSITÉ LAVAL
QUÉBEC
2011
© James Hodson, 2011
ii
Résumé
Cette thèse examine comment les perturbations naturelles et anthropiques façonnent la
répartition du lièvre d'Amérique (Lepus americanus). J'ai d‘abord étudié les variations
d'abondance du lièvre le long d‘un gradient de succession forestière, ainsi que l'influence de
la dynamique de trouées sur sa répartition à fine échelle en forêt ancienne. J'ai ensuite
évalué la réaction du lièvre à divers traitements sylvicoles, dont certains visaient à
maintenir la structure irrégulière des peuplements anciens. Pour ce faire, j‘ai étudié la
sélection de l'habitat du lièvre par l'approche d'isodars et reconstitué l'historique de
broutement. L'abondance du lièvre suivait une distribution bimodale avec l'âge des
peuplements, avec un premier pic 40-50 ans après perturbation et un second pic, moins
prononcé, en fin de succession. Les peuplements anciens étaient caractérisés par de
nombreuses trouées dans lesquelles l'abondance de nourriture du lièvre était relativement
élevée. Les comportements d'approvisionnement et de déplacement du lièvre indiquaient
toutefois qu'il percevait un risque de prédation plus élevé à l'intérieur des trouées. La
structure des peuplements anciens semble donc imposer un compromis entre l‘acquisition
de nourriture et l‘évitement des prédateurs. La réaction du lièvre à la coupe forestière
dépendait à la fois de l'intensité du traitement sylvicole et de la densité locale de la
population. Dans le cas des coupes avec une rétention d'arbres >50%, la préférence pour la
forêt non coupée disparaissait à mesure que la population locale augmentait. Au contraire,
pour les traitements avec une rétention d'arbres <20%, la préférence pour la forêt non
coupée s'intensifiait avec l'augmentation de la population locale. De même, au cours des
premières années après coupe, les patrons de broutement des tiges de bouleau (Betula
papyrifera) dans les traitements à rétention >50% sont demeurés similaires à ceux des
forêts non coupées, tandis que l'utilisation des tiges a diminué dans les traitements intensifs
(rétention <20%). Ces résultats démontrent que les traitements sylvicoles qui conservent la
structure des forêts anciennes peuvent aussi maintenir une répartition du lièvre
caractéristique de ce stade de succession. Cette thèse approfondit notre compréhension des
liens entre la répartition du lièvre et les régimes de perturbations régionales en forêt boréale
aménagée.
iii
Abstract
This thesis explores how different components of natural and human disturbance regimes
shape the distribution of a key boreal forest herbivore, the snowshoe hare (Lepus
americanus). I investigated both broad-scale changes in hare abundance during forest
succession and fine-scale responses to heterogeneity created by canopy gap dynamics in
old-growth forests. I then evaluated how hare respond to silvicultural treatments designed
to maintain the irregular structure of old-growth stands using patterns of density-dependent
habitat selection and browse history reconstruction. Snowshoe hare followed a bimodal
abundance distribution with stand age, with a pronounced peak in density between 40-50
years post-disturbance followed by a second more subtle increase phase during late-
succession. Within old-growth stands, canopy gaps offered areas of higher food
availability, but foraging and movement behaviours indicated that hares perceived a greater
risk of predation within openings. The structure of old-growth stands thus appears to
impose a trade-off between acquiring food and avoiding predation. The response of
snowshoe hare to forest harvesting depended on both disturbance intensity and local
population density. Preference for uncut forest stands over harvest treatments with >50%
tree retention quickly diminished as local populations increased. In contrast, preference for
uncut forests over treatments with <20% tree retention became more pronounced with
increasing local population density. Similarly, in the first years following harvesting,
browse use patterns of white birch (Betula papyrifera) stems in low intensity treatments
(>50% retention) remained similar to those in uncut old-growth forest stands, whereas
browse use declined rapidly in intensive harvest treatments (<20% retention) over the same
period. These findings suggest that silvicultural treatments that conserve old-growth forest
structure can also maintain distributions of hare that are characteristic of late-succession.
This thesis helps to further our understanding of the links between snowshoe hare
distribution and regional disturbance regimes in managed boreal forests.
iv
Avant-propos
Ce doctorat est présenté sous la forme d’une thèse avec insertion de quatre articles
scientifiques. La thèse inclut une introduction générale et une conclusion générale qui lient
l’ensemble des articles. En tant qu’auteur principal des quatre articles, j’ai élaboré les objectifs
de recherche, j’ai planifié et réalisé l'échantillonnage sur le terrain, j’ai effectué les analyses
statistiques et j’ai rédigé les manuscrits. Mon directeur de recherche, Daniel Fortin, a
largement contribué aux étapes de l'élaboration des objectifs et des protocoles
d'échantillonnage, ainsi qu'à l'analyse statistique et la rédaction d'articles. Mon co-directeur de
recherche, Louis Bélanger, a également participé à l'élaboration des objectifs et à la rédaction
des articles.
Le chapitre 2 a été publié dans Journal of Mammalogy, avec Daniel et Louis comme co-
auteurs.
Dans le chapitre 3, Mélanie-Louise Le Blanc m'a gracieusement fourni ses données sur les
captures de campagnols à dos roux et ses relevés de végétation dans les dispositifs de coupes
expérimentales. Elle était aussi impliquée dans la rédaction de l'article. Cet article est publié
dans Oecologia et Daniel, Louis et Mélanie sont les coauteurs.
Je dois également souligner la contribution d'Etienne Renaud-Roy, un étudiant au baccalauréat
que j'ai encadré avec Daniel Fortin au cours de son initiation à la recherche. Étienne a
contribué beaucoup à l'élaboration des méthodes pour les inventaires d'historique de
broutement présenté dans le chapitre 4, ainsi qu'à la saisie des données et aux analyses
statistiques préliminaires.
v
Remerciements
Tous ces travaux n‘auraient pas été possibles sans la contribution et le support de plusieurs
personnes tout au long de mon doctorat.
J‘aimerais tout d‘abord remercier ma copine, Kim Poitras, que j‘ai connue en tant
qu‘assistante au cours de ma première session de terrain d‘hiver en 2006. Kim m‘a aidé
énormément avec la conception et la mise en œuvre de mes expériences sur le
comportement d‘approvisionnement du lièvre, ce qui a nécessité de longues heures de
travail tant sur le terrain qu‘une fois de retour dans les camps forestiers en soirée. Kim a
vécu toutes mes frustrations et mes succès et m‘a toujours donné beaucoup
d‘encouragements et un support moral tout au long de mon doctorat. Elle m'a aussi donné
un coup de main énorme avec la traduction de l'introduction et de la conclusion de ma
thèse, ainsi qu'avec les résumés de chaque chapitre. Sans toi, je crois que je ne serais
jamais parvenu à finir. Merci Kim.
Merci à Daniel Fortin, mon directeur de recherche, pour m‘avoir motivé à pousser mon
projet de doctorat aussi loin que possible, pour nos échanges d'idées, pour ses conseils
statistiques et ses révisions de manuscrits souvent plus rapides que j'aurais cru possible, et
pour ses connaissances profondes et son enthousiasme pour l'écologie.
Un grand merci aussi à Louis Bélanger, mon co-directeur, pour ses révisions de manuscrits
et pour toutes les discussions intéressantes que nous avons eues dans son bureau au sujet du
lièvre et de l'aménagement écosystémique. J'ai vraiment apprécié le fait d'avoir eu des
directeurs de recherche qui avaient des intérêts et des points de vues différents mais
complémentaires.
Merci à Marc Mazerolle et à André Desrochers pour tous leurs conseils en statistiques.
vi
Les quatre étés et deux hivers que j‘ai passés sur le terrain ont nécessité beaucoup de main
d‘œuvre sous des conditions parfois très difficiles : beaucoup de pluie, beaucoup de neige,
beaucoup de mouches, de longues journées de travail et un terrain souvent plus accidenté
que nous l‘aurions souhaité. À travers tout ça, j‘ai eu beaucoup de plaisir avec tous mes
assistants de terrain et j‘espère qu‘ils ont vécu de belles expériences tout en apprenant
beaucoup sur la recherche en écologie terrestre. J‘aimerais les remercier : Krystel
Hammelin et Jean-François Poulin (été 2005), Jonathan Leclair (hiver 2006), Julie
Tremblay et Marc-Andrée Larose (été 2006), Olivier Deshaies et Jean-Simon Roy (hiver
2007), Etienne Renaud-Roy et Valérie Hébert-Gentille (été 2007), et finalement Sébastien
Lavoie et Marianne White (été 2008). Au cours de ces années de terrain, tous ces assistants
m‘ont aidé à compter un total de 53 632 crottins, 137 330 ramilles, 47 911 gaules, et 3 974
arbres matures!
Je tiens aussi à remercier tous les membres du labo Fortin pour avoir créé une excellente
atmosphère de travail tout au long de mon doctorat. Merci à Mélanie Le Blanc et Jérôme
Lemaître, mes compagnons de bureau, pour toutes les bonnes discussions en écologie et en
aménagement forestier, et pour les pauses d‘humour et de musique pendant les longues
journées de travail. Merci aussi à Nicolas Courbin et Guillaume Bastille-Rousseau pour
leur aide en géomatique et en statistiques, pour les matchs de squash et pour les 5 à 7. Un
gros merci à Nicolas pour m‘avoir accueilli chez lui pendant mon séminaire de doctorat et
pour m‘avoir supporté pendant les derniers moments de stress avant ma présentation.
Merci aussi à Guillaume pour le tour guidé de son village natal de Pohénégamook et pour
m‘avoir introduit à l‘ARGACÉ. Merci beaucoup à Sabrina Courant pour la relecture finale
de ma thèse et les corrections de fautes de français. Finalement, merci à Mélina Houle,
Jean-Sébastien Babin, Ermias Azeria, David Pinaud, Cheryl Johnson, Guillaume Moreau,
Pierre Etcheverry, Philippe Janssen, Mathieu Basille et Marie-Ève Fortin pour toutes les
bonnes conversations lors du lunch et des sorties de labo.
vii
Pour finir, un gros merci à mes parents pour leur support moral et financier tout au long de
mon doctorat et pour leur révision de mes manuscrits. Merci à mon père d'être venu me
visiter sur la Côte Nord pour vivre une expérience de terrain en hiver.
À Kim et mes parents ...
ix
Table des matières
Résumé ............................................................................................................................................. ii
Abstract .......................................................................................................................................... iii
Avant-propos ................................................................................................................................ iv
Remerciements ............................................................................................................................. v
Table des matières ...................................................................................................................... ix
Liste des tableaux ...................................................................................................................... xii
Liste des figures .......................................................................................................................... xv
Introduction ................................................................................................................................... 1 Écosystèmes forestiers et changements dans la gestion de la forêt ............................................. 1 Hétérogénéité et sélection de l’habitat ..................................................................................................... 3 Perturbations naturelles, hétérogénéité et répartition des animaux ........................................... 4 Utilisation des théories de sélection d’habitat pour comprendre la réaction animale aux perturbations environnementales .............................................................................................................. 7 Influence des prédateurs sur la sélection d'habitat des proies ...................................................... 8 Influence de l'hétérogénéité du risque de prédation sur les comportements d’approvisionnement et de déplacement des proies à fine échelle .............................................. 9 Forêts boréales, régimes de perturbation et succession écologique ......................................... 10
Espèce cible : le lièvre d'Amérique ............................................................................................... 14 Objectif et organisation de la thèse ............................................................................................... 16 Aire d’étude: la forêt boréale irrégulière de l’est du Québec ............................................... 18
Chapitre 1 ..................................................................................................................................... 20
Changes in relative snowshoe hare abundance across a 265-year gradient of boreal forest succession ......................................................................................................... 20
Résumé .................................................................................................................................................... 21 Abstract ................................................................................................................................................... 22 Introduction ........................................................................................................................................... 23 Methods ................................................................................................................................................... 26
Study Area .......................................................................................................................................................... 26 Site Selection ..................................................................................................................................................... 27 Pellet inventories ............................................................................................................................................ 28 Habitat structure ............................................................................................................................................. 29 Gap transects in mature and late-seral stands .................................................................................... 29 Statistical Analysis .......................................................................................................................................... 30 Changes in cover, browse, and hare abundance with time since disturbance ...................... 30 Variations in hare abundance with cover and browse availability ............................................ 31 Variations in hare abundance with canopy gap fraction in stands ≥80 years old ............... 32 Relative snowshoe hare abundance in stands regenerating from fire versus clearcutting ................................................................................................................................................................................ 32
Results ...................................................................................................................................................... 33 Changes in relative snowshoe hare abundance and habitat structure with stand age ...... 33 Changes in relative snowshoe hare abundance with habitat structure and food availability .......................................................................................................................................................... 34 Relative snowshoe hare abundance in fire- versus harvest-origin stands ............................. 35
x
Discussion ............................................................................................................................................... 35 Acknowledgements ............................................................................................................................. 40
Chapitre 2 ..................................................................................................................................... 54
Fine-scale disturbances shape space-use patterns of a boreal forest herbivore 54 Résumé .................................................................................................................................................... 55 Abstract ................................................................................................................................................... 57 Introduction ........................................................................................................................................... 58 Methods ................................................................................................................................................... 60
Study area........................................................................................................................................................... 60 Cover and browse availability within canopy gaps and under forest cover .......................... 61 Stand level habitat selection ....................................................................................................................... 62 Fine-scale movements .................................................................................................................................. 64 Giving-up densities ......................................................................................................................................... 66 Use of natural browse within canopy gaps .......................................................................................... 68
Results ...................................................................................................................................................... 68 Browse within canopy gaps ........................................................................................................................ 68 Winter habitat selection at the stand level ........................................................................................... 69 Fine scale movements ................................................................................................................................... 69 Giving-up densities ......................................................................................................................................... 70 Natural browse use ........................................................................................................................................ 71
Discussion ............................................................................................................................................... 72 Multi-trophic implications of habitat heterogeneity resulting from gap dynamics ............ 76
Acknowledgements ............................................................................................................................. 77
Chapitre 3 ..................................................................................................................................... 88
An appraisal of the fitness consequences of forest disturbance for wildlife using habitat selection theory .......................................................................................................... 88
Résumé .................................................................................................................................................... 89 Abstract ................................................................................................................................................... 91 Introduction ........................................................................................................................................... 92
Incorporating continuous habitat variables into isodar models: an example with forest disturbance ........................................................................................................................................................ 93
Methods ................................................................................................................................................... 97 Study Area .......................................................................................................................................................... 97 Experimental Harvest Blocks ..................................................................................................................... 98 Relative snowshoe hare density ............................................................................................................... 99 Relative red-backed vole density ........................................................................................................... 100 Measures of disturbance intensity and resource availability ..................................................... 101 Isodar analysis................................................................................................................................................ 102
Results ................................................................................................................................................... 102 Habitat disturbance ...................................................................................................................................... 102 Isodar analysis................................................................................................................................................ 103
Discussion ............................................................................................................................................ 105 Acknowledgements .......................................................................................................................... 110
Chapitre 4 ................................................................................................................................... 120
Browse history as an indicator of snowshoe hare response to silvicultural practices adapted for irregular boreal forests ............................................................. 120
xi
Résumé ................................................................................................................................................. 121 Abstract ................................................................................................................................................ 122 Introduction ........................................................................................................................................ 123 Methods ................................................................................................................................................ 125
Study Area ........................................................................................................................................................ 125 Experimental Blocks .................................................................................................................................... 126 Habitat structure and browse availability .......................................................................................... 127 Browse History .............................................................................................................................................. 128 Statistical Analysis ........................................................................................................................................ 130 Habitat structure and browse availability .......................................................................................... 130 Browse History .............................................................................................................................................. 131
Results ................................................................................................................................................... 132 Habitat structure and browse availability .......................................................................................... 132 Browse History .............................................................................................................................................. 133
Discussion ............................................................................................................................................ 134 Acknowledgements .......................................................................................................................... 137
Conclusion générale ............................................................................................................... 147 Changements d'abondance relative du lièvre au cours d’une succession forestière après feu et après coupe totale ............................................................................................................................ 148 Influence de la dynamique de trouées sur la répartition du lièvre en fin de succession 151 Régimes de perturbations boréales, aménagement écosystémique et influence de la récolte ligneuse sur la faune ..................................................................................................................... 154 Orientations des recherches futures ..................................................................................................... 159
Bibliographie ............................................................................................................................ 161
Appendice 1 ............................................................................................................................... 189 a) Sample photographs of different stand ages sampled within the post-harvest/post-fire forest chronosequence ............................................................................................................................... 189 b) Simulations to illustrate the effect of forest age-class distribution on snowshoe hare abundance under three disturbance regimes ................................................................................... 190
Appendice 2 ............................................................................................................................... 195 Sample photographs of canopy gaps originating from a) tree mortality and b) edaphic conditions ......................................................................................................................................................... 195
Appendice 3 ............................................................................................................................... 196 Sample photographs of the four browse stem architecture types considered during browse history surveys. ............................................................................................................................. 196
xii
Liste des tableaux
Table 1.1. Fit statistics for general additive models (GAMs) used to model snowshoe hare
pellet density, vertical cover, lateral cover, and browse availability as a function of
stand age in a 265 yr boreal forest chronosequence of stand development (n = 84
stands)........................................................................................................................42
Table 1.2. Competing models predicting the density of snowshoe hare pellets along a
chronosequence of forest stand development (n = 84) for stands aged between 3 and
265 years based on combinations of lateral cover, vertical cover, and browse
availability, including models with interactions between a dichotomous variable
(Dev_Phase: 0 = stands <80 years, 1 = stands >80 years) and lateral cover, vertical
cover, and browse availability to assess whether the importance of factors limiting
hare density varies between two phases of stand development.................................43
Table 1.3. Parameter estimates for the top ranking model predicting pellet density in a
forest chronosequence of stands (n = 84) varying in age between 3 and 265
years...........................................................................................................................44
Table 1.4. Competing models predicting the density of snowshoe hare pellets in stands >80
years old based on the fraction of all types of canopy gaps versus only the fraction
of mortality-origin canopy gaps................................................................................45
Table 1.5. Parameter estimates from the top-ranking model predicting snowshoe hare pellet
density as a function of mortality-origin canopy gap fraction in stands >80 years
old..............................................................................................................................46
Table 2.1. Mean (± 1 SE) deciduous browse density and conifer sapling density within
canopy gaps of edaphic and mortality origin in eastern Canadian boreal conifer
stands (80 to 200+ years), and Wilcoxon signed-rank tests (S) of paired differences
between browse and conifer density between gaps and adjacent forest
cover..........................................................................................................................78
Table 2.2. Competing models of resource selection by snowshoe hares using logistic
regression to compare points observed (n = 125) along winter snowshoe hare trails
to randomly located points (n = 184) within eastern Canadian boreal conifer stands
(>90 years).................................................................................................................79
Table 2.3. Model-averaged coefficients ( ) and unconditional standard errors (SE( )) for
habitat variables used in resource selection functions comparing points observed (n
=125) along winter snowshoe hare trails to randomly located points (n = 184), step-
selection functions for winter snowshoe hare trails (n = 105 observed step
segments), and analysis of movement speed by snowshoe hares along 10 bound
segments of winter trails in eastern Canadian boreal conifer stands (>90
years).........................................................................................................................80
xiii
Table 2.4. Mean (± 1 SE) values of habitat variables measured at points along single winter
snowshoe hare trails (n = 125) and randomly located points (n = 184) used in
resource selection functions within eastern Canadian boreal conifer stands (>90
years).........................................................................................................................81
Table 2.5. Competing models for step-selection functions along single winter snowshoe
hare trails (n = 105 observed step segments) in eastern Canadian boreal conifer
stands (>90 years)......................................................................................................82
Table 2.6. Mean (± 1 SE) values of habitat variables measured along 10-bound segments (n
= 105) and paired random segments from 16 single winter snowshoe hare trails, and
mean paired differences between values along observed and random segments used
in step-selection functions within eastern Canadian boreal conifer stands (>90
years).........................................................................................................................83
Table 2.7. Competing models of the influence of cover availability on movement speed,
estimated as the distance travelled in 10-bound segments (n = 105), along single
winter snowshoe hare trails (n = 16) in eastern Canadian boreal conifer stands (>90
years old)...................................................................................................................84
Table 3.1. Mean (range) canopy cover, deciduous browse availability and moss cover in
four types of silvicultural treatment and adjacent uncut forests measured within (a)
snowshoe hare pellet grids, and (b) along red-backed vole trap
lines.........................................................................................................................111
Table 3.2. Comparison based on Akaike‘s Information Criterion corrected for small sample
sizes (AICc) of isodar models predicting snowshoe hare (a) and red-backed voles
(b) density in uncut forests (NU) based on the density of hare and voles in harvested
stands (NH), disturbance intensity (D) measured as the percent difference in canopy
cover between uncut stands and adjacent harvested stands, as well as percent
differences in browse availability (browse) or moss cover (moss).........................113
Table 3.3. Parameter estimates and 95% confidence intervals (CI) for the top (ΔAICc <2)
isodar models describing snowshoe hare and red-backed vole distribution in pairs of
uncut and harvested boreal forest stands.................................................................114
Table 4.1. Comparison of different habitat characteristics among four silvicultural
treatments, and between cut and adjacent uncut forest stands by treatment type, in
four experimental blocks in the Côte-Nord region of Québec using mixed effects
analysis of variance.................................................................................................139
Table 4.2. Type III tests of fixed-effects, parameter estimates (β ±SE), and t-tests of
parameter estimates from a mixed-model logistic regression of the probability of
white birch stem use by snowshoe hare as a function of harvest treatment (SCPerm,
SCTemp, CPPTM, and CPRS; abbreviations described in legend for Figure 4.1),
harvest status (Cut = 1, Uncut = 0), and the year relative to when harvesting took
place (0-3 years, with 0 being the winter before harvesting), recorded from browse
xiv
history surveys in four experimental harvest blocks in the Côte-Nord region of
Québec.....................................................................................................................140
Table A1.1 Estimated snowshoe hare population size for a hypothetical 1000 ha forest
landscape with a 250-year fire cycle and a negative exponential forest age-class
distribution...............................................................................................................193
Table A1.2. Estimated snowshoe hare population size for a hypothetical 1000 ha forest
landscape under fully regulated even-aged management with a harvest rotation of
100 years..................................................................................................................194
Table A1.3. Estimated snowshoe hare population size for a hypothetical 1000 ha forest
landscape under cohort management assuming a 200-year fire cycle and a 100-year
maximum harvest rotation age for stands under even-aged management (following
Bergeron et al. 2002)...............................................................................................195
xv
Liste des figures
Figure I-1. Exemple hypothétique de la répartition d'individus entre deux habitats de qualité
différente (A et B) selon la théorie de la distribution idéale libre et l'isodar qui en
résulte (adapté de Fretwell and Lucas 1970 et Morris 1988)......................................6
Figure 1.1. Left panel: Map of the study area located in Québec‘s North Shore region
showing the location of harvest and fire origin stands that were sampled for relative
snowshoe hare abundance. Right panel: Pellet inventory grids used to measure
relative snowshoe hare abundance............................................................................47
Figure 1.2. General additive models (GAMs; solid lines) ± approximate 95% confidence
intervals (dotted lines) describing changes in snowshoe hare pellet density, vertical
cover, lateral cover, and browse availability with stand age in a boreal forest
chronosequence of stand development (n = 84 stands).............................................48
Figure 1.3. Predicted values (solid lines) ± 95% confidence intervals (dotted lines) of
snowshoe hare pellet density as a function of vertical cover, lateral cover between 1-
2m and browse availability in two phases of stand development (Dev_Phase: <80
years = 0, ≥80 years = 1) using parameter estimates from the model: (pellets/m²)0.5
= 0.104 + 0.446*Dev_Phase + 0.022*LatCov1-2 – 0.021*LatCov1-2×Dev_Phase +
0.016*VertCov – 0.017*VertCov×Dev_Phase + 0.050*Browse –
0.003*Browse²...........................................................................................................50
Figure 1.4. Predicted pellet density as a function of the proportion of stands in mortality-
origin canopy gaps in stands ≥80 years old...............................................................52
Figure 1.5. Boxplots of snowshoe hare pellet density in stands originating from forest fires
and clearcutting in four different stand age classes...................................................53
Figure 2.1. Predicted probability of jack pine bough use by snowshoe hares as a function of
habitat (Gap vs. Forest), the number of nights boughs were left within gaps and
adjacent forest, and the distance of boughs (n = 846 boughs) placed within canopy
gaps (n = 45 gaps) to the gap edge, in eastern Canadian boreal conifer stands (>90
years).........................................................................................................................85
Figure 2.2. Predicted probability (±1 SE) of natural browse use by snowshoe hares as a
function of habitat (Gap vs. Forest) and distance of stems (n = 1269 stems) to the
gap edge, within edaphic and mortality origin canopy gaps (n = 61 gaps) in eastern
Canadian boreal conifer stands (80 to >200 years)...................................................86
Figure 2.3. Snowshoe hare foraging behaviour captured from motion sensitive cameras
installed at canopy gaps with GUD experiments in eastern Canadian boreal conifer
stands (> 90 years)....................................................................................................87
xvi
Figure 3.1. Four scenarios of expected fitness (W)-density (N) functions (left-hand side)
and corresponding isodars (eq. [4]: NU = β0 + β1 D + β2 NH + β3[D×NH]; right-hand
side) for pairs of uncut (U) forest (solid line) and stands harvested (H) (dashed
lines) at different levels of disturbance intensity (D; the percent difference in canopy
cover [20-80%] relative to an adjacent uncut forest): a) if the effect of harvesting is
not related to measures of disturbance intensity, b) if disturbance has only
quantitative effects on habitat quality, c) if disturbance has only qualitative effects
on habitat quality, and d) if disturbance affects habitat quality both quantitatively
and qualitatively......................................................................................................115
Figure 3.2. Observed relative densities of (a) snowshoe hare (n = 27) and (b) red-backed
voles (n = 25) in pairs of uncut forest and four different silvicultural treatments
(CPRS, CPPTM, SCTemp, and SCPerm; abbreviations are defined in Table 1)
sampled in two consecutive years...........................................................................117
Figure 3.3. Estimated isodar curves (left side) and corresponding relative fitness vs. density
(right side) curves according to the mean percent difference in (a,b) canopy cover
(D) for snowshoe hare and, (c-f) moss cover (moss) for red-backed voles, between
four different silvicultural treatments and adjacent uncut forests...........................118
Figure 4.1. Sampling design for structural habitat features, browse availability and browse
history surveys within survey grids installed in pairs of uncut forest and stands cut
using four different silvicultural treatments............................................................142
Figure 4.2. Structural habitat features measured within pairs of uncut irregular boreal forest
stands (light grey bars) and stands harvested using four different types of
silvicultural treatment (dark grey bars; CPRS = cutting with protection of
regeneration and soils, CPPTM = irregular shelterwood cutting leaving small
merchantable stems, SCTemp = selection cutting with temporary trails, and SCPerm
= selection cutting with permanent trails)...............................................................143
Figure 4.3. Mean proportions of birch browse stems of each stem architecture type from
qualitative browse history surveys in four harvest treatment types and paired uncut
forests in four experimental harvest blocks in Québec‘s North Shore
region.......................................................................................................................145
Figure 4.4. Predicted probability (± 95% CI) of white birch stem use by snowshoe hare in
four different harvest treatments (SCPerm, SCTemp, CPPTM, CPRS) and paired
uncut forest stands from the winter before the harvest treatment took place (Year =
0) up until three years following cutting................................................................146
1
Introduction
Les écologistes reconnaissent depuis longtemps que les écosystèmes terrestres et
aquatiques sont des environnements dynamiques constamment restructurés par des
perturbations et par la colonisation et l‘extinction d‘organismes. Dans un contexte
écologique, une perturbation peut être définie comme « n‘importe quel évènement discret
dans le temps qui modifie la structure de l‘écosystème, de la communauté ou de la
population et qui modifie l‘abondance des ressources, la disponibilité du substrat ou
l‘environnement physique » (Pickett & White 1985). Les perturbations naturelles ont donc
un rôle déterminant dans la répartition des animaux parce qu‘elles créent une hétérogénéité
dans la répartition des ressources et des structures nécessaires pour leur survie et leur
reproduction (Sousa 1984). Cette hétérogénéité peut influencer tous les niveaux
d‘organisation étudiés par les écologistes, notamment le comportement des organismes, la
dynamique des populations, les interactions trophiques, la structure des communautés et la
biodiversité (Pickett et al. 1989). La plupart des espèces ont évolué dans des
environnements où plusieurs perturbations modifient la structure et la composition de leur
habitat, de sorte que certaines espèces bénéficient des conditions créées par les
perturbations (Sousa 1984). Toutefois les humains contribuent de plus en plus à la
perturbation des écosystèmes afin de répondre à une demande toujours grandissante en
ressources naturelles (Ellis et al. 2010).
Écosystèmes forestiers et changements dans la gestion de la forêt
Les écosystèmes forestiers hébergent la plus grande proportion de la biodiversité
terrestre (Hunter 2002). Ces écosystèmes subissent une pression immense due à
l‘exploitation des ressources forestières, à la déforestation à d‘autres fins économiques et
aux changements climatiques (Foley et al. 2005). La forêt boréale est le deuxième plus
grand biome forestier. Elle couvre 11% de la surface terrestre et contribue à la régulation
globale du climat et des cycles des nutriments (Bonan & Shugart 1989, Bonan 2008).
Historiquement, la gestion de la forêt boréale avait pour objectif principal la production
2
soutenue de produits de fibres de bois et de bois d'œuvre avec peu de considération pour ses
impacts sur l‘abondance et la distribution des organismes habitant ces forêts (Kuuluvainen
2002). À cause de la perception négative qu‘a le public des pratiques forestières comme la
coupe à blanc, et des soucis pour la conservation de la biodiversité, un nouveau paradigme
en matière de gestion de la forêt a vu le jour au cours des 20 dernières années. Ce
paradigme vise l‘exploitation des peuplements forestiers en imitant des perturbations
naturelles afin de mieux conserver la biodiversité (Attiwill 1994, Bergeron & Harvey 1997,
Angelstam 1998). La prémisse de ce paradigme est que les espèces se sont adaptées à ces
perturbations naturelles sur des milliers de générations et les populations devraient donc y
être relativement résilientes (Bunnell 1995, Buddle et al. 2006).
Un aspect important de cette nouvelle approche consiste à reconnaître que les forêts
sont sujettes à une gamme de perturbations naturelles qui varient considérablement en
intensité, en fréquence et en étendue selon les régions (Attiwill 1994, Bergeron et al. 2002).
Bien que des progrès considérables ont été réalisés dans la compréhension de l‘interaction
entre les régimes de perturbation naturelle et la dynamique de la végétation forestière
(Bartemucci et al. 2002, Pham et al. 2004, Harper et al. 2005, Hart & Chen 2006), notre
compréhension de l‘influence des régimes de perturbation sur la répartition de certains
animaux demeure encore fragmentaire (Bunnell 1995).
Les animaux ont des rôles essentiels dans le fonctionnement des écosystèmes
forestiers comme consommateurs de feuillage et de fibres, consommateurs et agents de
dispersion de pollen, de graines et de champignons, et comme ingénieurs écosystémiques
capables de modifier l‘architecture des plantes et des patrons de drainage locaux (Huntly
1991, Jones et al. 1994). Les forêts peuvent aussi fournir un couvert important contre les
prédateurs et les intempéries (Demarchi & Bunnell 1995, Beaudoin et al. 2004, Dussault et
al. 2005). Les espèces sont donc non seulement influencées par la dynamique de
végétation qui suit une perturbation, mais elles peuvent aussi l‘influencer en retour. De
plus, les interactions entre les herbivores et la végétation forestière peuvent être structurées
par les prédateurs qui influencent à la fois la répartition, la densité et le comportement de
leurs proies (Sinclair et al. 2000, Ripple & Beschta 2004). Si l‘on souhaite maintenir des
communautés d'animaux dans les écosystèmes aménagés qui soient semblables à celles
produites par les régimes de perturbations naturelles régionales, nous devons d‘abord
3
comprendre comment l‘hétérogénéité environnementale influence la répartition des
animaux à travers leur sélection d‘habitat. Cette étude examine comment les perturbations
naturelles et anthropiques façonnent la répartition d‘une espèce clé de la forêt boréale, le
lièvre d‘Amérique (Lepus americanus), dans une région caractérisée par des cycles de feu
prolongés et une proportion élevée de vieilles forêts structurées par une dynamique de
trouées à fine échelle.
Hétérogénéité et sélection de l’habitat
Tous les animaux subissent certaines variations spatiales et temporelles dans la
structure et la composition de leur environnement qui créent une hétérogénéité dans la
répartition de la nourriture et du couvert protecteur (Sousa 1984). La notion d‘habitat peut
être définie de façon générale comme l‘ensemble des ressources et des conditions présentes
à l‘intérieur d‘une région qui résultent en l‘occupation de cette région par une espèce (Hall
et al. 1997). De nombreux écologistes considèrent qu‘un habitat est un ensemble de
parcelles qui sont relativement homogènes au niveau des caractéristiques physiques et
biologiques les plus pertinentes pour le comportement et la survie d‘une espèce (Fretwell &
Lucas 1970), et à l‘intérieur desquelles la densité de la population et au moins un des
paramètres de croissance démographique sont différents de ceux des habitats adjacents
(Morris 2003b). La sélection d‘habitat serait donc le processus comportemental par lequel
les individus utilisent, ou occupent, préférentiellement un ensemble non aléatoire d‘habitats
parmi ceux qui sont disponibles (Morris 2003b). L‘habitat varie donc d‘une espèce à
l‘autre selon la façon dont les animaux perçoivent et répondent à la structure des parcelles
dans leur environnement (Kotliar & Wiens 1990). La plupart des animaux pourrait
percevoir une structure hiérarchique de parcelles, composée de petites parcelles nichées à
l‘intérieur de parcelles plus grandes à mesure que l‘échelle spatiale augmente (Wiens
1976). Il est pratique d'imaginer que la sélection d‘habitat a lieu à plusieurs échelles
spatiales incluant la sélection de domaines vitaux par les individus à l‘intérieur d‘un
paysage hétérogène à l‘échelle la plus étendue, la sélection de composantes de l‘habitat à
l‘intérieur du domaine vital, et la sélection de ressources à l‘intérieur des composantes
d‘habitat à une échelle encore plus fine (Johnson 1980, Senft et al. 1987). La sélection de
4
parcelles à l'échelle la plus fine pourrait expliquer les patrons de répartition d'animaux à
plus vaste échelle autant que la sélection à l'échelle la plus étendue pourrait restreindre la
sélection de l'habitat à plus fine échelle. Il est donc probable que la répartition animale
reflète une série de décisions prises à plusieurs échelles spatiales, où la sélection de l'habitat
à chaque échelle est influencée par des décisions prises à des échelles inférieures et
supérieures.
Perturbations naturelles, hétérogénéité et répartition des animaux
Les perturbations naturelles contribuent grandement à l‘hétérogénéité spatiale et
temporelle des habitats à travers les échelles de sélection (Pickett et al. 1989).
L‘hétérogénéité temporelle est causée par la formation continuelle de nouvelles parcelles
discrètes, ce qui crée des assemblages de parcelles voisines d'âges différents, tandis que
l‘hétérogénéité spatiale résulte d‘une variation dans l‘étendue et la forme des parcelles
créées par une perturbation (Pickett et al. 1989). Des perturbations intenses telles que les
feux de forêt, les ouragans, les raz-de-marée et les inondations peuvent redémarrer la
succession écologique sur de grandes étendues (Pickett & White 1985). La dynamique des
populations animales à l‘intérieur de ces étendues peut ensuite varier selon des
changements dans la structure et la composition de la végétation depuis la perturbation
(Vanhorne 1981, Gill et al. 1996b, Larson et al. 2004). Les perturbations à fine échelle
telles que la mortalité d'arbres individuels, le chablis, l'érosion et l‘action des vagues
peuvent aussi créer des parcelles plus petites où la disponibilité de nourriture et la structure
de l‘habitat sont modifiées. Ces plus petites perturbations peuvent ainsi influencer les
patrons d‘approvisionnement et de déplacement des animaux à l‘intérieur des plus grandes
parcelles (Irlandi et al. 1995, Cramer et al. 2002, Waterhouse et al. 2002, Macia &
Robinson 2005).
En réaction à l‘hétérogénéité environnementale, les individus devraient, à des
moments opportuns, sélectionner les milieux et les ressources qui maximisent leur chance
de survie et de reproduction (Stephens & Krebs 1986, Morris 2003b, Brown & Kotler 2004,
Haugen et al. 2006). À de vastes échelles, les patrons de répartition animale devraient
refléter l‘effort de chaque individu à maximiser son aptitude phénotypique («fitness») en
5
sélectionnant les habitats ayant la plus grande qualité (MacArthur & Pinka 1966, Fretwell
& Lucas 1970, Charnov 1976). La sélection d‘habitat dépend aussi de la densité
d‘individus car l'aptitude phénotypique individuelle à l‘intérieur d‘un habitat diminue à
mesure que la densité de population augmente, puisque les ressources sont partagées entre
un plus grand nombre d‘individus (Fretwell and Lucas 1970; Morris 1988). Les individus
devraient d‘abord se regrouper dans les meilleurs habitats, mais à mesure que l'aptitude
phénotypique diminue avec l‘augmentation de la densité de compétiteurs, quelques
individus devraient être en mesure d‘obtenir une aptitude phénotypique égale en se
déplaçant dans les habitats qui étaient initialement de qualité inférieure (Fretwell & Lucas
1970, Morris 2006). Selon la théorie de la distribution idéale libre, si les individus sont
libres de s‘installer dans l‘habitat qui maximise leur aptitude phénotypique, la répartition
d'une population entre les habitats à l‘équilibre devrait être telle qu‘aucun individu ne peut
améliorer son aptitude phénotypique en se déplaçant dans un autre habitat (Fretwell &
Lucas 1970). L'aptitude phénotypique moyenne devrait alors être équivalente dans tous les
habitats. Selon une telle répartition, la densité relative des individus dans les différents
habitats devrait être un bon indicateur des différences relatives de qualité entre ces habitats
d‘un point de vue adaptatif.
Les changements temporels de densité animale suite à une perturbation sur de
grandes superficies devraient refléter les changements de composition et de structure
végétale qui influencent la disponibilité de couvert protecteur et de nourriture (p. ex.
Beckwith 1954, Odum 1969, Kirkland 1990, Fisher & Wilkinson 2005). Une
compréhension de ces changements est essentielle puisqu‘elle nous permet de prédire
comment la répartition animale devrait changer selon la fréquence et l'étendue des
perturbations à vaste échelle (Larson et al. 2004). La composition végétale reflète la
diversité et l'abondance relative de différentes espèces végétales. Cette composition peut
donc avoir une influence importante sur la qualité et la quantité des ressources alimentaires
disponibles pour la faune, tout particulièrement pour les herbivores. La quantité et la
qualité des ressources alimentaires peuvent déterminer en grande partie l'aptitude
phénotypique maximale qui peut être atteinte dans un habitat lorsque la densité d'individus
est faible (Morris 1988, 1990). La structure de l'habitat représente, quant à elle,
l'arrangement horizontal et vertical des composantes biotiques et abiotiques d'un habitat.
6
Elle peut influencer l'efficacité avec laquelle les animaux sont capables d'extraire les
ressources (Jones et al. 2001), et ainsi déterminer le taux de déclin en aptitude
phénotypique lors d'une augmentation de densité des individus (Morris 1988, 1990).
Morris (1988) a développé une approche basée sur la théorie de la distribution
idéale libre (Fretwell & Lucas 1970) qui se sert des patrons de densité animale dans
plusieurs paires d'habitats adjacents pour révéler comment leurs différences peuvent réguler
la répartition des populations. Cette approche évalue la relation entre la densité d'individus
dans un habitat par rapport à la densité d‘individus dans un autre. L'isodar est la courbe de
régression qui représente la répartition d'individus à laquelle l'aptitude phénotypique est
égale dans les deux habitats (Figure I-1).
Figure I-1. Exemple hypothétique de la répartition d'individus entre deux habitats (A et B)
de qualité différente selon la théorie de la distribution idéale libre et l'isodar qui en résulte
(adapté de Fretwell and Lucas 1970 et Morris 1988).
L'ordonnée à l'origine de l'isodar indique la différence relative entre les deux
habitats en termes d'aptitude phénotypique maximale qui peut être atteinte à faible densité
(aspect quantitatif). La pente de l'isodar indique le taux relatif auquel l'aptitude
phénotypique diminue dans chaque habitat à mesure que la densité d'individus augmente
(aspect qualitatif) (Morris 1988). Des habitats adjacents de qualité équivalente devraient
produire un isodar avec une ordonnée à l'origine de zéro et une pente de un. Lorsque des
7
habitats diffèrent de manière quantitative ou qualitative, plusieurs scénarios de répartition
de la population sont possibles (Morris 1988). Par exemple, la densité de population dans
deux habitats adjacents peut converger lors d'une augmentation de la population locale
lorsque le taux de déclin en aptitude phénotypique est plus faible dans l'habitat qui est aussi
plus pauvre en ressources (Morris 1988). Ce dernier scénario souligne le fait que la
répartition d'individus dans une mosaïque d'habitats peut varier considérablement selon la
taille de la population locale. On pourrait donc arriver à des conclusions différentes au
sujet de la qualité relative des habitats, selon l‘abondance locale des populations. Les
espèces dont les populations sont cycliques sont donc particulièrement sujettes à de telles
variations dans les conclusions car la répartition d'individus entre habitats pourrait varier
considérablement selon la phase du cycle de la population (Wolff 1980).
Utilisation des théories de sélection d’habitat pour comprendre la réaction animale
aux perturbations environnementales
Les isodars ont été utilisés afin d‘évaluer si la sélection de l‘habitat change selon la
densité d'individus chez plusieurs espèces, notamment les petits mammifères, les oiseaux et
les poissons (Rodriguez 1995, Abramsky et al. 1997, Shenbrot 2004, Shochat et al. 2005).
Cette approche permet aussi de comprendre comment les patrons de sélection peuvent
varier selon les variations de structure d'habitat ou de disponibilité de nourriture provenant
de processus naturels ou anthropiques (Shenbrot & Krasnov 2000, Shenbrot 2004, Shochat
et al. 2005). Les isodars devraient ainsi fournir un cadre conceptuel utile avec lequel il est
possible de déterminer les conséquences de perturbations de l‘habitat en termes d'aptitude
phénotypique (Morris 1990, Pusenius & Schmidt 2002). Suite à une perturbation à grande
échelle, des individus qui adoptent une distribution idéale libre devraient se répartir entre
les habitats perturbés et non perturbés de façon à rééquilibrer l'aptitude phénotypique
moyenne entre les habitats (Pusenius & Schmidt 2002). L‘ ordonnée à l'origine et la pente
de l‘isodar entre des habitats perturbés et non perturbés devraient révéler comment
l'aptitude phénotypique maximale atteignable à faible densité et le taux de réduction dans
l'aptitude phénotypique avec l‘augmentation de la densité ont été modifiés par les
changements apportés à l‘habitat perturbé. Pour les espèces associées à une végétation
8
dense, ces changements pourraient être proportionnels à la quantité de végétation prélevée
ou détruite lors de la perturbation.
Influence des prédateurs sur la sélection d'habitat des proies
La prédation peut régir les adaptations comportementales et morphologiques des
proies, et déterminer ainsi leur répartition dans des environnements hétérogènes (Lima &
Dill 1990). Les prédateurs influencent la répartition des proies en consommant des
individus, de même qu‘en influençant leur sélection d‘habitat (Lima & Dill 1990, Lima
1998, Brown et al. 1999). Les refuges créés par l‘hétérogénéité dans la structure de
l‘habitat sont souvent d‘une importance primordiale pour le maintien des populations de
proies et de leurs prédateurs (Huffaker 1958, Hastings 1977, Lecomte et al. 2008).
Puisqu'il existe souvent une relation positive entre la disponibilité de nourriture et le risque
de prédation, la stratégie optimale pour les proies devrait être de choisir les habitats de sorte
que le gain marginal en aptitude phénotypique d‘éviter la prédation soit égal à la perte
marginale en aptitude phénotypique d‘une diminution de l‘accès aux ressources
(McNamara & Houston 1987, Fryxell & Lundberg 1998, Brown & Kotler 2004). Les
proies font souvent un tel compromis en sélectionnant des habitats moins risqués mais
offrant moins de nourriture (Sih 1980, Cerri & Fraser 1983). Les individus en quête
alimentaire acceptent aussi de plus grands risques si les bénéfices sont aussi plus grands
(Grand & Dill 1997, Brown & Kotler 2004). Les perturbations de l‘habitat devraient par
conséquent structurer le « paysage de la peur » (Laundré et al. 2001) et de la nourriture
(Searle et al. 2007) en enlevant du couvert végétal protecteur et en créant de nouvelles
parcelles de nourriture, causant ainsi des compromis potentiels entre nourriture et risque de
prédation. La variation dans la densité de proies dans des parcelles issues de perturbations
de vaste échelle devrait refléter comment le compromis entre nourriture et risque de
prédation varie avec les changements de structure et de composition de l‘habitat pendant la
succession écologique.
9
Influence de l'hétérogénéité du risque de prédation sur les comportements
d’approvisionnement et de déplacement des proies à fine échelle
La capacité de support d'un habitat, c'est à dire le nombre d'individus qu'il peut
supporter, peut dépendre de l'interaction entre la structure de l'habitat à fine échelle et le
comportement d‘approvisionnement des individus (Nonaka & Holme 2007). Cela implique
que deux habitats offrant une même abondance de ressources peuvent tout de même
supporter des densités de population très différentes, dépendamment de la répartition des
ressources à l‘intérieur de chacun. Ce phénomène s'explique par le fait que l'arrangement
spatial des parcelles de nourritures à l'intérieur d'un habitat peut influencer l'efficacité
d'approvisionnement des individus (Charnov 1976, Bernstein et al. 1991, Shipley &
Spalinger 1995). Les individus devraient se déplacer et s‘alimenter à l‘intérieur de leur
domaine vital de façon à maximiser le taux d‘acquisition d‘énergie par unité d‘espace et de
temps (Stephens & Krebs 1986, Vivas et al. 1991, Fortin 2003). Les perturbations à fine
échelle peuvent façonner la répartition des ressources alimentaires puisque la nourriture est
souvent concentrée dans les ouvertures créées par les perturbations (Kuijper et al. 2009).
Les individus pourraient concentrer leurs efforts d'approvisionnement dans ces zones pour
maximiser leurs gains en énergie. Cependant, cela implique que les animaux doivent
quitter le couvert végétal pour s'alimenter, ce qui pourrait créer un conflit entre l'accès à
une nourriture abondante et le maintien d'un faible risque de prédation.
