Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The responses of two cavity-nesting species to changes in habitat conditions and nest web community dynamics… Norris, Andrea Rose 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2007-0541.pdf [ 5.97MB ]
Metadata
JSON: 831-1.0074949.json
JSON-LD: 831-1.0074949-ld.json
RDF/XML (Pretty): 831-1.0074949-rdf.xml
RDF/JSON: 831-1.0074949-rdf.json
Turtle: 831-1.0074949-turtle.txt
N-Triples: 831-1.0074949-rdf-ntriples.txt
Original Record: 831-1.0074949-source.json
Full Text
831-1.0074949-fulltext.txt
Citation
831-1.0074949.ris

Full Text

THE RESPONSES OF TWO CAVITY-NESTING SPECIES TO CHANGES IN HABITAT CONDITIONS AND NEST WEB COMMUNITY DYNAMICS IN INTERIOR BRITISH COLUMBIA by Andrea Rose Norris B.Sc., University of Victoria, 2003 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in The Faculty of Graduate Studies (Forestry) THE UNIVERSITY OF BRITISH COLUMBIA May 2007 © Andrea Rose Norris, 2007 ABSTRACT Populations of small-bodied cavity nesters may be regulated by density dependence, interspecific interactions within the community and resource availability. My objectives were to determine how changes in community interactions and habitat conditions affected mountain chickadees and red-breasted nuthatches at the local, regional and nest-patch scales. I used point count surveys and vegetation surveys from 27 forest stands to examine how population densities of mountain chickadees and red-breasted nuthatches varied with: densities of black-capped chickadees, downy woodpeckers, red-naped sapsuckers and red squirrels; densities of aspen, beetle-infected lodgepole pine trees, and all trees; and proportion of edge habitat (naturally fragmented or harvested), at the stand and study area levels from 1997-2006. For mountain chickadees, populations increased following years of high densities of nuthatches and sapsuckers and low densities of squirrels (predators) but were strongly density dependent where densities of sapsuckers and squirrels were high, and at forest edges. For red-breasted nuthatches, populations increased with recent beetle-infected pine tree density and following years of high densities of downy woodpeckers, and decreased after low densities of downy woodpeckers and high densities of black-capped chickadees. I used vegetation surveys of 231 nests and available habitat to examine nuthatch nest-patch selection. As the beetle outbreak progressed, nuthatches selected nest-patches with significantly fewer suitable nest trees and significantly more beetle-infected pine trees. The impacts of a large-scale natural disturbance event had cascading effects on the community and subsequent cavity-nester populations. Thus beetle-management activities should include long-term monitoring programs to examine temporal variability in resources and the effects on community dynamics in order to manage for chickadee and nuthatch populations. ii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iii LIST OF TABLES v LIST OF FIGURES vii ACKNOWLEDGEMENTS x CO-AUTHORSHIP STATEMENT xi CHAPTER 1: GENERAL INTRODUCTION 1 Temporal variation in resource availability 1 Community dynamics 2 Study Area 5 Thesis predictions 6 Thesis overview 7 REFERENCES 9 CHAPTER 2: ECOLOGICAL SYNERGISMS IN SMALL CAVITY-NESTER COMMUNITIES: EFFECTS OF HABITAT, COMPETITION, FACILITATION, AND PREDATION ON MOUNTAIN CHICKADEE POPULATIONS 17 INTRODUCTION..... 17 METHODS 21 Study area 21 Population and habitat monitoring 22 Modeling population densities 24 Modeling approach 24 RESULTS 27 Changes in habitat characteristics 28 Changes in the community 28 Model selection 29 Model averaging 29 Habitat effects 29 Mountain chickadee density dependence 30 Other facilitator effects 31 DISCUSSION : 31 Habitat effects 32 Intra- and Interspecific Competition 33 Facilitation 34 Predation 35 REFERENCES.... 37 iii CHAPTER 3: CASCADING EFFECTS OF A MOUNTAIN PINE BEETLE OUTBREAK ON THE SMALL CAVITY-NESTER COMMUNITY INFLUENCES REGULATION OF RED-BREASTED NUTHATCH POPULATIONS 52 INTRODUCTION 52 METHODS 56 Study area 56 Population and habitat monitoring 56 Modeling population densities 58 Modeling approaches 59 RESULTS 61 Overall changes in populations at the study (regional) level 61 Concurrent correlations with nuthatch populations at time t at the stand level 62 Predictors of nuthatch populations at time t+1 at the stand level 62 DISCUSSION 64 Density dependence affected by interspecific interactions 64 Interspecific interactions 65 Habitat and predator effects 66 REFERENCES '. 68 CHAPTER 4: MOUNTAIN PINE BEETLE PRESENCE AFFECTS NEST PATCH CHOICE OF RED-BREASTED NUTHATCHES 82 INTRODUCTION... . 82 METHODS .84 Study area 84 Nest patch selection.... 