Les compromis entre maximiser les gains d'énergie et minimiser le risque de
prédation peuvent créer une variation spatiale considérable dans l'effort
d'approvisionnement (Brown 1988, Brown & Kotler 2004). Selon la théorie
d‘approvisionnement optimal et le théorème de la valeur marginale, les animaux devraient
cesser de s‘alimenter dans une parcelle lorsque les bénéfices liés à l‘exploitation de la
parcelle sont égaux aux coûts (Charnov 1976, Brown 1988). Ces coûts incluent les coûts
énergétiques, les coûts liés aux opportunités manquées en ne faisant pas d‘autres activités
qui pourraient aussi accroître l'aptitude phénotypique, et les coûts associés au risque de
prédation (Brown 1988). Les patrons d‘hétérogénéité spatiale dans le risque de prédation
peuvent être décrits en plaçant des parcelles de nourriture expérimentales dans différents
microhabitats (van der Merwe & Brown 2008). La quantité de nourriture laissée dans
chaque parcelle une fois que l‘animal cesse de s‘y nourrir ("Giving-up density" ou GUD)
10
devrait refléter la perception du risque dans chaque microhabitat puisque les dépenses
énergétiques durant l‘approvisionnement, de même que les coûts associés aux opportunités
manquées, peuvent être standardisés (Brown 1988). En théorie, l‘effort
d‘approvisionnement sera moindre dans les parcelles les plus risquées, de sorte que les
consommateurs laisseront davantage de nourriture à l'abandon dans la parcelle (c.-à-d.
GUDs plus élevés). Dans le cas des proies qui dépendent de la végétation pour se cacher
ou s'enfuir des prédateurs, plusieurs études ont démontré que les GUDs sont généralement
plus élevés en milieux ouverts que sous couvert de végétation et que les GUDs augmentent
généralement avec la distance qui sépare la parcelle du couvert protecteur (Brown et al.
1992, Hughes & Ward 1993). Cependant, la relation inverse peut être observée pour les
proies qui sont vulnérables aux prédateurs qui chassent par embuscade (Altendorf et al.
2001). Les patrons d‘exploitation de parcelles de nourriture peuvent donc révéler comment
le risque de prédation pourrait restreindre l'exploitation de la nourriture à l'intérieur
d'ouvertures créées par les perturbations.
Les déplacements d'individus peuvent aussi indiquer comment les individus
répondent à l'hétérogénéité structurale de leur environnement ainsi que les stratégies qu'ils
utilisent pour minimiser le risque de prédation. Par exemple, les proies sélectionnent
souvent des trajets de déplacement pour rester sous couvert végétal ou effectuent des
changements de vitesse pour traverser des milieux risqués plus rapidement (Vasquez et al.
2002, Fortin et al. 2005, Baker 2007). Les comportements d'approvisionnement et de
déplacement peuvent ainsi nous révéler les mécanismes par lesquels les perturbations à fine
échelle influencent la répartition des proies. L'évaluation de ces indicateurs
comportementaux devrait aussi améliorer notre compréhension de l'influence indirecte des
prédateurs sur la dynamique végétale après perturbation (Schmitz et al. 1997, Beyer et al.
2007, Ripple & Beschta 2007).
Forêts boréales, régimes de perturbation et succession écologique
L‘hétérogénéité spatiale et temporelle des écosystèmes forestiers dépend largement
de la fréquence, de la sévérité et de l‘étendue des perturbations. Ces évènements varient de
perturbations vastes et intenses (p. ex., feux de forêt et épidémies d'insectes) qui modifient
11
des peuplements en entier, jusqu‘à des évènements à fine échelle (p. ex., chablis ou
sénescence des arbres) qui ne modifient qu'une petite portion d'un peuplement à la fois.
(Pickett & White 1985, Niemela 1999). Dans la forêt boréale, le feu est considéré comme
la principale perturbation naturelle (Bergeron et al. 2001). Celle-ci interagit avec la
topographie et les conditions édaphiques pour créer une mosaïque complexe de
peuplements qui varient en âge, en composition et en structure (DeLong & Tanner 1996,
Bergeron et al. 2001). Les variations régionales dans le cycle de feux déterminent en
grande partie la proportion et la répartition des peuplements aux différents stades de
succession dans le paysage (Weir et al. 2000, Bergeron et al. 2001). Dans les régions
boréales plus sèches, l‘intervalle de temps entre deux feux de forêt peut être inférieur à 100
ans, ce qui crée des paysages dominés par de jeunes peuplements ayant une structure
régulière. Sous de telles conditions, peu de peuplements échappent au feu assez longtemps
pour qu'une structure typique des forêts anciennes se développe (Bergeron & Harper 2009).
On a longtemps supposé que les forêts anciennes étaient l‘exception et non la règle
parce que les cycles de feu sont relativement courts à travers la majeure partie de la forêt
boréale (Mosseler et al. 2003, Bergeron & Harper 2009). Il est maintenant reconnu que
dans les régions ayant des précipitations abondantes, le cycle de feu peut fréquemment
dépasser 200 ans, surpassant ainsi la longévité de la plupart des espèces d‘arbre boréales
(Bouchard et al. 2008, Bergeron & Harper 2009). Dans ces régions, une grande proportion
du paysage est composée de veilles forêts qui sont structurées par une dynamique de petites
trouées créées par la sénescence des arbres, le chablis et les maladies (Gauthier et al. 2000,
Pham et al. 2004, Aakala et al. 2007). Contrairement aux forêts tempérées et tropicales, le
développement des peuplements anciens en forêt boréale est caractérisé davantage par une
succession structurale que par des changements dans la composition des espèces d‘arbres
(Boucher et al. 2006, Bergeron & Harper 2009). Un peuplement à ce stade aura une
structure horizontale et verticale complexe créée par une canopée multi étagée, une grande
abondance de chicots et de débris ligneux et de nombreuses trouées à différents stades de
régénération (Mosseler et al. 2003, Wirth et al. 2009).
La répartition des populations animales en forêt boréale devrait refléter la mosaïque
de peuplements de différentes classes d‘âge car ceux-ci varient dans la disponibilité de la
nourriture et des abris qu‘ils offrent pour une espèce particulière (Bunnell 1995, Fisher &
12
Wilkinson 2005). Ainsi, la densité de plusieurs espèces animales varie de façon marquée
au cours du développement des peuplements, suivant leurs changements de structure et de
composition (Fisher & Wilkinson 2005, Schieck & Song 2006). Les études décrivant les
changements d‘abondance animale au cours d‘une succession ont souvent exclu les forêts
anciennes à cause de leur rareté dans plusieurs régions de la forêt boréale. De plus, bien
que la dynamique de trouées soit considérée comme une composante fondamentale des
forêts boréales anciennes (Bergeron & Harper 2009), peu d‘études ont évalué leurs impacts
sur la répartition de la faune à ce stade de succession.
La dynamique de trouées devrait contribuer aux variations dans la densité locale de
populations animales en fin de succession en influençant l‘abondance et la répartition de
nourriture et de couvert. La formation de trouées favorise l‘établissement et le dégagement
de la régénération en sous étage (Kneeshaw & Bergeron 1998, Pham et al. 2004, de Romer
et al. 2007). Des études réalisées en forêts tropicale et tempérée ont révélé que les trouées
créent pour les herbivores et les insectivores des zones où la nourriture est concentrée, et
que les herbivores peuvent à leur tour influencer les patrons de régénération végétale à
l‘intérieur des trouées en consommant les plants que l‘on y retrouve (Schreiner et al. 1996,
Gitzen & West 2002, Faccio 2003, Horn et al. 2005, Norghauer et al. 2008). En
comparaison, peu d‘études (quelques exemples pour les insectes sont cités dans Bouget &
Duelli 2004) ont évalué les interactions entre la dynamique naturelle de trouées et la
répartition à fine échelle de la faune dans les forêts boréales anciennes. Plusieurs espèces
de mammifères de la forêt boréale dépendent de la strate arbustive pour se nourrir et se
protéger, de sorte qu‘ils atteignent leur densité maximale dans les premiers stades de
succession forestière (Fisher & Wilkinson 2005). Ces mêmes espèces peuvent présenter un
deuxième pic d‘abondance dans les forêts anciennes si la formation de trouées et la
régénération en sous étage augmentent la disponibilité de nourriture et de couvert de façon
avantageuse pour l‘espèce (Sakai & Noon 1993, Buskirk et al. 1999). Lorsque la nourriture
est concentrée en milieu relativement ouvert, tel qu'à l‘intérieur des trouées nouvellement
formées, l‘herbivore peut faire face à un compromis entre nourriture et sécurité. Les
trouées peuvent ainsi structurer l‘hétérogénéité spatiale à fine échelle à la fois du risque de
prédation et de la disponibilité de nourriture. La proportion du peuplement occupée par les
13
trouées pourrait donc déterminer l‘abondance relative entre la nourriture et le couvert
protecteur et, ainsi, contribuer aux variations locales de densité des proies.
Dans les forêts aménagées, la coupe forestière a aussi une forte influence sur la
structure d'âge de la forêt, et par conséquent, sur la répartition des animaux. Une approche
écosystémique à l‘aménagement forestier en régions boréales caractérisées par de longs
cycles de feu pourrait nécessiter toute une gamme d‘approches sylvicoles afin de préserver
la structure et la composition typique de la forêt à l'échelle des peuplements et du paysage
(Bergeron 2004). Il faut aussi se demander si ces différentes approches sylvicoles sont
capables de reproduire des répartitions d‘animaux typiques de celles produites par les
perturbations naturelles (Thompson 2006). Par exemple, la coupe totale est souvent
proposée pour imiter les feux de forêt puisque les deux perturbations résultent en une
mortalité presque complète de la canopée sur de grandes étendues (McRae et al. 2001).
Cependant, plusieurs différences ont été observées au niveau de la composition et de la
structure des peuplements en régénération (McRae et al. 2001, Elson et al. 2007, Hart &
Chen 2008). Par conséquent, la composition des communautés d‘oiseaux, de petits
mammifères et d‘insectes dans les peuplements issus de ces deux types de perturbation
peuvent différer en début de succession, mais on observe tout de même une convergence
dans la composition et l'abondance des espèces au cours de la succession (Imbeau et al.
1999, Simon et al. 2002, Buddle et al. 2006, Schieck & Song 2006). Les ressemblances
dans les patrons d‘abondance après feu et après coupe à blanc restent toujours à être
évaluées pour plusieurs espèces fauniques.
Bien que les pratiques forestières au Canada changent graduellement afin de mieux
considérer la variabilité régionale des régimes de perturbations naturelles (Groot et al.
2005), la coupe totale demeure la méthode de récolte principale (CCFM 2010). Un défi
majeur pour l'aménagement forestier est que la période de rotation pour des récoltes
profitables par coupe totale est souvent plus courte que les cycles de feu régionaux, ce qui
peut causer une réduction importante dans l'aire occupée par les veilles forêts (Bergeron
2004). Des études récentes démontrent que l'étendue des forêts anciennes à déjà été réduite
bien en dessous des niveaux historiques observés dans certaines régions boréales (Boucher
et al. 2009, Cyr et al. 2009). Les coupes partielles ont été proposées afin de permettre la
récolte de bois tout en maintenant des peuplements ayant une structure typique des forêts
14
anciennes ainsi que les espèces fauniques qui y sont associées (Bergeron et al. 2001, Ruel
et al. 2007, Vanderwel et al. 2009). Les coupes partielles incluent tout traitement sylvicole
qui ne recueille qu‘une portion des arbres d‘un peuplement forestier (CFS 1999). Dans la
forêt boréale, il existe de nombreux types de coupe partielle incluant notamment les coupes
progressives qui protègent les tiges marchandes de faible diamètre (<15 cm), les coupes par
bandes, et les coupes de jardinage qui retiennent une plus grande proportion d‘arbres
matures (Groot et al. 2005, Ruel et al. 2007, Raymond et al. 2009).
Puisque l‘utilisation des coupes partielles est relativement récente dans plusieurs
régions de la forêt boréale, il est nécessaire d‘évaluer la capacité de nouveaux traitements
sylvicoles à retenir les espèces fauniques typiques des vieilles forêts comparativement aux
approches sylvicoles plus conventionnelles. Des revues récentes de la littérature indiquent
qu‘un taux de rétention de 70% des arbres matures pourrait être suffisant pour maintenir la
majorité des espèces de mammifères et d‘amphibiens associés aux peuplements matures et
anciens, alors qu‘une rétention <50% pourrait engendrer une diminution d‘abondance de
plusieurs espèces d‘oiseaux et de petits mammifères (Vanderwel et al. 2007, Vanderwel et
al. 2009). À ce jour, les études portant sur les coupes partielles se sont principalement
intéressées aux petits mammifères et aux oiseaux, de sorte qu‘il existe relativement peu
d‘information sur l‘impact de différentes sylvicultures sur des proies de plus grande taille,
comme le lièvre d'Amérique.
Espèce cible : le lièvre d'Amérique
Le lièvre d'Amérique fascine depuis longtemps les écologistes de part sa dynamique
de population qui montre un cycle de 10 ans relativement synchronisé à travers la forêt
boréale de l'Amérique du Nord (Keith & Windberg 1978, Smith 1983, Sinclair et al. 1993,
Krebs et al. 2001a). Des études à long terme ont démontré que ce cycle serait associé à la
fois aux effets des prédateurs sur la survie, le comportement et la taux de reproduction du
lièvre et aux interactions entre le lièvre et la végétation (Krebs et al. 1995, Sheriff et al.
2009). De plus, le lièvre est considéré comme une espèce clé de voûte de la forêt boréale
(Sinclair 2003), puisque les dynamiques des populations de plusieurs autres espèces
15
d‘herbivores et de carnivores lui sont intiment liées (Boutin et al. 1995). Il peut aussi avoir
un fort impact sur la croissance de la végétation en consommant des ramilles durant l‘hiver
et en enrichissant le sol par la déposition de fèces (Sinclair et al. 1988, Sinclair et al. 2000,
Butler & Kielland 2008).
La répartition spatiale du lièvre est directement liée aux variations dans la structure
de la forêt. Il existe une forte relation positive entre l'abondance du lièvre et la densité de la
strate arbustive sur l‘ensemble de l‘aire de répartition de l‘espèce (Litvaitis et al. 1985,
Wirsing et al. 2002, Hodges et al. 2009). L‘importance de la strate arbustive serait
notamment associée au couvert latéral qu‘elle crée et qui protège le lièvre contre ses
prédateurs, ces derniers étant responsables de plus de 75% de la mortalité du lièvre (Hodges
et al. 1999, Etcheverry et al. 2005). En effet, quelques études ont démontré que le taux de
survie du lièvre est plus élevé en milieu fermé qu'en milieu ouvert (Rohner & Krebs 1996,
Griffin & Mills 2009). Le lièvre dépend aussi de la strate arbustive pour se nourrir en
hiver, alors qu‘il consomme principalement des ramilles d'essences feuillues (Pease et al.
1979). Le couvert vertical créé par la végétation pourrait également protéger l‘espèce
contre les prédateurs aériens, mais une association positive entre le couvert vertical et
l'abondance du lièvre a été moins fréquemment observée (St-Laurent et al. 2008, Hodges et
al. 2009).
En raison de son importance dans les réseaux trophiques de la forêt boréale, le lièvre
est souvent ciblé pour évaluer les impacts de l'aménagement forestier, (p. ex. Monthey
1986, Ferron et al. 1998, De Bellefeuille et al. 2001, Homyack et al. 2007). Sa forte
association avec la strate arbustive fait en sorte qu'il est sensible aux changements de
structure de son habitat causés par les perturbations prenant place durant la succession
végétale (Thompson et al. 1989, Koehler 1991, Newbury & Simon 2005). Par exemple,
plusieurs études ont observé des réductions importantes d'abondance des populations de
lièvres suite aux coupes totales (Thompson et al. 1989, Koehler 1991, Newbury & Simon
2005). Une augmentation rapide de l'abondance du lièvre survient néanmoins lorsque la
régénération végétale vient à dépasser la surface de la neige pendant l'hiver, généralement
10-30 ans après la perturbation (Koehler 1991, Newbury & Simon 2005, Robinson 2006).
Les densités de lièvres devraient ensuite diminuer à mesure que la fermeture de la canopée
en vient à limiter la croissance de la végétation en sous étage qui fournit le couvert
16
protecteur et la nourriture hivernale (Fisher & Wilkinson 2005). La densité de lièvres
pourrait d‘ailleurs s‘accroître de nouveau dans les forêts anciennes suivant l‘augmentation
de densité de la strate arbustive associée à la formation de trouées (Buskirk et al. 1999).
Alors que deux études récentes ont observé des densités de lièvres similaires en forêt jeune
et en forêt ancienne (Griffin & Mills 2009, Hodges et al. 2009), l'hypothèse d'une
distribution bimodale de la densité du lièvre avec l'âge de la forêt n'a jamais été vérifiée en
échantillonnant une chronoséquence complète de succession forestière. De plus, dans une
optique d'aménagement écosystémique, aucune étude n'a évalué si les changements de
densité du lièvre sont semblables entre la succession des peuplements issus de coupes à
blanc et de feux.
Compte tenu que le lièvre dépend largement des couverts latéral et vertical pour se
protéger contre les prédateurs et pour s‘alimenter, il devrait être relativement sensible à
l'entremêlement de ces ressources créé par la dynamique spatio-temporelle des trouées en
fin de succession forestière. Puisque le lièvre subit un plus fort taux de mortalité dans les
milieux ouverts (Rohner & Krebs 1996, Griffin & Mills 2009), une augmentation de la
disponibilité de nourriture à l'intérieur des trouées pourrait créer un dilemme entre
l'acquisition de nourriture et le besoin de couvert de protection. La dynamique de trouées
dans les forêts anciennes pourrait donc structurer la quête alimentaire et les déplacements
au sein de ces peuplements. La taille des trouées et la proportion du peuplement en trouées
pourraient quant à elles contribuer aux variations d'abondance du lièvre en modulant à la
fois la disponibilité de la nourriture et le risque de prédation.
Objectif et organisation de la thèse
L‘objectif de cette thèse est de mieux comprendre comment les perturbations
naturelles et anthropiques influencent la répartition du lièvre d'Amérique dans un
écosystème de forêts boréales anciennes sous aménagement forestier. Pour ce faire, j‘ai
étudié les patrons d‘abondance, la sélection d‘habitat, le comportement
d‘approvisionnement et les déplacements du lièvre. Cette étude vise donc à acquérir des
connaissances sur le lien entre la dynamique des perturbations et les comportements de
17
sélection d‘habitat qui déterminent ultimement la répartition des animaux dans des
environnements hétérogènes.
Dans le chapitre 1, j‘évalue si l‘abondance du lièvre suit une distribution bimodale
avec l‘âge des peuplements en forêt boréale en échantillonnant une chronoséquence de
peuplements allant jusqu‘à 265 ans. Je m‘intéresse ensuite aux caractéristiques de l‘habitat
qui expliquent le mieux les changements d‘abondance du lièvre le long de cette succession
forestière. Je vérifie la réaction de l‘espèce aux variations de nourriture et de couvert au
sein de peuplements associés à différents stades de succession. La chronoséquence
forestière incluait à la fois des peuplements issus de feu et de coupe totale afin d‘évaluer si
ces deux types de perturbations engendrent des réactions semblables chez les populations
de lièvre.
Dans le deuxième chapitre, j‘examine l‘influence de la dynamique de trouées sur le
comportement d‘approvisionnement et les déplacements du lièvre dans des peuplements de
fin de succession. Je commence par déterminer si les trouées créent des parcelles plus
denses en nourriture hivernale pour le lièvre, ce qui entraînerait un compromis entre la
disponibilité de nourriture et de couvert. J‘utilise ensuite des suivis de pistes dans la neige
afin d‘évaluer comment l‘hétérogénéité dans la répartition de nourriture et de couvert
influence l‘utilisation que le lièvre fait des peuplements. Des expériences de densité à
l‘abandon et l‘évaluation de la consommation du brout naturel sont utilisées pour évaluer si
les lièvres perçoivent un plus grand risque de prédation à l‘intérieur des trouées que sous le
couvert forestier adjacent et si la perception du risque de prédation augmente avec la
distance de la lisière de la forêt.
Dans le troisième chapitre, je développe un cadre conceptuel basé sur la théorie de
sélection d‘habitat pour évaluer les conséquences, en termes d'aptitude phénotypique, que
peuvent avoir différents traitements sylvicoles sur des espèces forestières. Je teste des
prédictions théoriques en utilisant des mesures relatives de densité de lièvres et de
campagnols à dos roux (Myodes gapperi) dans des paires d‘habitat comprenant une forêt
non coupée adjacente à un peuplement traité suivant une des quatre approches sylvicoles
18
évaluées (27 à 100% des tiges marchandes récoltées, selon le traitement appliqué). Deux
de ces traitements sylvicoles ont été conçus afin de maintenir la structure complexe des
forêts anciennes, alors que les deux autres sont des traitements conventionnels largement
appliqués au Québec.
Dans le quatrième chapitre, je reconstitue l‘historique de broutement (Keigley et al.
2003) du lièvre dans des peuplements ayant subi différents traitements sylvicoles et dans
des forêts non coupées afin d‘examiner les patrons temporels d‘utilisation de l‘habitat. Des
inventaires d‘architecture de tiges (Keigley & Frisina 1998) de bouleau blanc (Betula
papyrifera) sont également utilisés pour déterminer s‘il y a des différences dans l‘utilisation
de l‘habitat suite à l‘application des différents traitements sylvicoles.
Aire d’étude: la forêt boréale irrégulière de l’est du Québec
Cette étude a été réalisée sur la Côte-Nord du Québec, une région qui reçoit des
précipitations abondantes à cause du climat maritime frais de l‘océan Atlantique. L‘aire
d‘étude était localisée autour des réservoirs hydroélectriques Manicouagan et Outardes, 50
à 200 km au nord de la ville de Baie Comeau. Une étude récente réalisée par Bouchard et
al. (2008) a établi que les cycles de feu de cette région varient de 270 à >500 ans, suivant
un gradient d‘ouest en est. Ce régime de feu prolongé a créé un paysage forestier dominé
par de vieux peuplements d‘épinettes noires (Picea marianna) et de sapins baumiers (Abies
balsamea) (Boucher et al. 2003). En début de succession après feu, les peuplements sont
généralement dominés par une combinaison d'espèces feuillues intolérantes à l‘ombre,
comme le bouleau blanc et le peuplier faux-tremble (Populus tremuloides), ou par l'épinette
noire ou le pin gris (Pinus banksiana) (De Grandpré et al. 2000). Le sapin baumier est une
espèce de conifère tolérante à l‘ombre qui s‘établit généralement tard durant la succession
végétale (Bouchard et al. 2008). Cette espèce peut demeurer opprimée dans le sous-étage
forestier jusqu‘à ce que les arbres de la première cohorte commencent à mourir (Kneeshaw
et al. 1998). Les peuplements feuillus et mixtes sont graduellement remplacés par des
peuplements mixtes d‘épinettes noires et de sapins baumiers, alors que les peuplements de
19
pins gris sont généralement remplacés par des peuplements à dominance d‘épinettes noires
(De Grandpré et al. 2000, Gauthier et al. 2010).
Les longs cycles de feu observés dans cette région signifient également que la
structure des peuplements est largement influencée par le chablis et la sénescence naturelle
(Boucher et al. 2003, Pham et al. 2004). Ces perturbations secondaires favorisent le
développement de peuplements pluriétagés caractérisés par une distribution irrégulière des
classes d‘âge et de diamètre des arbres (Boucher et al. 2003, Boucher et al. 2006). La strate
arbustive des peuplements anciens est dominée par l‘épinette noire et le sapin baumier, bien
que le bouleau blanc puisse être présent jusqu‘aux stades de fin de succession par la
régénération en trouée (Pham et al. 2004, Gauthier et al. 2010).
Cette région a aussi une longue histoire de coupes forestières pour fournir du bois
d‘œuvre et des fibres aux usines de sciage et de pâtes et papiers (Bouchard et al. 2008). Les
coupes forestières ayant pris place dans la première moitié du 20ème
siècle se concentraient
le long des cours d‘eau principaux. Seuls les plus gros arbres étaient alors abattus à la
main. La mécanisation de la coupe forestière à partir des années 1950 a permis
d‘augmenter la récolte annuelle dans cette région. Néanmoins, les forêts issues de
perturbations naturelles couvrent toujours environ 50% de l‘aire d‘étude. La région offre
donc une excellente opportunité pour évaluer les changements d‘abondance du lièvre le
long d‘une chronoséquence de peuplements forestiers issus de perturbations naturelles. La
présence d‘une proportion importante de peuplements forestiers non coupés permet
également d'évaluer l‘influence de la dynamique des trouées sur la répartition du lièvre à
l‘intérieur de vastes parcelles de forêts anciennes. Quatre dispositifs de récolte
expérimentale ont été établis à l‘intérieur des territoires de gestion de trois compagnies
forestières locales (Abitibi-Bowater, Arbec, et Kruger) afin de déterminer la faisabilité
économique et les défis opérationnels de deux nouveaux types de coupes par jardinage
adaptés aux forêts boréales irrégulières. Leurs impacts sur une gamme d‘espèces fauniques
seront aussi comparés avec ceux de traitements conventionnels.
Chapitre 1
Changes in relative snowshoe hare abundance across a
265-year gradient of boreal forest succession
James Hodson*, Daniel Fortin
*, and Louis Bélanger
†
*NSERC-Université Laval Industrial Research Chair in Silviculture and Wildlife,
Département de Biologie, Université Laval, Québec, QC, Canada, G1V 0A6 (JH, DF)
†Département des sciences du bois et de la forêt, Université Laval, Québec, QC, Canada,
G1V 0A6 (LB)
Résumé
Dans les régions de la forêt boréale ayant des cycles de feu prolongés, une grande
proportion des peuplements développent une structure de forêt ancienne à cause d'une
dynamique de petites trouées. Une régénération arbustive importante à l'intérieur des
trouées peut fournir des conditions d‘habitat favorables aux espèces animales typiques des
jeunes forêts, menant ainsi à une augmentation dans leur abondance en fin de succession.
Nous avons réalisé des inventaires de crottins le long d‘une chronoséquence de succession
après feu (17-265 ans) et après coupe (3-63 ans) dans la forêt boréale de l‘est du Canada
pour évaluer l‘hypothèse selon laquelle l‘abondance du lièvre d'Amérique (Lepus
americanus) suit une distribution bimodale avec l‘âge des peuplements. Une telle
distribution reflèterait les changements de nourriture et de couvert qui ont lieu au cours de
la succession. Un fort pic d'abondance relative du lièvre est apparu durant les 80 premières
années de succession, avec les densités de crottins les plus élevées se retrouvant dans les
peuplements de 40-50 ans. Les changements d'abondance du lièvre pendant cette période
ont été semblables dans les peuplements issus de coupe et de feu et ont suivi de près les
changements de couvert latéral et vertical. L'abondance relative du lièvre a augmenté de
nouveau dans les peuplements de >180 ans, mais les changements de densité de crottins en
fin de succession n'étaient pas synchronisés avec ceux du couvert latéral et vertical. Les
variations d‘abondance du lièvre dans les peuplements en fin de succession étaient liées à la
dynamique de trouées. En effet, les densités les plus élevées de crottins se trouvaient dans
les peuplements ayant une proportion intermédiaire de trouées issues de mortalité. Nos
résultats démontrent que le lièvre subit un fort pic d'abondance en début de succession,
suivi de changements moins prononcés au cours d'une période beaucoup plus longue
lorsque les peuplements développent une structure de forêt ancienne. Des changements
dans la distribution des classes d‘âges des peuplements induite par la gestion forestière
pourraient ainsi avoir des conséquences importantes sur la dynamique spatiotemporelle du
lièvre.
22
Abstract
In boreal forest regions with prolonged fire-cycles, a large proportion of forest stands
develop old-growth structure created by small canopy gap dynamics. Dense regeneration
within canopy gaps may provide habitat conditions suitable for wildlife typically associated
with early-seral forests leading to an increase in their abundance during late-succession.
We conducted pellet surveys in a chronosequence of post-fire (17-265 years) and post-
harvest (3-63 years) stand succession in Canada‘s eastern boreal forest to determine
whether snowshoe hares (Lepus americanus) followed a bimodal abundance distribution
with stand age. This bimodal distribution should reflect changes in food and cover during
post-disturbance succession. A strong peak in relative snowshoe hare abundance was
observed during the first 80 years of succession, with highest pellet densities occurring in
stands 40-50 years old. Changes in relative hare abundance during this period were similar
in both fire- and clearcut-origin stands and closely tracked changes in lateral and vertical
cover. Relative hare abundance increased again in stands >180 years, but changes in pellet
density during late-succession were not synchronous with variations in lateral and vertical
cover. Variation in hare abundance during late succession was partially mediated by gap
dynamics, with highest pellet densities occurring in stands occupied by an intermediate
proportion of mortality-origin canopy gaps. We showed that, relative to the length of
regional fire cycles, hares undergo rapid changes in density during early-succession
followed by a much longer period of subtle changes in density as stands develop old-
growth structure. Changes to forest age-class distribution induced by forest management
could therefore have important consequences for spatiotemporal dynamics of snowshoe
hares.
23
Introduction
Disturbance and succession are key processes creating spatiotemporal heterogeneity in
wildlife habitat. Following disturbance, changes in vegetation structure and composition
during succession lead to temporal changes in resource availability that influence animal
survival and reproduction (Sousa 1984, Pickett & White 1985, Brawn et al. 2001). Spatial
and temporal heterogeneity in forest ecosystems depends largely on the frequency, severity,
and extent of natural disturbances, which range from broad-scale stand-replacing events
such as wildfires and insect epidemics to fine-scale occurrences of individual tree mortality
from windfall, disease and senescence (Pickett & White 1985). In the boreal forest, fire is
the primary natural disturbance that interacts with topography and edaphic conditions to
create complex mosaics of stands varying in age, composition and structure (Bergeron et al.
2001, Bergeron et al. 2002). Regional variation in the rate of recurrence of fires largely
determines the proportion and distribution of different seral stages (Bergeron et al. 2001).
Broad-scale wildlife distribution and dynamics within boreal landscapes should thus reflect
the age-class distribution of forest stands determined by regional fire regimes (Bunnell
1995, Fisher & Wilkinson 2005).
Fire return intervals vary greatly across North America's boreal forest, ranging from 52-
813 years (Bergeron & Harper 2009). Following a stand-replacing fire the initial cohort of
trees generally grows at a similar rate, leading to relatively homogeneous stands with trees
of similar age and diameter. In drier regions with short fire cycles (<100 years), most
stands burn before trees die of natural senescence and landscapes are dominated by young
even-aged stands (Kneeshaw & Gauthier 2003, Bergeron & Harper 2009). In regions
where fire return intervals exceed the life span of most boreal tree species (generally 100-
200 yrs; Burns & Honkala 1990), gradual mortality of the first cohort of trees leads to the
development of stands with irregular tree diameter and age distributions through
recruitment and release of regeneration within canopy gaps (Bergeron & Harper 2009). For
example, fire return intervals exceed 250 years over much of eastern Canada's boreal forest
due to the humid maritime climate, resulting in landscapes dominated by structurally
complex late-seral stands (Foster 1983, Boucher et al. 2003, Bouchard et al. 2008).
24
Ecosystem-based forest management proposes that natural disturbance emulation and the
maintenance of historical forest age-class structure can provide a coarse filter approach to
conserving regional biodiversity (Bergeron et al. 2002, Armstrong et al. 2003). Recent
evidence suggests, however, that current forest management may be pushing some boreal
forest landscapes outside of their natural range of variability (Fall et al. 2004, Didion et al.
2007, Cyr et al. 2009). It is thus important to understand how wildlife species respond to
changes in habitat structure and resource availability throughout forest succession to be
able to successfully anticipate wildlife distribution under different disturbance regimes.
Although few boreal wildlife species are strictly associated with a specific seral-stage,
many undergo marked fluctuations in density over the course of forest stand development
(Fisher & Wilkinson 2005, Schieck & Song 2006). For example, species that depend on
food and cover provided by near-ground vegetation may follow a bimodal abundance
distribution with stand age, with a first peak in early-succession, followed by a low phase in
mature closed-canopy stands, and then a second increase phase during transition to old-
growth stand structure (e.g. Sakai & Noon 1993). Snowshoe hare (Lepus americanus) are
thought to follow such temporal changes in distribution (Buskirk et al. 1999) because there
is a strong positive association between hare density and the density of shrubs and saplings
providing near-ground lateral cover (Litvaitis et al. 1985, Wirsing et al. 2002, Hodges et al.
2009). This vegetation layer also provides hares with deciduous browse, their main winter
food (Pease et al. 1979). Canopy cover may also be important, but has been less
consistently linked with snowshoe hare abundance (Pietz & Tester 1983, Jacqmain et al.
2007, Hodges et al. 2009). The hypothesis of a bimodal distribution of hare abundance
with stand age has yet to be tested by measuring changes in their abundance throughout a
continuous sequence of forest succession.
Relationships between hare density and browse and cover availability may vary with
stand age according to which resource is more limiting during different phases of
succession. This is because changes in food and cover may be asynchronous due to
differences in the growth rate and light requirements of vegetation that provide these
resources. For example, following a stand-replacing disturbance, browse availability
should increase faster than lateral cover, because shade-intolerant deciduous vegetation
providing browse generally grows more quickly than conifers providing the majority of
25
lateral cover (Brassard & Chen 2006). As stands mature, both food and near-ground cover
should decrease once lower limbs begin to die on growing trees and canopy closure limits
light availability for understory vegetation (Buskirk et al. 1999). In late succession, lateral
cover may increase sooner than browse because shade-tolerant conifers such as balsam fir
can establish under a closed canopy (Duschesneau et al. 2001), whereas deciduous browse
species may depend on canopy gap formation for adequate light (Kneeshaw & Bergeron
1998). By obtaining simultaneous measures of food, cover and snowshoe hare abundance
over the course of succession, we should be able to assess whether the relative importance
of food and cover as factors limiting hare density varies during different phases of forest
succession.
During late-succession the balance between food and cover for herbivores may be
largely determined by canopy gap distribution. Although canopy gaps are considered a
fundamental feature of old-growth stands (Bergeron & Harper 2009), their influence on
wildlife distribution has rarely been evaluated in boreal forests. Openings in the canopy
can provide areas with higher concentrations of deciduous winter browse, but snowshoe
hares also appear to select areas with high canopy cover to reduce predation risk (Hodson et
al. 2010a). Hare abundance may therefore be highest at intermediate levels of canopy gap
abundance.
To date, most of our knowledge about changes in hare abundance during forest
succession has come from stands initiated by forest harvesting (Thompson et al. 1989,
Ferron et al. 1998, Newbury & Simon 2005). A question central to ecosystem-based
management is whether forest harvesting results in patterns of animal abundance over time
that are similar to those produced by natural disturbances. Clearcutting has often been
assumed to emulate fire because both types of disturbance result in almost complete
mortality of the canopy later (McRae et al. 2001). However, stand composition and
structure can differ substantially between fire- and clearcut-origin stands (Simon & Schwab
2005, Elson et al. 2007, Hart & Chen 2008) leading to different assemblages of birds and
insects during early-succession (Buddle et al. 2006, Schieck & Song 2006). These
differences are largely tied to disparities in the availability of snags and coarse woody
debris following disturbance (Imbeau et al. 1999, Buddle et al. 2006). Snowshoe hare
abundance, on the other hand, is more strongly linked to the structure of regenerating
26
vegetation. Trajectories of post-fire and post-harvest changes in hare abundance are
therefore likely to be similar (Fisher & Wilkinson 2005), but this theory remains to be
tested.
The objectives of this study were to: 1) determine whether food, cover and snowshoe
hare density follow bimodal distributions with time-since-disturbance in boreal forests, 2)
evaluate whether the relationship between relative snowshoe hare abundance and the
availability of food and cover vary across different seral stages, 3) determine whether
relative hare abundance in late-seral stands is influenced by the abundance of canopy gaps,
and 4) to characterize to what extent patterns of hare abundance following disturbance from
fire resemble those following clearcut harvesting.
Methods
Study Area
This study took place in the Côte-Nord region of Québec, Canada. The study area
covers approximately 26,600 km², starting 50 km north of the city of Baie-Comeau and
extending 190 km northward, centred on the Manicouagan and Outardes hydroelectric
reservoirs (Figure 1.1). The study area is located on the Canadian Shield and has a rolling,
hilly landscape with altitudes often surpassing 800 m, and a geology dominated by deposits
of glacial till. The regional climate is sub-humid, sub-polar, characterized by a very short
growing season with mean annual temperatures varying from 1.5oC in the south of the
study area (Baie Comeau) to -3.8oC in the north of the study area (Fermont), and abundant
annual precipitation ranging from 1014.4 mm in the south to 806.5 mm in the north, 35% of
which is snow (based on thirty year climate means [1971-2001] from Baie-Comeau and
Fermont; Environment Canada 2002).
Regional forests are dominated by black spruce (Picea marianna) and balsam fir
(Abies balsamea), with minor components of trembling aspen (Populus tremuloides), jack
pine (Pinus banksiana) and white birch (Betula papyrifera). Fire return intervals in the
region vary between 270 and >500 years (Bouchard et al. 2008), resulting in a forest
landscape composed of ~70% late-successional stands with irregular tree diameter and age
distributions (Boucher et al. 2003). In the early 1900‘s, logging in the region was mainly
27
selective and focused on large trees along waterways. Since the 1950‘s, road networks and
industrial mechanized harvesting, mainly clearcut logging, have progressed steadily
northwards. Further information on the region‘s disturbance history can be found in
Bouchard et al. (2008, 2009).
Site Selection
To assess whether snowshoe hare abundance follows a bimodal distribution with
time since disturbance, we sampled a chronosequence of 84 fire- and harvest-origin forest
stands ranging in age from 3 to 265 yrs (harvest-origin: 3-63 years; fire-origin: 17-265
years; Figure 1.1). Potential stands were identified in a geographic information system
(GIS) based on stand-age classes used on ecoforestry maps, updated with recent and
historical logging layers provided by local forestry companies. Stands within different age
classes that were accessible by road and >20 ha in size were selected such that sites from
each age class were distributed throughout the north-south gradient of our study area. We
sampled uncut fire-origin stands >70 years old from the two dominant overstory types in
the region: stands with >75% black spruce composition or mixed spruce-fir stands with
<75% black spruce and >25% balsam fir composition (Bouchard et al. 2008). Younger
fire- and harvest-origin stands varied from conifer-dominated (spruce-fir or jack-pine) to
mixedwood (mainly birch, trembling aspen and black spruce) and deciduous-dominated
(birch and trembling aspen with an understory of black spruce) species composition. All
three stand types succeed to either black spruce or mixed black spruce-balsam fir stands in
our study region (Bergeron et al. 2001, Brassard & Chen 2006, Bouchard et al. 2008).
Because fire- and harvest-origin stands from the same age class are spatially aggregated,
and the number of fire-origin sites <70 years old was limited by the long regional fire cycle,
we placed several sites separated by at least 500 m within the same fire or aggregate of
cuts, and sampled several fires or cut aggregations of the same age class spread throughout
the study region. Given this constraint we tested for the presence of spatial autocorrelation
among our samples (see Statistical Analysis below).
We restricted our sample of harvested stands to cuts <70 years old because earlier
logging methods were more representative of selective harvests than clearcuts and because
these sites mainly occurred in the southern portion of the study region in a different
28
ecozone. Stand ages for fires and cuts were obtained from updated ecoforestry maps for
stands <70 years old. We determined the age of stands from 70 to 182 years old based on
fire maps created by Bouchard et al. (2008). Because stands that burned prior to 1800 and
small fires between 1800-1900 could not be precisely delineated by Bouchard et al. (2008),
we took tree cores from five dominant canopy trees (at a height of 1 m from ground level)
within sites of unknown age, and then used the age of the oldest tree as a measure of
minimum stand age (these stands ranged in age from 131 to 265 years). Sample
photographs of stands from the forest chronosequence are provided in Appendix 1a.
Pellet inventories
We used faecal pellet counts as a measure of relative snowshoe hare abundance
because of the strong link between pellet density and hare density across the species range
(Krebs et al. 2001b, Mills et al. 2005, Homyack et al. 2006, McCann et al. 2008). We
installed a grid of pellet plots within each site (n = 84) between the summers of 2005-2007.
Most grids (n = 83) contained 19 large circular plots 1.5 m in radius (area: 7.07 m²), spaced
equidistantly at 75 m intervals in offset rows in the form of a hexagon (Figure 1.1). The
remaining site had 37 one-meter radius circular plots (area: 3.14 m²) with a 50 m
equidistant spacing between plots. All grids covered ca. 6 ha. Grids were rotated to fit
within stand boundaries while leaving a ≥50 m buffer from roads and adjacent stands of
different age classes. Pellets were counted and cleared from sites each summer between 1
June – 25 August and final pellet counts were conducted in the summer of 2008 (all plots
had been cleared at all sites by the end of summer 2007). We converted pellet counts to
pellet density (pellets/m²) and used mean pellet density at each site as an index of relative
hare abundance in our analyses. Although we were unable to visit sites on exactly the same
day in each year, differences in sampling interval (number of days between successive
visits to a site) among sites did not influence pellet density (P = 0.64), nor did it explain a
significant amount of variance in the residuals from models predicting pellet density (P >
0.38 for any model; see Statistical analysis below). We therefore simply used pellet
density as a measure of relative hare abundance to facilitate comparison of our results with
other studies.
29
Habitat structure
To quantify the availability of vegetative cover, we measured lateral and vertical
cover at every pellet plot within each grid (Figure 1.1). Lateral cover was estimated at each
pellet plot using a 2-m high profile board separated into four 50 cm segments of alternating
colour (Nudds 1977). The cover board was held 5 m away north and south from an
observer crouching at the plot centre. Visual obstruction of each 50 cm segment was
estimated in 10% classes, and readings from the two directions were averaged for a given
plot. We then used the average lateral cover between 0-1 m (LatCov0-1) and between 1-2
m (LatCov1-2) within each site to reflect near-ground cover available to hare in summer (0-
1 m) and winter (1-2 m) (Wolfe et al. 1982), as snow depth measured over 2 winters
(2006/2007) averaged 0.98 ± 0.02 m (n = 204) in the study region during the course of this
study (J.Hodson unpublished data). Vertical cover (VertCov) was estimated visually in
10% classes by an observer standing at the plot centre. We measured deciduous browse
availability (Browse) within 2 m × 10 m rectangular plots, oriented north-south, centred on
five pellet plots within each grid (always plots 4, 6, 10, 14, and 16; Figure 1.1). We
counted the number of twigs (terminal leaders ≥5 cm long) by species that were within 0-2
m above ground level within each plot (Potvin 1995). Deciduous browse species included
white birch, willow (Salix spp.), speckled alder (Alnus rugosa), green alder (Alnus crispa),
serviceberry (Amelanchier spp.), mountain ash (Sorbus spp.), red-osier dogwood (Cornus
stolonifera), mooseberry (Viburnum edule), pin cherry (Prunus pensylvanica) and mountain
maple (Acer spicatum). We also counted the number of twigs clipped by snowshoe hare
during the winter previous to each survey. Browse availability (twigs/m² between 0-2 m)
was calculated as the number of unclipped twigs plus the number of twigs clipped in the
winter previous to the survey.
Gap transects in mature and late-seral stands
We conducted line-intercept surveys (Runkle 1982, Pham et al. 2004) during
summer 2008 to measure the proportion of the forest in canopy gaps ("canopy gap
fraction") in mature and late-seral stands (≥80 years old; n = 34 sites). We focussed on
stands ≥80 years old because previous work (Bouchard et al. 2008) indicated that canopy
break-up and transition to uneven-aged stand structure generally begins at roughly 80 years
after fire in eastern boreal forests. Using a hip chain, we recorded the distance at which we
30
entered and exited gaps that intercepted a 300 m transect within each site (between plots 8
and 12 in each pellet grid; Figure 1.1) based on the distance between tree trunks delimiting
the gap edge ("expanded gap", sensu Runkle 1982). The total length of the transect within
canopy gaps was then used to estimate canopy gap fraction for each site. We considered all
openings where the height of regeneration was less than two thirds of the height of the
dominant canopy layer (Pham et al. 2004). We recorded whether each gap originated from
tree mortality, edaphic conditions or a combination of both, and we visually estimated the
cover of coniferous and deciduous regeneration within gaps using the Daubenmire (1959)
scale (0%, 1-5%, 5-25%, 25-50%, 50-75%, 75-95%, 95-100%). We also noted the
presence of four classes of tree mortality (Aakala et al. 2007) within canopy gaps: 1) intact
standing dead trees (no detectable fragmentation of their tops), 2) standing dead trees
snapped off above 1.3 m, 3) fallen trees snapped off or broken below 1.3 m, and 4)
uprooted trees.
Statistical Analysis
Changes in cover, browse, and hare abundance with time since disturbance
We used generalized additive models (GAMs) to describe changes in hare
abundance, habitat structure and browse availability as a function of stand age. GAMs are
a semi-parametric extension of generalized linear models (GLMs) capable of describing
highly non-linear and non-monotonic relationships between response and explanatory
variables using smoothing functions (Guisan et al. 2002). Prior to analysis, one pellet plot
was removed from the total count for an 18-yr-old clearcut site. This plot contained 1215
pellets, while 16 of the remaining plots at this site had 0 pellets, and two plots had 8 and 13
pellets respectively. This plot grossly inflated the site‘s mean pellet density and we
therefore considered it to be highly unrepresentative of the rest of the site. This was also
the highest individual pellet count out of all the pellet plots sampled in the chronosequence
(n = 1614). Mean pellet density for this site was thus based on the 18 remaining plots. A
square root transformation was used to normalize pellet density prior to analysis. Vertical
cover, lateral cover between 0-1 m and 1-2 m, and browse availability were not
31
transformed, because examination of the residuals from GAMs indicated that they were
homogenously distributed. GAMs were fitted with the MGCV package (Wood 2009) in R
9.0 (R Core Development Team 2009), using generalized cross-validation (GCV) to
determine the optimal amount of smoothing for each regression. Using the spatial
coordinates of the centroid of each pellet grid to measure distance between sites, we ran
Mantel tests to verify whether there was any spatial autocorrelation in snowshoe hare pellet
density or in the residuals from the GAM of pellet density predicted by stand age (Hodges
et al. 2009).
Variations in hare abundance with cover and browse availability
To test whether relative snowshoe hare abundance varied according to changes in
vertical cover, lateral cover, and browse availability, we used general linear models with
the square root of pellet density as the response variable. We used model comparison to
determine which combination of habitat features best explained variations in relative hare
abundance based on Akaikes Information Criterion adjusted for small sample size (AICc),
differences in AICc (Δi) and the weight of evidence (wi) for each model (Burnham and
Anderson 2002). To limit the size of our candidate model set, we first compared support
for linear versus quadratic models for vertical cover, lateral cover and browse availability
individually before building more complex models with combinations of these variables.
Linear and quadratic models for lateral cover and vertical cover had similar support from
the data (ΔAICc <1 in both cases), whereas a quadratic model of browse availability
provided a better fit to the data than a linear model (ΔAICc = 4.5). We therefore included
only simple effects of lateral and vertical cover and a quadratic effect of browse availability
in more complex models (Table 1.2). To assess whether the importance of food and cover
as factors limiting hare density varied between early- versus late-seral stands, we tested
additional models that included interactions between each habitat variable and a
dichotomous variable separating stands into two broad developmental stages
(―Dev_Stage‖), stands <80 years old and stands ≥80 years old. This division roughly
corresponds to the age at which canopy break-up and understory re-initiation generally
begins in eastern boreal forests (Bouchard et al. 2008). Multicollinearity was absent from
32
all candidate models, as variance inflation factors (VIF) for habitat variables were always
<2 (Graham 2003).
Variations in hare abundance with canopy gap fraction in stands ≥80 years old
To test whether hare abundance in stands ≥80 years old varied according to canopy
gap fraction, we used general linear models with the square root of pellet density as the
response variable and canopy gap fraction as a predictor. We used AICc, differences in
AICc (Δi ) and weight of evidence (wi) to compare the fit of candidate models with either
simple or quadratic effects of gap fraction for all gaps regardless of their origin (Gap
fraction) versus candidate models only considering the fraction of stands comprised of
mortality-origin gaps (Mortality gap fraction). These two alternatives were tested because
previous work indicated that snowshoe hare are more likely to forage in gaps with higher
densities of coniferous regeneration (Hodson et al. 2010a), and mortality-origin gaps had
higher levels of conifer regeneration than edaphic-origin and combination
edaphic/mortality-origin gaps (Kruskal-Wallis
2
2 = 172.5, P < 0.001, mean percent conifer
regeneration cover within gaps based on the mid-points of cover classes: mortality 42% [n
= 418], edaphic 13% [n = 118], combination 19% [n = 171]).
Relative snowshoe hare abundance in stands regenerating from fire versus clearcutting
To evaluate whether snowshoe hare pellet density varied similarly with time since
disturbance in fire- and harvest-origin stands, we compared pellet density between fires and
cuts within four 10-year stand age classes for which we had samples from both disturbance
types: 10-19 years (cut: 7 sites; fire: 2 sites), 30-39 years (cut: 6 sites; fire: 6 sites), 40-49
years (cut: 7 sites; fire: 4 sites), and 60-69 years (cut: 2 sites; fire: 2 sites). A two-way
ANOVA compared pellet density as a function of disturbance type, age class, and the
interaction between disturbance type and age class (model: [pellets/m²]0.5
= disturbance
type + age class + disturbance type × age class), to evaluate whether differences in pellet
density among age classes depended on disturbance type. Following a significant ANOVA,
Tukey‘s HSD post-hoc tests identified significant differences between age classes and
disturbance type.
33
Results
Changes in relative snowshoe hare abundance and habitat structure with stand age
Snowshoe hare pellets were found at 81 out of 84 sites in the forest chronosequence and
mean pellet density within sites varied from 0 to 11 pellets/m². Generalized additive
modelling indicated that snowshoe hare pellet densities followed a bi-modal distribution
with stand age. Pellet densities increased to a first peak in stands between 40 and 50 years
old, followed by a period of low pellet density from 80 to 180 years, after which point
pellet densities increased slightly in stands >180 years old (Figure 1.2). Stand age
explained 39% of the variation in snowshoe hare pellet density (Table 1.1). We detected no
spatial autocorrelation in observed pellet densities (Mantel test, Z = 0.024, P = 0.23) or in
the GAM residuals of snowshoe hare pellet density predicted by stand age (Mantel test, Z =
- 0.034, P = 0.79), suggesting that sampled sites were sufficiently far apart to be spatially
independent.
GAMs also revealed important non-linear changes in habitat structure with time
since disturbance. Vertical cover peaked slightly later than snowshoe hare pellet density, at
roughly 60 years since disturbance, and then decreased to remain at moderate levels (~ 50
%) from 80 to 265 years (Figure 1.2). Estimated trends in lateral cover between 0-1 m and
1-2 m from ground level both followed clear bimodal distributions with stand age, with
peaks in lateral cover between 1-2 m occurring roughly 10 years later than those observed
for lateral cover between 0-1 m (first peak: 30 vs. 40 years, second peak: >150 years vs.
>160 years; Figure 1.2). Lateral cover in late-seral stands (>150 years) reached levels that
were similar to those observed during the first peak in early-seral stands 30-40 years old.
Browse availability followed a unimodal distribution with stand age, reaching a peak at 30
years, and then decreasing to remain at relatively low levels from 80 years onwards (Figure
1.2).
34
Changes in relative snowshoe hare abundance with habitat structure and food availability
Variations in relative snowshoe hare abundance were best explained by a
combination of vertical cover, lateral cover between 1-2 m, and a quadratic effect of browse
availability, as well as interactions between cover and the developmental stage of forest
stands (Akaike weight [wi] = 0.92; Table 1.2). This model explained 57% (R² = 0.57) of the
variation in snowshoe hare pellet densities observed across the forest chronosequence.
Prediction of snowshoe hare pellet densities based on parameter estimates from this model
(Table 1.3) revealed that relative hare abundance increased with lateral cover between 1-2
m and vertical cover in stands <80 years old, whereas there was no relationship between
hare pellet density and these variables in stands ≥80 years old (Figure 1.3a,b). Pellet
density followed a curvilinear relationship with browse availability throughout the entire
forest chronosequence, and relative hare abundance tended to decrease when browse
availability surpassed 15 twigs/m² (Figure 1.3c). Models with lateral cover between 0-1 m
had very little support from the data (wi <0.001), suggesting that variation in cover within
this height stratum had little influence on relative hare abundance.
To determine which habitat features explained the most variation in pellet density in
stands <80 years old, we calculated partial R² values for vertical cover, lateral cover
between 1-2 m and browse availability for a model considering only this phase of stand
development: (pellets/m²)0.5
= VertCov + LatCov1_2 + Browse + Browse², n = 50 stands.
This model explained 53% of the variation in pellet density in stands <80 years old, with
vertical cover explaining the greatest amount of variation (partial r² = 0.27), followed by
lateral cover (partial r² = 0.16) and browse availability (Browse + Browse²: partial r² =
0.10). The same model applied to stands ≥80 years old (n = 34) explained only 5% (R² =
0.05) of the variation in pellet density, and 95% confidence intervals for parameter
estimates for all of the habitat variables included 0.
Canopy gap fraction in stands ≥80 years varied between 0.20 and 0.85. The majority
(59%) of recorded gaps (706 gaps encountered along 34 transects) originated from tree
mortality, with snapped standing and fallen dead trees being present in the largest
proportion of mortality-origin gaps (present in 95% and 82% of mortality-origin gaps
respectively). Intact standing dead trees and uprooted trees were present in only 57% and
25% of mortality-origin gaps, respectively. The proportion of stands in mortality-origin
35
gaps increased linearly with stand age (mortality-origin gap fraction = 0.13 + 0.003 Stand
age, P <0.001, R2 = 0.41, n = 34). There was a positive correlation between mortality-
origin gap fraction and lateral cover between 1-2m (Pearson's correlation: r = 0.73, P
<0.001, n = 34); however, there was no apparent relationship between the abundance of
mortality-origin gaps and browse availability (Pearson's correlation: r = 0.08, P = 0.57, n =
34). Comparison of models predicting snowshoe hare pellet density as a function of
canopy gap fraction in stands ≥80 years old revealed that a quadratic relationship between
mortality-origin gap fraction and pellet density had the most support from the data (wi =
0.61), whereas models including the fraction of all types of canopy gaps (edaphic,
mortality, and edaphic/mortality origin gaps) received even less support than the intercept-
only model (Table 1.4). The estimated curve based on parameter estimates from the top-
ranking model (Table 1.5) indicated that snowshoe hare pellet density peaked at
intermediate levels (40-50%) of mortality-origin gap fraction (Figure 1.4). This model
explained 22% (model R² = 0.22) of the variation in pellet density in stands ≥80 years old.
Relative snowshoe hare abundance in fire- versus harvest-origin stands
Pellet densities did not differ between fire- and harvest-origin stands across the four
stand age classes (disturbance type: F1,28 = 1.90, P = 0.18; disturbance type × age class:
F3,28 = 1.10, P = 0.36). Mean pellet density was 3.4 and 4.6 times higher in stands 40-49
years old than in stands 10-19 and 30-39 years old respectively (age class: F3,28 = 12.43, P
<0.001; Figure 1.5). To verify our conclusions about disturbance type, we also tested a
multiple regression model linking pellet density to stand age and disturbance type:
(pellets/m2)0.5
= disturbance type + stand age + stand age2 + disturbance type × stand age +
disturbance type × stand age2. Neither the simple effect of disturbance type (P = 0.44) nor
the interactions between stand age and disturbance type were significant (P > 0.50 in all
cases).
Discussion
36
A 265 year chronosequence of eastern boreal forest succession revealed that
snowshoe hare undergo a boom and bust in density during the first 80 years of succession,
followed by a second period of moderate increase in stands >180 years old. This finding
implies that significant changes in hare density occur within a fairly narrow temporal
window relative to fire return intervals that can exceed 500 years in eastern Canada
(Bouchard et al. 2008). We observed a clear bimodal pattern in near-ground cover with
stand age, but despite the well-known association between hare abundance and lateral cover
(e.g. Wirsing et al. 2002, Hodges et al. 2009), pellet densities in late-seral stands remained
almost 10 times lower than those observed during the first peak at 40-50 years post-
disturbance. These findings are novel in light of recent studies from Montana and
Yellowstone National Park that reported similar hare densities in early- (19-45 years old)
and late-seral stands (>150 years) with dense understories (Griffin & Mills 2009, Hodges et
al. 2009). These studies did not, however, explicitly assess which habitat features
explained these similarities. Models predicting pellet density based on habitat structure,
food availability and seral-stage revealed that protective cover was not a consistent
predictor of hare density in all phases of forest succession. Changes in pellet density
closely followed stand-level changes in lateral and vertical cover during early succession,
but not in stands >80 years old. Some variation in snowshoe hare abundance in forests ≥80
years old was, however, explained by the proportion of each stand that was occupied by
mortality-origin canopy gaps, with highest pellet densities occurring at intermediate canopy
gap fraction. These findings suggest that variation in hare abundance during late-
succession may be mediated by finer-scale heterogeneity created by canopy gap dynamics.
The temporal changes in relative snowshoe hare abundance that we observed during
the first 80 years of stand development are consistent with previous studies from boreal
regions that reported peak hare densities in early- to mid-successional stands (Thompson et
al. 1989, Koehler 1990, Paragi et al. 1997, Newbury & Simon 2005). The strong positive
association between hare density and vegetative cover in early-seral stands reflects the
importance of predation risk in shaping patterns of snowshoe hare distribution. The
majority of hare (>75%) die from predation (Hodges et al. 1999, Etcheverry et al. 2005)
and hare mortality is generally higher in open forests than in stands with dense understory
and canopy layers (Rohner & Krebs 1996, Griffin & Mills 2009). Accordingly, the timing
37
of the first peak in hare density along the forest chronosequence (~50 years) occurred in
between when the first maxima in lateral and vertical cover were observed, at roughly 40
and 60 years respectively. Predator avoidance frequently imposes a trade-off between
access to food and cover that creates a mismatch between food availability and herbivore
abundance when food-rich habitats are also the most risky (Mysterud et al. 1999). The
decreasing curvilinear relationship between relative hare abundance and browse availability
observed during early-succession reflects such a trade-off, because browse availability
peaked several years ahead of lateral and vertical cover along the chronosequence.
The modest increase in hare pellet density despite strong increases in understory
vegetation cover between 80 and 265 years post-disturbance suggests that food availability
may limit hare abundance during late-succession. Similarly, Crête and Courtois (1997)
suggested that low moose densities in the Côte-Nord region are due to the scarcity of
deciduous browse in old spruce-fir stands that dominate eastern boreal forests. Although
deciduous browse is concentrated within canopy gaps in late-seral stands (Hodson et al.
2010a) and canopy gap fraction generally increases with stand age (Harper et al. 2006),
browse availability did not follow the same bimodal distribution with stand age observed
for lateral cover and we did not observe any relationship between browse and the
abundance of mortality-origin gaps. This is likely because canopy gap regeneration is
dominated by black spruce and balsam fir in eastern old-growth boreal forests (Pham et al.
2004). Nevertheless, if food availability was truly limiting hare abundance in late-
succession, we should expect differences in browse availability between early- and late-
seral stands that are of the same order of magnitude as those observed for pellet density.
Whereas pellet density was roughly 10 times higher in stands 40-50 years old (mean ± s.d.:
7.17 ± 2.37 pellets/ m², n = 11) than in stands >180 years old (mean ± s.d: 0.78 ± 0.74
pellets/ m², n = 17), differences in the mean density of deciduous browse were less than 2-
fold (40-50 years: mean ± s.d. = 6.58 ± 5.71 twigs/m²; >180 years: 3.94 ± 2.96 twigs/m²).
The large difference in relative hare abundance between early- and late-successional stages
might therefore be explained by differences in food accessibility mediated by predation risk
rather than by absolute differences in browse availability.
Snowshoe hare generally select browse sites that are close to cover, presumably to
minimize the risk of being detected by predators or facilitate escape (Hodges & Sinclair
38
2005). Young boreal forest stands have finer-scale patterns of patch structure than older
stands (Harper et al. 2006), meaning that food and cover may be interspersed at a relatively
fine grain. Consequently, hare may never have to travel far from conifer cover to forage.
In older stands, however, the limited browse available is concentrated within canopy gaps,
and both foraging experiments and natural browse use surveys have shown that hare are
more reluctant to forage towards the centre of large openings (Hodson et al. 2010a).
Therefore, much of the browse within large gaps in old-growth forests may remain unused
by hare due to a high perception of predation risk.
Although hare obtain most of their browse from within gaps, they also selectively
travel under areas of greater canopy cover (Hodson et al. 2010a), meaning that increases in
canopy gap fraction coincide with reductions in areas providing safe travel corridors. Gap
dynamics should therefore mediate stand-level variation in hare density in old-growth
stages by influencing the balance between access to food and cover. Consistent with this
hypothesis, we observed that pellet density was highest in stands with an intermediate
abundance of mortality-origin canopy gaps. Foraging hares should benefit from the greater
amount of lateral cover in mortality-origin gaps, which may explain their stronger
association with these gaps than other types. Gaps with trees uprooted by windthrow also
provide microsites favourable for the establishment of preferred browse species such as
white birch (Carlton & Bazzaz 1998, Newbury & Simon 2005). Indeed, white birch stems
were present in a greater proportion of mortality-origin gaps that had uprooted trees (66%)
than those without (44%). Although uprooted trees were present in only 25% of mortality-
origin gaps, this type of tree mortality may nonetheless be an important determinant of
preferred browse availability for hare in late-seral stands. Overall, these findings
demonstrate that snowshoe hare abundance in old-growth boreal forests may be structured
by fine-scale patterns of food and cover interspersion, which largely depend on the
abundance and origin of canopy gaps.
Despite the importance of small gap dynamics in structuring snowshoe hare habitat
in humid boreal ecosystems, fire remains a key disturbance that periodically resets
succession over vast areas (Bouchard et al. 2008). An ecosystem-based approach to forest
management might suggest that portions of the landscape should be harvested with methods
that can emulate fire. Although others have described changes in post-fire and post-harvest
39
hare abundance (reviewed by; Fisher & Wilkinson 2005), this is the first study to
simultaneously compare patterns of hare abundance following these two disturbance types.
Similar patterns of hare abundance in fire- and harvest-origin stands during the first 70
years of stand development suggested that clearcutting created habitat conditions for hare
that were comparable to those produced by fire. We can wonder, however, whether the
application of pre-commercial thinning within a portion of the post-harvest chronosequence
could have influenced this conclusion. Pre-commercial thinning is increasingly practiced in
Québec (Pothier 2002), and has been shown to reduce hare densities in regenerating stands
(Ausband & Baty 2005, Griffin & Mills 2007, Homyack et al. 2007). Most cuts between
20-40 years old (9 out of 11 sites) had been thinned, and it is possible that hare abundance
within stands in this age range could have been reduced, causing a delay in the timing of
peak hare abundance in harvested stands. Although some studies have observed highest
pellet densities in stands 20-30 yrs (Thompson et al. 1989, 20-30 yrs; Newbury & Simon
2005) it is difficult to assess whether thinning actually caused a delay in the timing of the
first peak in our study because these other studies did not sample cuts >30 years.
Nevertheless, given that pellet densities were similar between fires and cuts in the
remaining age-classes that had not undergone thinning, it does not appear as though this
treatment was the cause of similarity between these two disturbance types. Furthermore,
the application of this treatment to a small segment of harvested stands did not influence
our ability to detect a bimodal distribution of hare abundance over the complete
chronosequence.
A more thorough understanding of how disturbance and succession shape animal
distribution in managed forests should help to inform the development of practices that
better maintain natural forest ecosystem dynamics. Snowshoe hare underwent pronounced
changes in abundance during the first 80 years of succession following stand-replacing
disturbance. This was followed by a much longer successional period (>150 yrs) during
which relatively subtle changes in hare abundance were mediated by fine-scale
disturbances. Although clearcut harvesting may establish early-seral habitat conditions for
snowshoe hare that are similar to fires, profitable harvest rotations for even-aged
management (<100 years) are generally shorter than fire return intervals in eastern boreal
forests (>250 yrs) (Bergeron et al. 2001, Harvey et al. 2002, Bouchard et al. 2008). The
40
continued use of this approach to forest management could therefore have important
consequences for the spatiotemporal distribution and dynamics of snowshoe hare by
increasing the proportion of seral stages in the landscape in which hare experience the most
pronounced changes in abundance. Because snowshoe hare follow a predictable pattern
with stand age, we should be able to forecast future patterns of hare distribution under
different management scenarios. To illustrate this, we used the GAM curve from the forest
chronosequence (Figure 1.2) to estimate snowshoe hare abundance in three hypothetical
1000 ha forest landscapes: 1) an unmanaged landscape with a 250-year fire cycle and a
negative exponential forest age-class distribution (Van Wagner 1978), 2) a fully regulated
landscape under even-aged management with a harvest rotation of 100 years and 3) a
landscape under cohort management proposed by Bergeron et al. (2002) based on a 200-
year fire cycle, whereby "stand-initiating" harvesting is used to recruit even-aged stands
<100 yrs (cohort 1) on 39% of the landscape, partial harvesting is used to move 24% of the
landscape into stands with an uneven or irregular structure (100-200 yrs; cohort 2) and
selection cutting is used to mimic gap dynamics in old-growth stands on 37% of the
landscape (200-300 yrs; cohort 3). Further details on landscape structure and calculations
of hare abundance are provided in Appendix 1b. Snowshoe hare abundance in the fully
regulated landscape under even-aged management was predicted to be 40% higher than in
the unmanaged landscape with a 250-year fire cycle. In contrast, the landscape under
cohort management was predicted to support only 6% more hares than the unmanaged
landscape. Assuming that silvicultural treatments retaining late-seral stand structure can
maintain similar densities of snowshoe hare to those observed in uncut forests >100 years
old, a greater use partial harvesting may be an appropriate strategy to maintain
characteristic distributions of snowshoe hare and their predators in regions with prolonged
fire cycles.
Acknowledgements
This work was supported by the NSERC-Laval University Industrial Research Chair in
Silviculture and Wildlife and its partners. We also would like to acknowledge funding
provided by the FQRNT and FCI. We gratefully acknowledge the many field assistants
41
whose dedicated efforts made this work possible: K. Hammelin, J.F. Poulin, J. Tremblay,
M.-A. LaRose, V. Hébert-Gentille, E. Renaud-Roy, M. White, S. Lavoie, K. Poitras, M.-L.
Le Blanc. We also thank our industrial partners Abitibi-Bowater, Kruger, and Arbec forest
industries for their financial and technical support.
42
Table 1.1. Fit statistics for general additive models (GAMs) used to model snowshoe hare
pellet density, vertical cover, lateral cover, and browse availability as a function of stand
age in a 265 yr boreal forest chronosequence of stand development (n = 84 stands).
Effective degrees of freedom (edf) and F-values to test significance of smoothing
parameters [s(stand age)] are provided.
Model edf F p adj. R²
Pellet density (pellets/m²) 6.86 7.78 <0.001 0.39
Vertical cover (%) 7.85 19.36 <0.001 0.65
Lateral cover 0-1 m (%) 6.37 4.89 <0.001 0.27
Lateral cover 1-2 m (%) 5.97 6.37 <0.001 0.31
Browse availability 0-2 m (twigs/m²) 6.28 2.68 0.019 0.14
43
Table 1.2. Competing models predicting the density of snowshoe hare pellets along a
chronosequence of forest stand development (n = 84) for stands aged between 3 and 265
years based on combinations of lateral cover, vertical cover, and browse availability,
including models with interactions between a dichotomous variable (Dev_Phase: 0 = stands
<80 years, 1 = stands ≥80 years) and lateral cover, vertical cover, and browse availability to
assess whether the importance of factors limiting hare density varies between two phases of
stand development.
Model K AICc ΔAICc wi
Dev_Phase + LatCov1-2 + LatCov1-2×Dev_Phase +
VertCov + VertCov×Dev_Phase + Browse + Browse² 9 164.17 0.00 0.92
Dev_Phase + LatCov1-2 + LatCov1-2×Dev_Phase +
VertCov + VertCov×Dev_Phase + Browse +
Browse×Dev_Phase + Browse² + Browse²×Dev_Phase 11 169.18 5.01 0.08
LatCov1-2 + VertCov 4 199.06 34.90 0.00
LatCov1-2 + VertCov + Browse + Browse² 6 199.81 35.64 0.00
LatCov1-2 3 205.81 41.65 0.00
VertCov 3 207.62 43.45 0.00
Browse + Browse² 4 215.76 51.60 0.00
LatCov0-1 3 217.52 53.36 0.00
intercept only 2 218.28 54.12 0.00
44
Table 1.3. Parameter estimates for the top ranking model predicting pellet density in a
forest chronosequence of stands (n = 84) varying in age between 3 and 265 years.
Dev_Phase is a dichotomous variable distinguishing stands in early phases of development
(<80 years, Dev_Phase = 0) from gap-phase stands (≥80 years, Dev_Phase = 1). Parameter
estimates for Dev_Phase in interaction with lateral cover between 1-2m (LatCov1-2) and
vertical cover (VertCov) indicate changes in the slope between pellet density and vertical
cover or lateral cover in stands ≥80 years old. Parameter estimates whose 95% or 90% CIs
do not include zero are indicated in bold.
Parameter Estimate 95% CI 90% CI
Intercept 0.104 (-0.303, 0.512) (-0.238, 0.446)
Dev_Phase 0.446 (-0.564, 1.456) (-0.401, 1.239)
LatCov1-2 0.022 (0.009, 0.035) (0.011, 0.033)
LatCov1-2×Dev_Phase -0.021 (-0.041, 0.000) (-0.038, -0.004)
VertCov 0.016 (0.009, 0.022) (0.010, 0.021)
VertCov×Dev_Phase -0.017 (-0.035, 0.002) (-0.032, -0.001)
Browse 0.050 (-0.013, 0.112) (-0.003, 0.102)
Browse² -0.003 (-0.006, -0.001) (-0.005, -0.001)
45
Table 1.4. Competing models predicting the density of snowshoe hare pellets in stands ≥80
years old based on the fraction of all types of canopy gaps versus only the fraction of
mortality-origin canopy gaps.
Model K AICc ΔAICc wi
Mortality gap fraction + Mortality gap fraction² 4 35.13 0.00 0.61
Mortality gap fraction 3 37.41 2.28 0.20
intercept only 2 38.56 3.43 0.11
Gap fraction + Gap fraction² 4 40.07 4.93 0.05
Gap fraction 3 40.94 5.80 0.03
46
Table 1.5. Parameter estimates from the top-ranking model predicting snowshoe hare pellet
density as a function of mortality-origin canopy gap fraction in stands ≥80 years old.
Parameter Estimate 95% CI
Intercept 0.321 (0.056, 0.586)
Mortality gap fraction 2.245 (0.589, 3.901)
Mortality gap fraction2
-2.305 (-4.376, -0.235)
47
Figure 1.1. Left panel: Map of the study area located in the Côte-Nord region of Québec
showing the location of harvest and fire origin stands that were sampled for relative
snowshoe hare abundance. Numbers indicate stand age (years) at sampled sites. Right
panel: Pellet inventory grids used to measure relative snowshoe hare abundance. Vertical
cover and lateral cover were measured at each of the 19 stations. Grey rectangles indicate
position of browse inventory plots (2 m ×10 m). Dotted line indicates the transect used for
canopy gap surveys in stands ≥ 80 years old.
48
49
Figure 1.2. General additive models (GAMs; solid lines) ± approximate 95% confidence
intervals (dotted lines) describing changes in snowshoe hare pellet density, vertical cover,
lateral cover, and browse availability with stand age in a boreal forest chronosequence of
stand development (n = 84 stands). "Cut-PCT" represents harvested stands that had
undergone pre-commercial thinning.
50
51
Figure 1.3. Predicted values (solid lines) ± 95% confidence intervals (dotted lines) of
snowshoe hare pellet density as a function of vertical cover, lateral cover between 1-2m and
browse availability in two phases of stand development (Dev_Phase: <80 years = 0, ≥80
years = 1) using parameter estimates from the model: (pellets/m²)0.5
= 0.104 +
0.446*Dev_Phase + 0.022*LatCov1-2 – 0.021*LatCov1-2×Dev_Phase + 0.016*VertCov –
0.017*VertCov×Dev_Phase + 0.050*Browse – 0.003*Browse². Predicted pellet densities
were calculated over the range of observed values of each habitat variable in each stage of
stand development, while holding the other variables at their mean in each developmental
stage. Values used to generate predicted curves for each developmental phase are as
follows: Stands <80 years: Vertical cover: mean = 37%, range = 0-83%; Lateral cover 1-
2m: mean = 31%, range = 3-60%; Browse availability: mean = 7.6 twigs/m², range = 0-30
twigs/m²; Stands ≥80 years: Vertical cover: mean = 51%, range = 18-70%; Lateral cover 1-
2m: mean = 26%, range = 6-56%; Browse availability: mean = 3.7 twigs/m², range = 0-11
twigs/m².
52
Figure 1.4. Predicted pellet density as a function of the proportion of stands in mortality-
origin canopy gaps ("mortality gap fraction") in stands ≥80 years old. Predicted values are
calculated from parameter estimates in Table 1.4, open circles are observed pellet densities.
53
Figure 1.5. Boxplots of snowshoe hare pellet density in stands originating from forest fires
and clearcutting in four different stand age classes. Solid lines within bars represent
median values. Sample sizes are indicated underneath each bar. Solid lines within bars
without whiskers (samples with n = 2) represent are equivalent to the mean. Groups of
fire/cut origin stands with different letters represent significant differences in pellet density
between age classes.
54
Chapitre 2
Fine-scale disturbances shape space-use patterns of a
boreal forest herbivore
James Hodson*, Daniel Fortin
*, and Louis Bélanger
†
*NSERC-Université Laval Industrial Research Chair in Silviculture and Wildlife,
Département de Biologie, Université Laval, Québec, QC, Canada, G1V 0A6 (JH, DF)
†Département des sciences du bois et de la forêt, Université Laval, Québec, QC, Canada,
G1V 0A6 (LB)
Article publié dans le Journal of Mammalogy 91 (3): 607-619
55
Résumé
Les perturbations naturelles ont une influence déterminante sur la structure et le
fonctionnement des écosystèmes. Les perturbations peuvent créer de nouvelles sources de
nourriture et modifier la structure de l‘habitat, générant par le fait même une hétérogénéité
spatiale qui affecte le compromis entre l‘acquisition de nourriture et l‘évitement des
prédateurs. Nous avons évalué comment la dynamique de trouées dans la forêt boréale
ancienne de l‘est du Canada affecte la répartition spatiale de la nourriture et du couvert
pour le lièvre d‘Amérique (Lepus americanus) et comment les lièvres réagissent à ces
patrons spatiaux. Nous avons d‘abord comparé la disponibilité de brout à l‘intérieur des
trouées avec celle dans la forêt avoisinante. Nous avons ensuite examiné la sélection
d‘habitat à fine échelle, les patrons de déplacement et les décisions alimentaires du lièvre
pendant l‘hiver. La perception du risque de prédation à l‘intérieur des trouées a été évaluée
à l‘aide d‘expériences d‘approvisionnement. La disponibilité de brout était quatre fois plus
grande dans les trouées que sous couvert forestier. Bien que les lièvres aient obtenu la
majorité de leur nourriture à partir des trouées pendant l‘hiver, leur utilisation de l‘espace
était influencée par la perception d‘un risque de prédation accru dans les trouées. Les
lièvres sélectionnaient les habitats ayant une plus grande fermeture de canopée, ce qui
suggère que leur utilisation des trouées se limite principalement aux activités
d‘alimentation. De plus, les lièvres orientaient généralement leurs déplacements afin
d‘éviter les trouées et ils augmentaient leur vitesse de déplacement dans les zones de faible
couverture végétale. Lors des expériences d‘approvisionnement, les lièvres ont consommé
les tiges de pin gris de façon plus intensive sous couvert forestier que dans les trouées, ce
qui met en lumière l‘existence d‘un compromis entre nourriture et sécurité. Les
expériences d‘approvisionnement et les relevés de tiges naturelles ont tous les deux indiqué
que, lorsque les lièvres s‘alimentaient dans les trouées, ils étaient moins enclins à utiliser
les tiges se trouvant loin du couvert forestier. Notre étude démontre comment la
dynamique des trouées dans les peuplements de forêt ancienne peut structurer
l‘organisation spatiale à fine échelle d‘une espèce clef de la forêt boréale en créant de
l‘hétérogénéité spatiale dans la répartition des sites risqués mais riches en nourriture. La
56
variation spatiale dans l‘utilisation du brout en réponse au risque de prédation peut à son
tour influencer les patrons de croissance et de survie des jeunes arbres dans les trouées.
Ainsi, la dynamique de trouées peut s‘avérer un processus fondamental qui structure les
interactions prédateur-proies dans les forêts boréales anciennes.
57
Abstract
Natural disturbance is a key determinant of ecosystem structure and function. Disturbances
can create novel resource patches and modify habitat structure, thereby inducing spatial
heterogeneity in the trade-off between food acquisition and predator avoidance by prey.
We evaluated how canopy gap dynamics in eastern Canadian old-growth boreal forest alter
the spatial distribution of food and cover for snowshoe hares (Lepus americanus) and how
hares responded to these spatial patterns. We first compared browse availability within
canopy gaps and the surrounding forest. We then examined fine-scale habitat selection,
movement patterns, and foraging decisions by hares during winter. Perception of risk
within canopy gaps was assessed using foraging experiments. We found that browse
availability was four times higher within gaps than under forest cover. Although hares
acquired most of their browse from gaps, their use of space during winter was influenced
by a greater perception of predation risk within gaps. Hares selectively used areas of higher
canopy closure suggesting that they restricted their use of gaps to foraging activities.
Furthermore, hares biased their movements away from gaps or increased their speed of
travel in areas of relatively low cover. Hares consumed experimental browse stems more
intensively under forest cover than in canopy gaps, indicating a trade-off between food and
safety. When foraging within canopy gaps, hares also were less likely to use both
experimental and natural food patches located far away from cover. Our study
demonstrates how gap dynamics in old-growth stands can structure the fine-scale spatial
organization of a key prey species of the boreal forest by creating spatial heterogeneity in
their landscapes of fear and food. Spatial variation in browse use in response to predation
risk may in turn influence patterns of sapling growth and survival within canopy gaps. Gap
dynamics therefore may be a fundamental process structuring predator-prey interactions in
old-growth boreal forests.
58
Introduction
Natural disturbances that vary in size, severity, and frequency play a fundamental role
in structuring aquatic and terrestrial ecosystems by creating heterogeneity at multiple
spatial and temporal scales (Sousa 1984, Pickett & White 1985). Habitat disturbance can
affect animal distribution by altering the composition and structure of vegetation that
provide food and cover, and many animals benefit from disturbances that create productive
conditions associated with areas undergoing regeneration (Sousa 1984). Although
infrequent broad-scale disturbances such as forest fires and tropical storms can influence
patterns of species occurrence at the landscape scale (Fisher & Wilkinson 2005, Willig et
al. 2007), frequent microhabitat disturbances such as tree-fall gaps, blowouts, and wave
action create fine-scale heterogeneity that also plays an important role in determining
species distribution (Paine & Levin 1981, Bouget & Duelli 2004, Cramer & Willig 2005).
Habitat heterogeneity can have a profound influence on trophic interactions. For
example, heterogeneity can promote the persistence of predator-prey populations by
reducing predator foraging efficiency, by creating spatial refuges for prey, or by creating
locally asynchronous population dynamics (Huffaker 1958, Hastings 1977, Holt & Hassell
1993). Recent investigations have shown that the functional response of both herbivores
and carnivores to food availability can depend on the spatial distribution of these resources
(Pitt & Ritchie 2002, Hobbs et al. 2003). Resource heterogeneity therefore can influence
the functional link among trophic levels. For herbivores, variation in the spatial
arrangement of plants can affect the rate at which they encounter food patches, thereby
influencing their rate of energy intake and dietary choice (Fortin et al. 2002, Hobbs et al.
2003). To increase their intake rate in heterogeneous environments herbivores should
concentrate on aggregations of food patches to reduce travel time between patches (Nonaka
& Holme 2007), but the most profitable food patches often are also the most risky (Brown
& Kotler 2004).
Fear of predation is a major force influencing movement and foraging decisions of prey
(Lima & Dill 1990), and disturbances that increase food resources also can remove habitat
structure that provides protection against predators. Given that predators may be more
efficient at detecting and capturing prey in certain habitats (Rohner & Krebs 1996), prey
59
often rely on habitat structure as a cue for risk (Brown & Kotler 2004). For example, they
may trade off food for safety by foraging less intensively in open habitats or with
increasing distance from protective cover (Hochman & Kotler 2007). During locomotion
prey also may attempt to mitigate risk by moving in areas of greater cover (Lagos et al.
1995, Fortin et al. 2005), or by adjusting their speed to quickly traverse areas where they
would be more conspicuous to predators (Vasquez et al. 2002). Slight variations in habitat
structure can result in relatively large changes in the perception of risk (van der Merwe &
Brown 2008). Therefore microhabitat disturbances should shape prey distribution by
continually changing the landscapes of food and fear (sensu Laundré et al. 2001) around
which prey species structure their home ranges.
Canopy gap dynamics in old-growth forests provide an interesting system in which to
evaluate how fine-scale disturbances influence the distribution of resources, prey, and their
interaction in the presence of predation risk. Old-growth boreal forests are characterized by
high structural heterogeneity due to fine-scale canopy disturbances such as windthrow,
insect outbreaks, disease, and tree senescence (McCarthy 2001). Because canopy closure
in mature boreal forest generally limits the availability of food resources for browsing
herbivores (Fisher & Wilkinson 2005), the establishment of early successional plants and
the release of advanced regeneration within canopy gaps could create resource-rich patches
within a matrix of low food availability. Gap disturbances also decrease the cover on
which such herbivores rely for protection from predators. Predation risk should influence
how far and intensively herbivores are willing to forage within canopy gaps. Foraging and
movement behaviors of herbivores can reveal how balancing food acquisition and predator
avoidance lead to their spatial distribution in forests structured by gap dynamics.
Our objective was to evaluate how canopy gaps in mature and old-growth boreal forests
influenced the fine-scale distribution of snowshoe hares (Lepus americanus). Snowshoe
hares are a key species of the boreal forest for multiple predators (Boutin et al. 1995).
Hares rely mainly on deciduous browse during winter (Pease et al. 1979), and they are
known to move and forage in proximity to cover as a response to predation risk (Hodges &
Sinclair 2005, Morris 2005). Snowshoe hares should be sensitive to variations in the
interspersion of food and cover created by canopy gaps, but little is known about their
response to fine-scale disturbances (<0.1 ha) that characterize old-growth boreal forest.
60
We first assessed whether browse availability was higher within gaps than under
surrounding forest cover, thereby creating a potential conflict between the search for food
and cover. We then tested whether heterogeneity in food and cover created by gap
dynamics influenced snowshoe hare habitat selection at the stand level, whether the
presence of gaps influenced movement decisions, and whether foraging behavior was
influenced by a relatively high perception of risk within canopy gaps. Perception of risk
was evaluated through giving-up density (GUD) experiments (Brown 1988) and surveys of
natural browse use within canopy gaps. GUD experiments are based on optimal foraging
theory, which predicts that foragers should leave a food patch when foraging gains no
longer exceed the sum of metabolic, missed-opportunity, and predation costs associated
with exploiting the patch (Brown 1988). Everything else being equal, prey should allocate
greater foraging effort to safe than risky patches, and the density of food left in different
patches can reveal their perception of risk (Brown 1988). We used GUD experiments to
test the predictions that, if hares trade-off food for safety, 1) consumption of experimental
food patches should be lower within canopy gaps than under forest cover, 2) foraging effort
should decrease with distance from cover (i.e. from the forest edge toward the center of
gaps), and 3) the probability of using experimental food patches should decline toward the
center of gaps.
Methods
Study area
The study was conducted in the boreal forest of the Côte-Nord region (49o50‘ –
51o30‘ N, 68
o30‘-69
o30‘ W) of Québec, Canada. The study area lies in the eastern black
spruce/moss bioclimatic region and has forest fire cycles between 270 and >500 years
(Bouchard et al. 2008). The region‘s long fire cycles have led to a forest landscape
composed of 70% irregularly structured old-growth stands dominated by black spruce
(Picea mariana) or mixed stands of balsam fir (Abies balsamea) and black spruce (Boucher
et al. 2003). Other common tree species include jack pine (Pinus banksiana), trembling
aspen (Populus tremuloides), white birch (Betula papyrifera), and eastern larch (Larix
laricina). The regional climate is subhumid, and subpolar, with a mean annual temperature
61
of -2.5oC and abundant annual precipitation (1,000-1,300 mm), 35% of which is snow
(Robitaille & Saucier 1998).
Cover and browse availability within canopy gaps and under forest cover
We sampled four gaps from each of 28 sites in spruce and spruce-fir stands during
the summer of 2007 to determine whether browse and availability of lateral cover within
canopy gaps differed from the surrounding forest. We used fire maps created by Bouchard
et al. (2008) to identify stands ranging from 80 years, the age at which canopy gap
formation and transition to irregular stand structure begins (Bouchard et al. 2008), to >200
years. Gaps were classified as being either of primarily edaphic origin or originating from
mortality of canopy trees (sample photographs of edaphic- and mortality-origin canopy
gaps are provided in Appendix 2). At each site we selected one canopy gap in each of four
size classes (50-100 m², 100-200 m², 200-300 m², >300 m²) based on gaps typical of
eastern boreal forests (Pham et al. 2004). We measured the length and width of each gap to
estimate gap area as an ellipse (Runkle 1981). We sampled the first gap encountered of
each size class along a 300-m transect starting and finishing within the stand. Additional
transects were walked if we did not encounter all gap size classes on the first transect. If
we were unable to find gaps >300 m2
(n = 5 sites), we sampled a second gap from either the
100-200 m2 or 200-300 m
2 size class to obtain 4 gaps per site.
Near-ground lateral cover is provided mainly by coniferous saplings, and the
terminal twigs of deciduous saplings and shrubs constitute the main source of browse for
hares during winter (Pease et al. 1979, Litvaitis et al. 1985). To measure cover and browse
availability within gaps we counted the number of coniferous saplings (>50 cm in height
and <9 cm diameter at breast height) and the number of deciduous twigs (terminal shoots
>5 cm long) between 0-2 m above ground level within a 1-m buffer on either side of the
long axis of each gap. Each stem was identified to species and classified according to its
height: 0.5-1 m, 1-2 m, 2-3 m, or >3 m. We also measured the distance of each sapling
(conifer and deciduous) to the gap edge in 1-m intervals. The main deciduous browse
species included white birch, willow (Salix spp.), speckled alder (Alnus incana rugosa),
green alder (Alnus viridis crispa), serviceberry (Amelanchier spp.), and mountain ash
(Sorbus spp.). We did not count the number of black spruce and balsam fir twigs (these 2
62
species represented 99% of conifer stems in our gap regeneration surveys), as these species
are rarely browsed by hares (Newbury & Simon 2005, St-Laurent et al. 2008)
To compare browse and cover availability within gaps to surrounding forests we
extended the gap‘s transect by 5 m into the forest at either end of the gap (n = 57 gaps). In
some cases canopy gaps were too frequent to sample 5 m of intact forest adjacent to each
gap so we either moved one of the 5 m plots to one of the ends of the wide axis (n = 35
gaps), extended the long axis by 10 m in one direction (n =15 gaps), or sampled the next
first 10 m of intact forest following the gap along our gap inventory transect (n =5 gaps).
We used Wilcoxon signed-rank tests to compare browse and cover availability within gaps
and adjacent forests (Lehmann 1998).
Stand level habitat selection
To evaluate how snowshoe hares respond to heterogeneity in the distribution of
browse and cover created by canopy gaps we compared habitat characteristics at points
along single winter snowshoe hare trails to randomly located points within four conifer
stands (>90 years) during March and April of 2007. This information was used to estimate
resource selection functions (RSFs: Boyce et al. 2002, Manly et al. 2002). We focused on
winter habitat use as tracks left in the snow permitted a fine-scale assessment of habitat
selection. Single winter trails represented tracks left in the snow by the passage of a single
hare moving in one direction. Fifty random points were generated within each stand using
ArcView GIS software (version 3.2, ESRI Inc., Redlands, California). Random points were
≥20 m from each other and from the edge of stand boundaries. To obtain a random sample
of snowshoe hare trails we followed a path linking the random points within each stand and
sampled snowshoe hare trails that intersected this random trajectory as we encountered
them. We sampled points at 20-m intervals along each encountered trail, following the
hare‘s direction of travel, up to a maximum of five points per trail. The coordinates of each
observed point were recorded with a GPS to make sure that all sampled trails were at least
20 m apart, as for random points. Sampled trail segments were sufficiently long (80 m) to
occur both within gaps and under canopy cover. We sampled a total of 125 points from 25
single trails (n = 7, 7, 5, and 6 trails within each of the four sampled stands, respectively)
63
and 184 random points before access to sites was limited by road closure for the spring
thaw.
To quantify habitat structure at observed and random points we measured cover and
browse availability within circular plots around each point. Canopy closure was estimated
visually in 10% classes at each point and 5 m away in four opposite directions, and we used
the mean of the five readings in subsequent analyses. We estimated lateral visual
obstruction at each point in 10% classes by observing a 0.5 2 m (width height) cover
board (Nudds 1977) from 5 m away in four opposite directions and used the average of the
4 readings in subsequent analyses. To further quantify cover availability we counted the
number of conifer stems within a 4-m radius circle (50-m² plots), and each stem was
classified into one of two cover classes based on its lateral visual obstruction between 0-1
m from the snow surface. Class 1 stems included bare trunks and trunks with dead lateral
branches (mainly mature stems and snags), whereas class 2 stems included trees with live
green branches, saplings completely covered with snow, and recently fallen trees with
green branches that would completely obstruct vision. Browse availability was measured
as the number of deciduous stems within each plot that had twigs available between 0-1 m
of the snow surface.
RSF models were estimated using mixed-effects logistic regressions, with sites
included as a random effect. A set of candidate models was produced based on
combinations of canopy closure, lateral visual obstruction, conifer stem density by cover
class, and browse availability. Candidate models were compared based on Akaike‘s
Information Criterion (AIC), differences in AIC (Δi ) and Akaike weights (wi) (Burnham &
Anderson 2002). As none of our candidate models had wi >0.90, we used multimodel
inference based on average coefficients, and associated unconditional standard errors and
95% confidence intervals (Burnham & Anderson 2002). Multicollinearity was absent from
candidate RSFs, as variance inflation factors (VIF) were always <2 (Graham 2003).
Evaluation of candidate models with similarly strong empirical support (those with delta
AIC <2.0, Burnham & Anderson 2002) was performed using k-fold cross validation (Boyce
et al. 2002). Models were built by randomly selecting 70% of observed locations as a
training set and withholding 30% of the data for model evaluation (test set). Random
locations were ranked according to RSF scores calculated from the models and were binned
64
into 10 approximately equal sized groups. The number of observed locations from the
evaluation set within each bin was tallied, and we calculated a Spearman-rank correlation
(rs) between the frequency of test-set observed locations within each bin and bin number to
evaluate the predictive success of each model. This process was repeated 100 times for
each model, and the averages ( sr ) are reported. Mixed-effects logistic regressions were
performed with R 2.9.0 software (R Development Core Team 2006) using the lme4
package (Bates & Sarkar 2006), and k-fold cross validation was run using SAS 9.1 (SAS
Institute Inc. 2003).
Fine-scale movements
Snowshoe hares could use two movement tactics to minimize risk associated with
the reduced protective cover characterizing canopy gaps: they could 1) bias movements
away from openings towards greater cover, or 2) increase movement speed to reduce time
spent in openings. To assess whether snowshoe hares adjust their movements to fine-scale
habitat structure we used step-selection functions (Fortin et al. 2005). A step was defined
as a 10-bound segment along single winter snowshoe hare trails based on fresh tracks left in
the snow. Predator tracks following the observed trails were absent, meaning that observed
movements did not reflect responses to active pursuit by predators. Each observed step was
paired with two random segments originating from the same point of departure. Lengths
and turning angles of random steps were drawn from the distributions of observed steps.
An initial sample of observed step lengths and turning angles was necessary before we
could start measuring habitat attributes along observed and associated random steps. Each
new observed step length and turning angle was added to the pooled distribution from
which random steps were drawn. Kolmogorov-Smirnov two-sample tests (Sokal & Rohlf
1995) confirmed that the distribution of observed and random step lengths and turning
angles were similar (step lengths: P = 0.23; turning angles: P = 0.27), thereby reducing
potential risk of bias (Fortin et al. 2005).
Along observed and random steps we made a visual assessment of canopy closure in
10% classes at the start, middle, and end of each step segment. The proportion of the step
that occurred within a canopy gap was estimated in 10% classes. Lateral cover was
65
estimated from the number of coniferous tree stems by cover class (Class 1 or 2, as
described above) within 1 m on either side of the step. Browse availability was estimated
by counting all deciduous twigs by species located <1 m above the snow surface within 1 m
on either side of the step. A total of 105 steps were surveyed along 16 snowshoe hare trails.
Observed and associated random steps were compared using conditional logistic regression
(Fortin et al. 2005). Pairs of observed and random steps were included as individual strata.
To account for nonindependence of multiple steps along a given trail, series of successive
steps were included as individual clusters in the model, and robust variance was calculated
on the basis of independent clusters (Fortin et al. 2005). We used model comparison based
on the quasi-likelihood under independence criterion (QIC: Craiu et al. 2008) to compare
candidate models with different combinations of canopy closure, conifer stem density, and
browse availability. Model averaging was then used to calculate parameter estimates,
unconditional SEs, and 90% and 95% CIs. Conditional logistic regressions were run using
the PHREG procedure in SAS 9.1 (SAS Institute Inc. 2002).
To evaluate whether snowshoe hares responded to variations in cover availability by
changing their speed we used general linear mixed models with the distance traveled in 10
bounds, an index of movement speed, as the dependent variable and combinations of
canopy closure, conifer stem density, and browse availability as independent variables. We
did not include the proportion of segments within gaps as a variable (‗proportion in gap‘) in
candidate models because almost half of the observed trail segments (47 out of 105; 45%)
were completely under canopy cover (i.e., 0% of the trail segment was within a gap). This
variable also did not capture variation in canopy cover that was due to changes in
interstitial spacing between trees (average canopy closure along segments without canopy
gaps varied between 27% and 77%, but average closure along segments with gaps varied
between 3% and 73%). Individual trails nested within sites were considered as random
effects, and we used an autoregressive (order 1) correlation structure to account for
autocorrelation between successive trail segments. We used AIC corrected for small
sample size (AICc) to rank candidate models and multimodel inference to calculate
coefficients for variables with 90% and 95% CIs. To evaluate the accuracy of top ranking
models (ΔAICc <2.0) we calculated marginal R² values for each model (Orelien & Edwards
66
2008). General linear mixed-models were run using the MIXED procedure in SAS 9.1
(SAS Institute Inc. 2002).
Giving-up densities
We selected 88 gaps distributed within 20 different sites (1-11 gaps/site) in spruce
and spruce-fir dominated stands (>90 years old). Gaps were sampled during the winters of
2006 (65 gaps) and 2007 (23 gaps). Selected gaps were free of coniferous regeneration that
could provide cover and of deciduous regeneration that could provide alternative foraging
opportunities. Length and width of gaps were used to estimate gap size as the area of an
ellipse, and sizes ranged from 20 m² to 942 m². Within the gaps, giving-up densities were
measured using jack pine boughs as experimental food patches, consistent with methods
developed by Morris (2005). Jack pine is a preferred browse species for snowshoe hare
(Bergeron & Tardif 1988) and was absent in the understory of stands in which we
conducted gap surveys and GUD experiments. Jack pine boughs thus represented attractive
food patches for hares within these stands. Furthermore, we had access to a 30-year-old
fire-origin stand of regenerating jack pine that gave us a vast source of boughs from trees of
similar age and height, helping to reduce sources of variability in the quality of boughs used
in the experiment. Changes in protein and fiber content are such that the digestibility and
energetic value of boughs should decrease as stems get thicker toward their bases (Palo et
al. 1992). Therefore the rate of energy gain should decrease as hares clip progressively
larger diameter segments. The diameter at point of browse thus provides an estimate of
GUD, with smaller browse diameters indicating higher GUDs (Morris 2005). We cut
terminal jack pine boughs to a length of 50 cm and removed all cones and lateral branches.
The basal diameter of each bough was measured to the nearest 0.02 mm with calipers to
account for variations in branch morphology. Then boughs were inserted 10 cm into the
snow in pairs at 1-m intervals, starting at the center of the gap and extending 4 m into the
adjacent forest along the wide axis, with a pair positioned at the gap edge. We placed
between 2-11 branch pairs within gaps according to gap width. Boughs were left in place
between 4-26 days to allow sufficient time for hares to encounter the gaps and revisit
branches over several nights. At the end of each sampling period we removed boughs and
measured the diameter at point of browse and the residual length of all browsed stems.
67
Motion-sensitive digital cameras (Reconyx Silent Image, La Crosse, Wisconsin) were
installed at some gaps to observe foraging behavior.
Diameter at point of browse was compared between canopy gaps and continuous
forests where foraging had occurred in both the gap and the adjacent forest. To test
whether GUDs differed by habitat (Gap vs. Forest) and increased with distance from the
gap edge within gaps we used a linear mixed-effects model with habitat (Gap = 1, Forest =
0) and a habitat distance interaction as fixed effects. The basal diameter of jack pine
stems (ln-transformed) was included as a covariate to account for variation in branch
morphology. Because the amount of time branches were left in place varied from gap to
gap, the natural log of the number of nights (―no_nights‖) also was included in the model,
both as a simple effect to test whether diameter at point of browse increased with time that
branches were left in place and in a triple interaction with habitat and distance [habitat
distance ln(no_nights)] to test whether branches farther from cover within gaps were
browsed to larger diameters the longer they were left in place. We included sites and gaps
nested within sites as random effects to account for our hierarchical sampling design of
branches grouped within gaps, and gaps grouped within sites. Random site effects also
accounted for potential site level differences in snowshoe hare abundance. We used the
Kenward-Roger method (Kenward & Roger 1997) to calculate denominator degrees of
freedom for the fixed effects because the number of branch pairs within gaps varied
according to gap size, thereby creating an unbalanced design. Linear mixed-effects models
were run using the MIXED procedure in SAS 9.1 (SAS Institute Inc. 2002) and Type III
contrasts were used to test the significance of fixed effects.
All gaps with at least one clipped bough in either the forest or gap were used to test
the probability of bough use in forests versus gaps and, once in gaps, the effect of distance
of branches to the gap edge. To model the probability of branch use (browsed = 1,
nonbrowsed = 0) we used a mixed-model logistic regression with habitat (Gap = 1, Forest =
0) and a habitat distance interaction as fixed effects and sites and gaps nested within sites
as random effects. We also included the natural log of the number of nights branches were
left in place as a simple effect to test whether branches were more likely to be browsed the
longer they were left in place, and in a triple interaction with habitat and distance [habitat
distance log(no_nights)] to determine if branches that were farther from cover within
68
gaps were more likely to be used the longer they were left in place. The mixed-model
logistic regression was run using the GLIMMIX procedure in SAS 9.1 (SAS Institute Inc.
2002).
Use of natural browse within canopy gaps
Signs of browsing by snowshoe hares were recorded during surveys of browse
availability within canopy gaps. We counted the number of twigs browsed by snowshoe
hares during the winter (2007) previous to our survey (summer 2007) to estimate browsing
intensity as a proportion of used versus available twigs. Each stem (including conifers) was
also classified as browsed or nonbrowsed based on the presence of any twigs clipped by
snowshoe hares. As hares mainly consume woody browse during winter, browse surveys
reflected patterns of winter habitat use. Based on areas where deciduous stems were
present in both the gap and adjacent forest, we modeled the probability of stem use as a
function of habitat (Gap vs. Forest) and, once in gaps, the distance of stems to the gap edge.
We used a mixed-model logistic regression with habitat (Gap = 1, Forest = 0) and a habitat
distance interaction as fixed effects, and sites and gaps nested within sites as random
effects. The Kenward-Roger degrees of freedom correction was applied to account for
spatial variations in numbers of stems at different distances from the gap edge. As conifer
regeneration within gaps may provide cover for hares, we tested a second model that also
included the density of conifer regeneration within gaps. This model included a habitat
conifer sapling density interaction and the three-way interaction habitat conifer density
distance of browse stems to the gap edge. Mixed-model logistic regressions were run using
the GLIMMIX procedure in SAS 9.1 (SAS Institute Inc. 2002).
Results
Browse within canopy gaps
Of the 112 canopy gaps sampled 99 had browse available within the gap, including
61 gaps with browse found in both the gap and the adjacent forest. Gaps originated more
frequently from the mortality of canopy trees (n = 71; 63%) than from edaphic conditions
(n = 41; 37%). The density of deciduous browse was greater within gaps of both edaphic
69
and mortality origin than under adjacent forest cover (Table 2.1). The density of coniferous
saplings was lower within edaphic origin gaps than adjacent forest, whereas no difference
was detected between mortality origin gaps and adjacent forest.
Winter habitat selection at the stand level
Among the competing models explaining snowshoe hare selection for winter trail
location, three RSFs received similarly strong empirical support (ΔAIC <2; Table 2.2). K-
fold cross-validation indicated that all three models had good predictive success, with
sr ranging between 0.86 and 0.91. Model averaging of parameter estimates revealed that
canopy closure and browse availability had the strongest influence on selection for winter
trail locations, as these two habitat attributes were the only ones with 95 % CIs that
excluded zero (Table 2.3). Snowshoe hares selected areas with greater canopy closure
( Canopy closure = 0.064, 95% CI = 0.043 - 0.085) and browse availability ( Browse availability =
0.085, 95% CI = 0.002 - 0.169) compared to random locations within stands (Table 2.4).
Fine scale movements
Model comparison of step-selection functions did not provide overwhelming
support for a particular model (ΔQIC <2 for five models, Table 2.5). Model averaging of
the parameter estimates revealed that the proportion of steps made within canopy gaps was
lower than expected by chance alone ( Proportion in gap = -0.005, 95% CI = -0.009 - -0.001;
Table 2.3). Unconditional 90% confidence intervals also indicated that hares tended to
move selectively in areas with higher canopy closure ( Canopy closure = 0.022, 90% CI =
0.000 - 0004). However, little evidence was found that hares selectively moved along areas
with higher conifer stem density or greater browse availability (Tables 2.3, 2.6).
The distance traveled by hares in 10 bounds, an index of movement speed, varied
from 3.4 m to 16.9 m. Several competing models received similarly high support, with
ΔAICc <2 (Table 2.7). Model averaging (Table 2.3) revealed that snowshoe hares reduced
their speed in areas with greater canopy closure ( Canopy closure = -0.044, 95% CI = -0.081 - -
70
0.008) and greater densities of Class 1 conifer stems ( Class 1 conifer stem density = -1.545, 95%
CI = -2.618 - -0.472). Hares also tended to reduce speed in areas with greater densities of
Class 2 conifer stems ( Class 2 conifer stem density = -1.161, 90% CI = -2.217 - -0.105). Although
these habitat features explained a statistically significant portion of the variation in the
distance hares covered in ten bounds, this portion remained rather low for all candidate
models (R2 < 0.15 for all regressions used in model averaging).
Giving-up densities
Snowshoe hares visited (i.e., ≥ one branch clipped) 45 of the 88 canopy gaps used
for GUD experiments. Visited gaps were 4-16 m in width (i.e., between 2 and 8 branch
pairs placed within the gap) and 22-440 m² in area. Of those, 36 gaps had branches clipped
by hares in both the gap and the adjacent forest. The diameter at which hares clipped
boughs within gaps did not vary as a function of distance from cover (Habitat Distance;
F1,598 = 1.02, P = 0.31) or as a function of distance to cover and time (Habitat Distance
ln(no_nights); F1,598 = 0.42, P = 0.52). Inferences were thus based on a model investigating
whether the diameter at point of browse varied between gaps and the adjacent forest
(variable: Habitat) while controlling for basal stem diameter and time; i.e., Diameter at
point of browse = Habitat + ln (Basal stem diameter) + ln(no_nights), where habitat was a
class variable. Variations in branch morphology had a strong influence on diameter at
point of browse (βln basal diameter = 2.50; F1,609 = 74.54, P <0.0001), and diameter at point of
browse also increased with the time that boughs were left in place (βln no_nights= 0.82; F1,41.1
= 11.56, P = 0.002). This model further revealed that hares clipped boughs to larger
diameters under forest cover than within gaps (Habitat: F1,587 = 12.67, P = 0.0004, n = 36
gaps). Based on the least squared means of the mixed model (based on a mean basal
branch diameter of 8.08 mm and a mean time of 13 nights), hares clipped boughs to a mean
diameter of 5.08 mm ± 0.24 under forest cover while those within gaps were clipped to
4.84 mm ± 0.24.
We also found that hares were less likely to clip experimental branches within gaps
as the distance from the forest edge increased (βHabitat Distance = -1.11; F1,841 = 9.38, P =
0.002, n = 45 gaps with ≥ one branch clipped). However, boughs located farther within
71
gaps were more likely to be browsed the longer they were left in place (βHabitat Distance
ln(no_nights) = 0.33; F1,841 = 6.08, P = 0.014; Figure 2.1). To determine whether a threshold
distance could be identified where the probability of branch use became significantly lower
within gaps than in adjacent forests, we used a mixed-effects model with distance as a class
variable (forest = 0, gap = 1-6+ m; distances 6-8 m were pooled due to low number of
replicates) and time as a covariate and compared the probability of use at each distance
with that of the forest. At distances of ≥4 m, the probability of branch use was
systematically lower within gaps than under forest cover (P < 0.05 for all cases).
Natural browse use
Similar proportions of deciduous stems had signs of browsing by snowshoe hares
(current or previous years) within gaps (42%, n = 1337 stems) and forest adjacent to gaps
(37%, n = 251 stems). We did not observe any signs of browsing by snowshoe hare on
coniferous saplings within either gaps (n = 1233 stems) or under adjacent forest cover (n =
634 stems). The proportion of available terminal twigs that were browsed during the last
winter season (2007) was low in both habitats (Forest: 1.8%, Gap: 2.3%). Consistent with
GUD experiments, we found a decreasing probability of use by hares of natural browse
stems located farther within gaps (Habitat Distance: F1,1257 = 7.98, P = 0.005, n = 61
gaps; Figure 2.2). Using distance as a class variable, we also found that browsing in gaps
was significantly less likely than under adjacent forests at distances ≥ 7 m from cover
within gaps (P < 0.005). Including the density of conifer regeneration within gaps in
logistic regressions did not change the probability of browse stem use within gaps, as
neither the 3-way interaction of Habitat Conifer Density Distance nor the 2-way
interaction Habitat Conifer Density were significant (P > 0.40). However, when we
included only the density of conifer regeneration > 2 m in height, we found that it had a
positive effect on the probability that deciduous stems within gaps would be browsed
(βHabitat Conifer sapling density >2m height = 2.609; F1,130.8 = 4.63, P = 0.03), but it did not change the
pattern that stems at greater distances from the forest edge within gaps remained less likely
to be browsed (βHabitat Distance = -0.1381; F1,1265 = 7.90, P = 0.005, after removing the
72
nonsignificant 3-way interaction of Habitat Distance Conifer sapling density >2-m
height).
Discussion
Movement and foraging behaviors revealed that fine-scale disturbances in old-
growth boreal forest shape space-use patterns of snowshoe hares by creating heterogeneity
in their landscapes of fear and food. Canopy gaps created areas of higher browse density
compared to closed-canopy conditions, but hares perceived these openings as relatively
risky. Hares responded to spatial variation in food and safety by selecting areas within
stands that had both higher canopy closure and higher browse availability than random
locations. Furthermore, hares adjusted their movements and foraging behavior to minimize
time spent in openings. To our knowledge this is the first study linking snowshoe hare
distribution to habitat heterogeneity induced by fine-scale canopy gap dynamics. The
process of gap formation, regeneration, and closure should create a shifting mosaic of food
and cover for hares, which in turn should shape their interactions with predators in old-
growth boreal forests.
Gap dynamics induced by fine-scale disturbances create a ―foodscape‖ (Searle et al.
2007) for snowshoe hares that is constantly changing over time and space. We observed
that hares acquire most of their winter food within canopy gaps. Although hares harvested
similar proportions of twigs available within gaps and under forest cover, they consumed
considerably more twigs from gaps because these openings offered nearly four times more
browse (mean = 4.5 twigs/m² in all gaps; mean = 1.15 twigs/m² under canopy cover).
Therefore gap dynamics should increase browse supply for hares as forest stands undergo a
transition from mature to old-growth structure. The spatial and temporal distribution of
food resources for hares in these stands should depend largely on the rate of gap formation
and gap closure. New gaps in old-growth boreal forests form at a rate of approximately 1%
of stand area per year (McCarthy 2001), and these gaps can take between 50-200 years to
close (Lertzman & Krebs 1991). Gaps thus accumulate and expand faster than they close,
such that the gap fraction within old-growth stands increases with time (Harper et al. 2006)
until the next major stand-replacing disturbance occurs. The process of gap closure also
73
appears to depend on the origin of canopy gaps. Compared to gaps originating from tree
mortality, edaphic gaps were characterized by little to no coniferous regeneration. These
gaps likely persist from the previous stand-initiating disturbance and should remain open
for long time periods because poor germination beds and competition by shrubs do not
generally facilitate tree establishment (Mallik 2003, Harper et al. 2006). Consequently, the
spatial and temporal distribution of gaps within old-growth stands should remain fairly
constant when edaphic gaps predominate, whereas stands dominated by gaps from tree
mortality should have a spatial distribution of food that varies dynamically over shorter
time scales. These processes also determine the snowshoe hare‘s landscape of fear (sensu
Laundré et al. 2001).
Prey need to balance resource acquisition with safety to realize their potential fitness
(Brown & Kotler 2004). When prey are more vulnerable to predation in areas of reduced
vegetation cover they may structure their movements to reduce time spent in openings. For
example, in the presence of predators degus (Octodon degus) select travel routes that follow
the distribution of shrub cover and increase their speed when crossing openings. In the
absence of predators, however, they increase their use of open habitats (Lagos et al. 1995,
Vasquez et al. 2002). Hares appear to be more vulnerable to predation in open habitats
(Rohner & Krebs 1996), and we found that hares selected areas within mature and old-
growth stands that had higher than average canopy closure. They also made fine-scale
adjustments to reduce the proportion of their trajectory that occurred within gaps and sped
up in areas of reduced canopy closure. These behavioral adjustments suggest that
snowshoe hares spend most of their time under closed canopy cover and that the use of
gaps is largely restricted to foraging activities. Their fear of predators also appears to
constrain their foraging behavior in gaps.
In the presence of predators prey may forego foraging in resource-rich habitats in
return for greater safety (Wirsing et al. 2007). Numerous studies, where patches of
vegetation cover are embedded in an open matrix, have demonstrated that small mammals
accept reduced rates of energy intake for the greater safety of exploiting food patches under
cover (for review see Brown & Kotler 2004). In our system canopy gaps represented open
patches embedded in a matrix of vegetative cover. We observed that snowshoe hares
clipped experimental jack pine boughs to larger diameters (lower GUDs) under forest cover
74
than within gaps, presumably accepting a lower rate of energy intake by foraging more
intensively under the safety of canopy cover. Although prey often display higher GUDs
(i.e., lower foraging efforts) as distance from cover increases (Hughes & Ward 1993,
Hochman & Kotler 2007), snowshoe hares did not appear to diminish their foraging effort
toward the center of gaps. These findings are consistent with Hodges and Sinclair (2005)
but contrary to Morris (2005), who observed that hares clipped jack pine boughs to smaller
diameters at greater distances from cover along sharp ecotones between shrub habitat and
abandoned agricultural fields. The lack of change in browse diameter with distance from
cover in gaps could be the result of weak diminishing returns for hares browsing jack pine
boughs. If hares experienced a relatively flat harvest rate curve while consuming boughs,
meaning little decrease in the rate of energy gain with increasing diameter, this would have
limited our capacity to detect fine-scale variation in perception of risk. Accordingly,
information on protein and fiber content of jack pine boughs at increasing stem diameters
would be necessary to quantify harvest rate curves, which in turn would facilitate the
interpretation of GUD experiments on hare foraging behavior. Our results also could be
explained by a foraging tactic displayed by snowshoe hare. When foraging away from
cover, prey must balance exposure time against foraging efficiency, and they often choose
to carry items back to protective cover rather than consume them in the open (Lima 1985,
Hughes & Ward 1993). Our motion-sensitive cameras revealed that hares can clip large
segments of branches and carry them back to the forest cover (Figure 2.3). In such cases
hares would have been consuming boughs in the same place with the same risk, regardless
of where the bough was initially placed. The diameter at point of browse would then no
longer reflect time spent in the open harvesting a series of successively larger diameter
segments of diminishing energetic value.
Although distance to cover might not influence the diameter at point of browse
when harvesting a branch, herbivores may remain reluctant to venture far from cover to
browse. Foragers should accept greater risk only for greater rewards (Kotler & Blaustein
1995). When presented with similar food patches, foragers should select the safest patches
first. Consistently, we found that hares were less likely to use experimental food patches as
their distance from the safety of canopy cover increased. Moreover, the probability of
natural browse use also declined as stems were located farther within gaps. Overall, hares
75
were significantly less likely to use natural browse stems that were >7 m from cover (i.e.,
near the center of gaps of >14 m in diameter). The landscape of fear is therefore shaped by
variations in the size of canopy openings. Although gap formation can improve habitat
quality for hares by increasing food availability, browse in the center of large gaps
essentially could be unavailable to hares. Although most gaps in old-growth boreal stands
are <100 m² in area (<12 m diameter), gaps may cover >80 % of stands (Pham et al. 2004).
The accumulation and expansion of many small gaps therefore could have important stand-
level implications for habitat quality as the matrix of continuous canopy cover offering safe
travel corridors becomes increasingly fragmented. Hares were more likely to use browse
within gaps with greater densities of coniferous regeneration tall enough (>2 m) to provide
cover above the snow during winter. Succession within gaps should contribute to
spatiotemporal heterogeneity in the distribution of risk for hares.
Trade-offs between food and safety also can vary according to population density
(China et al. 2008). Snowshoe hares display cyclical population dynamics (Krebs et al.
2001a), with up to 182-fold changes in density in some regions (Krebs et al. 1986). Wolff
(1980) observed that snowshoe hares increased their use of open food-rich habitats and
clipped deciduous twigs to larger diameters (>1 cm) towards the peak phase of their cycle.
The patterns of browse use observed in our study could vary according to the phase of the
snowshoe hare cycle. Furbearer harvest data suggest that snowshoe hare populations are
cyclical in our study region (Godbout 1998, Bourbonnais 1999), but cycles are of much
lower amplitude (9-10 fold changes in density) than those reported in the West (17-182 fold
changes in density; Keith & Windberg 1978, Krebs et al. 1986). Previous population peaks
in the study region occurred in 1980-81 and 1988-89 (Bourbonnais 1999), and St-Laurent et
al. (2008) reported that hare were at their peak in 1998-99 in an adjacent region. Our study
should have occurred during the peak phase of the cycle, assuming an 8-9 year periodicity.
Pellet count data from 18 stands >80 years old, each sampled over three consecutive years
(2006-2008), seem to confirm this. We recorded mean pellet densities of 0.31 pellets/m² in
2006, 0.50 pellets/m² in 2007, and 0.39 pellets/m² in 2008, suggesting that the peak
occurred in 2007 (J. Hodson, pers. obs.). These pellet densities would correspond to hare
densities of roughly 0.03-0.05 hares/ha based on regression equations developed by Krebs
et al. (2001b). These estimates are lower than most hare densities recorded during the low
76
phase of population cycles in other regions [range = 0.03 – 1.70 hare/ha, mean = 0.62
hare/ha—Murray (2003)]. The low proportion of terminal twigs consumed by hares (1.8-
2.3%) suggests that they were not faced with a food shortage, whereas other hare
populations can consume 80-100% of available browse during population peaks (Wolff
1980, Smith et al. 1988). This suggests that fine-scale spatial patterns of browse use might
not change considerably over the course of low-amplitude population cycles in eastern old-
growth boreal forests.
Multi-trophic implications of habitat heterogeneity resulting from gap dynamics
Despite increasing emphasis on the maintenance of old-growth stands in managed
boreal landscapes (Mosseler et al. 2003), we still understand little about how gap dynamics
in these forests influence the fine-scale distribution of boreal wildlife. Our study indicates
that gap dynamics could have multitrophic level consequences by creating spatial
heterogeneity in the landscapes of fear and food for snowshoe hare, a key prey species of
boreal ecosystems. Nonlethal effects of predators on their prey can have major
repercussions on ecosystems. For example, the evasive games played between herbivores
and their predators can have cascading effects on vegetation growth triggered by spatial
variations in browsing intensity (Schmitz et al. 1997, Beyer et al. 2007). Traditionally,
models of vegetation succession following disturbance have not considered the roles of
herbivores (Wisdom et al. 2006), but studies suggest that forest herbivores can shape
competitive vegetation interactions by preferentially browsing certain tree species (Schmitz
2005). These interactions may be further modified by spatial variation in predation risk.
For example, moose (Alces alces) preferentially browse deciduous vegetation that competes
with regenerating conifers in clearcuts, but their use of browse declines from the forest
edge towards the center of cuts because of increased predation risk (Schmitz 2005). Spatial
heterogeneity in risk-sensitive foraging by hares similarly could influence patterns of
vegetation succession within canopy gaps. Furthermore, fine-scale disturbances such as
canopy gap dynamics may shape predator-prey ―shell games‖ by determining where food
occurs for prey who must balance patch use with remaining elusive to predators, and by
shaping the movement of predators that may focus their search for prey in areas where their
prey‘s resources are most concentrated (Mitchell & Lima 2002, Andruskiw et al. 2008).
77
Gap dynamics therefore may be a fundamental process structuring predator-prey
interactions in old-growth boreal forests, with cascading implications across several trophic
levels.
Acknowledgements
This work was supported by the NSERC-Laval University Industrial Research Chair in
Silviculture and Wildlife and its partners. We also would like to acknowledge funding
provided by the FQRNT and FCI. We gratefully acknowledge the many field assistants
whose dedicated efforts made this work possible: K. Poitras, J. Leclair, O. Deshaies, J.S.
Roy, E. Renaud-Roy, and V. Hébert-Gentille. We also thank our industrial partners
Abitibi-Bowater, Kruger, and Arbec forest industries for their financial and technical
support. Finally we acknowledge A. Desrochers and M. Mazerolle for valuable statistical
advice and D. Morris for guidance on GUD experiments.
78
Table 2.1. Mean (± 1 SE) deciduous browse density and conifer sapling density within
canopy gaps of edaphic and mortality origin in eastern Canadian boreal conifer stands (80
to 200+ years), and Wilcoxon signed-rank tests (S) of paired differences between browse
and conifer density between gaps and adjacent forest cover. Positive differences indicate a
higher browse or conifer sapling density within canopy gaps than adjacent forest.
Gap
Origin
Browse and
Cover Gap Forest (Gap-Forest) S P
Edaphic
(n = 41)
Deciduous browse
(twigs/m²) 4.93 ± 0.98 1.17 ± 0.29 3.76 ± 0.95 298 <0.001
Conifer saplings
(stems/m²) 0.30 ± 0.05 0.82 ± 0.09 -0.52 ± 0.11 -319 <0.001
Mortality
(n = 71)
Deciduous browse
(twigs/m²) 4.09 ± 0.72 1.14 ± 0.35 2.95 ± 0.71 736 <0.001
Conifer saplings
(stems/m²) 0.75 ± 0.06 0.70 ± 0.06 0.05 ± 0.06 157 0.36
79
Table 2.2. Competing models of resource selection by snowshoe hares using logistic
regression to compare points observed (n = 125) along winter snowshoe hare trails to
randomly located points (n = 184) within eastern Canadian boreal conifer stands (>90
years).
Model K AIC Δ AIC wi
Canopy closure + Browse availability 3 364.8 0.0 0.42
Canopy closure + Lateral Visual Obstruction 0-2m
+ Browse availability 4 366.1 1.3 0.22
Canopy closure + Class 1 conifer stem density +
Class 2 conifer stem density + Browse availability 5 366.8 2.0 0.15
Canopy closure 2 367.3 2.5 0.12
Canopy closure + Lateral Visual Obstruction 0-2m 3 368.6 3.8 0.06
Canopy closure + Class 1 conifer stem density +
Class 2 conifer stem density 4 369.9 5.1 0.03
Class 1 conifer stem density + Class 2 conifer stem
density + Browse availability 4 390.5 25.7 0.00
Class 1 conifer stem density + Class 2 conifer stem
density 3 397.7 32.9 0.00
Browse availability 2 414.1 49.3 0.00
Lateral Visual Obstruction 0-2m + Browse
availability 3 416.0 51.2 0.00
Lateral Visual Obstruction 0-2m 2 422.9 58.1 0.00
80
Table 2.3. Model-averaged coefficients ( ) and unconditional standard errors (SE( )) for habitat variables used in resource selection
functions comparing points observed (n = 125) along winter snowshoe hare trails to randomly located points (n = 184), step-selection
functions for winter snowshoe hare trails (n = 105 observed step segments), and analysis of movement speed by snowshoe hares along
10 bound segments of winter trails in eastern Canadian boreal conifer stands (>90 years). Coefficients are in bold when their 95% (†)
or 90% confidence intervals excluded zero.
Resource selection functions
(RSF)
Step-selection functions
(SSF)
Movement speed
Variable ± SE( ) ± SE( ) ± SE( )
Canopy closure (%) 0.064† ± 0.011 0.022 ± 0.013 -0.044† ± 0.019
Proportion in gap (%) N/A -0.005† ± 0.003 N/A
Browse availabilitya
0.085† ± 0.043 -0.028 ± 0.059 0.110 ± 0.154
Class 1 conifer stem densityb 0.015 ± 0.012 0.078 ± 0.305 -1.545† ± 0.547
Class 2 conifer stem densityb -0.003 ± 0.028 -0.476 ± 0.780 -1.161 ± 0.642
Lateral visual obstruction 0-2 m (%) -0.008 ± 0.010 N/A N/A
a Measured as the density of deciduous stems per 50 m² for RSFs and as the density of deciduous twigs between 0-1 m above the snow
per m² for SSFs and movement speed
b Measured as the number of conifer stems per 50 m² for RSFs and as the number of stems per m² for SSFs and movement speed
81
Table 2.4. Mean (± 1 SE) values of habitat variables measured at points along single winter
snowshoe hare trails (n = 125) and randomly located points (n =184) used in resource
selection functions within eastern Canadian boreal conifer stands (>90 years).
Variable Observed Random
Canopy closure (%) 54.10 ± 1.09 41.76 ± 1.07
Lateral visual obstruction 0-2m (%) 18.41 ± 1.10 18.98 ± 0.97
Class 1 conifer stem density (stems/50 m²) 22.71 ± 1.35 15.40 ± 0.82
Class 2 conifer stem density (stems/50 m²) 4.14 ± 0.43 5.48 ± 0.39
Browse availability (stems/50 m²) 2.20 ± 0.38 1.08 ± 0.17
82
Table 2.5. Competing models for step-selection functions along single winter snowshoe
hare trails (n = 105 observed step segments) in eastern Canadian boreal conifer stands (>90
years).
Models K QIC Δ QIC wi
Canopy closure 1 230.1 0.0 0.21
Proportion in gap 1 230.3 0.2 0.19
Canopy closure + Browse availability 2 230.7 0.6 0.15
Proportion in gap + Browse availability 2 230.9 0.8 0.14
Browse availability 1 231.2 1.1 0.12
Canopy closure + Class 1 conifer stem density + Class
2 conifer stem density 3 233.4 3.3 0.04
Proportion in gap + Class 1 conifer stem density +
Class 2 conifer stem density 3 233.7 3.6 0.03
Class 1 conifer stem density + Class 2 conifer stem
density 2 233.9 3.8 0.03
Canopy closure + Class 1 conifer stem density + Class
2 conifer stem density + Browse availability 4 234.0 3.9 0.03
Class 1 conifer stem density + Class 2 conifer stem
density + Browse availability 3 234.3 4.2 0.03
Proportion in gap + Class 1 conifer stem density +
Class 2 conifer stem density + Browse availability 4 234.3 4.2 0.03
83
Table 2.6. Mean (± 1 SE) values of habitat variables measured along 10-bound segments (n
= 105) and paired random segments from 16 single winter snowshoe hare trails, and mean
paired differences between values along observed and random segments used in step-
selection functions within eastern Canadian boreal conifer stands (>90 years).
Variables Observed Random Paired
Difference
Proportion in gap (%) 28.76 ± 3.20 31.24 ± 2.41 -2.48 ± 2.41
Canopy closure (%) 47.62± 1.63 46.41 ± 1.17 1.21 ± 0.76
Class 1 conifer stem density (stems/m²) 0.63 ± 0.05 0.62 ± 0.03 0.01 ± 0.03
Class 2 conifer stem density (stems/m²) 0.36 ± 0.03 0.38 ± 0.03 -0.02 ± 0.02
Browse availability (twigs/m²) 0.26 ± 0.13 0.31 ± 0.11 -0.05 ± 0.05
84
Table 2.7. Competing models of the influence of cover availability on movement speed,
estimated as the distance travelled in 10-bound segments (n = 105), along single winter
snowshoe hare trails (n = 16) in eastern Canadian boreal conifer stands (>90 years old).
Model K AICc Δ AICc wi
Canopy closure + Class 1 conifer stem density +
Class 2 conifer stem density 5 428.8 0.0 0.33
Class 1 conifer stem density + Class 2 conifer
stem density 3 428.8 0.0 0.33
Canopy closure + Class 1 conifer stem density +
Class 2 conifer stem density + Browse availability 6 430.1 1.3 0.17
Class 1 conifer stem density + Class 2 conifer
stem density + Browse availability 5 430.3 1.5 0.16
Canopy closure 3 441.1 12.3 0.00
Canopy closure + Browse availability 4 441.8 13.0 0.00
Intercept-only 2 450.4 21.6 0.00
Browse availability 3 450.7 21.9 0.00
85
Figure 2.1 Predicted probability of jack pine bough use by snowshoe hares as a function of
habitat (Gap vs. Forest), the number of nights boughs were left within gaps and adjacent
forest, and the distance of boughs (n = 846 boughs) placed within canopy gaps (n = 45
gaps) to the gap edge, in eastern Canadian boreal conifer stands (>90 years).
86
Figure 2.2 Predicted probability (±1 SE) of natural browse use by snowshoe hares as a
function of habitat (Gap vs. Forest) and distance of stems (n = 1269 stems) to the gap edge,
within edaphic and mortality origin canopy gaps (n = 61 gaps) in eastern Canadian boreal
conifer stands (80 to >200 years).
87
Figure 2.3 Snowshoe hare foraging behavior captured from motion sensitive cameras
installed at canopy gaps with GUD experiments in eastern Canadian boreal conifer stands
(>90 years). Photographs show a hare clipping a large jack pine bough segment (indicated
by arrows) in the gap and returning with it to forest cover.
88
Chapitre 3
An appraisal of the fitness consequences of forest
disturbance for wildlife using habitat selection theory
James Hodson*, Daniel Fortin
*, Mélanie-Louise Le Blanc
* and Louis Bélanger
†
*NSERC-Université Laval industrial research chair in silviculture and wildlife,
Département de Biologie, Université Laval, Québec, Québec, Canada, G1V 0A6
†Département des sciences du bois et de la forêt, Université Laval, Québec, QC, Canada,
G1V 0A6
Article publié dans Oecologia 164(1): 73-86.
Résumé
La théorie des isodars peut révéler les conséquences des perturbations de l‘habitat sur
l'aptitude phénotypique d‘espèces fauniques grâce à l‘évaluation des changements dans la
sélection de l‘habitat en fonction de l‘abondance des conspécifiques. Nous avons démontré
comment il est possible d‘intégrer des mesures d‘intensité de perturbation de l‘habitat ou de
variations de disponibilité des ressources dans les fonctions d'aptitude phénotypique-densité
pour établir la forme fonctionnelle des isodars attendus selon différentes relations entre
perturbations et aptitude phénotypique. À partir de ce cadre conceptuel, nous avons étudié
les influences de coupes forestières d‘intensités variées et de la disponibilité des ressources
sur la qualité de l‘habitat du lièvre d‘Amérique (Lepus americanus) et du campagnol à dos
roux (Myodes gapperi). Les isodars de ces deux espèces avaient des ordonnées à l'origine
positives indiquant que l'aptitude phénotypique maximale potentielle était inférieure dans
les peuplements coupés que dans ceux laissés intacts. La sélection de l‘habitat par le lièvre
était influencée à la fois par sa densité locale et par les différences dans la fermeture de la
canopée entre les peuplements coupés et non coupés. Nos modèles prédisent que l‘aptitude
phénotypique du lièvre devrait diminuer avec l‘augmentation de la densité d‘individus dans
tous les milieux. Cependant, le taux relatif auquel l'aptitude phénotypique diminuait dans
chaque habitat dépendait de l'intensité de la perturbation. Dans le cas des traitements ayant
préservé >50% des arbres, les courbes d'aptitude phénotypique-densité estimées pour les
coupes convergeaient avec celle de la forêt non coupée, tandis qu'une divergence des
courbes était prédite dans le cas des traitements ayant préservé <20% des arbres. La
sélection pour les forêts non coupées devenait donc moins prononcée avec l‘augmentation
de la taille de la population lorsque l'intensité de perturbation était faible. Les campagnols
étaient influencés par les différences du couvert de mousses entre les habitats, ce qui
reflèterait l‘importance de l‘humidité près du sol pour l‘espèce. Un couvert de mousses
plus faible dans les peuplements coupés réduirait l'aptitude phénotypique potentielle
maximale pouvant y être atteinte et augmenterait la vitesse du déclin de l'aptitude
phénotypique avec l‘accroissement de la densité de campagnols. Nos modèles prédisent
une réduction des différences de densités de campagnols entre les peuplements coupés et
90
non coupés à mesure que la taille des populations augmente. Cette étude démontre
l‘importance de prendre en considération les variations comportementales associées aux
changements de densité des populations lors de l‘évaluation de l‘impact des perturbations
d'habitat sur la répartition des animaux.
91
Abstract
Isodar theory can help to unveil the fitness consequences of habitat disturbance for wildlife
through an evaluation of adaptive habitat selection using patterns of animal abundance in
adjacent habitats. By incorporating measures of disturbance intensity or variations in
resource availability into fitness-density functions, we can evaluate the functional form of
isodars expected under different disturbance-fitness relationships. Using this framework,
we investigated how a gradient of forest harvesting disturbance and differences in resource
availability influenced habitat quality for snowshoe hare (Lepus americanus) and red-
backed voles (Myodes gapperi) using pairs of logged and uncut boreal forest. Isodars for
both species had positive intercepts, indicating reductions to maximum potential fitness in
logged stands. Habitat selection by hare depended on conspecific density and differences
in canopy cover between harvested and uncut stands. Fitness-density curves for hare in
logged stands were predicted to shift from diverging to converging with those in uncut
forest across a gradient of high to low disturbance intensity. Selection for uncut forests
thus became less pronounced with increasing population size at low levels of logging
disturbance. Voles responded to differences in moss cover between habitats which
reflected moisture availability. Lower moss cover in harvested stands either reduced
maximum potential fitness or increased the relative rate at which fitness declined with
increasing population density. Differences in vole densities between harvested and uncut
stands were predicted, however, to diminish as populations increased. Our findings
underscore the importance of accounting for density-dependent behaviors when evaluating
how changing habitat conditions influence animal distribution.
92
Introduction
Natural and anthropogenic disturbances varying in intensity, frequency, and spatial
extent play a fundamental role in the structure and function of ecosystems. To successfully
manage biodiversity we need to be able to predict how different types of disturbance will
influence the distribution of animal populations in space and time. The impacts of
disturbance on wildlife should be evaluated by measuring the fitness consequences of
habitat alteration on animals (Kight & Swaddle 2007). Individual fitness, however, can be
rather difficult to measure in the field (Morris 1987). Fortunately, ecological theories can
provide an assessment of the effects of disturbance on fitness through an evaluation of
adaptive animal behaviors (Gill et al. 1996a, Morris 2003b, Brown & Kotler 2004).
Habitat selection is a density and frequency dependent process such that evolutionary
fitness should decline with increasing population density and fitness depends on habitat
choices of conspecifics (Morris 2006). Individuals maximizing their fitness should initially
congregate in the best habitat, but as density increases they should begin to occupy lower-
quality habitats (Fretwell & Lucas 1970, Morris 1988). These principles constitute the
foundation of ―the ideal free distribution (IFD)‖ (Fretwell & Lucas 1970), which proposes
that population distribution among habitats should be such that no individual can improve
its fitness by moving to another habitat. This simple theory has received empirical support
(e.g. Haugen et al. 2006), and has provided the basis for isodar theory (Morris 1988).
Assuming that animals can move freely between pairs of adjacent habitats to equalize
mean fitness, differences in habitat quality can be inferred by plotting the density of
individuals in each habitat over a range of population sizes (Morris 1988). The isodar is the
curve, plotted in density space (N2 vs. N1), where fitness is equal in each habitat, but along
which fitness varies (Morris 1988, 2003b). The intercept of the isodar represents the
difference in maximum potential fitness that can be attained in each habitat at low density,
and represents ―quantitative‖ differences between the two habitats usually associated with
resource availability. The slope of the isodar indicates the relative rate at which fitness
declines with increasing density in each habitat (or the ratio of the slopes of the fitness-
density functions for each habitat) and reflects ―qualitative‖ differences (Morris 1988).
Qualitative differences may be due to disparities in habitat structure or resource quality that
93
affect the efficiency with which individuals convert resources into descendants (Morris
1990). Isodars have been successfully used to evaluate how spatial and temporal variation
in food and cover influences habitat selection and population distribution (e.g. Pusenius &
Schmidt 2002, Ramp & Coulson 2002, Shenbrot 2004). Isodars should thus provide a
powerful approach to evaluate the relative impacts of disturbance on habitat quality (Morris
1990, 2003a).
Usually, isodar analysis is based simply on the density of individuals in replicated pairs
of contrasting habitats (e.g. Morris 1992, 1996), but additional insights can be gained by
adding terms into the isodar that directly reflect habitat quality (Morris & Kingston 2002).
In this study, we develop a framework based on isodar theory to assess the impact of
changing habitat conditions on animal populations. This framework is relevant to any
change in wildlife habitat, but here we illustrate the approach by considering the impact of
different levels of forest harvesting disturbance.
Incorporating continuous habitat variables into isodar models: an example with forest
disturbance
Disturbance can modify habitat quality by altering both resource availability
(quantitative effects) and habitat structure (qualitative effects) (Morris 1990, 2003a). The
magnitude of quantitative and qualitative changes to disturbed habitats is likely to depend
on the intensity of habitat alteration. By measuring relative changes to habitat structure or
resource availability and directly incorporating these measures into fitness density
equations and, subsequently, into isodars we should be able to assess how fitness changes
over a gradient of disturbance intensity. Shenbrot and Krasnov (2000) proposed a
"paraisodar" approach to evaluate habitat selection along continuous environmental
gradients. Paraisodars compare population densities at two points in time (high vs. low
density) at several sites along an environmental gradient, and the intercept and slope of the
paraisodar reveals whether changes in population distribution reflect qualitative and/or
quantitative differences between habitats. The approach that we propose here is based on
disturbances that create discrete habitat patches and uses patterns of animal density in pairs
of adjacent disturbed and undisturbed habitats to reveal how density-dependent habitat
94
selection may change along a gradient of disturbance intensity. Our approach differs from
paraisodars in that we directly incorporate measures of habitat contrast between disturbed
and undisturbed habitats into the isodar regressions.
In managed forest landscapes, silvicultural treatments vary according to the volume
and distribution of trees removed from a forest stand. Trees provide both food and shelter
for a wide variety of forest wildlife, and we could, for example, use changes in canopy
cover in harvested stands as a measure of habitat disturbance intensity (D). Using uncut
forests stands as a reference, we can calculate disturbance intensity as the percent
difference in canopy cover between adjacent harvested and uncut stands: D = [(% canopy
cover in uncut forest - % canopy cover harvested forest) / % canopy cover in uncut forest]
× 100%. On this basis, the response of populations to disturbance (D > 0) may take
multiple functional forms. Following the approach taken by Morris (1989) and Fortin et al.
(2008), we begin by incorporating additional terms into the simple fitness-density function
proposed by Morris (1988: Eq. 6) for a single species occupying a single habitat. The
fitness (WH) of individuals in a harvested stand (H) should be a function of disturbance
intensity (D) and conspecific density (NH), and the effect of disturbance on fitness may also
be density-dependent:
WH = wH + β1HD + β2HNH + β3H(D×NH). (1)
If we plot fitness as a function of conspecific density, the intercept is given by (wH + β1D),
where wH represents the maximum individual fitness at low population density in harvested
habitats and β1H represents changes to maximum fitness associated with the percent change
in canopy cover relative to an uncut forest (D). The slope of the fitness-density curve is
given by (β2H + β3HD) where β2H (usually negative) represents the rate of decline in
individual fitness with increasing density in the harvested habitat, and β3H accounts for the
change in the rate that fitness declines with density according to the level of disturbance.
We can assess the impact of different intensities of forest harvesting by contrasting
the distribution of animals in pairs of adjacent habitats, the first habitat type being uncut
stands (U) and the second harvested stands (H) with varying levels of canopy cover
reduction. If individuals are free to move between habitats, fitness maximization should
95
lead to a population distribution such that WU = WH (Fretwell & Lucas 1970, Morris 1988),
which implies that:
wU + 2UNU = wH + β1HD + β2H NH + β3H(D×NH) (2)
By solving for NU, we get the isodar that yields the equilibrium density in the uncut forest
habitat as a function of the density in the harvested habitat:
(3)
We assume that whenever there is a harvesting disturbance (D > 0), maximum fitness at
low density and the rate of decline in fitness with increasing density in the harvested habitat
may differ from uncut forests (and therefore wH wU and β2H 2U) and that these
differences may also be proportional to disturbance intensity (D). This isodar can be
estimated with a multiple regression taking the general form:
NU = β0 + β1 D + β2 NH + β3(D×NH) (4)
Where:
0 wH wU
2U
,
1 1H
2U
,
2 2H
2U
, and
3 3H
2U
.
Quantifying Eq. 4 from empirical observations can reveal how quantitative and
qualitative differences between logged and uncut forest habitats vary according to
disturbance intensity. We only present hypothetical scenarios for species associated with
closed-canopy forests where logging reduces habitat quality, and assume that fitness
declines linearly with both disturbance intensity and conspecific density, producing linear
isodars. More complex terms can be incorporated into isodars that will bend them into
curved or non-linear forms (Morris 2003b), and residuals from the isodar should be
inspected to determine the adequacy of a linear model. Developing non-linear isodars
would require the same general steps but would simply begin with non-linear fitness
NU wH wU 2U
1H
2U
D 2H
2U
NH 3H
2U
(D NH )
96
functions. Regardless of whether linear or non-linear effects of disturbance are expected,
the logic behind the proposed scenarios remains the same. We also assume that the range
of population sizes investigated is large enough to be able to detect the negative association
between fitness and density (i.e., β2 > 0). Habitats that are quantitatively and qualitatively
similar should yield isodars with an intercept of 0 and a slope of 1. Assuming that adjacent
uncut forest stands are of similar quality prior to disturbance, if harvesting disturbance has
no effect on habitat quality (wU = wH and 2U = β2H), then, according to Eq. 4, we would
obtain an isodar for which β0 = β1 = β3 = 0, and β2 = 1. If disturbance influences habitat
quality (wU ≥ wH and/or 2U ≠ β2H,) but the effects of harvesting are not related to
differences in canopy cover, then we would obtain an isodar for which β1 = β3 = 0, together
with β0 ≥ 0 and (or) β2 ≠ 1 (Figure 3.1a). If harvesting only causes quantitative differences
between habitats that are proportional to changes in canopy cover, then we would expect β0
≥ 0, β1 > 0, β2 = 1, and β3 = 0 (Figure 3.1b). If harvesting only causes qualitative
differences that are proportional to changes in canopy cover, then we would expect β0 = β1
= 0, β2 > 0, β3 > 0 (Figure 3.1c; similar to Figure 4 in Morris 1990). Finally, if disturbance
induces both negative quantitative and qualitative changes that are proportional to changes
in canopy cover, then we would expect β0 ≥ 0, β1 > 0, β2 >0, and β3 > 0 (Figure 3.1d;
similar to Figure 3 in Morris 1990). Because different responses to disturbance translate
into distinct forms of isodars, we can gain insights into the fitness consequences of habitat
alteration by quantifying Eq. 4 using field data and by testing for the statistical relevance of
including each model parameter.
We applied this framework to evaluate the response of snowshoe hare (Lepus
americanus) and southern red-backed voles (Myodes gapperi) to a gradient of habitat
disturbance resulting from different silvicultural treatments that removed 27-100% of
merchantable trees in old-growth boreal forest. Both snowshoe hare and red-backed voles
are recognized as important prey species of the boreal forest that support a diverse predator
community (Boutin et al. 1995, Cheveau et al. 2004), and their presence within harvested
stands may have a strong impact on ecosystem dynamics.
Both species are density-dependent habitat selectors (Morris 1996, 2005) that are
known to respond to harvesting disturbances that reduce tree cover (Ferron et al. 1998,
Klenner & Sullivan 2009). The density of both species has also been observed to increase
97
linearly with canopy cover (Pietz & Tester 1983, Klenner & Sullivan 2009) and we
therefore expect that the fitness consequences of harvesting on snowshoe hare and voles
should vary linearly with disturbance intensity according to one of the scenarios outlined
above. The proposed approach is not only adequate to evaluate the effect of habitat
disturbance on evolutionary fitness, but also to assess the consequences of spatial variations
in other potentially important habitat covariates. We demonstrate this possibility by
assessing whether each species was responding to variations in habitat features that might
not be strictly linked to changes in canopy cover resulting from logging. For snowshoe
hare, we tested additional isodar models including the availability of deciduous browse
(i.e., scenarios in Figure 3.1b-3.1d, where D is replaced by the percent difference in browse
availability between cut and uncut stands), which represents their main source of food
during winter (Pease et al. 1979). Red-backed voles have high water requirements (Getz
1968) and are most commonly associated with mesic forest habitats with moist
microclimates provided by coarse woody debris, shrubs and moss cover (Morris 1996,
Orrock et al. 2000). We therefore tested additional isodar models for voles that included
moss cover as a measure of microhabitat moisture (i.e., scenarios in Figure 3.1b-3.1d,
where D is replaced by moss cover).
Methods
Study Area
This study was conducted in the Côte-Nord region (N 50o36‘ - 51
o28‘, W 67
o98‘ -
69o37‘) of Québec, Canada. This region is characterized by a rolling, hilly landscape with
altitudes often surpassing 800 m and a geology dominated by deposits of glacial till. The
regional climate is sub-humid, sub-polar, characterized by a very short growing season with
a mean annual temperature of -2.5oC and abundant annual precipitation (1000-1300 mm),
35% of which is snow (Robitaille & Saucier 1998). The study area lies in the eastern black
spruce/moss bioclimatic region and has an estimated fire return cycle between 270 and
>500 years (Bouchard et al. 2008). The long fire cycle in this region has led to a forest
landscape composed of 70% irregularly structured late-successional stands dominated
mainly by black spruce (Picea mariana) or mixed stands of black spruce-balsam fir (Abies
98
balsamea) (Boucher et al. 2003). Other tree species common to this region include jack
pine (Pinus banksiana), trembling aspen (Populus tremuloides), white birch (Betula
papyrifera), and eastern larch (Larix laricina).
Experimental Harvest Blocks
We investigated density-dependent habitat selection by snowshoe hare and red-
backed voles in an experimental silvicultural system with a 27-100% gradient in
merchantable tree removal that was achieved by harvesting old-growth forest stands with
four logging practices: 1) clearcutting with protection of regeneration and soils (CPRS), 2)
irregular shelterwood cutting leaving small merchantable stems (known as CPPTM in
Québec; Groot 2002), 3) selection cutting with temporary harvest trails (SCTemp), and 4)
selection cutting with permanent harvest trails (SCPerm). CPRS cuts attempt to minimize
disturbance to regeneration and soils by using evenly spaced harvest trails, protecting
regeneration between trails with a diameter at breast height (DBH) of less than 9 cm (Ruel
et al. 2007). CPPTM cuts aim to protect advanced regeneration and small merchantable
stems between 9 and 15 cm DBH dispersed throughout the cut (Groot 2002). The SCTemp
treatment uses 5 m wide harvest trails spaced every 30 m, with 50% of the initial basal area
of merchantable stems harvested within a 5 m band on either side of each harvest trail. A
15 m wide band of uncut forest is left between each partially harvested band. The SCPerm
treatment uses harvesting trails spaced at 35 m intervals which will be re-used during
subsequent rotations. Regularly spaced secondary harvest trails perpendicular to the
permanent trails are then used to harvest 25% of the initial basal area of merchantable
stems within bands between the permanent trails. Based on pre- and post-logging
inventories in each harvested stand, the basal area of merchantable stems (>9 cm diameter
at breast-height) was reduced by >90% in CPRS cuts, by 77-83% in CPPTM cuts, by 31-
43% in SCTemp cuts, and by 27-40% in SCPerm cuts (Ruel et al. 2007).
Four experimental harvest blocks were established during 2004 and 2005.
Individual blocks were comprised of the four silvicultural treatments, each covering ca. 20
ha. Treatments were paired with an adjacent patch of uncut old-growth forest of similar or
larger area such that snowshoe hare and red-backed voles had free access to both cut and
99
uncut forests. Two selection cutting treatments did not have directly adjacent uncut forests,
and were thus inappropriate for inclusion in isodar analysis. Our disturbance gradient
therefore consisted of 14 pairs of harvested and uncut forest.
Relative snowshoe hare density
We used fecal pellet inventories to evaluate the relative density of hare in each pair
of harvested and uncut forest (Krebs et al. 1987, Krebs et al. 2001b). We installed grids of
19 pellet plots within each harvested stand and adjacent uncut forest (i.e. 19 plots/habitat).
Pellet plots were equidistantly spaced by 75 m (i.e. equilateral triangles with 75 m per side)
such that each pellet grid covered an area of 6 ha. The average minimum distance between
the edges of pellet grids in adjacent habitats was 145 m. We used large circular pellet plots
with a 1.5 m radius (area = 7.07 m²/plot), to increase the probability of encountering pellets
within plots under low hare density (Murray et al. 2002). Pellets were cleared from all
plots in summer 2006, and new pellets were counted and cleared in the summers of 2007
and 2008, except for four grids (one CPPTM and one CPRS cut paired with uncut stands)
that we installed in the summer of 2007. For these sites we separated pellets into new and
old pellets based on color (Krebs et al. 1987, Newbury & Simon 2005), using new pellets
from the grids cleared in previous years as a reference, and only included new pellets from
these grids in analysis. While the distinction of old versus new pellets based on color can
be inconsistent among observers (Hodges & Mills 2008), the distinction between old and
new pellets at each plot was agreed upon by 3 observers. Furthermore, the removal of old
pellets from the total pellet count for these sites did not influence the outcome of the isodar
analysis. In 2008, one of the habitat pairs (CPPTM treatment) for snowshoe hare was
omitted because the uncut habitat had been harvested, and the isodar analysis was therefore
based on a total of 14 sites in 2007 and 13 in 2008.
We used the mean density of pellets (calculated as pellets/m²) in each habitat
(average of 19 plots/habitat) as an index of relative snowshoe hare abundance as a strong
link has been made between pellet density and hare abundance across several regions
(Krebs et al. 2001b, Murray et al. 2002, McCann et al. 2008). Pellet counts from both years
varied between 0 and 130 pellets/plot. One plot, however, in an uncut forest grid in 2007
had 330 pellets due to a fallen black spruce which overhung the plot and had been
100
extensively browsed by hare. This plot greatly inflated the mean pellet density for this grid
because the remaining 18 plots each had ≤3 pellets/plot. This plot was therefore removed
and the mean pellet density for this grid was thus based on the remaining 18 plots, while all
remaining replicates (n = 26) were based on the mean of 19 plots/grid.
Relative red-backed vole density
Small mammals were captured within each habitat pair along two parallel transects
separated by 100 m that were placed perpendicular to the forest edge and extended 125 m
into harvested and uncut forest. Transects were installed such that trap lines within the
harvested habitat were perpendicular to the orientation of harvest trails, in order for traps to
fall within a variety of microhabitats (trails, selectively cut strips, and uncut strips). The
configuration of one of the selection harvest/uncut habitat pairs (SCTemp treatment) did
not allow us to place trap lines perpendicular to the orientation of harvest trails and was
therefore omitted from analysis. We thus used a total of 13 habitat pairs for red-backed
voles. Live traps (7.7 8.8 23.0 cm; Sherman Traps, Tallahassee, Fla.) were placed every
10 m along each transect, starting at 5 m from the border in each habitat (12 traps/transect
for a total of 24 traps/habitat, and 48 traps per habitat pair). Red-backed voles were
captured in each habitat pair for three consecutive nights, and traps were inspected and re-
set at dawn. Captured voles were ear-tagged with a unique tag number (style 1005-1;
National Band & Tag , Newport, Ky.) before being released. All red-backed voles were
captured during July and August in 2006 and 2007. Animals were captured and handled
following protocols approved by the Université Laval Animal Welfare Committee and the
Ministère des Ressources Naturelles et de la Faune du Québec.
We calculated relative red-backed vole abundance as the minimum number known
alive (MNA) per 100 trap-nights corrected for sprung traps, because a large proportion of
sprung traps were empty in each year (46% in 2006, 26% in 2007) (Beauvais & Buskirk
1999). The number of different individuals captured in each habitat varied from 0-40
individuals in harvested stands and 0-25 individuals in uncut stands before correcting for
variations in sampling effort due to empty sprung traps. Individuals caught in both habitats
were counted in both habitats as in Morris (2005). One habitat pair (SCTemp treatment)
did not have any red-backed vole captures in 2006, and this replicate was not included in
101
analyses because it provided no information about habitat selection. The final analysis was
thus based on 12 sites in 2006 and 13 in 2007.
Measures of disturbance intensity and resource availability
To characterize disturbance intensity (D = [(% Canopy cover Uncut – % Canopy
cover Harvested)/ % Canopy cover Uncut] × 100%) of each of the harvested forest stands,
we used measures of canopy cover specific to each of the snowshoe hare pellet grids and
small mammal trapping transects. Canopy cover (% closure) at each pellet plot was
measured with a convex densiometer with readings taken in the four cardinal directions at 1
m above ground level. The mean canopy cover across the 19 plots of each grid in cut and
uncut habitats was then used to estimate disturbance intensity. For small mammal trap
lines, canopy cover was measured at 20 and 100 m from the uncut/harvested forest edge
along each transect in each habitat using the same technique as for snowshoe hare, and the
mean of the four measures in each habitat was used to estimate disturbance intensity.
We measured deciduous browse availability for snowshoe hare at each pellet plot
based on the density of deciduous saplings and shrubs >50 cm in height and <9 cm DBH.
Deciduous species included white birch, willow (Salix spp.), speckled alder (Alnus rugosa),
green alder (Alnus crispa), serviceberry (Amelanchier spp.) and mountain ash (Sorbus
spp.). The mean density (stems/m²) of deciduous saplings in the 19 plots in each habitat
was then used to calculate differences in browse availability between habitat pairs. For red-
backed voles, percent cover of live (green) moss (%) was estimated visually within 1 m²
plots placed at 20, 60, and 120 m along each trapping transect within each habitat. We
used the average of the six measures to represent moss cover in each habitat. The most
abundant moss species included Pleurozium schreberi, Ptilium crista-castrensis, and
Sphagnum spp. We then used the percent difference in browse availability and moss cover
between uncut and harvested habitats in our candidate isodar models, i.e., (uncut –
harvested)/(uncut) 100%, with positive values indicating higher levels of browse
availability or moss cover in uncut than harvested stands, and negative values indicating
higher levels in harvested stands.
102
Isodar analysis
For each species, we tested the four candidate isodar models (Figure 3.1a-d) using
mixed-effects multiple regressions. We compared models using disturbance intensity as a
continuous variable based on percent differences in canopy cover between habitat pairs,
and three further models using percent differences in browse availability for snowshoe hare
and percent differences in moss cover for red-backed voles. We used a square-root
transformation of snowshoe hare pellet density (pellets/m²) and red-backed vole density
(MNA/100 trap-nights) in harvested and uncut stands (NU and NH) to improve normality of
the data prior to isodar analysis. As suggested by Knight and Morris (1996), we inspected
residuals from the isodar models for non-linearities, but none were apparent. While non-
linear relationships could emerge with a larger sample size, our current analyses
nonetheless expose the main impacts of logging on hares and voles within the range of
disturbance evaluated. Because each site was sampled over two consecutive years, we
included habitat pairs as a random effect in the model to account for non-independence
between repeated measures at individual sites.
The various candidate models were compared based on differences in Akaike‘s
Information Criterion adjusted for small sample sizes (ΔAICc) and the weight of evidence
of each model (wi) (Burnham & Anderson 2002). We evaluated the significance of each of
the parameters retained in the best models for snowshoe hare and red-backed voles based
on whether their 95 % confidence intervals excluded 0. The fit of the best isodar models
for each species was evaluated using marginal R² values (Orelien & Edwards 2008).
Results
Habitat disturbance
The four silvicultural treatments resulted in a gradient in disturbance intensity that
led to percent differences in canopy cover between harvested and uncut stands ranging
from 19 - 97% (Table 3.1a). Differences in browse availability between uncut and
harvested forest stands were not correlated with disturbance intensity (Pearson‘s
correlation: r = 0.09, p = 0.74), and browse availability was actually higher within six of
the harvested stands (3 CPPTM and 3 SCPerm) than in adjacent uncut stands. Canopy
103
cover along small mammal trap lines within harvested stands was 11 - 100% lower than in
adjacent uncut forests (Table 3.1b). Percent differences in moss cover in harvested stands
tended to increase with disturbance intensity (Pearson‘s correlation: r = 0.47, p = 0.10), but
moss cover was higher within three harvested stands than in adjacent uncut stands (Table
3.1b). The discrepancy between the gradient in relative differences in canopy cover (11-
100%) and the level of removal of merchantable timber within harvested stands (27-100%)
was likely due to slight differences in canopy cover between adjacent stands that existed
prior to harvesting disturbance and because measures of merchantable timber (>9 cm DBH)
removal did not account for protected regeneration (<9 -15 cm DBH) that also contributed
to canopy cover.
Isodar analysis
Snowshoe hare pellet densities were generally higher within uncut forests than
within harvested stands (Figure 3.2a). Comparison of candidate isodar models for
snowshoe hare based on AICc indicated that the model retaining the density of snowshoe
hare in harvested stands and the interaction between hare density and disturbance intensity
(model: NU = β0 + β2NH + β3D×NH) received the most support with a weight of evidence of
72% (Table 3.2a). Confidence intervals (95%) for parameter estimates from the top-
ranking model confirmed a positive isodar intercept together with a positive interaction
between disturbance intensity and hare density in harvested stands (β3 D×NH), indicating
that the isodar slope increased with disturbance intensity (Table 3.3a). This model
explained 58% of the variation in hare density in uncut stands (marginal R² = 0.58). Model
comparison also indicated that percent differences in browse availability had little effect on
habitat selection between cut and uncut stands.
Red-backed vole densities were also generally higher within uncut forests, but some
replicates from CPPTM and SCPerm treatments had higher densities within harvested
stands than adjacent uncut forests (Figure 3.2b). Three potential vole isodar models with
similar support from the data (ΔAICc <2) were identified from model comparisons, two of
which included moss cover (Table 3.2b). Models including differences in canopy cover
received essentially no empirical support (wi < 0.01 in all cases). The high sum of Akaike
weights (Burnham and Anderson 2002) for models including moss cover, w+ = 0.8,
104
indicated that moss cover was an important habitat feature influencing vole distribution.
Moreover, the 95% confidence intervals associated with moss always excluded 0 (Table
3.3b). We therefore considered the two top-ranking models with moss cover in more detail
to examine how variation in moss cover influenced the response of voles to disturbance
(Table 3.3b). Parameter estimates from these models indicated that either the isodar
intercept (equal to β0 + β1moss) would increase as the percent difference in moss cover
between harvested and uncut stands increased (i.e. when moss in uncut > moss in
harvested), or that the isodar slope (equal to β2 + β3moss) would increase as the percent
difference in moss cover between habitats increased (i.e. when moss in uncut > moss in
harvested). Both models explained a considerable proportion of the variation in vole
density in uncut forests with similar marginal R² values of 0.67.
To illustrate the predicted effects of disturbance intensity and differences in habitat
structure on the form of top-ranking isodars models for each species, we calculated isodar
intercepts and slopes at the mean levels of disturbance intensity for snowshoe hare (Table
3.2a), and the mean percent differences in moss cover for red-backed voles (Table 3.2b),
for each of the four silvicultural treatments. According to the top-scoring isodar model for
snowshoe hare (NU = β0 + β2NH + β3D×NH), the estimated isodar slope at a given level of
disturbance is determined by (β2 + β3D) which is equal to 0.151 + 0.017 D (Figure 3.3a).
For red-backed voles, the first model (NU = β0 + β1moss + β2NH) indicated that the isodar
intercept would change according to differences in moss cover between uncut and
harvested stands (intercept = β0 + β1moss = 1.084 + 0.016moss; Figure 3.3c). The second
isodar model for red-backed voles (NU = β0 + β2NH + β3moss ×NH) indicated that the isodar
slope would vary with differences in moss cover (slope = β2 + β3moss = 0.620 +
0.004moss; Figure 3.3e).
To illustrate how these estimated isodars would translate into relative fitness-density
curves for hare and voles in uncut forest and each silvicultural treatment, we used uncut
forest as a reference category and assigned values of wU = β 0 + max(NH), and β2U = -1 for
the intercept and slope of the fitness-density curve (WU = wU + β2UNU) in the uncut habitat
(Figure 3.3b, 3.3d, and 3.3f). By substituting wU and β2U into Eq. 4, we can calculate the
relative values of wH, β1H, β2H, and β3H, and use these in the fitness-density functions for
harvested habitats (Eq. 1: WH = wH + β1HD + β2HNH + β3H[D×NH]) to calculate the relative
105
intercept and slopes of the fitness-density functions for average D and moss values
associated with each silvicultural treatment. For snowshoe hare Eq.1 simplifies to WH = wH
+ β2HNH + β3H[D×NH], while for red-backed vole the two possible fitness density functions
for harvested habitats are: WH = wH + β1Hmoss + β2HNH and WH = wH + β2HNH +
β3H[moss×NH]. Plots of the relative fitness-density curves reveal that the density of
snowshoe hare in lightly disturbed stands (SCPerm and SCTemp) should converge with the
density of individuals in uncut forest stands as populations increase, while the density of
individuals in more severely disturbed stands (CPPTM and CPRS cuts) should remain
consistently lower than in uncut forests (Figure 3.3b). For red-backed voles, the first isodar
suggests that the reduction in maximum fitness at low density in harvested stands relative
to uncut stands becomes greater as moss cover within cuts decreases relative to uncut
forests (Figure 3.3d). As populations increase, however, the density of voles in harvested
stands should converge on the density in uncut forests. The second vole isodar indicates
that the density of voles in harvested stands should converge more quickly with that of
uncut stands when the percent difference in moss cover between the two habitats decreases
(Figure 3.3e).
Discussion
Our study illustrates how isodar theory can reveal density-dependent consequences of
habitat disturbance on animal distribution and thereby expose fitness-related effects of
human activities. This theoretical approach is applicable to a wide range of species and
ecosystems that are subject to both natural and anthropogenic disturbances. A number of
habitat covariates reflecting important aspects of the ecology of species, such as resource
requirements or inter-specific interactions, might be incorporated into theoretical fitness-
density functions (Shenbrot & Krasnov 2000, Morris 2005, Fortin et al. 2008b). These
functions can then be used to generate hypotheses about the form of isodars to be expected
under different levels of habitat disturbance. Here we compared models using a simple
measure of disturbance intensity based on differences in canopy cover between harvested
stands and adjacent uncut forest with models including continuous habitat covariates that
reflected differences in resource availability or microhabitat conditions (as reflected by
106
moss cover) between pairs of habitats. We found that habitat selection by snowshoe hare
depended on both the relative difference in canopy cover between harvested and uncut
stands and local hare density. Red-backed vole habitat selection was also density-
dependent, but selection between harvested and uncut stands was influenced more by
differences in moss cover, which likely reflected moisture availability, than by differences
in canopy cover.
The top isodar model for snowshoe hare most closely matched a scenario based on
quantitative and qualitative effects of disturbance (i.e., similar to Figure 3.1d, but with β1 =
0). The isodar intercept was significant overall (β0 > 0), but the quantitative effects of
harvesting were not specifically linked to the difference in canopy cover (β1 = 0) within the
range of disturbance evaluated. Quantitative differences between habitats may be linked to
either lower resource availability or increased predation risks. Snowshoe hare isodar
models including browse availability received very little support, suggesting that
quantitative effects of harvesting were not due to differences in food availability. Although
isodar studies generally associate quantitative differences between habitats with a change in
food availability (e.g. Ramp & Coulson 2002, Krasnov et al. 2003, Shenbrot 2004), the
removal of protective cover can sometimes have an even stronger influence on habitat
quality, and hence animal distribution (Lin & Batzli 2002, Pusenius & Schmidt 2002).
Mismatches between consumer and resource distribution are most likely to occur when
fitness depends on predation risk as well as resource intake (Grand & Dill 1997, Morris
2005). Prey should base their assessment of habitat quality on the trade-off between
foraging opportunities and predation risk, and under an ideal free distribution more
individuals should occupy the safe habitat when feeding opportunities are similar in both
safe and risky options (Brown & Kotler 2004). Hugie ad Dill (1994) also proposed that
when both predators and prey are free to select habitats in a manner that maximizes their
fitness, prey distribution should reflect differences in the inherent riskiness of habitats and
should be relatively insensitive to resource distribution. Morris (2005) shows that
differences in predator densities and attack rates between habitats can change the isodar
intercept for prey. Consistently, Abramsky et al. (1997) found that manipulations of
predation risk in pairs of desert habitats resulted in quantitative differences between
habitats for Allenby‘s gerbils (Gerbillus allenbyi). Hare may therefore have favored uncut
107
forests at low-density based on the greater availability of vegetative cover from predators.
Preference for safe habitats should also depend on local population density. As increasing
population density reduces per capita food availability in the safe habitat, individuals may
be willing to accept greater risk in exchange for access to food in habitats with lower cover
(Brown & Kotler 2004). For example, Wolff (1980) observed that snowshoe hare
distribution expands into riskier open habitats as food resources are depleted in refuge
habitats during the increase phase of their population cycle. Furthermore, per capita
predation risk may also decrease as prey populations increase, and the relative risk in
different habitats may then depend on local population size (China et al. 2008). Therefore
changes in the distribution of individuals between disturbed and undisturbed habitats with
increasing population size should reflect the concurrent effects of increasing density on
dilution of risk and on competition for food resources.
A significant positive regression slope of animal density between adjacent habitats
provides a strong indication of density-dependent habitat selection (Ramp and Coulson
2002). Snowshoe hare pellet densities were observed to increase in both uncut and
harvested habitats as local density increased, but the distribution of hare between the two
habitats also depended on the level of habitat alteration of the harvested stand. We found
that harvesting should cause the hare isodar slope (equal to: β2 + β3D×NH; Eq. 4) to vary
above and below unity according to disturbance intensity. Deviations of the slope from
unity reflect differences between habitats due to changes in habitat structure or resource
quality that influence how efficiently individuals can extract, consume, and convert
resources into descendants (Morris 2003a, Shenbrot 2004). These qualitative differences
thereby influence the relative rate at which fitness declines with density in each habitat, and
result in diverging densities when isodar slopes are >1, and converging densities when
isodar slopes are <1 (Morris 1988). Within the range of disturbance evaluated, the isodar
model for hare suggests that canopy cover reductions >40% should reduce resource use
efficiency by hare relative to uncut forests (isodar slope >1), whereas levels of canopy
removal from 20-40% should result in increased efficiency (isodar slope <1). This finding
yields a highly testable hypothesis which could be assessed using foraging experiments
such as giving-up densities (Morris 2003a, China et al. 2008). According to the relative
fitness-density curves for hare (Figure 3.3b), the range of harvesting disturbance studied
108
here initially reduced habitat quality, but at low levels of disturbance, the selection for
uncut stands should become less pronounced as local hare density increases.
The top-ranking isodar models for red-backed voles suggested either purely
quantitative effects of disturbance (similar to Figure 3.1b) or a combination of quantitative
and qualitative effects (similar to Figure 3.1d, but with β1 = 0). Both models had positive
isodar intercepts (β0 > 0) indicating quantitative effects of harvesting that were independent
of disturbance intensity based on changes in canopy cover. The first vole isodar model
(Table 3.3b) also revealed that some of the quantitative effects of harvesting could be
explained by differences in moss cover between habitats (β1 > 0, Figure 3.3c). Red-backed
voles are known for their high water requirements and their preference for mesic habitats
(Getz 1968, Morris 1996). The removal of canopy cover and soil disturbance from
harvesting can reduce live moss cover relative to uncut stands (Deans et al. 2003), and
decreases in vole density following harvesting disturbance have been attributed to
reductions in the availability of moist microhabitats (Sullivan et al. 2008). Accordingly,
the isodar intercept (equal to: β0 + β1moss) was predicted to increase as the percent
difference in moss cover between harvested and uncut stands increased (i.e. when uncut >
harvested; Figure 3.3c). This suggests that the maximum potential fitness for voles in
harvested stands was further reduced when moss cover was lower in cut than uncut forests.
The second vole isodar model suggested that vole habitat selection depended on both
the degree of contrast in moss cover between harvested and uncut forests and local vole
density. The slope of the isodar (equal to: β2 + β3moss×NH) was predicted to increase as
differences in moss cover between habitats increased (i.e. when uncut > harvested; Figure
3.3e). Despite the qualitative effects of variations in moss cover, isodar slopes were
predicted to remain below unity within range of contrast in moss cover between habitats
that we observed (Figure 3.3e). These qualitative effects of disturbance (isodar slopes <1
for both models) suggest disparities in the relative efficiency of resource use by voles in
each habitat such that differences vole density between harvested and uncut stands should
diminish as populations grow.
Our interpretation of the isodar models rests on the assumption that individuals of each
species were able to access both habitats to assess their relative quality. Snowshoe hare
home ranges in the north-eastern portion of their range often surpass the size of our harvest
109
treatments (~20 ha), and their mean daily movements can exceed 300 m, with occasional
exploratory movements of up to 1200 m (Ferron et al. 1998, St-Laurent et al. 2008). We
are therefore confident that hare were capable of moving between habitats to assess their
relative quality. Red-backed voles are also capable of movements of >200 m (Bowman et
al. 2001), which exceeds the length of live capture transects used in each habitat (125 m),
and some individuals were indeed captured in both habitats in each year (n = 7 in 2006, n =
11 in 2007). The density of each species in harvested and uncut habitats should therefore
reflect an active selection between habitats based on an assessment of their relative quality.
We also assumed that individuals of each species were free to settle in the habitat that
maximized their fitness. While snowshoe hare may sometimes display agonistic
interactions and dominance hierarchies at high density (Quenette et al. 1997), they are not
territorial (Krebs et al. 2001a) and individuals were therefore likely to have free access to
both habitats. Red-backed voles are generally territorial (Perrin 1979), but other types of
ideal distribution resulting from territoriality such as ideal-despotic or site-dependent
habitat selection will also generate isodars (Morris 2003b). Territorial individuals will
reduce the apparent quality of the best habitat, and isodars then reflect densities in each
habitat such that the perceived fitness of individuals is equal between habitats but mean
fitness in each habitat actually differs (Morris 2003b). The isodar intercept in such cases
should be lower than if individuals followed an ideal-free distribution (Morris 1994).
Regardless of whether or not populations followed an ideal-free distribution, isodar
intercepts > 0 and slopes ≠ 1 indicate differences in maximum potential fitness at low
density and in the relative rate of decline of fitness with density in each habitat.
Conclusions about the direction of fitness consequences of disturbance from isodar analysis
should therefore be robust to deviations from the assumption of an ideal-free distribution.
The observed patterns of hare and vole distribution among harvested and uncut forest
habitats highlight the fact that we must account for density-dependent behaviors of habitat
selection when assessing the impacts of disturbance. The distribution of individuals in
harvested and uncut habitats depended on both disturbance intensity and local population
size. For snowshoe hare the negative effects of high intensity harvest treatments were
predicted to become more pronounced as populations increased, whereas differences in
density between uncut forests and low intensity treatments were predicted to diminish with
110
population growth. Differences in vole density between harvested and uncut stands were
predicted to attenuate more quickly with population increases at lower levels of contrast in
moss cover between habitats. This finding is particularly significant for species that
display cyclical or fluctuating population dynamics (Krebs et al. 1995, Cheveau et al. 2004,
Boonstra & Krebs 2006), because conclusions about the effects of disturbance may change
according to when studies are conducted during their population cycles (Morris 1990). For
example, Klenner and Sullivan (2009) did not find any differences in abundance of red-
backed voles between uncut forests and sites with different levels of forest harvesting
disturbance in years of peak vole abundance, but significant differences emerged as
populations declined globally. The density-dependent response of red-backed voles to
harvesting that we observed may explain conflicting findings as to the short-term effects of
forest harvesting on red-backed voles (Kirkland 1990). Studies that simply compare the
mean abundance of individuals in disturbed and undisturbed areas may therefore require
several years of data to detect the global impacts of habitat alteration if the effects of
disturbance depend on overall population size. The experimental design used in this study
largely bypasses this problem by using pairs of adjacent harvested and uncut forest stands
where individuals have free access to both habitats. The relative distribution of individuals
in each habitat over a range of local densities can then reveal density-dependent behaviors
of habitat selection associated with different levels of disturbance. Adaptive behaviors
such as habitat selection thus provide a strong basis for evaluating the consequences of
habitat alteration on wildlife.
Acknowledgements
This work was supported by the NSERC-Laval University industrial research chair in
silviculture and wildlife and its partners. We gratefully acknowledge the many field
assistants whose dedicated efforts made this work possible including: K. Hammelin, J.-F.
Poulin, J. Tremblay, M.-A. Larose, E. Renaud-Roy, V. Hébert-Gentille, M. White, S.
Lavoie, K. Poitras, A. Beaulieu-Lemieux, P. Etcheverry, G., Gingras, F. Lesmerises , R.
Roy, M. Skelling and É. Vachon-Hamel.
111
Table 3.1. Mean (range) canopy cover, deciduous browse availability and moss cover in
four types of silvicultural treatment and adjacent uncut forests measured within (a)
snowshoe hare pellet grids, and (b) along red-backed vole trap lines. Mean (range) percent
differences between habitat pairs are also provided for each silvicultural treatment:
clearcutting with protection of regeneration and soils, CPRS; irregular shelterwood cutting
leaving small merchantable stems, CPPTM; selection cutting with temporary harvest trails,
SCTemp; and selection cutting with permanent harvest trails, SCPerm.
Treatment
Harvested forest Uncut forest Percent differencea
a) Snowshoe hare pellet grids
Canopy cover (%)
CPRS 8.6 (2.6, 17.4) 78.6 (69.0, 85.0) 89.1 (79.6, 96.9)
CPPTM 22.2 (17.4, 27.5) 77.2 (71.3, 81.0) 71.1 (61.4, 78.3)
SCTemp 50.8 (46.8, 56.9) 74.6 (64.8, 84.7) 31.0 (23.4, 44.8)
SCPerm 62.1 (58.6, 66.4) 80.1 (78.2, 81.8) 22.5 (18.8, 27.0)
Browse availability (deciduous saplings/m²)
CPRS 0.12 (0.10, 0.14) 0.26 (0.13, 0.51) 36.7 (0.0, 76.8)
CPPTM 0.32 (0.05, 0.80) 0.15 (0.02, 0.26) -106.3 (-208.6, 64.0)
SCTemp 0.16 (0.15, 0.17) 0.34 (0.28, 0.39) 53.8 (45.9, 62.3)
SCPerm 0.10 (0.05, 0.17) 0.07 (0.03, 0.13) -64.8 (-150, -16.7)
b) Red-backed vole trap lines
Canopy cover (%)
CPRS 1.5 (0.0, 3.4) 68.7 (61.4, 77.3) 97.8 (94.7, 100.0)
CPPTM 9.1 (6.0, 12.1) 68.3 (60.1, 75.7) 86.5 (79.9, 91.7)
SCTemp 32.9 (29.9, 36.0) 60.4 (56.8, 64.0) 45.6 (43.8, 47.4)
SCPerm 50.7 (42.0, 60.7) 60.5 (55.3, 70.6) 16.4 (11.0, 24.1)
Moss cover (%)
CPRS 22.5 (12.5, 37.5) 56.5 (50.8, 62.5) 60.9 (40.0, 77.4)
CPPTM 31.1 (21.7, 39.6) 44.4 (27.9, 69.2) 16.2 (-41.8, 68.7)
SCTemp 33.8 (28.3, 39.2) 42.9 (38.3, 47.5) 21.8 (17.5, 26.1)
SCPerm 43.6 (39.6, 50.4) 45.7 (42.5, 47.5) 4.5 (-7.1, 16.7)
112
a Percent difference between values in uncut and harvested forest = [(Uncut – Harvested)/Uncut] × 100%
113
Table 3.2. Comparison based on Akaike‘s Information Criterion corrected for small sample
sizes (AICc) of isodar models predicting snowshoe hare (a) and red-backed voles (b)
density in uncut forests (NU) based on the density of hare and voles in harvested stands
(NH), disturbance intensity (D) measured as the percent difference in canopy cover between
uncut stands and adjacent harvested stands, as well as percent differences in browse
availability (browse) or moss cover (moss).
Model AICc ΔAICc wi
a) Snowshoe hare
NU = β0 + β2NH 12.00 4.52 0.08
NU = β0 + β1D + β2NH 10.74 3.26 0.14
NU = β0 + β1D + β2NH + β3D ×NH 13.05 5.57 0.04
NU = β0 + β2NH + β3D×NH 7.48 0.00 0.72
NU = β0 + β1browse + β2NH 16.77 9.29 0.01
NU = β0 + β1browse + β2NH + β3browse ×NH 22.73 15.25 0.00
NU = β0 + β2NH + β3browse ×NH 16.29 8.82 0.01
b) Red-backed vole
NU = β0 + β2NH 85.55 1.62 0.19
NU = β0 + β1D + β2NH 95.15 11.22 < 0.01
NU = β0 + β1D + β2NH + β3D×NH 102.37 18.44 < 0.01
NU = β0 + β2NH + β3D×NH 94.95 11.02 < 0.01
NU = β0 + β1moss + β2NH 83.93 0.00 0.43
NU = β0 + β1moss + β2NH + β3moss ×NH 89.18 5.23 0.03
NU = β0 + β2NH + β3moss×NH 84.39 0.46 0.34
114
Table 3.3. Parameter estimates and 95% confidence intervals (CI) for the top (ΔAICc <2)
isodar models describing snowshoe hare and red-backed vole distribution in pairs of uncut
and harvested boreal forest stands.
Parameter Estimate 95% CI
a) Snowshoe hare
NU = β0 + β2NH + β3D×NH
Β0 0.245 (0.106, 0.383)
Β2 0.151 (-0.257, 0.558)
Β3 0.017 (0.007, 0.027)
b) Red-backed vole
NU = β0 + β1moss + β2NH
Β0 1.084 (0.061, 2.107)
Β1 0.016 (0.003, 0.029)
Β2 0.705 (0.487, 0.922)
NU = β0 + β2NH + β3moss×NH
Β0 1.552 (0.694, 2.410)
Β2 0.620 (0.406, 0.833)
Β3 0.004 (0.001, 0.007)
115
116
Figure 3.1 Four scenarios of expected fitness (W)-density (N) functions (left-hand side) and
corresponding isodars (eq. [4]: NU = β0 + β1 D + β2 NH + β3[D×NH]; right-hand side) for
pairs of uncut (U) forest (solid line) and stands harvested (H) (dashed lines) at different
levels of disturbance intensity (D; the percent difference in canopy cover [20-80%] relative
to an adjacent uncut forest): a) if the effect of harvesting is not related to measures of
disturbance intensity, b) if disturbance has only quantitative effects on habitat quality, c) if
disturbance has only qualitative effects on habitat quality, and d) if disturbance affects
habitat quality both quantitatively and qualitatively. In these scenarios it is assumed that the
intercept and/or slope of fitness-density functions decrease with increasing levels of
disturbance. In each case (a-d) an isodar between two uncut forest stands (with equal
canopy cover) with an intercept of 0 and slope of 1 is provided as a reference
117
Figure 3.2 Observed relative densities of (a) snowshoe hare (n = 27) and (b) red-backed
voles (n = 25) in pairs of uncut forest and four different silvicultural treatments (CPRS,
CPPTM, SCTemp, and SCPerm; abbreviations are defined in Table 3.1) sampled in two
consecutive years. Relative snowshoe hare density is based on the square root of pellet
density (pellets/m²). Red-backed vole densities are expressed as the square root of the
minimum number alive (MNA) per 100 trap-nights. Diagonal lines in each figure provide a
reference line indicating equal density in each habitat
118
Figure 3.3 Estimated isodar curves (left side) and corresponding relative fitness vs. density
(right side) curves according to the mean percent difference in (a,b) canopy cover (D) for
snowshoe hare and, (c-f) moss cover (moss) for red-backed voles, between four different
silvicultural treatments and adjacent uncut forests. Mean percent differences in canopy
119
cover and moss cover between pairs of uncut forest and each harvest treatment are
indicated in brackets in the figure keys (see Table 3.1 for the definition of treatment
abbreviations)
120
Chapitre 4
Browse history as an indicator of snowshoe hare
response to silvicultural practices adapted for irregular
boreal forests
James Hodson*, Daniel Fortin
*, Louis Bélanger
†, and Etienne Renaud-Roy
*
*NSERC-Université Laval industrial research chair in silviculture and wildlife,
Département de Biologie, Université Laval, Québec, QC, Canada, G1V 0A6
†Département des sciences du bois et de la forêt, Université Laval, Québec, QC, Canada,
G1V 0A6
121
Résumé
Nous avons caractérisé l‘historique de broutement du bouleau blanc (Betula papyrifera)
afin d‘évaluer la réaction du lièvre d‘Amérique (Lepus americanus) à quatre types de
traitement sylvicole variant en intensité de 0-70% de rétention d'arbres marchands.
L‘utilisation des différents traitements sylvicoles a été évaluée en comparaison avec des
forêts non coupées adjacentes à chaque peuplement coupé. Nous avons d‘abord effectué
des inventaires architecturaux de tiges de bouleaux afin de comparer la prévalence de
formes structurales indiquant un changement dans la pression de broutement. Nous avons
ensuite identifié les années auxquelles un sous-échantillon de tiges de bouleaux ont été
broutées par les lièvres afin de comparer la probabilité d‘utilisation du brout dans le temps
à partir de l‘hiver qui a précédé la coupe forestière jusqu‘à 2-3 années après coupe. Une
plus grande proportion de tiges de bouleaux avait une architecture indiquant une réduction
des niveaux de broutement dans les peuplements coupés, peu importe l‘approche sylvicole
appliquée. Une analyse détaillée des tiges a toutefois indiqué que les changements dans la
pression de broutement variaient considérablement entre les traitements. Bien que les tiges
de bouleau avaient la même probabilité d‘être consommée avant l‘application des
traitements sylvicoles, cette probabilité a diminué rapidement au cours des 2-3 années qui
ont suivi la récolte des peuplements pour les traitements ayant maintenu moins de 25% des
arbres (surface terrière ≤3 m²/ha), alors que la probabilité ne changeait pas
significativement dans les forêts non coupées adjacentes. Pendant cette même période, les
tiges de bouleau avaient la même probabilité d‘être broutées par le lièvre dans les
peuplements traités par une coupe de jardinage qui maintenait >50% des arbres (surface
terrière ≥ 15 m²/ha) que dans les forêts non coupées. L'inventaire de l'historique de
broutement représente une approche efficace pour décrire des variations temporelles dans
l'utilisation de l'habitat par les herbivores. L‘approche permet donc de comparer
l‘utilisation de sites avant et après une perturbation, même si l‘échantillonnage ne se fait
qu‘une fois le milieu perturbé. À court terme, les lièvres d'Amérique semblent utiliser les
coupes de jardinage et les forêts boréales anciennes non coupées de façon similaire.
122
Abstract
We used browse history surveys of white birch (Betula papyrifera) stems to evaluate the
response of snowshoe hare (Lepus americanus) to four harvest treatments that varied in
intensity from 0-70% retention of merchantable trees. The use of harvest treatments was
assessed relative to uncut forests adjacent to each logged stand. Stem architecture surveys
were used to compare the prevalence of growth forms of birch to assess broad changes in
browsing pressure following harvesting. We also identified the years in which individual
birch stems were browsed by hares to compare probability of stem use over time, from the
winter preceding harvesting until 2-3 years after logging. Architecture surveys revealed a
higher proportion of birch stems with released type growth forms in logged stands,
suggesting a reduction in browsing pressure by snowshoe hare following all types of
harvest treatment. Detailed stem analysis suggested, however, that changes in browsing
pressure varied considerably among treatments. Hares browsed individual stems with
equal probability in all sites prior to harvesting. Probability of stem use declined rapidly in
the two treatments with <25% tree retention (basal area ≤3 m²/ha) relative to adjacent uncut
forests in the 2-3 years following harvesting. In contrast, birch stems in selection cutting
treatments with >50% tree retention (basal area ≥15 m²/ha) were just as likely to be
browsed by hare as those in adjacent uncut forests during the same period. Browse history
surveys provide an efficient means to describe temporal trends in habitat use by herbivores.
They also allowed the comparison of habitat use before and after disturbance, even if sites
were not accessible prior to disturbance. Over the short-term, snowshoe hares appear to
make similar use of low-intensity selection cutting treatments and uncut old-growth boreal
forest stands.
123
Introduction
A dominant paradigm of contemporary forest management involves the emulation of
natural disturbance regimes to mitigate the impacts of logging on forest ecosystems
(Attiwill 1994, Bergeron & Harvey 1997, Ruel et al. 2007). The premise is that by using
silvicultural practices that maintain forest stand and landscape attributes within regional
ranges of natural variability, the wildlife communities and ecological processes that are
shaped by local disturbance regimes should also be maintained (Angelstam 1998,
Kuuluvainen 2002, Buddle et al. 2006, Bergeron et al. 2007). In boreal forests, forest stand
and landscape structure depend on the frequency, intensity and spatial extent of natural
disturbances that range from broad-scale stand-replacing fires, to fine-scale occurrences of
individual tree mortality affecting only small areas within a stand (Kuuluvainen 2002).
Despite increasing recognition of the regional variability of natural disturbance regimes
(Bergeron et al. 2001), clearcutting continues to be a dominant type of forest harvesting
(CCFM 2010). This approach could be somewhat consistent with ecosystem-based
management in areas where frequent and severe fires create forest landscapes dominated by
even-aged stands (Bergeron et al. 2002, Fenton et al. 2009). However, fire regimes vary
broadly with local climate across the boreal forest range (Bergeron et al. 2001), with fire
cycles exceeding 500 years in some regions (Bouchard et al. 2008). The absence of
frequent fires leads to landscapes dominated by old-growth forests that are shaped by fine-
scale disturbances such as insect damage, windthrow, and natural senescence (Boucher et
al. 2003). Small gap dynamics created by these disturbances result in the development of
structurally complex stands characterized by heterogeneous canopies, abundant dead-wood,
and dense understory regeneration (hereafter referred to as "irregular" stands; Pham et al.
2004, Bergeron & Harper 2009, Raymond et al. 2009). Current even-aged management,
based on short fire cycles, tends to eliminate old-growth forests from the landscape
(Bergeron 2004). An ecosystem management approach for regions with prolonged fire
cycles may therefore require alternative silvicultural strategies to maintain the prevalence
of irregularly structured stands and their associated wildlife (Bergeron et al. 2002, Harvey
124
et al. 2002). Different partial harvesting systems have been proposed as a means to extract
timber while maintaining late-seral stand structure (Ruel et al. 2007, Raymond et al. 2009).
However, their potential to retain some wildlife species associated with old-growth boreal
forest has yet to be evaluated (Vanderwel et al. 2009).
Whereas stand-replacing harvest treatments such as clearcutting tend to shift wildlife
composition towards early-successional associates (Fisher & Wilkinson 2005, Schieck &
Song 2006), the retention of mature trees within harvested stands can help to maintain
many small mammal and bird species associated with late-seral forests (Sullivan et al.
2001, Gitzen & West 2002, Klenner & Sullivan 2003, Fuller et al. 2004, Gitzen et al.
2007). Levels of tree retention as high as 70%, however, may be necessary to maintain
some late-sucessional species (Vanderwel et al. 2007, Vanderwel et al. 2009). Although
there is a growing body of literature on the response of small mammals, birds, and insects
to partial harvesting (see reviews by Vanderwel et al. 2007, Rosenvald & Lohmus 2008,
Vanderwel et al. 2009, Zwolak 2009), little information exists about the impacts of these
treatments on larger-bodied species such as snowshoe hare (Lepus americanus) (but see
Fuller & Harrison 2005).
Snowshoe hares are often selected to assess the impacts of forest harvesting disturbance
due to their large trophic influence as forest herbivores and prey for numerous predators
(Boutin et al. 1995, Dlott & Turkington 2000). Hares should be sensitive to the level of
live tree retention within different silvicultural treatments because they favour habitats with
abundant lateral and vertical cover providing protection from predators (Wolff 1980,
Litvaitis et al. 1985, Beaudoin et al. 2004). Clearcut harvesting greatly reduces snowshoe
hare abundance over the short term by reducing the availability of year-round cover (Ferron
et al. 1998, De Bellefeuille et al. 2001, Newbury & Simon 2005). Snowshoe hares are
generally present at low to moderate densities in late-successional forests across their range
(Thompson et al. 1989, Newbury & Simon 2005, Griffin & Mills 2009), and their use of
recently harvested stands may serve as an indicator of silvicultural practices that can
maintain or accelerate a return to late-seral habitat conditions.
Developing management techniques that can accelerate a return to original conditions
implies prior knowledge of ecosystem function. The evaluation of changes in animal
distribution due to habitat alteration is often based on surveys conducted before and after a
125
perturbation (Underwood 1994). Unfortunately, ecologists are frequently asked to assess
the impacts of a disturbance once it has already taken place. Dendrochronology and
browse history surveys can, in such instances, provide valuable indirect indices of
herbivore population trends (Keigley et al. 2003, Klvana et al. 2004, Payette et al. 2004,
Vila et al. 2004). Information on browse use by herbivores can be collected in a single
field season and can provide multi-year records of shifting habitat use patterns. Browse
history techniques have been developed by Keigley et al. (1998, 2003) to assess changing
trends in browse use by characterizing the stem architecture of browse species, as well as
by identifying the specific years in which stems were consumed. Such techniques have
been used, for example, to determine how spatial variation in risk-sensitive foraging by elk
(Cervus elaphus) has influenced patterns of aspen (Populus spp.) and willow (Salix spp.)
recovery following re-introduction of wolves (Canis lupus) into Yellowstone National Park
(Ripple & Beschta 2003, 2006, 2007, Halofsky et al. 2008). As snowshoe hare rely mainly
on woody browse during the winter season (Pease et al. 1979), their distinctive browsing
scars on stems retained within harvested areas can be used to reconstruct patterns of habitat
use before and after disturbance.
In this paper, we use browse history inventories of white birch (Betula papyrifera)
stems to evaluate temporal patterns of snowshoe hare habitat use in four different harvest
treatments paired with uncut irregular stands. Harvest treatments ranged from almost
complete overstory removal (<10% retention of merchantable tree basal area) to low-
intensity selection cutting (60-70% retention). We specifically tested whether browse use
was similar among all stands prior to disturbance, and whether post-logging differences in
browse use between cut and uncut stands varied among treatments according to harvest
intensity.
Methods
Study Area
This study was conducted in the Côte-Nord region (N 50o36‘ - 51
o28‘, W 67
o98‘ -
69o37‘) of Québec, Canada. The region is characterized by a rolling, hilly landscape with
altitudes often surpassing 800 m, and a geology dominated by deposits of glacial till. The
126
regional climate is sub-humid, sub-polar, with a very short growing season, a mean annual
temperature of -2.5oC, and abundant annual precipitation (1000-1300 mm), 35% of which
falls as snow (Robitaille & Saucier 1998). The study area lies in the eastern black
spruce/moss bioclimatic region (Région Écologique 6i - Saucier et al. 1998), with an
estimated fire return cycle between 270 and >500 years (Bouchard et al. 2008). The long
fire cycle in this region has led to a forest landscape composed of 70% late-seral stands
dominated mainly by black spruce (Picea mariana) or mixed stands of black spruce and
balsam fir (Abies balsamea) (Boucher et al. 2003). Other tree species common to this
region include jack pine (Pinus banksiana), trembling aspen (Populus tremuloides), white
birch, and eastern larch (Larix laricina). Forest harvesting is the major source of
anthropogenic forest disturbance.
Experimental Blocks
Four experimental harvest blocks were established during 2004 and 2005.
Individual blocks were comprised of four ca. 20 ha silvicultural treatments that included:
clearcutting with protection of regeneration and soils (CPRS, Groot et al. 2005), irregular
shelterwood cutting leaving small merchantable stems (CPPTM in Québec, for equivalent
acronyms in other regions see Groot et al. 2005), selection cutting with temporary harvest
trails (SCTemp), and selection cutting with permanent harvest trails (SCPerm). The four
harvest treatments differ in intensity according to the proportion of live trees harvested, and
in the distribution, quantity, and size of residual vegetation retained within the harvested
area. CPRS cuts attempt to minimize disturbance to regeneration and soils by using evenly
spaced harvest trails, protecting regeneration between trails with a diameter at breast height
(DBH) of less than 10 cm (Ruel et al. 2007). CPPTM cuts aim to protect small
merchantable stems between 9 and 15 cm DBH dispersed throughout the cut (Groot 2002).
The SCTemp treatment uses 5 m wide harvest trails spaced every 30 m, with 50% of the
initial basal area of merchantable stems harvested within a 5 m band on either side of each
harvest trail. A 15 m wide band of uncut forest is left between each partially harvested
band. Uncut bands will then be harvested during the subsequent rotation. The SCPerm
treatment uses harvesting trails spaced at 35 m intervals which will be re-used during
subsequent rotations. Regularly spaced secondary harvest trails perpendicular to the
127
permanent trails are used to harvest 25% of the initial basal area of merchantable stems
within bands between the permanent trails (for additional details see Liu et al. 2007, Ruel et
al. 2007, Cimon-Morin et al. 2010). Pre- and post-logging inventories indicated that levels
of retention of merchantable stems in each of the treatments were <10% in CPRS, 17-23%
in CPPTM, 57-69% in SCTemp, and 60-73% in SCPerm (Ruel et al. 2007).
We evaluated how snowshoe hare used each of the four harvest treatments relative
to uncut forests. Each harvested stand was paired with a directly adjacent (n = 14 habitat
pairs) or nearby (~850 m; n = 2 habitat pairs) uncut irregularly structured stand of similar
size. This paired approach controlled for variations in snowshoe hare abundance among
experimental blocks. Such a paired design is particularly effective at detecting differences
in habitat quality because individuals that can access both harvested and uncut habitats
should preferentially use the higher quality habitat to maximize their fitness (Morris
2003a). Differences in browse use by snowshoe hares between pairs of harvested and
uncut forest stands should thus reflect the level or use of each habitat based on their
perception of relative habitat quality.
Habitat structure and browse availability
Habitat attributes important for snowshoe hare, such as protective vegetative cover
and browse availability, were evaluated in all harvested and uncut stands between June-
August 2007 (Figure 4.1). We measured canopy closure and lateral visual obstruction at 19
points in each cut and adjacent uncut stand. These points corresponded to snowshoe hare
pellet plots that were installed in each treatment pair (see Hodson et al. 2010b for details).
Points were spaced equidistantly by 75 m (i.e., equilateral triangles with 75 m per side) in
each habitat such that sampling grids covered 6 ha. Canopy closure (%) was estimated
using a convex densiometer held at 1 m above ground level, with measures taken in the
four cardinal directions. Lateral visual obstruction (hereafter ―lateral cover‖) was measured
with a 2 m high profile board (Nudds 1977) placed 5 m away north and south from an
observer standing at each of the 19 stations. Visual obstruction of the cover board was
estimated in 10% classes. We measured the basal area of live merchantable trees (>9cm
DBH) using a 2-factor prism at three stations separated by 130 m. Finally, browse
availability was measured within 2 10 m quadrats at 9 of the 19 stations in each site (i.e.
128
even numbered stations). Given that hare can reach 50-75 cm above snow level (Pease et
al. 1979, Wirsing & Murray 2002) and that snow depth in our study region can exceed 1 m
during winter, we measured browse availability as the number of twigs (>5 cm long
terminal shoots) from deciduous shrubs and trees between 0-2 m above ground level in
each plot. The main deciduous browse species included white birch, willow (Salix spp.),
speckled alder (Alnus rugosa), green alder (Alnus crispa), serviceberry (Amelanchier spp.)
and mountain ash (Sorbus spp.). Because white birch was the focus of our browse history
surveys, we separated browse availability into two groups: white birch and other deciduous
species.
Browse History
In 2007, we evaluated browse history at all sites by conducting stem architecture
surveys of white birch and by identifying years in which individual white birch stems were
browsed by snowshoe hares. White birch was selected because it is a preferred browse
species for snowshoe hare (Newbury & Simon 2005) and because it was present in all
experimental blocks. Twigs browsed by hares are easily distinguished from other
herbivores based on the clean 45 angle at which they are clipped. Snowshoe hare were the
only species observed to clip white birch stems in our surveys. To qualitatively assess
changing levels of browse use in uncut and treated blocks, we recorded the frequency of
different browse stem architecture types along six 6 × 75 m quadrats located in each
silvicultural treatment and in their paired uncut stands. The six quadrats were located along
every third transect between stations within pellet sampling grids placed in each pair of
harvested and uncut stand (Figure 4.1). We recorded the architecture types (Keigley et al.
2003) of all white birch stems falling within 3 m on either side of the 75-m transects that
met the following criteria: 1) stems had to be at least 5 years old (determined by counting
terminal bud scars; see below) so that they could provide information on pre-treatment use
by hare, and 2) stems also had to be at least 50 cm in height 5 years prior to the survey
(determined using dating methods described below) so that they would be accessible to
hare above the snow during a large part of each winter. Four architecture types were
considered based on those identified by Keigley et al. (2003; see illustrations provided
therein): stems with uninterrupted growth, arrested-type stems, released-type stems, and
129
retrogressed-type stems. Each type represents a different level of intensity and temporal
trend in browse use. Uninterrupted stems included those with no signs of browsing or
which had experienced light to moderate browsing by snowshoe hare that did not impede
the vertical growth of the main stem. Arrested-type stems resulted from intense browsing
over several years, which stunted vertical growth at the height accessible to hare during
winter. Released-type stems were characterized by a shift from intense browsing (arrested
form) to a light or moderate level of use that allowed vertical growth to resume beyond the
maximum height available to hare. Finally, retrogressed-type growth forms represented a
shift from a light to moderate level of browse use (uninterrupted form) to an intense level
of use, whereby most of the lateral stems available to hare along the main vertical stems
were recently browsed. Sample photographs of each stem architecture type are provided in
Appendix 3. We used the proportion of the total number of stems of each architecture type
recorded over the six quadrats at each site to evaluate changes in browse intensity. For
example, a high proportion of released-type stems should indicate a decline in habitat use
by snowshoe hare, whereas high proportions of retrogressed type stems should indicate an
increase in habitat use.
To further assess changing temporal trends in browse use we also re-constructed the
browse history of a subsample of white birch stems in each harvested and uncut stand.
During winter, snowshoe hare generally clip the terminal leaders of the previous summer‘s
growth, which kills the terminal bud (Pease et al. 1979). The following spring, vertical
growth resumes from a dormant lateral bud further down the stem (Keigley & Frisina
1998). In contrast, vertical growth of non-browsed stems resumes from the terminal bud,
leaving a bud scar for each year of uninterrupted growth. The number of years of growth
following the mortality of a terminal leader due to browsing can then be determined by
counting the number of terminal bud scars along the stem originating from a lateral bud
which resumed vertical growth (Keigley & Frisina 1998). We can thus identify the year in
which each twig was clipped by counting the bud scars backwards from the current year‘s
growth to the browsed segment. Bud scars on white birch are generally discernable for up
to five years of previous growth. When stems are intensively browsed over several years,
dense clusters of clipped twigs may form at the height accessible to hare during winter
(arrested-type growth forms). Because it is difficult to identify the year in which each twig
130
within a cluster was browsed, we followed the method outlined by Keigley et al. (2003) and
cut stems just below the cluster of browsed twigs. We then counted the number of growth
rings using a 10× pocket loupe, and considered the stem as browsed in each previous year
for a number of years equal to the number of growth rings – 1 (for the current year‘s
growth). For the purposes of this study, we were interested in the previous four years of
growth, which includes the winter preceding the earliest cuts (2004) in the experimental
blocks.
We sampled the first three birch stems encountered along each stem architecture
survey quadrat (Figure 4.1) that had at least one browse mark so that they could provide a
historical record of use. This yielded a total of 18 stems sampled in each cut and uncut
stand (i.e. 36 stems/treatment pair). We used stems from the following transect if we were
unable to find three stems along a transect that met these criteria. In some cases we added
extra transects until we found a total of 18 stems within the stand. For each stem, we
identified the year in which twigs were clipped by snowshoe hare starting from the winter
before harvest treatments took place (i.e. the winter of 2003/2004 or 2004/2005) until the
winter of 2006/2007 (2-3 years after harvesting took place). We then used the presence or
absence of scars in each year to classify each stem as used or unused for each of the
previous four years.
Statistical Analysis
Habitat structure and browse availability
We used mixed-effects analysis of variance to compare habitat attributes important
to snowshoe among harvest treatments, and between harvested and uncut stands within
each treatment. To account for our paired cut/uncut forest design, experimental blocks and
harvest treatments nested within experimental blocks were included as random effects.
Analysis considered the type of harvest treatment applied (i.e., pairs of CPRS/uncut forest,
CPPTM/uncut forest, SCTemp/uncut forest, and SCPerm/uncut forest), the harvest status of
stands within a pair (i.e. cut vs. uncut), and the harvest treatment harvest status
interaction. This approach ensured that post hoc contrasts following a significant harvest
treatment harvest status effect, maintained the paired structure of our data when testing
131
for differences between cut and uncut stands within each treatment type. We used an α of
0.10 for all tests of fixed effects to reduce the possibility of type II errors. When the
harvest treatment harvest status interaction had P ≤0.10, we used tests of the simple effect
of harvest treatment on the least square means for cut and uncut stands 1) to determine
whether habitat attributes within harvested stands varied among treatments and 2) to verify
whether the habitat structure of paired uncut stands was consistent across treatments. We
used t-tests of the least-squared means of the mixed model as post hoc comparisons to
identify the significant differences. Percent canopy cover and lateral cover were arcsine
transformed, and browse availability was square root transformed to approximate a normal
distribution. Tree basal area met the assumption of normality.
Browse History
To compare the proportion of birch stems of each of the four architectural types
between harvested and uncut stands and between harvest treatment types, we used a mixed
model multivariate analysis of variance (MANOVA). The proportions of each architecture
type were the dependent variables, and harvest treatment (cut/uncut pair), harvest status
(cut vs. uncut), and the interaction between harvest treatment and harvest status were the
fixed effects. Experimental blocks, and silvicultural treatments nested in experimental
blocks were included as random effects to account for our paired design. The Hotelling-
Lawley trace statistic was used to test the global significance of the fixed effects on the
proportions of different architecture types. When the overall MANOVA was significant,
(P ≤0.10) individual ANOVA‘s were used as a post hoc test to evaluate the influence of
fixed effects on the proportion of each architecture type. Proportions of stem architecture
types at individual sites were weighted by the natural logarithm of the total number of
stems inventoried at each site to give less weight to sites with low total stem counts.
To assess whether snowshoe hare used sites with similar intensity before harvesting
took place, we first tested a model predicting the probability of white birch browse use by
snowshoe hare in the winter before harvesting occurred. We used a mixed effects logistic
regression with the type of harvest treatment (cut/uncut pairs grouped by treatment type),
and harvest status (cut = 1, uncut = 0) as fixed effects. To account for the hierarchical
structure of our sampling design, and the non-independence between the 18 stems selected
132
within each site, we included random effects that accounted for the fact that stems were
nested within cut/uncut pairs, cut/uncut pairs were nested within harvest treatments, and
harvest treatments were nested within experimental blocks. To model temporal changes in
the probability of birch stem browsing by snowshoe hare we used a second mixed effects
logistic regression for repeated measures. We considered browsing from the winter before
harvest treatments (year =0) took place until 2 to 3 years after logging. The type of harvest
treatment (cut/uncut pairs grouped by treatment type), harvest status (cut = 1, uncut = 0),
and year were included as fixed effects, as well as two-way interactions between harvest
treatment and harvest status, harvest treatment and year, harvest status and year, and a
three-way interaction between harvest treatment, harvest status, and year. Birch stems were
considered the experimental units upon which repeated measures were taken (i.e. each stem
was classified as used or unused during each year) and we used an autoregressive (order 1)
correlation structure to account for the fact that the probability of stem use was more likely
to be similar in successive years than 2 or 3 years apart in time. We used the same random
effects structure as specified for our pre-harvest model.
Results
Habitat structure and browse availability
Despite the different patterns of logging applied in the two selection cutting
treatments (SCTemp and SCPerm), structural attributes of residual vegetation remained
largely similar (Table 4.1, Figure 4.2). Residual vegetation structure was also similar
among stands harvested with protection of advanced regeneration (CPRS) and with
protection of small merchantable stems (CPPTM) (Table 4.1, Figure 4.2). However,
important differences were detected between selection cutting (SCTemp and SCPerm) and
conventional (CPRS and CPPTM) harvest treatments (Table 4.1, Figure 4.2). Canopy
cover retained within selection cuts (SCTemp: 55%, SCPerm: 62%) was greater than in
CPPTM (22%) and CPRS (9%), and mean residual live tree basal area was also more than
5 times higher in the two types of selection cutting (SCTemp: 15 m²/ha, SCPerm: 17
m²/ha) than in CPPTM (3 m²/ha) or CPRS (1 m²/ha). More importantly, selection cutting
133
treatments (SCTemp and SCPerm) retained several habitat features at levels similar to
adjacent uncut forest stands (Figure 4.2). For example, lateral cover between 0-2 m in the
two selection cutting treatments (SCTemp: 41%, SCPerm: 40%) remained similar to that
in adjacent uncut forests (mean 44%), whereas lateral cover in CPPTM (23%) and CPRS
(19%) was roughly half that in uncut forests (mean 45%) despite the protection of advanced
regeneration and small merchantable stems.
Some consistent differences were detected in the availability of white birch browse
between uncut and harvested stands (Table 4.1). The mean availability of white birch
browse between 0-2 m above ground level was generally higher in uncut forests (mean ±
se: 1.15 ± 0.16 twigs/m²) than in harvested stands across all treatments (mean ± se: 0.78 ±
0.20 twigs/m²). No such differences between cut and uncut stands were detected for other
deciduous browse (Table 4.1), although post hoc tests following a significant harvest
treatment effect revealed that the density of browse was generally higher within SCTemp
(mean ± se: 2.69 ± 0.56 twigs/m²) and CPRS (mean ± se: 2.35 ± 1.06 twigs/m²) treatment
pairs (i.e. cut and uncut habitats combined) than in SCPerm (mean ± se: 0.51 ± 0.19
twigs/m²) treatment pairs. The mean availability of other deciduous browse species in
CPPTM treatment pairs was 1.54 ± 1.04 twigs/m². Apart from variations in browse
availability, we detected no other significant differences in habitat features among uncut
stands, indicating that the structure of uncut forest stands was similar across all treatments
(Table 4.1).
Browse History
Stem architecture surveys of white birch within harvested and uncut stands
indicated a generally low intensity of browse use, with the majority of stems in each habitat
showing an uninterrupted growth form (73% in cuts and 75% in uncut stands; Figure 4.3).
We did not detect any differences in the proportions of stem architecture types among
treatments (harvest treatment: Hotelling-Lawley Trace U = 1.53, F9,12.87 = 1.61, P = 0.21,
and harvest treatment harvest status interactions: Hotelling-Lawley Trace U = 1.12,
F9,12.87 = 1.18, P = 0.38). Overall differences were detected, however, in the proportions of
each architecture type between uncut forests cut and all harvested stands combined (harvest
status: Hotelling-Lawley Trace U = 1.91 F3,10 = 6.37, P = 0.01). Post hoc comparisons of
134
individual architecture types between cut and uncut stands indicated differences in the
proportion of stems with a released type architecture (F1,12 = 17.44, P = 0.001) and the
proportion of arrested type stems (F1,12 = 8.82, P = 0.01), but not for stems with
uninterrupted or retrogressed architectures (P > 0.9). The proportion of stems with an
arrested growth form was almost twice as high in uncut stands (mean ± se: 0.194 ± 0.038)
as in harvested stands (mean ± se: 0.089 ± 0.024), whereas released type stems accounted
for a greater proportion of stems in harvested stands (mean ± se: 0.151 ± 0.036) than in
uncut stands (mean ± se: 0.033 ± 0.012).
The winter before harvest treatments were applied, birch stems had a similar
probability of being browsed by snowshoe hare within stands that remained uncut and
stands that were to be logged (Habitat effect: F1, 543 = 0.07, P = 0.79), regardless of the
future type of silviculture applied (Harvest treatment: F3,9 = 1.22, P = 0.36; harvest
treatment harvest status: F3, 543 = 1.47, P = 0.22). Following harvesting, temporal
patterns of browse use varied between cut and uncut stands according to the intensity of the
harvest treatment and the time since disturbance (harvest status x time: F1,1973 = 21.58, P <
0.001, and harvest treatment harvest status time: F3,1973 = 5.52, P = 0.001; Table 4.2).
The probability of birch stem use remained similar between selection cuts (SCTemp and
SCPerm treatments) and associated uncut stands (Figure 4.4) in the 2-3 years following
harvesting. In contrast, the probability of birch stem use decreased over time in CPPTM
and CPRS compared to associated uncut stands (Figure 4.4).
Discussion
White birch stems provided a temporal record of browse use by snowshoe hare that
allowed a comparison of the relative use of adjacent cut and uncut forest stands, from
before harvesting took place until 2-3 years following logging, based entirely on post-
disturbance inventories. Stem architecture surveys provided a broad assessment of shifting
patterns of habitat use by snowshoe hare based on relative differences in the prevalence of
growth forms indicating either sustained or altered levels of browsing pressure. We
detected higher proportions of released type stems in all logged stands, indicating changes
in the distribution of hare activity following all types of harvest disturbance. These
135
released growth forms reflect a decrease in browsing pressure that allowed intensively
browsed stems to resume vertical growth beyond the height accessible to hare. More
detailed stem analyses based on dating clipped twigs indicated that hare browsed white
birch stems with equal probability among treatment pairs prior to harvesting. The
probability of birch stem use in low retention (<25% retention, mean basal area ≤ 3 m²/ha)
treatments (CPRS and CPPTM) declined to almost zero in the 2-3 years following
harvesting, whereas birch stems in both types of selection cutting (>50% tree retention,
mean basal area ≥15 m²/ha) remained as likely to be used as those in uncut stands. Thus,
changes in browsing pressure were not equivalent across all harvest treatments and, unlike
CPRS and CPPTM cuts, selection cutting treatments may be able to maintain habitat use by
snowshoe hare at levels that approach those in uncut forest stands with irregular structure,
at least over the short-term.
We assume that browse use patterns by herbivores such as snowshoe hares can
provide an index of their relative use of nearby forest patches. This assumption is
reasonable in the present study because we sampled sites composed of old-growth spruce-
fir stands meaning that habitat pairs would have been of similar composition and structure
prior to logging and should thus have offered comparable levels of both browse and cover
prior to disturbance. Furthermore, food and cover resources are interspersed at a fine-scale
in old-growth boreal forests due to heterogeneity created by small canopy gap dynamics
(Hodson et al. 2010a). Consistently, Ferron and Ouellet (1992) observed that feeding and
resting sites for snowshoe hare overlap extensively within stands, meaning that patterns of
browse use on stems dispersed throughout old-growth stands should provide a reliable
index of the relative use of such stands by hare. The different patterns of browse use
detected among the four harvest treatments should therefore reflect differences in the
overall use of these areas by snowshoe hares. Caution should be taken, however, in
applying this approach in situations where different broad-scale patches provide
complementary resources such as food and cover (e.g. Tufto et al. 1996, Massé & Côté
2009), because lower browse use in areas of greater vegetative cover might not reflect the
use of such areas for avoiding predators or adverse climatic conditions.
The different patterns of browse use observed among low and high intensity harvest
treatments were most likely driven by differences in the retention of vegetative cover
136
providing protection from predators. Prey frequently trade-off food for safety by foregoing
foraging opportunities or decreasing foraging effort in open areas (Brown & Kotler 2004,
Ripple & Beschta 2006, Hodson et al. 2010a). Snowshoe hare mortality is generally higher
in open habitats (Rohner & Krebs 1996, Griffin & Mills 2009). Hares also tend to select
browse sites that are in close proximity to vegetative cover (usually <1m, Hodges &
Sinclair 2005) and are more likely to use browse in areas with higher canopy closure
(Rogowitz 1987). Although white birch browse availability was lower in all cuts, we only
detected decreases in the probability that hare would use remaining birch stems in CPRS
and CPPTM treatments that also had greatly reduced levels of vegetative cover relative to
uncut forests. Furthermore, the fact that the availability of other browse species remained
similar between logged and uncut stands suggests that food availability was not driving
post-harvest patterns of habitat use. Lateral cover, provided mainly by conifer
regeneration, remained at similar levels to uncut forests in both types of selection cutting,
consistent with recent findings that these treatments maintain many aspects of old-growth
forest structure (Cimon-Morin et al. 2010). This similarity was likely due to the more
limited movement of harvest machinery and the retention of uncut bands within selection
cutting treatments which resulted in a greater protection of the sapling layer relative to
CPRS and CPPTM. The high variation in canopy gap abundance typical of old-growth
spruce-fir stands (Pham et al. 2004) also meant that in some cases vertical cover in
selection cuts was only slightly lower than in adjacent uncut stands.
The decreased use of CPRS cuts is consistent with previous studies reporting much
lower hare densities in recent clearcuts than in uncut forests (Ferron et al. 1998, De
Bellefeuille et al. 2001, Newbury & Simon 2005, Potvin et al. 2005). Our study further
reveals that shelterwood treatments protecting small merchantable stems (CPPTM) do not
appear to retain sufficient additional cover to sustain their use by hares (Fortin et al. 2011).
The differences in birch browse use among the four harvest treatments are also consistent
with trends from snowshoe hare pellet surveys conducted in the same survey grids 3-4
years post-harvest (Hodson et al. 2010b). Patterns of pellet density in the paired cut and
uncut stands indicated that hare density in selection cuts quickly converged with that in
uncut forests as local abundance increased, whereas hare density remained consistently
lower in CPRS and CPPTM cuts than in adjacent uncut forests across a range of local
137
population size. The advantage to using browse history surveys is that we were able to
obtain additional information on pre-harvest habitat use that was not available from pellet
inventories as well as a longer temporal record of habitat use based on data collected in a
single year. Browse history surveys also allow us to conclude with greater certainty that
patterns of snowshoe hare pellet density observed in cut and uncut forests were attributable
to harvesting and not to pre-existing differences in habitat quality.
Our study indicates that, over the short-term, selection cutting systems retaining
>50% of initial stand basal area can maintain habitat use by a key prey species such as
snowshoe hare at levels similar to those observed in uncut old-growth stands. Recent
studies from Maine also indicate that low-intensity partial harvests (post-harvest basal areas
of 12.8-37.7 m2/ha) maintain comparable hare densities to uncut forests for up to 20 years
post-harvest (Fuller & Harrison 2005, Robinson 2006). These same studies found,
however, that snowshoe hare densities were more than twice as high in regenerating
clearcuts 15-30 years old. Although clearcutting might create better snowshoe hare habitat
in the future, maintaining representative proportions of irregularly structured stands with
snowshoe hare densities that are typical of these stands might be more consistent with an
ecosystem-based approach to managing boreal forests under long fire cycles. In addition to
their operational and economic feasibility (Liu et al. 2007), selection cutting treatments also
appear to maintain assemblages of small mammals and birds typical of old-growth boreal
forests (Le Blanc et al. 2010). They may therefore represent a promising silvicultural
approach to reconcile timber harvesting with the maintenance of irregular boreal forest
stands and their associated wildlife. Further wildlife surveys conducted before and after
subsequent harvest interventions occur within the uncut portions of the selection cutting
treatments will help to reveal their longer-term contribution to ecosystem-based
management.
Acknowledgements
This work was supported by the NSERC-Laval University industrial research chair in
silviculture and wildlife and its partners. Funding was also provided by the FQRNT and
138
FCI. We gratefully acknowledge the dedicated assistance of K. Hammelin, J.-F. Poulin, J.
Tremblay, M.-A. Larose, V. Hébert-Gentille, M. White, S. Lavoie, M.L. Le Blanc, and K.
Poitras.
139
Table 4.1. Comparison of different habitat characteristics among four silvicultural treatments, and between cut and adjacent uncut
forest stands by treatment type, in four experimental blocks in the Côte-Nord region of Québec using mixed effects analysis of
variance. Fixed effects include Harvest treatment (pairs of cut/uncut forests grouped by the harvest treatment type applied to the cut
stand: CPRS, CPPTM, SCTemp, and SCPerm; abbreviations described in legend for Figure 4.2), Harvest status (Cut vs. Uncut
stands), and the interaction between Harvest treatment and Harvest status. Tests of simple effects of Harvest Treatment on each level
of Harvest status are presented to indicate differences among harvest treatment types in cut and uncut stands.
Harvest
treatment Harvest status
Harvest
treatment
Harvest status
Simple effects of Harvest
treatment on Harvest status
Cut Uncut
Habitat characteristics F3,9 P F1,12 P F3,12 P F3,12 p F3,12 P
Canopy cover (%) 20.04 <0.01 344.67 <0.01 32.05 <0.01 48.21 <0.01 0.23 0.88
Live tree (>9cm DBH) basal
area (m²/ha) 5.20 0.02 47.42 <0.01 3.15 0.07 6.67 <0.01 2.07 0.16
Lateral cover 0-2m (%) 8.6 <0.01 43.42 <0.01 8.94 <0.01 16.02 <0.01 1.52 0.26
White birch browse 0-2 m
(twigs/²) 0.85 0.50 3.70 0.08 0.73 0.55 0.82 0.51 0.76 0.54
Other deciduous browse 0-2 m
(twigs/m²) 2.90 0.09 0.07 0.80 0.22 0.88 1.12 0.38 2.07 0.16
140
Table 4.2. Type III tests of fixed-effects, parameter estimates (β ± SE), and t-tests of
parameter estimates from a mixed-model logistic regression of the probability of white
birch stem use by snowshoe hare as a function of harvest treatment (SCPerm, SCTemp,
CPPTM, and CPRS; abbreviations described in legend for Figure 4.1), harvest status (Cut =
1, Uncut = 0), and the year relative to when harvesting took place (0-3 years, with 0 being
the winter before harvesting), recorded from browse history surveys in four experimental
harvest blocks in the Côte-Nord region of Québec. Parameter estimates equal to 0 for
CPRS treatment indicate that CPRS cuts were used as the reference treatment in the
analysis.
Type III tests of fixed
effects Parameter estimates
Effect F Value P
Harvest
treatment β DF
t
value P
Intercept 0.682 ± 0.403 3 1.69 0.19
Harvest
treatment
F[3,9] = 0.83 0.51 SCPerm -0.680 ± 0.543 9 -1.25 0.24
SCTemp -0.790 ± 0.540 9 -1.46 0.18
CPPTM -0.460 ± 0.541 9 -0.85 0.42
CPRS 0.0 . . .
Harvest status F[1,1973] = 0.07 0.79 -0.563 ± 0.492 1973 -1.14 0.25
Harvest
treatment
Harvest status
F[3,1973] = 0.75 0.53 SCPerm 0.761 ± 0.693 1973 1.1 0.27
SCTemp 0.841 ± 0.692 1973 1.22 0.22
CPPTM 0.918 ± 0.697 1973 1.32 0.19
CPRS 0.0 . . .
Year F[1,1973] = 3.4 0.07 -0.277 ± 0.133 1973 -2.08 0.04
Harvest
treatment
Year
F[3,1973] = 0.74 0.53 SCPerm 0.265 ± 0.187 1973 1.42 0.16
SCTemp 0.209 ± 0.185 1973 1.13 0.26
CPPTM 0.153 ± 0.186 1973 0.83 0.41
CPRS 0.0 . . .
Harvest status
Year F[1,1973] = 21.58 <0.01 -0.685 ± 0.219 1973 -3.13 <0.01
141
Harvest
treatment
Harvest status
Year
F[3,1973] = 5.52 <0.01 SCPerm 0.724 ± 0.286 1973 2.54 0.01
SCTemp 0.473 ± 0.286 1973 1.65 0.10
CPPTM -0.336 ± 0.309 1973 -1.09 0.28
CPRS 0.0 . . .
142
Figure 4.1. Sampling design for structural habitat features, browse availability and browse
history surveys within survey grids installed in pairs of uncut forest and stands cut using
four different silvicultural treatments. Lateral and vertical cover was measured at each of
the 19 stations in the grids. Live tree basal area was measured at stations indicated by large
circles. Browse availability was measured in 2 x 10 m plots located at every second survey
station (rectangular symbols). White birch stem architecture surveys were conducted in 6 x
75 m belt transects located on every third transect within the survey grid (thick lines; 6 per
survey grid).
143
CPRS CPPTM SCTemp SCPerm
Clo
sure
(%
)
020
40
60
80
100
Canopy cover
aa
b
c
********
CPRS CPPTM SCTemp SCPerm
Vis
ual obstr
uction (
%)
020
40
60
80
100
Lateral cover 0-2 m
****
aa
bb
CPRS CPPTM SCTemp SCPerm
m²/ha
010
20
30
40
Tree basal area
aa
bb
****
CutUncut
* p < 0.05
** p < 0.01
CPRS CPPTM SCTemp SCPerm
Clo
sure
(%
)
020
40
60
80
100
Canopy cover
aa
b
c
********
CPRS CPPTM SCTemp SCPerm
Vis
ual obstr
uction (
%)
020
40
60
80
100
Lateral cover 0-2 m
****
aa
bb
CPRS CPPTM SCTemp SCPerm
m²/ha
010
20
30
40
Tree basal area
aa
bb
****
CutUncut
* p < 0.05
** p < 0.01
CPRS CPPTM SCTemp SCPerm
Clo
sure
(%
)
020
40
60
80
100
Canopy cover
aa
b
c
********
CPRS CPPTM SCTemp SCPerm
Vis
ual obstr
uction (
%)
020
40
60
80
100
Lateral cover 0-2 m
****
aa
bb
CPRS CPPTM SCTemp SCPerm
Vis
ual obstr
uction (
%)
020
40
60
80
100
Lateral cover 0-2 m
****
aa
bb
CPRS CPPTM SCTemp SCPerm
m²/ha
010
20
30
40
Tree basal area
aa
bb
****
CPRS CPPTM SCTemp SCPerm
m²/ha
010
20
30
40
Tree basal area
aa
bb
****
CutUncut
* p < 0.05
** p < 0.01
CutUncut
* p < 0.05
** p < 0.01
CutUncut
* p < 0.05
** p < 0.01
Figure 4.2. Structural habitat features measured within pairs of uncut irregular boreal
forest stands (light grey bars) and stands harvested using four different types of silvicultural
treatment (dark grey bars; CPRS = cutting with protection of regeneration and soils,
CPPTM = irregular shelterwood cutting leaving small merchantable stems, SCTemp =
144
selection cutting with temporary trails, and SCPerm = selection cutting with permanent
trails). Asterisks above habitat pairs indicate significant differences between cut and uncut
stands within each harvest treatment type. Bars with different letters indicate significant
differences between harvested stands of each treatment type or between uncut stands
associated with each harvest treatment.
145
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Cut (n=175) Uncut
(n=363)
Cut (n=240) Uncut
(n=439)
Cut (n=119) Uncut
(n=345)
Cut (n=82) Uncut
(n=485)
SCPerm SCTemp CPPTM CPRS
Pro
port
ion o
f to
tal birch s
tem
s
Arrested
Retrogressed
Released
Uninterrupted
Figure 4.3. Mean proportions of birch browse stems of each stem architecture type from
qualitative browse history surveys in four harvest treatment types and paired uncut forests
in four experimental harvest blocks in Québec‘s North Shore region. The total number of
birch stems sampled in cut and uncut stands for each silvicultural treatments (n = 4 habitat
pairs per treatment) is indicated in brackets (total number of stems encountered in six 75 m
long x 6 m wide plots at each site, 4 uncut and 4 cut sites for each silvicultural treatment).
146
CPRS
0.0
0.2
0.4
0.6
0.8
1.0
CPPTM Cut
Uncut
SCTemp
0 1 2 3
Pro
babili
ty o
f ste
m u
se
0.0
0.2
0.4
0.6
0.8
1.0
SCPerm
Year
0 1 2 3Cut Cut
CPRS
0.0
0.2
0.4
0.6
0.8
1.0
CPPTM Cut
Uncut
SCTemp
0 1 2 3
Pro
babili
ty o
f ste
m u
se
0.0
0.2
0.4
0.6
0.8
1.0
SCPerm
Year
0 1 2 3Cut Cut
Figure 4.4. Estimated probability (± 95% CI) of white birch stem use by snowshoe hare in
four different harvest treatments (SCPerm, SCTemp, CPPTM, CPRS) and paired uncut
forest stands from the winter before the harvest treatment took place (Year = 0) up until
three years following cutting. Grey arrows indicate the summer when harvesting took
place in treated stands.
147
Conclusion générale
Cette thèse intègre des théories de sélection d'habitat avec le risque de prédation,
l'utilisation de parcelles de nourriture et les patrons de déplacement pour mieux comprendre
comment les perturbations naturelles et anthropiques structurent la répartition d'une espèce
clé de la forêt boréale, le lièvre d'Amérique. J'ai d‘abord démontré que les populations de
lièvres varient avec l'âge de la forêt suivant une distribution bimodale, avec un pic
d'abondance très prononcé environ 40-50 ans après perturbation, suivi d'une deuxième
phase d'augmentation, plus légère cette fois, dans les peuplements de >180 ans. Ensuite,
j'ai observé que le lièvre évitait de se déplacer dans les ouvertures des peuplements matures
et anciens, et réduisait son utilisation des parcelles de nourriture vers le centre des trouées.
Ces résultats suggèrent que la dynamique de trouées influence la répartition du lièvre à fine
échelle durant la succession, alors que les peuplements développent une structure
irrégulière. Dans le chapitre 3, j'ai démontré que l'impact de la coupe forestière sur la
qualité de l'habitat du lièvre et du campagnol à dos roux dépendait à la fois du niveau
d'altération de l'habitat et de la densité de leurs populations locales. Les modèles d'isodars
obtenus pour les deux espèces indiquaient que lorsque l'intensité de coupe ou la différence
de disponibilité de ressources entre l'habitat coupé et non coupé était faible, l'effet de la
coupe sur la qualité de l'habitat devenait moins prononcé à mesure que la taille de la
population locale augmentait. Finalement, une reconstitution de l'historique de broutement
du bouleau blanc par le lièvre indique que l'utilisation des peuplements ayant subi une
coupe par jardinage est demeurée similaire à celle des forêts non coupées et ce, durant une
période d‘au moins 2-3 ans. En revanche, j‘ai observé une diminution marquée dans la
probabilité d'utilisation des tiges de bouleaux dans les peuplements ayant subi un traitement
plus intensif (CPRS et CPPTM).
Les résultats présentés dans cette thèse nous permettent d'approfondir notre
compréhension du lien entre les régimes de perturbations naturelles et les changements
d'abondance du lièvre à travers la succession forestière. Ils peuvent aussi servir comme
point de référence sur la répartition du lièvre dans une région de la forêt boréale qui se
distingue écologiquement par un long cycle de feux et une forte abondance de forêts
148
anciennes. Cette information pourrait s'avérer utile pour évaluer comment la répartition du
lièvre à l'échelle du paysage pourrait changer selon différents scénarios d'aménagement
forestier. Cette thèse contribue aussi à l'évaluation de l'impact de nouvelles approches
sylvicoles conçues pour maintenir la structure et la composition des forêts anciennes et les
espèces fauniques qui y sont associées. Mes travaux devraient aider au développement
d‘un aménagement durable pour la forêt boréale irrégulière.
Changements d'abondance relative du lièvre au cours d’une succession forestière
après feu et après coupe totale
Il y a de plus en plus de reconnaissance de l‘importance de maintenir des forêts
anciennes pour conserver la biodiversité animale et végétale (Mosseler et al. 2003).
Puisque l'utilisation à vaste échelle d'un aménagement équien a tendance à éliminer les
peuplements forestiers en fin de succession écologique (Bergeron 2004), cette
problématique a une importance particulière dans les régions où le cycle de feux est
relativement long (Bergeron et al. 2001). Étant donné que l'industrie forestière a déjà causé
des changements marqués du paysage dans certaines régions de la forêt boréale (Boucher et
al. 2009, Cyr et al. 2009), il est essentiel de bien comprendre les changements de répartition
animale tout au long de la succession forestière pour apprécier les conséquences
potentielles d'une réduction des aires de forêts anciennes. Dans le chapitre 1, j'ai testé
l'hypothèse émise par Buskirk et al. (1999) selon laquelle l'abondance du lièvre devrait
suivre une distribution bimodale avec l'âge des peuplements, avec des pics d'abondance
dans les peuplements au stade de mi-succession et dans les peuplements anciens.
L'occurrence de ces deux pics devrait correspondre à des maximums de densité de la strate
arbustive qui fournit au lièvre une couverture protectrice contre les prédateurs. Mes
travaux représentent la première évaluation explicite de cette hypothèse puisque la plupart
des études jusqu'à présent ont seulement comparé les densités de lièvre en début de
succession avec celles dans les stades matures ou surannés (Thompson et al. 1989,
Newbury & Simon 2005, Hodges et al. 2009). Mon étude représente aussi une première
description des changements simultanés de la structure des peuplements et de l'abondance
relative du lièvre tout au long d'une chronoséquence de succession complète.
149
J'ai observé que l'abondance relative du lièvre suivait une distribution bimodale
avec l'âge de la forêt; cependant, l'augmentation de l'abondance du lièvre en fin de
succession était subtile en comparaison au couvert latéral qui, lui, atteignait des niveaux
similaires à ceux observés pendant le premier pic d'abondance. Ce chapitre de ma thèse a
révélé plusieurs résultats marquants. Premièrement, malgré l'importance reconnue du
couvert latéral, la généralité de l'association positive entre l'abondance du lièvre et le
couvert latéral ne s'étendait pas à tous les stades de succession. Le couvert latéral est sans
doute important pour réduire le risque de prédation au cours de toutes les phases de
succession, mais j'ai aussi détecté une forte association positive entre l'abondance du lièvre
et la fermeture de la canopée durant la croissance de la première cohorte d'arbres.
Cependant, ces deux attributs structuraux des peuplements n'ont pas expliqué la variabilité
dans l'abondance du lièvre pendant la période où les peuplements font une transition vers
une structure irrégulière. Au premier regard, la faible densité de lièvres en fin de
succession semble être attribuable à un manque de nourriture à l'échelle du peuplement
puisque la densité moyenne de ramilles feuillues était presque deux fois moindre que celle
observée durant le premier pic d'abondance entre 40-50 ans après perturbation. Cependant,
je n'ai pas détecté de forte relation entre la disponibilité du brout et la densité de crottins en
fin de succession. De plus, les inventaires d'historique de broutement (chapitre 4) ont
indiqué qu'environ 70% des tiges de bouleau blanc avaient une forme de croissance
ininterrompue, indiquant que la pression de broutement par le lièvre était plutôt faible dans
les peuplements anciens. Cela suggère que la faible disponibilité de nourriture ne serait pas
le seul facteur responsable des faibles densités de lièvres dans les peuplements de forêt
ancienne.
Les inventaires de trouées dans les peuplements de ≥80 ans ont indiqué une relation
curviligne entre la densité de crottins et la proportion du peuplement occupé par les trouées
résultant de la mortalité d'arbres. Les plus hautes densités de crottins correspondaient à une
proportion intermédiaire de trouées. Cette relation suggère que la qualité de l‘habitat du
lièvre en fin de succession pourrait dépendre du compromis à fine échelle entre
l'accessibilité à des milieux de plus grand couvert vertical et celle à des ouvertures qui
contiennent une plus forte concentration de brout. Les différences à fine échelle dans
l'entremêlement du couvert et de la nourriture créé par la dynamique de trouées pourraient
150
donc contribuer à la variabilité dans la densité du lièvre à l'échelle du peuplement dans les
forêts de fin de succession.
L'étude d‘une chronoséquence forestière présentée au chapitre 1 représente aussi
une contribution importante pour l'écologie et l'aménagement forestier parce qu‘elle nous
permet de faire des liens entre la répartition du lièvre à vaste échelle et les régimes
régionaux de perturbations. La structure des paysages forestiers qui provient de la
fréquence et de l'étendue des perturbations majeures comme les feux de forêts semble avoir
une forte influence sur les interactions entre la végétation, le lièvre et ses prédateurs. Par
exemple, une dynamique source-puits dans les paysages hétérogènes du sud de l'aire de
répartition du lièvre expliquerait en partie l'absence du cycle des populations de lièvres
dans cette région (Griffin & Mills 2009). L'étude de Griffin et Mills (2009) démontre que
les habitats de forêts fermées servent de sources démographiques qui supportent les
populations de lièvres dans les milieux ouverts. La faible disponibilité de forêts fermées
dans ces régions fait en sorte que les individus sont obligés de passer à travers les milieux
ouverts plus risqués durant leurs déplacements quotidiens. Les taux de reproduction et de
survie plus faibles dans les milieux ouverts crée un puits qui pourrait donc limiter la
croissance des populations à l'échelle régionale (Griffin & Mills 2009).
En forêt boréale, les régimes de feux influenceraient l‘amplitude du cycle de
population du lièvre (Ferron & St-Laurent 2008), de même que les défenses anti-herbivores
du bouleau blanc (Bryant et al. 2009). Dans les régions de l'ouest canadien, les cycles de
feux sont généralement courts (<100 ans) et on observe une forte amplitude dans le cycle
du lièvre. Par exemple, Boutin et al. (1995) ont observé des changements de densité par un
facteur de 26 - 44 durant une période de 20 ans dans le Yukon. Le bouleau blanc poussant
dans ces régions investit également davantage dans les défenses anti-herbivores. Les feux
de forêts fréquents font en sorte qu' une grande proportion du paysage est composée de
jeunes forêts denses pouvant faciliter une croissance exponentielle des populations de lièvre
(Ferron & St-Laurent 2008). Ceci causerait de grandes fluctuations de densité pendant le
cycle du lièvre, ainsi qu'une forte pression sélective pour l‘investissement en défenses
contre les herbivores chez le bouleau blanc à cause de la pression de broutement élevée
durant les pics de densité du lièvre (Bryant et al. 2009). En revanche, dans l'est du Canada,
on observe des cycles de feux plus prolongés (> 200 ans; Bouchard et al. 2008, Bergeron &
151
Harper 2009) et des grandes proportions de vieilles forêts (Boucher et al. 2003). Les
variations cycliques des populations de lièvres sont beaucoup moins prononcées dans cette
région (Godbout 1998; Bourbonnais 1999), de sorte que le bouleau blanc investit
relativement peu pour se défendre contre les herbivores (Bryant et al. 2009). Mon étude
démontre que les lièvres subissent de grands changements de densité au cours des 80
premières années de succession, suivi d'une longue période (>100 ans) pendant laquelle la
densité n‘augmente que légèrement. Si la productivité du lièvre est relativement faible dans
les forêts matures et anciennes, les grandes surfaces couvertes par ces peuplements dans les
régions de l'est pourraient avoir un effet tampon sur les fluctuations de densité de lièvre
pendant son cycle. Peu importe les mécanismes responsables des variations régionales du
cycle du lièvre, des fréquences de récoltes forestières plus courtes que les cycles de feux
locaux pourraient induire des variations spatiotemporelles d'abondance de lièvre plus
marquées, tout simplement parce qu'il y aurait une plus grande proportion de jeunes forêts
dans le paysage à tout moment. Les simulations présentées au chapitre 1 démontrent
également que les populations de lièvres pourraient essentiellement doubler suite à un
aménagement forestier basé sur la coupe totale avec une rotation de récolte inférieure au
cycle de feu régional. La plus forte proportion de jeunes forêts qu‘entraînerait un tel
aménagement pourrait aussi favoriser les populations d'orignaux, une espèce qui semble
aussi limitée par la faible disponibilité de brout en essences feuillues observée dans les
forêts anciennes (Crête et Courtois 1997).
Influence de la dynamique de trouées sur la répartition du lièvre en fin de succession
La dynamique de trouées est un processus clé des forêts anciennes parce qu'elle est
à l‘origine de la complexité structurelle caractéristique de ce stade de succession (Bergeron
& Harper 2009). Malgré l‘importance de ce processus, peu d‘études se sont intéressées à
l‘influence de la dynamique de trouées sur la densité et l‘abondance de la faune en forêt
boréale. Des études en forêts tropicale et tempérée ont démontré l'importance des trouées
comme parcelles de ressources alimentaires concentrées pour les insectes, les oiseaux et les
petits mammifères (Blake & Hoppes 1986, Menzel et al. 1999, Beck et al. 2004, Horn et al.
2005). Les herbivores peuvent à leur tour influencer la régénération à l‘intérieur des
152
trouées en influençant la croissance et la survie des essences végétales qu'ils consomment
(Pedersen & Wallis 2004, Norghauer et al. 2008, Royo et al. 2010).
Dans le chapitre 2, j‘ai démontré que l'hétérogénéité spatiale dans la fermeture de la
canopée ainsi que la régénération végétale à l‘intérieur des trouées dans les peuplements
anciens influencent la répartition du lièvre. Des inventaires de régénération ont confirmé
que le brout en essences feuillues était concentré à l‘intérieur des trouées créant ainsi des
parcelles de forte densité de nourriture pour le lièvre. Dans le but d‘augmenter l‘efficacité
de leur quête de nourriture, les lièvres auraient pu ajuster leurs déplacements et leur effort
d‘approvisionnement afin de tirer avantage de la concentration de brout feuillu dans les
trouées. Toutefois, les inventaires de pistes réalisés en hiver ont montré que les lièvres
sélectionnaient des milieux ayant une fermeture de canopée supérieure à la moyenne et
qu‘ils ajustaient leurs déplacements afin d‘éviter les trouées ou de les traverser plus
rapidement. Les expériences de densités à l‘abandon ont indiqué que les lièvres
percevaient un plus grand risque de prédation dans les trouées et qu‘ils étaient moins
susceptibles de brouter des tiges localisées relativement loin du couvert forestier, vers le
centre de la trouée. Les trouées semblaient donc créer un compromis entre nourriture et
sécurité.
La perception de risque plus importante dans les trouées est probablement
attribuable à la plus grande vulnérabilité du lièvre aux prédateurs aériens, comme le grand
duc d'Amérique (Bubo virginianus), dans les milieux ouverts (Rohner and Krebs 1986).
Cependant, sa vulnérabilité aux prédateurs terrestres des vielles forêts, comme la martre
d'Amérique (Martes americana), ne devrait pas varier de façon considérable entre les
trouées et les milieux sous la canopée, puisque la capacité de la martre à détecter et capturer
un lièvre ne devrait pas être influencée par le couvert vertical. Néanmoins, le lièvre
pourrait avoir une plus forte probabilité de rencontrer une martre à l‘intérieur des trouées.
En effet, les martres s‘aventurent souvent dans les trouées où elles utilisent les débris
ligneux pour accéder aux proies logeant sous la neige, comme les campagnols à dos roux
(Andruskiw et al. 2008).
Bien que les lièvres semblent percevoir un risque de prédation plus important dans
les trouées, ils obtenaient tout de même la majorité de leur nourriture hivernale à l‘intérieur
de celles-ci. La faible abondance relative de lièvres observée dans les vieilles forêts
153
comparativement à celle dans les forêts de mi-succession pourrait donc être causée par
l‘accessibilité réduite du brout vers le centre des grandes trouées et par la faible
disponibilité de nourriture dans les vieux peuplements. Ces résultats démontrent que les
perturbations de faible superficie peuvent avoir une forte influence sur la répartition fine
des proies. La perception du risque de prédation du lièvre était fonction des dimensions de
la trouée, de la distance au couvert forestier et du stade de régénération à l'intérieur des
trouées. Cette perception du risque déterminait en partie son effort d'approvisionnement
qui pourrait influencer, à son tour, les patrons de régénération de la végétation au sein des
peuplements anciens.
Il est aussi possible que l'utilisation des parcelles de nourriture à l'intérieur des
trouées soit influencée par une plus forte accumulation de neige dans les ouvertures, ce qui
augmenterait les dépenses énergétiques lors des déplacements. Cependant, ce phénomène
ne devrait pas avoir fortement influencé les expériences de densité à l'abandon, puisque les
branches de pin gris étaient placées le long de transects où la neige avait été compactée par
nos déplacements. Les dépenses énergétiques seraient donc demeurées les mêmes pour les
lièvres qui se déplaçaient sur le transect de la forêt vers le centre de la trouée. De plus, les
lièvres ne se déplaçaient pas le long de trajets où l'enfoncement était moindre qu‘attendu de
façon aléatoire (profondeur d‘enfoncement : segments observés [moyenne ± s.e.] = 8.17 ±
0.26 cm; segments aléatoires = 8.15 ± 0.24 cm; t = 0.19, df = 104, P = 0.85). L'influence
des trouées sur les déplacements et l'approvisionnement du lièvre semble donc surtout liée
aux variations de perception du risque de prédation.
Contrairement aux perturbations comme le feu qui créent de vastes parcelles de
végétation qui subissent des changements graduels de composition et de structure, les
trouées se forment fréquemment et se régénèrent lentement (Lertzman & Krebs 1991,
McCarthy 2001). Ceci signifie que plusieurs animaux peuvent subir, au cours de leur vie,
des changements considérables dans la structure de leur habitat à fine échelle. La
dynamique de trouées pourrait être ainsi un processus clé déterminant la répartition de la
faune dans les régions occupées par des grandes aires de forêt anciennes.
154
Régimes de perturbations boréales, aménagement écosystémique et influence de la
récolte ligneuse sur la faune
La transition vers un aménagement forestier plus écosystémique a nécessité le
développement d'une gamme d'approches sylvicoles ayant pour but de maintenir ou de
recréer les structures de peuplements issus de la dynamique naturelle des forêts (Groot et al.
2005, Raymond et al. 2009). Un des objectifs principaux de ces nouvelles approches
sylvicoles est de réconcilier la récolte de bois avec le maintien des populations animales
associées à différents stades de succession (Vanderwel et al. 2009). Il est donc nécessaire
d'évaluer la capacité de ces approches sylvicoles à maintenir une répartition de la faune qui
est semblable à celle créée par les perturbations naturelles.
Bien que les cycles de feu dans l'est du Canada soient beaucoup plus prolongés que
ceux dans l'ouest (Bergeron & Harper 2009), il reste que le feu est le type de perturbation
naturelle le plus important afin de ré-initier la succession forestière sur de grandes surfaces
(Bouchard et al. 2008). Par conséquent, des approches sylvicoles comme la coupe totale
peuvent avoir leur place au sein d'un régime d'aménagement écosystémique de la forêt
boréale de l'est. Même s'il peut y avoir des différences importantes au niveau de la
composition et de la structure de la régénération suite à une coupe et à un feu (Elson et al.
2007, Hart & Chen 2008), les différences initiales observées au niveau de la composition et
de l'abondance des insectes, des petits mammifères et des oiseaux ont tendance à
disparaître au cours des 30 premières années de croissance de la forêt (Simon et al. 2002,
Buddle et al. 2006, Schieck & Song 2006). L'étude de la chronoséquence après feu et après
coupe présentée dans le chapitre 1 suggère que les coupes totales peuvent recréer des
conditions d'habitat pour le lièvre qui sont semblables à celles produites par le feu. Je n'ai
détecté aucune différence significative dans les patrons d'abondance relative du lièvre
durant les six premières décennies de succession après feu et après coupe. Malgré les
ressemblances au niveau de l'abondance relative du lièvre, la coupe totale ne conserve pas
plusieurs des attributs de peuplements issus de feux, dont certains sont importants pour les
espèces d'oiseaux nichant en cavité et les insectes saproxyliques (Imbeau et al. 1999,
Buddle et al. 2006, Boulanger & Sirois 2007). La capacité des coupes totales à imiter les
feux semble donc limitée pour certains groupes d‘animaux.
155
Il est aussi évident que l'application de coupes totales sur de vastes superficies est
incompatible avec le maintien de la représentativité des forêts anciennes dans certaines
régions (Cyr et al. 2009). L'expansion de l'aménagement forestier sur la Côte-Nord a déjà
réduit la superficie des forêts anciennes sous les limites attendues par les perturbations
naturelles (Bouchard & Pothier 2011). Une approche qui pourrait permettre de conserver le
niveau actuel d'abondance des forêts anciennes consisterait à prolonger la rotation entre les
récoltes de façon à mieux refléter la longueur des cycles de feux caractérisant la région
(Burton et al. 1999). Cependant, cette approche causerait certainement des pertes
significatives dans la possibilité annuelle de coupes et des pertes de tiges marchandes
compte tenu de la courte longévité de la plupart des espèces d'arbres des écosystèmes
boréaux (Bergeron et al. 2001, Bergeron et al. 2002). La coupe partielle est proposée
comme une approche alternative pour maintenir des peuplements de structure irrégulière et
leur faune tout en permettant une récolte de bois. Durant les deux dernières décennies
plusieurs études ont démontré que la coupe partielle peut maintenir des espèces de petits
mammifères, d'oiseaux et d'insectes qui sont associées aux vieilles forêts, mais le taux de
rétention d'arbres nécessaire peut varier considérablement entre espèces (Vanderwel et al.
2007, Rosenvald & Lohmus 2008, Vanderwel et al. 2009, Zwolak 2009, Work et al. 2010).
Dans le chapitre 3, j'ai développé un cadre conceptuel basé sur la théorie des isodars
pour déterminer si l'impact d'une perturbation dépendait à la fois de la densité locale
d'individus et de l'intensité de l'altération de l'habitat. Les patrons de densité d'individus
dans des paires d'habitats perturbés et non perturbés adjacents indiquent les conséquences
de différentes intensités de perturbation en termes d'aptitude phénotypique. J'ai testé cette
approche en utilisant la répartition du lièvre et du campagnol à dos roux dans des paires de
forêts non coupées et de forêts perturbées par quatre types de récolte. Les traitements
sylvicoles incluaient tout d‘abord deux types de coupe de jardinage conçues pour maintenir
la structure irrégulière des peuplements et ayant tous deux un taux de rétention d'arbres de
>50%. Ils incluaient également deux types de récolte conventionnelle (<20% de rétention
d'arbres) qui sont couramment appliqués à la grandeur du Québec. J'ai trouvé que la
sélection entre les forêts coupées et non coupées par le lièvre et le campagnol à dos roux
dépendait à la fois de l'intensité de perturbation et de la taille de la population locale. À
faible densité, le lièvre et le campagnol à dos roux ont tous les deux préféré les forêts non
156
coupées, ce qui suggère que l'aptitude phénotypique potentielle dans les peuplements
coupées était réduite, et ce, pour tous les traitements sylvicoles. À mesure que la taille de la
population augmentait, la répartition d'individus entre les forêts coupées et non coupées
dépendait de l'intensité de perturbation ou des différences de disponibilité des ressources.
Dans le cas des coupes par jardinage, la préférence du lièvre pour l'habitat non coupé
diminuait avec l'augmentation de la population locale. Ceci implique que les différences de
densité entre l'habitat coupé et non coupé devraient s'atténuer rapidement à mesure que la
population augmente. Par contre, pour les deux traitements plus intensifs (CPRS et
CPPTM), la différence de densité du lièvre entre l'habitat coupé et non coupé augmentait
avec la taille de la population locale, indiquant que le déclin en fitness avec la densité était
plus rapide dans ces deux types de coupe qu'en forêt non coupée. La répartition du
campagnol à dos roux dans les habitats coupés et non coupés était davantage influencée par
les différences dans le recouvrement de mousse que par le niveau de réduction dans la
fermeture de la canopée. Les deux isodars possibles identifiés pour le campagnol suggèrent
qu'une réduction de recouvrement de mousse associée à la coupe réduisait la qualité de
l'habitat. Cependant, la densité de campagnols dans les coupes convergeait rapidement vers
celle dans l'habitat non coupé à mesure que la population locale augmentait.
Ces résultats démontrent que, sans tenir compte que la sélection de l'habitat peut
dépendre de la densité d'individus, nos conclusions sur l'impact de différentes intensités de
coupes forestières pourraient varier selon la taille de la population locale. Cela suggère que
des études d'impacts de perturbation de l'habitat peuvent être très sensibles au moment où
l'on réalise les inventaires, surtout dans le cas d'espèces cycliques comme le lièvre. Pour
les deux espèces étudiées, les différences de densité entre les habitats coupés et non coupés
devaient s'atténuer rapidement lorsque la perturbation était de faible intensité. Le cadre
conceptuel utilisé devrait être particulièrement utile pour révéler quand les effets de
perturbation dépendent de la densité locale d'individus, et pour détecter les seuils de
perturbations auxquels les animaux perçoivent un peuplement coupé comme étant
équivalent à une forêt non coupée. Morris (1990) a démontré aussi comment les isodars
peuvent être appliqués aux études de chronoséquence pour déterminer à quel moment une
forêt en régénération pourrait devenir équivalente à une forêt mature pour une espèce
157
animale donnée. Cette approche est donc d'une grande utilité pour répondre à plusieurs
questions pertinentes à l'aménagement forestier.
Pour que l'on puisse conclure avec plus de certitude qu‘un changement de densité
animale est dû à une perturbation de l'habitat, il peut être préférable de se baser sur une
comparaison des patrons de l'utilisation de l'habitat avant et après perturbation. Dans le
chapitre 4, j'ai utilisé des inventaires d'historique de broutement pour tracer un portrait des
patrons temporels de l'utilisation relative des peuplements coupés et non coupés allant de
l'hiver précédent l‘application des traitements sylvicoles jusqu'à 2-3 années après la coupe.
Cette approche m'a permis de décrire des patrons d'utilisation d'habitat pendant une plus
longue période de temps qu'il n‘aurait été possible avec les inventaires de crottins. De plus,
en identifiant les années d'origine des cicatrices provenant de broutement par le lièvre sur
les bouleaux blancs, j'ai pu déterminer que les tiges de bouleaux dans toutes les paires
d‘habitat avaient une probabilité d'utilisation équivalente avant la coupe. Cela suggère que
le niveau d'activité du lièvre était semblable à travers toutes les paires de forêts avant que
les traitements ne soient appliqués.
La plus grande proportion de tiges avec des architectures relâchées observée dans
les peuplements coupés suggérait une réduction dans l'utilisation des milieux coupés suite à
l'application de tous les types de traitements. Cependant, des inventaires de tiges plus
détaillés ont révélé que le niveau de changement d'activité du lièvre après coupe n'était pas
le même pour chaque traitement. À chaque hiver après coupe, les lièvres étaient aussi
susceptibles d‘utiliser les tiges de bouleau dans les coupes de jardinage de faible intensité
(>50% de rétention) que celles dans les forêts non coupées adjacentes. En revanche, la
probabilité d'utilisation des tiges de bouleaux a diminué de façon marquée dans les
traitements intensifs (CPRS et CPPTM) par rapport aux forêts non coupées durant la même
période.
Les résultats combinés des chapitres 3 et 4 indiquent que les coupes de jardinage
retiennent un couvert suffisant pour maintenir l'utilisation de ces parterres de coupe par le
lièvre. En plus du potentiel pour le maintien des prédateurs du lièvre, l'utilisation soutenue
des coupes de jardinage par le lièvre pourrait aussi contribuer à une réduction de la
compétition entre la régénération d'essences feuillues et celle d'essences commerciales
comme l'épinette noire. Par exemple, la rétention de bandes de forêt non coupée afin de
158
fournir un couvert protecteur pour l'orignal à l'intérieur des coupes totales a permis une
utilisation plus uniforme du brout feuillu à travers les aires de coupe, réduisant ainsi la
compétition entre les feuillus et les conifères (Schmitz 2005).
Les travaux présentés dans les chapitres 3 et 4 complémentent bien d'autres études
récentes qui indiquent que les coupes de jardinage maintiennent la structure irrégulière des
forêts boréales anciennes ainsi que les assemblages de petits mammifères et d'oiseaux
typiques de ces peuplements (Le Blanc 2009, Cimon-Morin et al. 2010). Bien que les
coupes totales en régénération peuvent supporter des densités de lièvres beaucoup plus
élevées que les coupes partielles ou les vieilles forêts (Fuller & Harrison 2005, Robinson
2006), la création d'habitat de haute qualité pour le lièvre n'est peut-être pas désirable dans
le contexte d'un aménagement écosystémique dans des régions caractérisées par des longs
cycles de feux. L'étude de la chronoséquence forestière du chapitre 1 suggère que le lièvre
devrait se trouver à des densités plutôt faibles ou modérées sur la majorité du territoire de la
forêt boréale de l'est, puisque le cycle de feu a créé un paysage dominé par des forêts
anciennes (Boucher et al. 2003). Une application plus importante de coupes partielles
pourrait mieux conserver la dominance de peuplements à structure irrégulière et maintenir
ainsi une répartition de lièvre qui est plus typiques de cette région.
Bien que ces résultats indiquent que les coupes de jardinage de faible intensité
pourraient maintenir des communautés fauniques semblables à celles des forêts anciennes
non coupées, il est peut-être trop tôt pour conclure que cette sylviculture n‘a aucun impact
sur la faune. Les peuplements traités par jardinage que j'ai étudiés étaient toujours
adjacents à des peuplements non-coupés. La matrice forestière (p. ex. forte ou faible
proportion de coupes totales) dans laquelle ces peuplements se trouvent pourrait toutefois
avoir un impact majeur sur les populations locales d‘animaux. De plus, des travaux
supplémentaires seront nécessaires pour déterminer si les coupes de jardinage permettent
également de maintenir les prédateurs occupant les vieilles forêts.
L'application de coupes partielles de faible intensité sur de grandes superficies a
aussi certaines limitations en tant que stratégie pour le maintien des peuplements à
structures irrégulières dans les forêts boréales de l'est. Pour obtenir un volume de bois
comparable à ce qu'on obtient avec la coupe totale, il faudrait un réseau routier plus vaste,
ce qui ne serait pas souhaitable pour une espèce comme le caribou des bois (Rangifer
159
tarandus caribou) qui semble nécessiter de grandes aires de forêt non fragmentée (Dyer et
al. 2001, Courtois et al. 2004, Fortin et al. 2008a). L'utilisation des coupes partielles n'est
pas la seule stratégie disponible pour maintenir les forêts anciennes dans les paysages
forestiers sous aménagement. Une autre possibilité pourrait être la création de grandes
aires protégées sur une portion du territoire et l'application d'un aménagement plus intensif
sur une autre portion du territoire (Messier et al. 2003, Côté et al. 2010). D‘un autre côté,
peut-être faut-il accepter une baisse de la possibilité forestière pour conserver les forêts
anciennes et leur faune, surtout dans les régions avec des cycles de feux prolongés. Il
faudrait donc s‘interroger quant aux changements de structure et composition des
écosystèmes forestiers que la société est prête à accepter.
Orientations des recherches futures
Le travail accompli dans les quatre derniers chapitres amène presque autant de
questions qu‘il n‘en répond. Bien que j‘aie décrit les changements d‘abondance relative du
lièvre au cours d‘une succession forestière et évalué l‘influence de la dynamique de trouées
sur sa répartition en fin de succession, nous en connaissons relativement peu sur les
changements de paramètres démographiques du lièvre pendant la succession et les facteurs
qui les expliquent. L‘augmentation modeste d‘abondance du lièvre en fin de succession
malgré des niveaux élevés de couvert latéral est en quelque sorte un mystère compte tenu
des densités semblables de lièvres observées dans les jeunes et vieilles forêts d'autres
régions (Griffin & Mills 2009, Hodges et al. 2009). Notre compréhension des interactions
trophiques qui influencent la dynamique de population et la répartition du lièvre provient en
grande partie d‘études à long terme conduites dans l‘ouest canadien (e.g. Boutin et al. 1995,
Krebs et al. 2001a). Il y a plusieurs différences évidentes entre les communautés de
prédateurs et de proies des forêts de l‘est et de l‘ouest canadien qui pourraient influencer la
dynamique de population régionale et la répartition du lièvre. Par exemple, les études
réalisées au Yukon ont montré que le coyote (Canis latrans) et le lynx (Lynx canadensis)
sont majoritairement responsables de la régulation des populations du lièvre (Krebs et al.
1995, O'Donoghue et al. 1998). Les coyotes sont absents des forêts boréales du nord-est
canadien et je n‘ai observé aucune trace de lynx dans les peuplements anciens au cours de
deux hivers d‘inventaires de pistes dans mon aire d‘étude. La communauté de petits
160
mammifères des vieilles pessières à mousse de l‘est est dominée par le campagnol à dos
roux (Lemaitre 2009) alors que cette espèce et un de ses prédateurs principaux, la martre
d‘Amérique, sont beaucoup moins abondants au Yukon (Boutin et al. 1995). Le lièvre
représente aussi une composante importante du régime alimentaire hivernal de la martre
(Cumberland et al. 2001). Cette dernière pourrait donc être un des principaux prédateurs du
lièvre dans les forêts anciennes.
Il serait intéressant de réaliser des études démographiques à plus long terme afin de
déterminer le rôle que jouent des prédateurs comme le lynx et la martre dans la régulation
des populations de lièvre durant différents stades de succession de la forêt boréale de l‘est.
Par exemple, alors qu‘une végétation dense peut offrir une protection contre les prédateurs
tels que le lynx, offre-t-elle la même protection contre un prédateur de plus petite taille
comme la martre? De plus, des prédateurs généralistes pourraient être impliqués dans le
gradient régional d‘amplitude des cycles de population du lièvre (Murray 2000, Klemola et
al. 2002). Peut-être que la martre joue un rôle dans la réduction de l‘amplitude des cycles
de population du lièvre parce qu'elle ne dépend pas strictement du lièvre comme source de
nourriture. Nous en connaissons aussi très peu sur la répartition des prédateurs aviaires
dans la forêt boréale de l‘est mais des études dans l‘ouest canadien suggèrent que des
espèces tel que le grand duc d‘Amérique peuvent être responsables d'une partie importante
de la mortalité du lièvre (Rohner & Krebs 1996). Des études portant sur le taux de survie,
la reproduction et les causes de mortalité du lièvre dans différents stades de succession des
forêts boréales de l‘est pourraient grandement améliorer notre compréhension des processus
qui contribuent aux variations dans la dynamique de population de lièvre à travers son aire
de répartition. Une approche qui utilise des exclos pour les prédateurs ou l'ajout de
nourriture, similaire à celle appliquée au Yukon (Krebs et al. 2001a), pourrait nous
informer sur les mécanismes qui limitent la densité de lièvre en fin de succession. Une
meilleure compréhension de ces facteurs nous aiderait aussi à prédire les changements
potentiels dans la dynamique des écosystèmes boréaux induits par des changements de
régimes de perturbation associés à l'aménagement forestier ou au réchauffement climatique.
161
Bibliographie
Aakala, T., Kuuluvainen, T., De Grandpré, L. and Gauthier, S. 2007. Trees dying standing
in the northeastern boreal old-growth forests of Quebec: spatial patterns, rates, and
temporal variation. - Canadian Journal of Forest Research 37: 50-61.
Abramsky, Z., Rosenzweig, M. L. and Subach, A. 1997. Gerbils under threat of owl
predation: isoclines and isodars. - Oikos 78: 81-90.
Altendorf, K. B., Laundre, J. W., Gonzalez, C. A. L. and Brown, J. S. 2001. Assessing
effects of predation risk on foraging behavior of mule deer. - Journal of
Mammalogy 82: 430-439.
Andruskiw, M., Fryxell, J. M., Thompson, I. D. and Baker, J. A. 2008. Habitat-mediated
variation in predation risk by the American marten. - Ecology 89: 2273-2280.
Angelstam, P. K. 1998. Maintaining and restoring biodiversity in European boreal forests
by developing natural disturbance regimes. - Journal of Vegetation Science 9: 593-
602.
Armstrong, G. W., Adamowicz, W. L., Beck, J. A., Cumming, S. G. and Schmiegelow, F.
K. A. 2003. Coarse filter ecosystem management in a nonequilibrating forest. -
Forest Science 49: 209-223.
Attiwill, P. M. 1994. The disturbance of forest ecosystems - the ecological basis for
conservative management. - Forest Ecology and Management 63: 247-300.
Ausband, D. E. and Baty, G. R. 2005. Effects of precommercial thinning on snowshoe hare
habitat use during winter in low-elevation montane forests. - Canadian Journal of
Forest Research 35: 206-210.
Baker, L. 2007. Effect of corridors on the movement behavior of the jumping spider
Phidippus princeps (Araneae, Salticidae). - Canadian Journal of Zoology 85: 802-
808.
Bartemucci, P., Coates, K. D., Harper, K. A. and Wright, E. F. 2002. Gap disturbances in
northern old-growth forests of British Columbia, Canada. - Journal of Vegetation
Science 13: 685-696.
Bates, D. and Sarkar, D. 2006. lme4: linear mixed-effects models using S4 classes. - R
package version 0.9975-7.
162
Beaudoin, C., Crête, M., Huot, J., Etcheverry, P. and Côté, S. D. 2004. Does predation risk
affect habitat use in snowshoe hares? - Ecoscience 11: 370-378.
Beauvais, G. P. and Buskirk, S. W. 1999. Modifying estimates of sampling effort to
account for sprung traps. - Wildlife Society Bulletin 27: 39-43.
Beck, H., Gaines, M. S., Hines, J. E. and Nichols, J. D. 2004. Comparative dynamics of
small mammal populations in treefall gaps and surrounding understorey within
Amazonian rainforest. - Oikos 106: 27-38.
Beckwith, S. L. 1954. Ecological succession on abandoned farm lands and its relationship
to wildlife management. - Ecological Monographs 24: 349-376.
Bergeron, J. M. and Tardif, J. 1988. Winter browsing preferences of snowshoe hares for
coniferous seedlings and its implication in large-scale reforestation programs. -
Canadian Journal of Forest Research 18: 280-282.
Bergeron, Y. 2004. Is regulated even-aged management the right strategy for the Canadian
boreal forest? - Forestry Chronicle 80: 458-462.
Bergeron, Y., Drapeau, P., Gauthier, S. and Lecomte, N. 2007. Using knowledge of natural
disturbances to support sustainable forest management in the northern Clay Belt. -
Forestry Chronicle 83: 326-337.
Bergeron, Y., Gauthier, S., Kafka, V., Lefort, P. and Lesieur, D. 2001. Natural fire
frequency for the eastern Canadian boreal forest: consequences for sustainable
forestry. - Canadian Journal of Forest Research 31: 384-391.
Bergeron, Y. and Harper, K. A. 2009. Old-growth forests in the Canadian boreal: the
exception rather than the rule? - In: Old-Growth Forests. Springer-Verlag, pp. 285-
300.
Bergeron, Y. and Harvey, B. 1997. Basing silviculture on natural ecosystem dynamics: an
approach applied to the southern boreal mixedwood forest of Quebec. - Forest
Ecology and Management 92: 235-242.
Bergeron, Y., Leduc, A., Harvey, B. D. and Gauthier, S. 2002. Natural fire regime: a guide
for sustainable management of the Canadian boreal forest. - Silva Fennica 36: 81-
95.
163
Bernstein, C., Kacelnik, A. and Krebs, J. R. 1991. Individual decisions and the distribution
of predators in a patchy environment. II. The influence of travel costs and structure
of the environment. - Journal of Animal Ecology 60: 205-225.
Beyer, H. L., Merrill, E. H., Varley, N. and Boyce, M. S. 2007. Willow on Yellowstone's
northern range: evidence for a trophic cascade? - Ecological Applications 17: 1563-
1571.
Blake, J. G. and Hoppes, W. G. 1986. Influence of resource abundance on use of tree-fall
gaps by birds in an isolated woodlot. - Auk 103: 328-340.
Bonan, G. B. 2008. Forests and climate change: forcings, feedbacks, and the climate
benefits of forests. - Science 320: 1444-1449.
Bonan, G. B. and Shugart, H. H. 1989. Environmental factors and ecological processes in
boreal forests. - Annual Review of Ecology and Systematics 20: 1-28.
Boonstra, R. and Krebs, C. J. 2006. Population limitation of the northern red-backed vole in
the boreal forests of northern Canada. - Journal of Animal Ecology 75: 1269-1284.
Bouchard, M. and Pothier, D. 2011. Long-term influence of fire and harvesting on boreal
forest age structure and forest composition in eastern Québec. - Forest Ecology and
Management 261: 811-820.
Bouchard, M., Pothier, D. and Gauthier, S. 2008. Fire return intervals and tree species
succession in the North Shore region of Eastern Quebec. - Canadian Journal of
Forest Research 38: 1621-1633.
Bouchard, M., Pothier, D. and Ruel, J. C. 2009. Stand-replacing windthrow in the boreal
forests of eastern Quebec. - Canadian Journal of Forest Research 39: 481-487.
Boucher, D., De Grandpré, L. and Gauthier, S. 2003. Development of a stand structure
classification systems and comparison of two lichen-spruce woodlands in Quebec. -
Forestry Chronicle 79: 318-328.
Boucher, D., Gauthier, S. and De Grandpré, L. 2006. Structural changes in coniferous
stands along a chronosequence and a productivity gradient in the northeastern
boreal forest of Quebec. - Ecoscience 13: 172-180.
Boucher, Y., Arseneault, D., Sirois, L. and Blais, L. 2009. Logging pattern and landscape
changes over the last century at the boreal and deciduous forest transition in Eastern
Canada. - Landscape Ecology 24: 171-184.
164
Bouget, C. and Duelli, P. 2004. The effects of windthrow on forest insect communities: a
literature review. - Biological Conservation 118: 281-299.
Boulanger, Y. and Sirois, L. 2007. Postfire succession of saproxylic arthropods, with
emphasis on coleoptera, in the North Boreal forest of Quebec. - Environmental
Entomology 36: 128-141.
Bourbonnais, N. 1999. Bilan de la situation du lièvre d'Amérique sur la Côte Nord de 1979-
1980 à 1997-1998. - Faune et Parcs Québec, Service de l'aménagement et de
l'exploitation de la faune, région Côte Nord (09), p. 36.
Boutin, S., Krebs, C. J., Boonstra, R., Dale, M. R. T., Hannon, S. J., Martin, K. and
Sinclair, A. R. E. 1995. Population-changes of the vertebrate community during a
snowshoe hare cycle in Canada's boreal forest. - Oikos 74: 69-80.
Bowman, J., Forbes, G. J. and Dilworth, T. G. 2001. Distances moved by small woodland
rodents within large trapping grids. - Canadian Field-Naturalist 115: 64-67.
Boyce, M. S., Vernier, P. R., Nielsen, S. E. and Schmiegelow, F. K. A. 2002. Evaluating
resource selection functions. - Ecological Modelling 157: 281-300.
Brassard, B. W. and Chen, H. Y. H. 2006. Stand structural dynamics of North American
boreal forests. - Critical Reviews in Plant Sciences 25: 115-137.
Brawn, J. D., Robinson, S. K. and Thompson III, F. R. 2001. The role of disturbance in the
ecology and conservation of birds. - Annual Review of Ecology and Systematics 32:
251-276.
Brown, J. S. 1988. Patch use as an indicator of habitat preference, predation risk, and
competition. - Behavioral Ecology and Sociobiology 22: 37-47.
Brown, J. S. and Kotler, B. P. 2004. Hazardous duty pay and the foraging cost of predation.
- Ecology Letters 7: 999-1014.
Brown, J. S., Laundré, J. W. and Gurung, M. 1999. The ecology of fear: optimal foraging,
game theory, and trophic interactions. - Journal of Mammalogy 80: 385-399.
Brown, J. S., Morgan, R. A. and Dow, B. D. 1992. Patch use under predation risk .2. A test
with fox squirrels, Sciurus niger. - Annales Zoologici Fennici 29: 311-318.
Bryant, J. P., Clausen, T. P., Swihart, R. K., Landhäusse, S. M., Stevens, M. T., Hawkins,
C. D. B., Carrière, S., Kirilenko, A. P., Veitch, A. M., Popko, R. A., Cleland, D. T.,
Williams, J. H., Jakubas, W. J., Carlson, M. R., Bodony, K. L., Cebrian, M., Paragi,
165
T. F., Picone, P. M., Moore, J. E., Packee, E. C. and Malone, T. 2009. Fire drives
transcontinental variation in tree birch defense against browsing by snowshoe hares.
- American Naturalist 174: 13-23.
Buddle, C. M., Langor, D. W., Pohl, G. R. and Spence, J. R. 2006. Arthropod responses to
harvesting and wildfire: implications for emulation of natural disturbance in forest
management. - Biological Conservation 128: 346-357.
Bunnell, F. L. 1995. Forest-dwelling vertebrate faunas and natural fire regimes in British-
Columbia - patterns and implications for conservation. - Conservation Biology 9:
636-644.
Burnham, K. P. and Anderson, D. R. 2002. Model selection and inference: a practical
information-theoretic approach. - Springer-Verlag.
Burns, R. M. and Honkala, B. H. 1990. Silvics of North America. - U.S. Department of
Agriculture, Forest Service.
Burton, P. J., Kneeshaw, D. D. and Coates, K. D. 1999. Managing forest harvesting to
maintain old growth in boreal and sub-boreal forests. - Forestry Chronicle 75: 623-
631.
Buskirk, S. W., Ruggiero, L. F., Aubry, K. B., Pearson, D. E., Squires, J. R. and McKelvey,
K. S. 1999. Comparative ecology of lynx in North America. - In: L.F. Ruggiero,
K.B. Aubry, S.W. Buskirk, G.M. Koehler, C.J. Krebs, K.S. McKelvey and J.R.
Squires (eds.), Ecology and conservation of lynx in the United States. - U.S.
Department of Agriculture, Forest Service, pp. 397-417.
Butler, L. G. and Kielland, K. 2008. Acceleration of vegetation turnover and element
cycling by mammalian herbivory in riparian ecosystems. - Journal of Ecology 96:
136-144.
Carlton, G. C. and Bazzaz, F. A. 1998. Regeneration of three sympatric birch species on
experimental hurricane blowdown microsites. - Ecological Monographs 68: 99-120.
CCFM. 2010. National Forestry Database. Canadian Council of Forest Ministers. Available
from http://nfdp.ccfm.org/data/comp_62e.html
Cerri, R. D. and Fraser, D. F. 1983. Predation and risk in foraging minnows: balancing
conflicting demands. - The American Naturalist 121: 552-561.
166
CFS. 1999. Silvicultural Terms in Canada, 2nd
ed. Canadian Forest Service. Natural
Resources Canada. Available from http://nfdp.ccfm.org/terms/terms_e.php
Charnov, E. L. 1976. Optimal foraging, marginal value theorem. - Theoretical Population
Biology 9: 129-136.
Cheveau, M., Drapeau, P., Imbeau, L. and Bergeron, Y. 2004. Owl winter irruptions as an
indicator of small mammal population cycles in the boreal forest of eastern North
America. - Oikos 107: 190-198.
China, V., Kotler, B. P., Shefer, N., Brown, J. S. and Abramsky, Z. 2008. Density-
dependent habitat and patch use in gerbils: consequences of safety in numbers? -
Israel Journal of Ecology & Evolution 54: 373-388.
Cimon-Morin, J., Ruel, J.-C. and Darveau, M. 2010. Short term effects of alternative
silvicultural treatments on stand attributes in irregular balsam fir-black spruce
stands. - Forest Ecology and Management 260: 907-914.
Côté, P., Tittler, R., Messier, C., Kneeshaw, D. D., Fall, A. and Fortin, M.-J. 2010.
Comparing different forest zoning options for landscape-scale management of the
boreal forest: Possible benefits of the TRIAD. - Forest Ecology and Management
259: 418-427.
Courtois, R., Ouellet, J. P., Dussault, C. and Gingras, A. 2004. Forest management
guidelines for forest-dwelling caribou in Québec. - The Forestry Chronicle 80: 598-
607.
Craiu, R. V., Duchesne, T. and Fortin, D. 2008. Inference methods for the conditional
logistic regression model with longitudinal data. - Biometrical Journal 50: 97-109.
Cramer, M. J. and Willig, M. R. 2005. Habitat heterogeneity, species diversity and null
models. - Oikos 108: 209-218.
Cramer, M. J., Willig, M. R. and O'Shea, T. J. 2002. Habitat heterogeneity, habitat
associations, and rodent species diversity in a sand-Shinnery oak landscape. -
Journal of Mammalogy 83: 743-753.
Crête, M. and Courtois, R. 1997. Limiting factors might obscure population regulation of
moose (Cervidae: Alces alces) in unproductive boreal forests. - Journal of Zoology
242: 765-781.
167
Cumberland, R. E., Dempsey, J. A. and Forbes, G. J. 2001. Should diet be based on
biomass? Importance of larger prey to the american marten. - Wildlife Society
Bulletin 29: 1125-1130.
Cyr, D., Gauthier, S., Bergeron, Y. and Carcaillet, C. 2009. Forest management is driving
the eastern North American boreal forest outside its natural range of variability. -
Frontiers in Ecology and the Environment 7: 519-524.
Daubenmire, R. 1959. A canopy-cover method of vegetation analysis. - Northwest Science
33: 43-64.
De Bellefeuille, S., Bélanger, L., Huot, J. and Cimon, A. 2001. Clear-cutting and
regeneration practices in Quebec boreal balsam fir forest: effects on snowshoe hare.
- Canadian Journal of Forest Research 31: 41-51.
De Grandpré, L., Morissette, J. and Gauthier, S. 2000. Long-term post-fire changes in the
northeastern boreal forest of Quebec. - Journal of Vegetation Science 11: 791-800.
de Romer, A. H., Kneeshaw, D. D. and Bergeron, Y. 2007. Small gap dynamics in the
southern boreal forest of eastern Canada: do canopy gaps influence stand
development? - Journal of Vegetation Science 18: 815-826.
Deans, A. M., Malcolm, J. R., Srnith, S. M. and Carleton, T. J. 2003. A comparison of
forest structure among old-growth, variable retention harvested, and clearcut
peatland black spruce (Picea mariana) forests in boreal northeastern Ontario. -
Forestry Chronicle 79: 579-589.
DeLong, S. C. and Tanner, D. 1996. Managing the pattern of forest harvest: lessons from
wildfire. - Biodiversity and Conservation 5: 1191-1205.
Demarchi, M. W. and Bunnell, F. L. 1995. Forest cover selection and activity of cow
moose in summer. - Acta Theriologica 40: 23-36.
Didion, M., Fortin, M. J. and Fall, A. 2007. Forest age structure as indicator of boreal forest
sustainability under alternative management and fire regimes: a landscape level
sensitivity analysis. - Ecological Modelling 200: 45-58.
Dlott, F. and Turkington, R. 2000. Regulation of boreal forest understory vegetation: the
roles of resources and herbivores. - Plant Ecology 151: 239-251.
168
Duschesneau, R., Lesage, I., Messier, C. and Morin, H. 2001. Effects of light and
intraspecific competition on growth and crown morphology of two size classes of
understory balsam fir saplings. - Forest Ecology and Management 140: 215-225.
Dussault, C., Ouellet, J. P., Courtois, R., Huot, J., Breton, L. and Jolicoeur, H. 2005.
Linking moose habitat selection to limiting factors. - Ecography 28: 619-628.
Dyer, S., O'Neill, J., Wasel, S. and Boutin, S. 2001. Avoidance of industrial development
by woodland caribou. - Journal of Wildlife Management 65: 531-542.
Ellis, E. C., Klein Goldewijk, K., Siebert, S., Lightman, D. and Ramankutty, N. 2010.
Anthropogenic transformation of the biomes, 1700 to 2000. - Global Ecology and
Biogeography 19: 589-606.
Elson, L. T., Simon, N. P. P. and Kneeshaw, D. 2007. Regeneration differences between
fire and clearcut logging in southeastern Labrador: a multiple spatial scale analysis.
- Canadian Journal of Forest Research 37: 473-480.
Environment Canada. 2002. Canadian climate normals. Environment Canada.
Meteorological Service of Canada. Available from
http://www.climate.weatheroffice.ec.ca/climateData/canada_e.html
Etcheverry, P., Crete, M., Ouellet, J. P., Rivest, L. P., Richer, M. C. and Beaudoin, C. 2005.
Population dynamics of snowshoe hares in relation to furbearer harvest. - Journal of
Wildlife Management 69: 771-781.
Faccio, S. D. 2003. Effects of ice storm-created gaps on forest breeding bird communities
in central Vermont. - Forest Ecology and Management 186: 133-145.
Fall, A., Fortin, M. J., Kneeshaw, D. D., Yamasaki, S. H., Messier, C., Bouthillier, L. and
Smyth, C. 2004. Consequences of various landscape-scale ecosystem management
strategies and fire cycles on age-class structure and harvest in boreal forests. -
Canadian Journal of Forest Research 34: 310-322.
Fenton, N. J., Simard, M. and Bergeron, Y. 2009. Emulating natural disturbances: the role
of silviculture in creating even-aged and complex structures in the black spruce
boreal forest of eastern North America. - Journal of Forest Research 14: 258-267.
Ferron, J. and Ouellet, J. P. 1992. Daily partitioning of summer habitat and use of space by
the snowshoe hare in southern boreal forest. - Canadian Journal of Zoology 70:
2178-2183.
169
Ferron, J., Potvin, F. and Dussault, C. 1998. Short-term effects of logging on snowshoe
hares in the boreal forest. - Canadian Journal of Forest Research 28: 1335-1334.
Ferron, J. and St-Laurent, M. H. 2008. Forest-fire regime: the missing link to understand
snowshoe hare population fluctuations? - In: P.C. Alves, N. Ferrand and K.
Hacklander (eds.), Lagomorph Biology: Evolution, Ecology and Conservation.
Springer-Verlag, pp. 141-152.
Fisher, J. T. and Wilkinson, L. 2005. The response of mammals to forest fire and timber
harvest in the North American boreal forest. - Mammal Review 35: 51-81.
Foley, J. A., DeFries, R., Asner, G. P., Barford, C., Bonan, G., Carpenter, S. R., Chapin, F.
S., Coe, M. T., Daily, G. C., Gibbs, H. K., Helkowski, J. H., Holloway, T., Howard,
E. A., Kucharik, C. J., Monfreda, C., Patz, J. A., Prentice, I. C., Ramankutty, N. and
Snyder, P. K. 2005. Global consequences of land use. - Science 309: 570-574.
Fortin, D. 2003. Searching behavior and use of sampling information by free-ranging bison
(Bos bison). - Behavioral Ecology and Sociobiology 54: 194-203.
Fortin, D., Beyer, H. L., Boyce, M. S., Smith, D. W., Duchesne, T. and Mao, J. S. 2005.
Wolves influence elk movements: behavior shapes a trophic cascade in Yellowstone
National Park. - Ecology 86: 1320-1330.
Fortin, D., Courtois, R., Etcheverry, P., Dussault, C. and Gingras, A. 2008a. Winter
selection of landscapes by woodland caribou: behavioural response to geographical
gradients in habitat attributes. - Journal of Applied Ecology 45: 1392-1400.
Fortin, D., Fryxell, J. M. and Pilote, R. 2002. The temporal scale of foraging decisions in
bison. - Ecology 83: 970-982.
Fortin, D., Hébert, C., Légaré, J.-P., Courbin, N., Swiston, K., Hodson, J., Le Blanc, M.-L.,
Dussault, C., Pothier, D., Ruel, J.-C. and Couturier, S. 2011. Partial harvesting in
old-growth boreal forests and the preservation of animal diversity from ants to
woodland caribou. - In: E.B. Wallace (ed.) Woodlands: Ecology, Management and
Conservation. Nova Science Publishers, Inc., pp. 115-136.
Fortin, D., Morris, D. W. and McLoughlin, P. D. 2008b. Habitat selection and the evolution
of specialists in heterogeneous environments. - Israel Journal of Ecology &
Evolution 54: 311-328.
170
Foster, D. R. 1983. The history and pattern of fire in the boreal forest of southeastern
Labrador. - Canadian Journal of Botany 61: 2459-2471.
Fretwell, S. D. and Lucas, H. L. J. 1970. On territorial behavior and other factors
influencing habitat distribution in birds. - Acta Biotheoretica 14: 16-36.
Fryxell, J. and Lundberg, P. 1998. Individual Behavior and Community Dynamics. -
Chapman and Hall.
Fuller, A. K. and Harrison, D. J. 2005. Influence of partial timber harvesting on American
martens in north-central Maine. - Journal of Wildlife Management 69: 710-722.
Fuller, A. K., Harrison, D. J. and Lachowski, H. J. 2004. Stand scale effects of partial
harvesting and clearcutting on small mammals and forest structure. - Forest Ecology
and Management 191: 373-386.
Gauthier, S., Boucher, D., Morissette, J. and De Grandpre, L. 2010. Fifty-seven years of
composition change in the eastern boreal forest of Canada. - Journal of Vegetation
Science 21: 772-785.
Gauthier, S., De Grandpre, L. and Bergeron, Y. 2000. Differences in forest composition in
two boreal forest ecoregions of Quebec. - Journal of Vegetation Science 11: 781-
790.
Getz, L. L. 1968. Influence of water balance and microclimate on local distribution of
redback vole and white-footed mouse. - Ecology 49: 276-&.
Gill, J. A., Sutherland, W. J. and Watkinson, A. R. 1996a. A method to quantify the effects
of human disturbance on animal populations. - Journal of Applied Ecology 33: 786-
792.
Gill, R. M. A., Johnson, A. L., Francis, A., Hiscocks, K. and Peace, A. J. 1996b. Changes
in roe deer (Capreolus capreolus L) population density in response to forest habitat
succession. - Forest Ecology and Management 88: 31-41.
Gitzen, R. A. and West, S. D. 2002. Small mammal response to experimental canopy gaps
in the southern Washington Cascades. - Forest Ecology and Management 168: 187-
199.
Gitzen, R. A., West, S. D., Maguire, C. C., Manning, T. and Halpern, C. B. 2007. Response
of terrestrial small mammals to varying amounts and patterns of green-tree retention
in Pacific Northwest forests. - Forest Ecology and Management 251: 142-155.
171
Godbout, G. 1998. Détermination de la présence d'un cycle de population du lièvre
d'Amérique (Lepus americanus) au Québec et des méthodes de suivi applicables à
cette espèce. Université de Québec à Rimouski. 119 pp.
Graham, M. H. 2003. Confronting multicollinearity in ecological multiple regression. -
Ecology 84: 2809-2815.
Grand, T. C. and Dill, L. M. 1997. The energetic equivalence of cover to juvenile coho
salmon (Oncorhynchus kisutch): ideal free distribution theory applied. - Behavioral
Ecology 8: 437-447.
Griffin, P. C. and Mills, L. S. 2007. Precommercial thinning reduces snowshoe hare
abundance in the short term. - Journal of Wildlife Management 71: 559-564.
Griffin, P. C. and Mills, L. S. 2009. Sinks without borders: snowshoe hare dynamics in a
complex landscape. - Oikos 118: 1487-1498.
Groot, A. 2002. Is uneven-aged silviculture applicable to peatland black spruce (Picea
mariana) in Ontario, Canada? - Forestry 75: 437-442.
Groot, A., Lussier, J. M., Mitchell, A. K. and Macsaac, D. A. 2005. A silvicultural systems
perspective on changing Canadian forestry practices. - Forestry Chronicle 81: 50-
55.
Guisan, A., Edwards, T. C. and Hastie, T. 2002. Generalized linear and generalized additive
models in studies of species distributions: setting the scene. - Ecological Modelling
157: 89-100.
Hall, L. S., Krausman, P. R. and Morrison, M. L. 1997. The habitat concept and a plea for
standard terminology. - Wildlife Society Bulletin 25: 173-182.
Halofsky, J. S., Ripple, W. J. and Beschta, R. L. 2008. Recoupling fire and aspen
recruitment after wolf reintroduction in Yellowstone National Park, USA. - Forest
Ecology and Management 256: 1004-1008.
Harper, K. A., Bergeron, Y., Drapeau, P., Gauthier, S. and De Grandpré, L. 2005.
Structural development following fire in black spruce boreal forest. - Forest
Ecology and Management 206: 293-306.
Harper, K. A., Bergeron, Y., Drapeau, P., Gauthier, S. and De Grandpré, L. 2006. Changes
in spatial pattern of trees and snags during structural development in Picea mariana
boreal forests. - Journal of Vegetation Science 17: 625-636.
172
Hart, S. A. and Chen, H. Y. H. 2006. Understory vegetation dynamics of North American
boreal forests. - Critical Reviews in Plant Sciences 25: 381-397.
Hart, S. A. and Chen, H. Y. H. 2008. Fire, logging, and overstory affect understory
abundance, diversity, and composition in boreal forest. - Ecological Monographs
78: 123-140.
Harvey, B. D., Leduc, A., Gauthier, S. and Bergeron, Y. 2002. Stand-landscape integration
in natural disturbance-based management of the southern boreal forest. - Forest
Ecology and Management 155: 369-385.
Hastings, A. 1977. Spatial heterogeneity and stability of predator-prey systems. -
Theoretical Population Biology 12: 37-48.
Haugen, T. O., Winfield, I. J., Vollestad, L. A., Fletcher, J. M., James, J. B. and Stenseth,
N. C. 2006. The ideal free pike: 50 years of fitness-maximizing dispersal in
Windermere. - Proceedings of the Royal Society B-Biological Sciences 273: 2917-
2924.
Hobbs, N. T., Gross, J. E., Shipley, L. A., Spalinger, D. E. and Wunder, B. A. 2003.
Herbivore functional response in heterogeneous environments: a contest among
models. - Ecology 84: 666-681.
Hochman, V. and Kotler, B. P. 2007. Patch use, apprehension, and vigilance behavior of
Nubian Ibex under perceived risk of predation. - Behavioral Ecology 18: 368-374.
Hodges, K. E., Krebs, C. J. and Sinclair, A. R. E. 1999. Snowshoe hare demography during
a cyclic population low. - Journal of Animal Ecology 68: 581-594.
Hodges, K. E. and Mills, L. S. 2008. Designing fecal pellet surveys for snowshoe hares. -
Forest Ecology and Management 256: 1918-1926.
Hodges, K. E., Mills, L. S. and Murphy, K. M. 2009. Distribution and abundance of
snowshoe hares in Yellowstone National Park. - Journal of Mammalogy 90: 870-
878.
Hodges, K. E. and Sinclair, A. R. E. 2005. Browse site selection by snowshoe hares: effects
of food supply and predation risk. - Canadian Journal of Zoology 83: 280-292.
Hodson, J., Fortin, D. and Bélanger, L. 2010a. Fine-scale disturbances shape space-use
patterns of a boreal forest herbivore. - Journal of Mammalogy 91: 607-619.
173
Hodson, J., Fortin, D., Le Blanc, M.-L. and Bélanger, L. 2010b. An appraisal of the fitness
consequences of forest disturbance for wildlife using habitat selection theory. -
Oecologia 164: 73-86.
Holt, R. D. and Hassell, M. P. 1993. Environmental heterogeneity and the stability of host
parasitoid interactions. - Journal of Animal Ecology 62: 89-100.
Homyack, J. A., Harrison, D. J. and Krohn, W. B. 2007. Effects of precommercial thinning
on snowshoe hares in Maine. - Journal of Wildlife Management 71: 4-13.
Homyack, J. A., Harrison, D. J., Litvaitis, J. A. and Krohn, W. B. 2006. Quantifying
densities of snowshoe hares in Maine using pellet plots. - Wildlife Society Bulletin
34: 74-80.
Horn, S., Hanula, J. L. and Ulyshen, M. D. 2005. Abundance of green tree frogs and insects
in artificial canopy gaps in a bottomland hardwood forest. - American Midland
Naturalist 153: 321-326.
Huffaker, C. B. 1958. Experimental studies on predation: dispersion factors and predator-
prey oscillations. - Hilgardia 27: 343-383.
Hughes, J. J. and Ward, D. 1993. Predation risk and distance to cover affect foraging
behavior in Namib Desert gerbils. - Animal Behaviour 46: 1243-1245.
Hugie, D. M. and Dill, L. M. 1994. Fish and game - a game-theoretic approach to habitat
selection by predators and prey. - Journal of Fish Biology 45: 151-169.
Hunter, M. L. 2002. Fundamentals of conservation biology. - Blackwell Science, Inc.
Huntly, N. 1991. Herbivores and the dynamics of communities and ecosystems. - Annual
Review of Ecology and Systematics 22: 477-503.
Imbeau, L., Savard, J. P. L. and Gagnon, R. 1999. Comparing bird assemblages in
successional black spruce stands originating from fire and logging. - Canadian
Journal of Zoology 77: 1850-1860.
Irlandi, E. A., Ambrose, W. G. and Orlando, B. A. 1995. Landscape ecology and the
marine-environment - how spatial configuration of seagrass habitat influences
growth and survival of the bay scallop. - Oikos 72: 307-313.
Jacqmain, H., Bélanger, L., Hilton, S. and Bouthillier, L. 2007. Bridging native and
scientific observations of snowshoe hare habitat restoration after clearcutting to set
174
wildlife habitat management guidelines on Waswanipi Cree land. - Canadian
Journal of Forest Research 37: 530-539.
Johnson, D. H. 1980. The comparison of usage and availability measurements for
evaluating resource preference. - Ecology 61: 65-71.
Jones, C. G., Lawton, J. H. and Shachak, M. 1994. Organisms as ecosystem engineers. -
Oikos 69: 373-386.
Jones, M., Mandelik, Y. and Dayan, T. 2001. Coexistence of temporally partitioned spiny
mice: Roles of habitat structure and foraging behavior. - Ecology 82: 2164-2176.
Keigley, R. B. and Frisina, M. R. 1998. Browse evaluation by analysis of growth form. -
Montana Fish, Wildlife, and Parks, p. 149.
Keigley, R. B., Frisina, M. R. and Fager, C. 2003. A method for determining the onset year
of intense browsing. - Journal of Range Management 56: 33-38.
Keith, L. B. and Windberg, L. A. 1978. A demographic analysis of the snowshoe hare
cycle. - Wildlife Monographs 58: 6-70.
Kenward, M. G. and Roger, G. H. 1997. Small sample inference for fixed effects from
restricted maximum likelihood. - Biometrics 53: 983 - 997.
Kight, C. R. and Swaddle, J. P. 2007. Associations of anthropogenic activity and
disturbance with fitness metrics of eastern bluebirds (Sialia sialis). - Biological
Conservation 138: 189-197.
Kirkland, G. L. 1990. Patterns of initial small mammal community change after
clearcutting of temperate North-American forests. - Oikos 59: 313-320.
Klemola, T., Tanhuanpaa, M., Korpimaki, E. and Ruohomaki, K. 2002. Specialist and
generalist natural enemies as an explanation for geographical gradients in
population cycles of northern herbivores. - Oikos 99: 83-94.
Klenner, W. and Sullivan, T. P. 2003. Partial and clear-cut harvesting of high-elevation
spruce-fir forests: implications for small mammal communities. - Canadian Journal
of Forest Research 33: 2283-2296.
Klenner, W. and Sullivan, T. P. 2009. Partial and clearcut harvesting of dry Douglas-fir
forests: implications for small mammal communities. - Forest Ecology and
Management 257: 1078-1086.
175
Klvana, I., Berteaux, D. and Cazelles, B. 2004. Porcupine feeding scars and climatic data
show ecosystem effects of the solar cycle. - American Naturalist 164: 283-297.
Kneeshaw, D., Bergeron, Y. and De Grandpré, L. 1998. Early response of abies balsamea
seedlings to artificially created openings. - Journal of Vegetation Science 9: 543-
550.
Kneeshaw, D. and Gauthier, S. 2003. Old growth in the boreal forest: a dynamic
perspective at the stand and landscape level. - Environmental Reviews 11: S11-
S114.
Kneeshaw, D. D. and Bergeron, Y. 1998. Canopy gap characteristics and tree replacement
in the southeastern boreal forest. - Ecology 79: 783-794.
Knight, T. W. and Morris, D. W. 1996. How many habitats do landscapes contain? -
Ecology 77: 1756-1764.
Koehler, G. M. 1990. Population and habitat characteristics of lynx and snowshoe hares in
north central Washington. - Canadian Journal of Zoology 68: 845-851.
Koehler, G. M. 1991. Snowshoe hare, Lepus americanus, use of forest successional stages
and population-changes during 1985-1989 in North-Central Washington. - Canadian
Field-Naturalist 105: 291-293.
Kotler, B. P. and Blaustein, L. 1995. Titrating food and safety in a heterogeneous
environment: when are the risky and safe patches of equal value? - Oikos 74: 251-
258.
Kotliar, N. B. and Wiens, J. A. 1990. Multiple scales of patchiness and patch structure - a
hierarchical framework for the study of heterogeneity. - Oikos 59: 253-260.
Krasnov, B. R., Khokhlova, I. S. and Shenbrot, G. I. 2003. Density-dependent host
selection in ectoparasites: an application of isodar theory to fleas parasitizing
rodents. - Oecologia 134: 365-372.
Krebs, C. J., Boonstra, R., Boutin, S. and Sinclair, A. R. E. 2001a. What drives the 10-year
cycle of snowshoe hares? - Bioscience 51: 25-35.
Krebs, C. J., Boonstra, R., Nams, V., O'donoghue, M., Hodges, K. E. and Boutin, S. 2001b.
Estimating snowshoe hare population density from pellet plots: a further evaluation.
- Canadian Journal of Zoology 79: 1-4.
176
Krebs, C. J., Boutin, S., Boonstra, R., Sinclair, A. R. E., Smith, J. N. M., Dale, M. R. T.,
Martin, K. and Turkington, R. 1995. Impact of food and predation on the snowshoe
hare cycle. - Science 269: 1112-1115.
Krebs, C. J., Gilbert, B. S., Boutin, S. and Boonstra, R. 1987. Estimation of snowshoe hare
population-density from turd transects. - Canadian Journal of Zoology 65: 565-567.
Krebs, C. J., Gilbert, B. S., Boutin, S., Sinclair, A. R. E. and Smith, J. N. M. 1986.
Population biology of snowshoe hares .1. Demography of food-supplemented
populations in the Southern Yukon, 1976-84. - Journal of Animal Ecology 55: 963-
982.
Kuijper, D. P. J., Cromsigt, J., Churski, M., Adam, B., Jedrzejewska, B. and Jedrzejewski,
W. 2009. Do ungulates preferentially feed in forest gaps in European temperate
forest? - Forest Ecology and Management 258: 1528-1535.
Kuuluvainen, T. 2002. Natural variability of forests as a reference for restoring and
managing biological diversity in boreal Fennoscandia. - Silva Fennica 36: 97-125.
Lagos, V. O., Contreras, L. C., Meserve, P. L., Gutierrez, J. R. and Jaksic, F. M. 1995.
Effects of predation risk on space use by small mammals: a field experiment with a
Neotropical rodent. - Oikos 74: 259-264.
Larson, M. A., Thompson III, F. R., Millspaugh, J. J., Dijak, W. D. and Shifley, S. R. 2004.
Linking population viability, habitat suitability, and landscape simulation models
for conservation planning. - Ecological Modelling 180: 103-118.
Laundré, J. W., Hernandez, L. and Altendorf, K. B. 2001. Wolves, elk, and bison:
reestablishing the "landscape of fear" in Yellowstone National Park, USA. -
Canadian Journal of Zoology 79: 1401-1409.
Le Blanc, M.-L. 2009. Réaction des micromammifères et des oiseaux à différentes
intensités de coupe en forêt boréale irrégulière. M.Sc. Université Laval. Québec. 50
pp.
Le Blanc, M.-L., Fortin, D., Darveau, M. and Ruel, J.-C. 2010. Short term response of
small mammals and forest birds to silvicultural practices differing in tree retention
in irregular boreal forests. - Écoscience 17: 334-342.
177
Lecomte, N., Careau, V., Gauthier, G. and Giroux, J. F. 2008. Predator behaviour and
predation risk in the heterogeneous Arctic environment. - Journal of Animal
Ecology 77: 439-447.
Lehmann, E. L. 1998. Nonparametrics: statistical methods based on ranks. - Prentice Hall.
fore Québec. 150 pp.
Lertzman, K. P. and Krebs, C. J. 1991. Gap-phase structure of a sub-alpine old-growth
forest. - Canadian Journal of Forest Research 21: 1730-1741.
Lima, S. L. 1985. Maximizing feeding efficiency and minimizing time exposed to predators
- a trade-off in the Black-Capped Chickadee. - Oecologia 66: 60-67.
Lima, S. L. 1998. Nonlethal effects in the ecology of predator-prey interactions - what are
the ecological effects of anti-predator decision-making? - Bioscience 48: 25-34.
Lima, S. L. and Dill, L. M. 1990. Behavioral decisions made under the risk of predation - a
review and prospectus. - Canadian Journal of Zoology 68: 619-640.
Lin, Y. T. K. and Batzli, G. O. 2002. The cost of habitat selection in prairie voles: an
empirical assessment using isodar analysis. - Evolutionary Ecology 16: 387-397.
Litvaitis, J. A., Sherburne, J. A. and Bissonette, J. A. 1985. Influence of understory
characteristics on snowshoe hare habitat use and density. - Journal of Wildlife
Management 49: 866-873.
Liu, C., Ruel, J. C. and Zhang, S. Y. 2007. Immediate impacts of partial cutting strategies
on stand characteristics and value. - Forest Ecology and Management 250: 148-155.
MacArthur, R. H. and Pinka, E. R. 1966. On optimal use of a patchy environment. - The
American Naturalist 100: 603-609.
Macia, S. and Robinson, M. P. 2005. Effects of habitat heterogeneity in seagrass beds on
grazing patterns of parrotfishes. - Marine Ecology-Progress Series 303: 113-121.
Mallik, A. U. 2003. Conifer regeneration problems in boreal and temperate forests with
ericaceous understory: role of disturbance, seedbed limitation, and keytsone species
change. - Critical Reviews in Plant Sciences 22: 341-366.
Manly, B. F. J., McDonald, L. L., Thomas, D. L., McDonald, T. L. and Erickson, W. P.
2002. Resource selection by animals: statistical analysis and design for field studies.
- Kluwer Press.
178
Massé, A. and Côté, S. D. 2009. Habitat selection of a large herbivore at high density and
without predation: trade-off between forage and cover? - Journal of Mammalogy
90: 961-970.
McCann, N. P., Moen, R. A. and Niemi, G. J. 2008. Using pellet counts to estimate
snowshoe hare numbers in Minnesota. - Journal of Wildlife Management 72: 955-
958.
McCarthy, J. W. 2001. Gap dynamics of forest trees: a review with particular attention to
boreal forests. - Environmental Reviews 9: 1-59.
McNamara, J. M. and Houston, A. I. 1987. Starvation and predation as factors limiting
population size. - Ecology 68: 1515-1519.
McRae, D. J., Duchesne, L. C., Freedman, B., Lynham, T. J. and Woodley, S. 2001.
Comparisons between wildfire and forest harvesting and their implications in forest
management. - Environmental Reviews 9: 223-260.
Menzel, M. A., Ford, W. M., Laerm, J. and Krishon, D. 1999. Forest to wildlife opening:
habitat gradient analysis among small mammals in the southern Appalachians. -
Forest Ecology and Management 114: 227-232.
Messier, C., Bigué, B. and Bernier, L. 2003. Using fast-growing plantations to promote
forest ecosystem protection in Canada. - Unasylva 54-63.
Mills, L. S., Griffin, P. C., Hodges, K. E., Mckelvey, K., Ruggiero, L. and Ulizio, T. 2005.
Pellet count indices compared to mark-recapture estimates for evaluating snowshoe
hare density. - Journal of Wildlife Management 69: 1053-1062.
Mitchell, W. A. and Lima, S. L. 2002. Predator-prey shell games: large-scale movement
and its implications for decision-making by prey. - Oikos 99: 249-259.
Monthey, R. W. 1986. Responses of snowshoe hares, Lepus americanus, to timber
harvesting in northern Maine. - The Canadian Field-Naturalist 100: 568-570.
Morris, D. W. 1987. Ecological scale and habitat use. - Ecology 68: 362-369.
Morris, D. W. 1988. Habitat-dependent population regulation and community structure. -
Evolutionary Ecology 2: 253-269.
Morris, D. W. 1989. Habitat-dependent estimates of competitive interaction. - Oikos 55:
111-120.
179
Morris, D. W. 1990. Temporal variation, habitat selection and community structure. -
Oikos 59: 303-312.
Morris, D. W. 1992. Scales and costs of habitat selection in heterogeneous landscapes. -
Evolutionary Ecology 6: 412-432.
Morris, D. W. 1994. Habitat matching - Alternatives and implications to populations and
communities. - Evolutionary Ecology 8: 387-406.
Morris, D. W. 1996. Coexistence of specialist and generalist rodents via habitat selection. -
Ecology 77: 2352-2364.
Morris, D. W. 2003a. How can we apply theories of habitat selection to wildlife
conservation and management? - Wildlife Research 30: 303-319.
Morris, D. W. 2003b. Toward an ecological synthesis: a case for habitat selection. -
Oecologia 136: 1-13.
Morris, D. W. 2005. Habitat-dependent foraging in a classic predator-prey system: a fable
from snowshoe hares. - Oikos 109: 239-254.
Morris, D. W. 2006. Ecology - Moving to the ideal free home. - Nature 443: 645-646.
Morris, D. W. and Kingston, S. R. 2002. Predicting future threats to biodiversity from
habitat selection by humans. - Evolutionary Ecology Research 4: 787-810.
Mosseler, A., Thompson, I. and Pendrel, B. A. 2003. Overview of old-growth forests in
Canada from a science perspective. - Environmental Reviews 11: 1-7.
Murray, D. L. 2000. A geographic analysis of snowshoe hare population demography. -
Canadian Journal of Zoology 78: 1207-1217.
Murray, D. L. 2003. Snowshoe hare and other hares. - In: G.A. Feldhamer, B.C. Thompson
and J.A. Chapman (eds.), Wild Mammals of North America: Biology, Management
and Conservation. John Hopkins University Press, pp. 147-175.
Murray, D. L., Roth, J. D., Ellsworth, E., Wirsing, A. J. and Steury, T. D. 2002. Estimating
low-density snowshoe hare populations using fecal pellet counts. - Canadian Journal
of Zoology 80: 771-781.
Mysterud, A., Lian, L. B. and Hjermann, D. O. 1999. Scale-dependent trade-offs in
foraging by European roe deer (Capreolus capreolus) during winter. - Canadian
Journal of Zoology 77: 1486-1493.
180
Newbury, T. L. and Simon, N. P. P. 2005. The effects of clearcutting on snowshoe hare
(Lepus americanus) relative abundance in central Labrador. - Forest Ecology and
Management 210: 131-142.
Niemela, J. 1999. Management in relation to disturbance in the boreal forest. - Forest
Ecology and Management 115: 127-134.
Nonaka, E. and Holme, P. 2007. Agent-based model approach to optimal foraging in
heterogeneous landscapes: effects of patch clumpiness. - Ecography 30: 777-788.
Norghauer, J. M., Malcolm, J. R. and Zimmerman, B. L. 2008. Canopy cover mediates
interactions between a specialist caterpillar and seedlings of a neotropical tree. -
Journal of Ecology 96: 103-113.
Nudds, T. D. 1977. Quantifying the vegetative structure wildlife cover. - Wildlife Society
Bulletin 5: 113-117.
O'Donoghue, M., Boutin, S., Krebs, C. J., Zuleta, G., Murray, D. L. and Hofer, E. J. 1998.
Functional responses of coyotes and lynx to the snowshoe hare cycle. - Ecology 79:
1193-1208.
Odum, E. 1969. The strategy of ecosystem development. - Science 164: 262-270.
Orelien, J. G. and Edwards, L. J. 2008. Fixed-effect variable selection in linear mixed
models using R-2 statistics. - Computational Statistics & Data Analysis 52: 1896-
1907.
Orrock, J. L., Pagels, J. F., McShea, W. J. and Harper, E. K. 2000. Predicting presence and
abundance of a small mammal species: the effect of scale and resolution. -
Ecological Applications 10: 1356-1366.
Paine, R. T. and Levin, S. A. 1981. Inter-tidal landscapes - Disturbance and the dynamics
of pattern. - Ecological Monographs 51: 145-178.
Palo, R. T., Bergstrom, R. and Danell, K. 1992. Digestibility, distribution of phenols, and
fiber at different twig diameters of birch in winter - implication for browsers. -
Oikos 65: 450-454.
Paragi, T. F., Johnson, W. N., Katnik, D. D. and Magoun, A. J. 1997. Selection of post-fire
seres by lynx and snowshoe hares in the Alaskan Taiga. - Northwestern Naturalist
78: 77-86.
181
Payette, S., Boudreau, S., Morneau, C. and Pitre, N. 2004. Long-term interactions between
migratory caribou, wildfires and nunavik hunters inferred from tree rings. - Ambio
33: 482-486.
Pease, J. L., Vowles, R. H. and Keith, L. B. 1979. Interaction of snowshoe hares and woody
vegetation. - Journal of Wildlife Management 43: 43-60.
Pedersen, B. S. and Wallis, A. M. 2004. Effects of white-tailed deer herbivory on forest gap
dynamics in a wildlife preserve, Pennsylvania, USA. - Natural Areas Journal 24:
82-94.
Perrin, M. R. 1979. Roles of reproduction, survival, and territoriality in the seasonal
dynamics of Clethrionomys-gapperi populations. - Acta Theriologica 24: 475-500.
Pham, A. T., De Grandpré, L., Gauthier, S. and Bergeron, Y. 2004. Gap dynamics and
replacement patterns in gaps of the northeastern boreal forest of Quebec. - Canadian
Journal of Forest Research 34: 353-364.
Pickett, S. T. A., Kolasa, J., Armesto, J. J. and Collins, S. L. 1989. The ecological concept
of disturbance and its expression at various hierarchical levels. - Oikos 54: 129-136.
Pickett, S. T. A. and White, P. S. 1985. The ecology of natural disturbance and patch
dynamics. - Academic Press Inc.
Pietz, P. J. and Tester, J. R. 1983. Habitat selection by snowshoe hares in north central
Minnesota. - Journal of Wildlife Management 47: 686-696.
Pitt, W. C. and Ritchie, M. E. 2002. Influence of prey distribution on the functional
response of lizards. - Oikos 96: 157-163.
Pothier, D. 2002. Twenty-year results of precommercial thinning in a balsam fir stand. -
Forest Ecology and Management 168: 177-186.
Potvin, F. 1995. L'inventaire du brout: revue des methodes et description des deux
techniques. - Gouvernement du Quebec, Ministere de l'environnement et de la
faune, Direction de la faune et des habitats, p. 70.
Potvin, F., Breton, L. and Courtois, R. 2005. Response of beaver, moose, and snowshoe
hare to clear-cutting in a Quebec boreal forest: a reassessment 10 years after cut. -
Canadian Journal of Forest Research 35: 151-160.
Pusenius, J. and Schmidt, K. A. 2002. The effects of habitat manipulation on population
distribution and foraging behavior in meadow voles. - Oikos 98: 251-262.
182
Quenette, P. Y., Ferron, J. and Sirois, L. 1997. Group foraging in snowshoe hares (Lepus
americanus): aggregation or social group? - Behavioural Processes 41: 29-37.
R Core Development Team. 2009. R: a language and environment for statistical computing.
Version 9.0. R Foundation for Statistical Computing. Vienna, Austria.
Ramp, D. and Coulson, G. 2002. Density dependence in foraging habitat preference of
eastern grey kangaroos. - Oikos 98: 393-402.
Raymond, P., Bedard, S., Roy, V., Larouche, C. and Tremblay, S. 2009. The irregular
shelterwood system: review, classification, and potential application to forests
affected by partial disturbances. - Journal of Forestry 107: 405-413.
Ripple, W. J. and Beschta, R. L. 2003. Wolf reintroduction, predation risk, and cottonwood
recovery in Yellowstone National Park. - Forest Ecology and Management 184:
299-313.
Ripple, W. J. and Beschta, R. L. 2004. Wolves and the ecology of fear: can predation risk
structure ecosystems? - Bioscience 54: 755-766.
Ripple, W. J. and Beschta, R. L. 2006. Linking wolves to willows via risk-sensitive
foraging by ungulates in the northern Yellowstone ecosystem. - Forest Ecology and
Management 230: 96-106.
Ripple, W. J. and Beschta, R. L. 2007. Restoring Yellowstone's aspen with wolves. -
Biological Conservation 138: 514-519.
Robinson, L. 2006. Ecological relationships among partial harvesting, vegetation,
snowshoe hares, and Canada lynx in Maine. M.Sc. University of Maine. Orono,
ME. 204 pp.
Robitaille, A. and Saucier, J.-P. 1998. Paysages régionaux du Québec méridional. -
Gouvernement du Québec, Ministère des ressources naturelles.
Rodriguez, M. A. 1995. Habitat-specific estimates of competition in stream salmonids - a
field-test of the isodar model of habitat selection. - Evolutionary Ecology 9: 169-
184.
Rogowitz, G. L. 1987. Forage quality and use of reforested habitats by snowshoe hares. -
Canadian Journal of Zoology 66: 2080-2083.
Rohner, C. and Krebs, C. J. 1996. Owl predation on snowshoe hares: consequences of
antipredator behaviour. - Oecologia 108: 303-310.
183
Rosenvald, R. and Lohmus, A. 2008. For what, when, and where is green-tree retention
better than clear-cutting? A review of the biodiversity aspects. - Forest Ecology and
Management 255: 1-15.
Royo, A. A., Collins, R., Adams, M. B., Kirschbaum, C. and Carson, W. P. 2010. Pervasive
interactions between ungulate browsers and disturbance regimes promote temperate
forest herbaceous diversity. - Ecology 91: 93-105.
Ruel, J. C., Roy, V., Lussier, J. M., Pothier, D., Meek, P. and Fortin, D. 2007. Development
of a silviculture adapted to the irregular boreal forest. - Forestry Chronicle 83: 367-
374.
Runkle, J. R. 1981. Gap regeneration in some old-growth forests of the Eastern-United-
States. - Ecology 62: 1041-1051.
Runkle, J. R. 1982. Patterns of disturbance in some old-growth mesic forests of eastern
North-America. - Ecology 63: 1533-1546.
Sakai, H. F. and Noon, B. R. 1993. Dusky-footed woodrat abundance in different-aged
forests in northwestern California. - Journal of Wildlife Management 57: 373-382.
SAS Institute Inc. 2003. SAS. Version 9.1. SAS Institute Inc. Cary, Norht Carolina.
Saucier, J.-P., Bergeron, J.-F., Grondin, P. and Robitaille, A. 1998. Les régions écologiques
du Québec méridional (3e version) : un des éléments du système hiérarchique de
classification écologique du territoire mis au point par le ministère des Ressources
naturelles du Québec. L'Aubelle, pp. S1-S12.
Schieck, J. and Song, S. J. 2006. Changes in bird communities throughout succession
following fire and harvest in boreal forests of western North America: literature
review and meta-analyses. - Canadian Journal of Forest Research 36: 1299-1318.
Schmitz, O. J. 2005. Scaling from plot experiments to landscapes: studying grasshoppers to
inform forest ecosystem management. - Oecologia 145: 225-234.
Schmitz, O. J., Beckerman, A. P. and Obrien, K. M. 1997. Behaviorally mediated trophic
cascades: effects of predation risk on food web interactions. - Ecology 78: 1388-
1399.
Schreiner, E. G., Krueger, K. A., Happe, P. J. and Houston, D. B. 1996. Understory patch
dynamics and ungulate herbivory in old-growth forests of Olympic National Park,
Washington. - Canadian Journal of Forest Research 26: 255-265.
184
Searle, K. R., Hobbs, N. T. and Gordow, I. J. 2007. It's the "foodscape", not the landscape:
using foraging behavior to make functional assessments of landscape condition. -
Israel Journal of Ecology & Evolution 53: 297-316.
Senft, R. L., Coughenour, M. B., Bailey, D. W., Rittenhouse, L. R., Sala, O. E. and Swift,
D. M. 1987. Large herbivore foraging and ecological hierarchies. - Bioscience 37:
789-&.
Shenbrot, G. 2004. Habitat selection in a seasonally variable environment: test of the isodar
theory with the fat sand rat, Psammomys obesus, in the Negev Desert, Israel. -
Oikos 106: 359-365.
Shenbrot, G. and Krasnov, B. 2000. Habitat selection along an environmental gradient:
theoretical models with an example of Negev Desert rodents. - Evolutionary
Ecology Research 2: 257-277.
Sheriff, M. J., Krebs, C. J. and Boonstra, R. 2009. The sensitive hare: sublethal effects of
predator stress on reproduction in snowshoe hares. - The Journal of animal ecology
78: 1249-1258.
Shipley, L. A. and Spalinger, D. E. 1995. Influence of size and density of browse patches
on intake rates and foraging decisions of young moose and white-tailed deer. -
Oecologia 104: 112-121.
Shochat, E., Patten, M. A., Morris, D. W., Reinking, D. L., Wolfe, D. H. and Sherrod, S. K.
2005. Ecological traps in isodars: effects of tallgrass prairie management on bird
nest success. - Oikos 111: 159-169.
Sih, A. 1980. Optimal behavior: can foragers balance two conflicting demands? - Science
210: 1041-1043.
Simon, N. P. P. and Schwab, F. E. 2005. The response of conifer and broad-leaved trees
and shrubs to wildfire and clearcut logging in the boreal forests of central labrador. -
Northern Journal of Applied Forestry 22: 35-41.
Simon, N. P. P., Stratton, C. B., Forbes, G. J. and Schwab, F. E. 2002. Similarity of small
mammal abundance in post-fire and clearcut forests. - Forest Ecology and
Management 165: 163-172.
185
Sinclair, A. R. E. 2003. Mammal population regulation, keystone processes and ecosystem
dynamics. - Philosophical Transactions of the Royal Society of London Series B-
Biological Sciences 358: 1729-1740.
Sinclair, A. R. E., Gosline, J. M., Holdsworth, G., Krebs, C. J., Boutin, S., Smith, J. N. M.,
Boonstra, R. and Dale, M. 1993. Can the solar-cycle and climate synchronize the
snowshoe hare cycle in canada - evidence from tree rings and ice cores. - American
Naturalist 141: 173-198.
Sinclair, A. R. E., Krebs, C. J., Fryxell, J. M., Turkington, R., Boutin, S., Boonstra, R.,
Seccombe-Hett, P., Lundberg, P. and Oksanen, L. 2000. Testing hypotheses of
trophic level interactions: a boreal forest ecosystem. - Oikos 89: 313-328.
Sinclair, A. R. E., Krebs, C. J., Smith, J. N. M. and Boutin, S. 1988. Population biology of
snowshoe hares .3. Nutrition, plant secondary compounds and food limitation. -
Journal of Animal Ecology 57: 787-806.
Smith, C. H. 1983. Spatial trends in canadian snowshoe hare, Lepus americanus,
population-cycles. - Canadian Field-Naturalist 97: 151-160.
Smith, J. N. M., Krebs, C. J., Sinclair, A. R. E. and Boonstra, R. 1988. Population biology
of snowshoe hares .2. Interactions with winter food plants. - Journal of Animal
Ecology 57: 269-286.
Sokal, R. R. and Rohlf, F. J. 1995. Biometry: the principles and practice of statistics in
biological research. - W.H. Freeman.
Sousa, W. P. 1984. The role of disturbance in natural communities. - Annual Review of
Ecology and Systematics 15: 353-391.
St-Laurent, M. H., Cusson, M., Ferron, J. and Caron, A. 2008. Use of residual forest by
snowshoe hare in a clear-cut boreal landscape. - Northeastern Naturalist 15: 497-
514.
Stephens, D. W. and Krebs, J. R. 1986. Foraging Theory. - Princeton University Press.
Sullivan, T. P., Sullivan, D. S. and Lindgren, P. M. F. 2001. Influence of variable retention
harvests on forest ecosystems. I. Diversity of stand structure. - Journal of Applied
Ecology 38: 1221-1233.
186
Sullivan, T. P., Sullivan, D. S. and Lindgren, P. M. F. 2008. Influence of variable retention
harvests on forest ecosystems: plant and mammal responses up to 8 years post-
harvest. - Forest Ecology and Management 254: 239-254.
Thompson, I. D. 2006. Monitoring of biodiversity indicators in boreal forests: a need for
improved focus. - Environmental Monitoring and Assessment 121: 261-271.
Thompson, I. D., Davidson, I. J., Odonnell, S. and Brazeau, F. 1989. Use of track transects
to measure the relative occurrence of some boreal mammals in uncut forest and
regeneration stands. - Canadian Journal of Zoology 67: 1816-1823.
Tufto, J., Andersen, R. and Linnell, J. 1996. Habitat use and ecological correlates of home
range size in a small cervid: the roe deer. - Journal of Animal Ecology 65: 715-724.
Underwood, A. J. 1994. On beyond BACI: sampling designs that might reliably detect
environmental disturbances. - Ecological Applications 4: 3-15.
van der Merwe, M. and Brown, J. S. 2008. Mapping the landscape of fear of the Cape
ground squirrel (Xerus Inauris). - Journal of Mammalogy 89: 1162-1169.
Van Wagner, C. E. 1978. Age-class distribution and the forest fire cycle. - Canadian
Journal of Forest Research 8: 220-227.
Vanderwel, M. C., Malcolm, J. R. and Mills, S. C. 2007. A meta-analysis of bird responses
to uniform partial harvesting across North America. - Conservation Biology 21:
1230-1240.
Vanderwel, M. C., Mills, S. C. and Malcolm, J. R. 2009. Effects of partial harvesting on
vertebrate species associated with late-successional forests in Ontario's boreal
region. - Forestry Chronicle 85: 91-104.
Vanhorne, B. 1981. Demography of Peromyscus-maniculatus populations in seral stages of
coastal coniferous forest in southeast Alaska. - Canadian Journal of Zoology 59:
1045-1061.
Vasquez, R. A., Ebensperger, L. A. and Bozinovic, F. 2002. The influence of habitat on
travel speed, intermittent locomotion, and vigilance in a diurnal rodent. - Behavioral
Ecology 13: 182-187.
Vila, B., Torre, F., Guibal, F. and Martin, J. L. 2004. Can we reconstruct browsing history
and how far back? Lessons from Vaccinium parvifolium Smith in Rees. - Forest
Ecology and Management 201: 171-185.
187
Vivas, H. J., Saether, B. E. and Andersen, R. 1991. Optimal twig-size selection of a
generalist herbivore, the moose alces-alces - implications for plant herbivore
interactions. - Journal of Animal Ecology 60: 395-408.
Waterhouse, F. L., Harestad, A. S. and Ott, P. K. 2002. Use of small streams and forest
gaps for breeding habitats by winter wrens in coastal British Columbia. - Northwest
Science 76: 335-346.
Weir, J. M. H., Johnson, E. A. and Miyanishi, K. 2000. Fire frequency and the spatial age
mosaic of the mixed-wood boreal forest in western Canada. - Ecological
Applications 10: 1162-1177.
Wiens, J. A. 1976. Population responses to patchy environments. - Annual Review of
Ecology and Systematics 7: 81-120.
Willig, M. R., Bloch, C. P., Brokaw, N., Higgins, C., Thompson, J. and Zimmermann, C.
R. 2007. Cross-scale responses of biodiversity to hurricane and anthropogenic
disturbance in a tropical forest. - Ecosystems 10: 824-838.
Wirsing, A. J., Heithaus, M. R. and Dill, L. M. 2007. Fear factor: do dugongs (Dugong
dugon) trade food for safety from tiger sharks (Galeocerdo cuvier)? - Oecologia
153: 1031-1040.
Wirsing, A. J. and Murray, D. L. 2002. Patterns in consumption of woody plants by
snowshoe hares in the northwestern United States. - Ecoscience 9: 440-449.
Wirsing, A. J., Steury, T. D. and Murray, D. L. 2002. A demographic analysis of a southern
snowshoe hare population in a fragmented habitat: evaluating the refugium model. -
Canadian Journal of Zoology 80: 169-177.
Wirth, C., Messier, C., Bergeron, Y., Frank, D. and Fankhänel, A. 2009. Old-growth forest
definitions: a pragmatic view. - In: Old-Growth Forests. Springer-Verlag, pp. 11-33.
Wisdom, M. J., Vavra, M., Boyd, J. M., Hemstrom, M. A., Ager, A. A. and Johnson, B. K.
2006. Understanding ungulate herbivory-episodic disturbance effects on vegetation
dynamics: knowledge gaps and management needs. - Wildlife Society Bulletin 34:
283-292.
Wolfe, M. L., Debyle, N. V., Winchell, C. S. and Mccabe, T. R. 1982. Snowshoe hare
cover relationships in Northern Utah. - Journal of Wildlife Management 46: 662-
670.
188
Wolff, J. O. 1980. The role of habitat patchiness in the population dynamics of snowshoe
hares. - Ecological Monographs 50: 111-130.
Wood, S. 2009. mgcv: GAMs with GCV/AIC/REML smoothness estimation and GAMMs
by PQL. - R package version 1.5-5.
Work, T. T., Jacobs, J. M., Spence, J. R. and Volney, W. J. 2010. High levels of green-tree
retention are required to preserve ground beetle biodiversity in boreal mixedwood
forests. - Ecological Applications 20: 741-751.
Zwolak, R. 2009. A meta-analysis of the effects of wildfire, clearcutting, and partial harvest
on the abundance of North American small mammals. - Forest Ecology and
Management 258: 539-545.
189
Appendice 1
a) Sample photographs of different stand ages sampled within the post-harvest/post-
fire forest chronosequence
11 yr-old clearcut-origin stand
24 yr-old clearcut-origin stand
34 yr-old clearcut-origin stand
43 yr-old clearcut-origin stand
84 yr-old fire-origin stand
186 yr-old fire origin stand
190
b) Simulations to illustrate the effect of forest age-class distribution on snowshoe hare
abundance under three disturbance regimes
Snowshoe hare abundance was estimated for three hypothetical 1000 ha landscapes with
forest age-class structures generated from three different disturbance regimes: 1) a 250-year
forest fire-return-interval with no human management, 2) fully-regulated even-aged
management with a 100-year harvest rotation, and 3) cohort management (Bergeron et al.
2002) based on a 200-year fire cycle and a maximum rotation age of 100 years for stands
under even-aged management. Each 1000 ha landscape was broken down into 10-year age
classes. Snowshoe hare pellet density in each 10-year age class was assigned based on the
GAM curve describing the relationship between snowshoe hare pellet density and stand age
(Figure 1.2). Stands ≥ 260 years were grouped into one category ("260+") because the
GAM function estimated pellet density only up to 265 years (which was the oldest stand
sampled in the forest chronosequence). Pellet densities estimated for each 10-year age
class were then converted into snowshoe hare density (D, hares/ha) based on the regression
equation developed by Krebs et al. 2001:
D = 1.567 × e[-1.203 + 0.889 x ln(pellets/m² * 0.155 m²)]
The total hare population for each hypothetical landscape was then calculated by taking the
sum of hare density estimated in each 10-year age class multiplied by the total area of the
landscape in each age class.
The first forest landscape was based on a negative exponential age-class distribution
generated by a 250-year fire cycle using the formula developed by Van Wagner (1978),
where the cumulative proportion of the landscape up to a given age x is given by:
Σf(x) = 1 - e-px
191
where p equals fire probability (in this case p = 0.004, for 1 fire in every 250 yrs), and x =
age class x in 10 yr intervals.
Under the negative exponential distribution, mean stand age is equal to the fire cycle
length, and 63.2% of the stands are younger than the mean stand age. For a landscape with
a 250-year fire cycle this means that 37% of the landscape is >250 years old, and roughly
68% is >100 years old. The total population size estimated for this landscape was 89
snowshoe hares (Table A1.1).
The second landscape represents a fully regulated forest under even-aged management with
a harvest rotation of 100 years. Under this management scenario, there is an even
distribution among age-classes, meaning that there is 10% of the landscape in each 10-year
age class up to 100 years, with no forests >100 years. The total population size estimated
for this landscape was 143 snowshoe hares, roughly 40% more than in Landscape 1 (Table
A1.2).
The third landscape was generated using the cohort management approach proposed by
Bergeron et al. (2002). The age-class distribution for this hypothetical 1000 ha landscape is
based the recommended landscape proportions presented in Table 1 in Bergeron et al.
(2002) for a landscape with a 200-year fire cycle and a 100-year maximum rotation age for
stands under even-aged management. In this scenario "stand-initiating" harvesting is used
to recruit even-aged stands <100 years (cohort 1) on 39% of the landscape, partial
harvesting is used to move 24% of the landscape into stands with an uneven or irregular
structure (100-200 years; cohort 2) and selection cutting is used to mimic gap dynamics in
old-growth stands on 37% of the landscape (200-300 years; cohort 3). In this scenario an
even distribution of the landscape among 10-year stand age classes within each cohort was
assumed. The total population size estimated for this landscape was 95 snowshoe hares,
only 6% more than in Landscape 1 (Table A1.3).
192
Table A1.1. Estimated snowshoe hare population size for a hypothetical 1000 ha forest
landscape with a 250-year fire cycle and a negative exponential forest age-class
distribution.
Forest age-
class (years)
Estimated
pellet
density
(pellets/m2) Hares/ha
Cumulative
% of
landscape
Total area
per 10-year
age-class
(ha)
Total no. of
hares per
forest age-
class
10 0.30 0.03 3.92 39.21 1.19
20 0.95 0.09 7.69 37.67 3.22
30 2.21 0.18 11.31 36.20 6.56
40 3.59 0.28 14.79 34.78 9.72
50 3.92 0.30 18.13 33.41 10.09
60 3.03 0.24 21.34 32.10 7.72
70 1.80 0.15 24.42 30.84 4.66
80 0.89 0.08 27.39 29.63 2.39
90 0.45 0.04 30.23 28.47 1.26
100 0.34 0.03 32.97 27.36 0.94
110 0.35 0.04 35.60 26.28 0.94
120 0.38 0.04 38.12 25.25 0.97
130 0.39 0.04 40.55 24.26 0.93
140 0.34 0.03 42.88 23.31 0.81
150 0.28 0.03 45.12 22.40 0.66
160 0.25 0.03 47.27 21.52 0.57
170 0.27 0.03 49.34 20.68 0.58
180 0.35 0.04 51.32 19.86 0.70
190 0.49 0.05 53.23 19.09 0.91
200 0.67 0.06 55.07 18.34 1.16
210 0.88 0.08 56.83 17.62 1.41
220 1.05 0.09 58.52 16.93 1.59
230 1.14 0.10 60.15 16.26 1.64
240 1.10 0.10 61.71 15.63 1.52
250 0.95 0.09 63.21 15.01 1.29
260+ 0.75 0.07 100.00 367.88 25.66
Total hare population: 89.09
193
Table A1.2. Estimated snowshoe hare population size for a hypothetical 1000 ha forest
landscape under fully regulated even-aged management with a harvest rotation of 100
years.
Forest age-
class (years)
Estimated
pellet density
(pellets/m2) Hares/ha
Cumulative
% of
landscape
Total
area per
10-year
age-class
(ha)
Total no.
of hares
per
forest
age-class
10 0.30 0.03 10 100 3.03
20 0.95 0.09 20 100 8.55
30 2.21 0.18 30 100 18.13
40 3.59 0.28 40 100 27.96
50 3.92 0.30 50 100 30.19
60 3.03 0.24 60 100 24.06
70 1.80 0.15 70 100 15.12
80 0.89 0.08 80 100 8.05
90 0.45 0.04 90 100 4.44
100 0.34 0.03 100 100 3.44
Total hare population: 142.97
194
Table A1.3. Estimated snowshoe hare population size for a hypothetical 1000 ha forest
landscape under cohort management assuming a 200-year fire cycle and a 100-year
maximum harvest rotation age for stands under even-aged management (following
Bergeron et al. 2002).
Cohort
Forest
age-class
(years)
Estimated
pellet
density
(pellets/m2) Hares/ha
Cumulative
% of
landscape
Total area
per 10-
year age-
class (ha)
Total no.
of hares
per forest
age-class
1
10 0.30 0.03 3.90 39.0 1.18
20 0.95 0.09 7.80 39.0 3.33
30 2.21 0.18 11.70 39.0 7.07
40 3.59 0.28 15.60 39.0 10.91
50 3.92 0.30 19.50 39.0 11.78
60 3.03 0.24 23.40 39.0 9.38
70 1.80 0.15 27.30 39.0 5.90
80 0.89 0.08 31.20 39.0 3.14
90 0.45 0.04 35.10 39.0 1.73
100 0.34 0.03 39.00 39.0 1.34
2
110 0.35 0.04 41.40 24.0 0.86
120 0.38 0.04 43.80 24.0 0.92
130 0.39 0.04 46.20 24.0 0.92
140 0.34 0.03 48.60 24.0 0.83
150 0.28 0.03 51.00 24.0 0.70
160 0.25 0.03 53.40 24.0 0.64
170 0.27 0.03 55.80 24.0 0.68
180 0.35 0.04 58.20 24.0 0.85
190 0.49 0.05 60.60 24.0 1.14
200 0.67 0.06 63.00 24.0 1.52
3
210 0.88 0.08 66.70 37.0 2.95
220 1.05 0.09 70.40 37.0 3.47
230 1.14 0.10 74.10 37.0 3.72
240 1.10 0.10 77.80 37.0 3.61
250 0.95 0.09 81.50 37.0 3.18
260+ 0.75 0.07 100.00 185.0 12.90
Total hare population: 94.64
195
Appendice 2
Sample photographs of canopy gaps originating from a) tree mortality and b) edaphic
conditions
a) Mortality-origin canopy gaps
b) Edaphic-origin canopy gaps
196
Appendice 3
Sample photographs of the four browse stem architecture types considered during
browse history surveys.
Uninterrupted
(photo: A. Allard-Duchene)
Arrested
(Photo: A. Allard-Duchene)
Retrogressed
(photo: J.Hodson)
Released
(photo: J.Hodson)