84 Statistical analyses 85 RESULTS 86 DISCUSSION 87 Nest patch selection in pre-outbreak years 87 Nest patch selection in outbreak years 88 Implications for populations 89 MANAGEMENT IMPLICATIONS 90 REFERENCES .....90 CHAPTER 5: GENERAL DISCUSSION .97 Conservation and management implications 100 REFERENCES 101 iv LIST OF TABLES Table 1.1. Study sites in the Cariboo-Chilcotin area of interior British Columbia showing survey years for each site where no harvesting took place (Controls) and where a harvesting treatment was applied (Post-cut). Fixed radius point counts for breeding birds and vegetation surveys at permanent sample plots centred at each point count station were conducted annually at each of 27 study sites. Sites selected for harvesting were monitored in pre-cut conditions for two to eight years and monitored in post-cut conditions for one to nine years. Habitat variables measured for each stand: tree density, densities of mountain pine beetle-infected lodgepole pine trees (presence of mountain pine beetle was a proxy for a range of forest insects that comprised changes in food availability), and proportion of stand within 50m of an edge 13 Table 2,1. General linear mixed models predicting population densities of mountain chickadees in Interior British Columbia from 1997 to 2006. Variables examined were mountain chickadee (at the local scale: M; and at the regional scale: RegM), proportion of beetle-infected pine trees (Bi), red-breasted nuthatch (Rb), downy woodpecker (D), red-naped sapsucker (Rn), red squirrel (Sq) abundance, tree density (St) and its quadratic form (St2) and percentage edge (E), and whether the site was harvested (C). Site and year were included as random variables. For each model, the number of parameters (K), the number of observations (N: all site/year combinations for 27 sites and 10 years), the maximum log-likelihood ratio (logLik), the Akaike's Information Criterion corrected for small samples (AICc), the difference in AICc compared to the model with the least AICc value (AAICc), and its weight (AICc wt). Shown are 14 of 35 models examined 42 Table 2.2. Model-averaged parameter estimates for the explanatory variables of mountain chickadee population densities at time, t+1: mountain chickadee (at the local scale: M; and at the regional scale: RegM), mountain pine beetle (Bi), red-breasted nuthatch (Rb), downy woodpecker (D), red-naped sapsucker (Rn), red squirrel (Sq) abundance, tree density (St) and its quadratic transformed variable (St2) and percentage edge (E), at time, t and all two-way interaction terms. Standard errors (SE) and the number of models of 35 examined, in which each of the effects appear (N) are given. Parameters have a significant effect on mountain chickadee growth rates at the a = 0.05 level where t-value > /l .98/, and at the a = 0.1 level where t-value > / l .65/. Sign of estimate indicates a positive or negative effect on mountain chickadee densities in subsequent years 43 Table 3.1. Parameter estimates for the explanatory variables for red-breasted nuthatch densities of at time, t: mountain chickadee (M), black-capped chickadee (Be), downy woodpecker (D), red squirrel (Sq), recent beetle-infected pine (Bi), aspen trees (At), proportion of edge habitat (E), tree density (St) and its quadratic form (St2), and harvested sites (H), with significant correlations with red-breasted nuthatch densities at the 0.05 level of confidence, bolded. 73 Table 3.2. General linear mixed models predicting densities of red-breasted nuthatches, at time, t+1, in central British Columbia from 1997 to 2006. Variables examined were densities of red-breasted nuthatches (Rb), mountain chickadees (M), black-capped chickadees (Be), downy woodpeckers (D), red squirrels (Sq), recent beetle-infected pine trees (Bi), aspen trees (At), all trees (St, and its quadratic form, St2), proportion of edge habitat (E), and instances of harvesting (H), at time, t. Site and year were included as random effects. For each model, the number of parameters (K), the number of observations (n), the maximum log-likelihood ratio (loglik), the Akaike's Information Criterion corrected for small samples (AICc), the difference in AICc compared to the model with the least AICc value (AAICc), and its weight (wt)..... ..74 Table 3.3. Model-averaged parameter estimates for the explanatory variables for red-breasted nuthatch densities (at time t+1) from top two models, where AAICc < 4. Variables examined were densities of red-breasted nuthatches (Rb; at time t), mountain chickadees (M), black-capped chickadees (Be), downy woodpeckers (D), recent beetle-infected pine trees (Bi), aspen trees (At) and an interaction of red-breasted nuthatches with downy woodpeckers (Rb*D). Significant effects on red-breasted nuthatch densities at the 0.05 level of confidence are bolded..: 75 Table 4.1. Generalized linear mixed effects model parameter estimates of predicted red-breasted nuthatch nest patch use with the following fixed effects: Total aspen density, beetle-infected lodgepole pine density, and suitable nest tree density in central British Columbia, Canada •. 94 vi LIST OF FIGURES Figure 1.1. Population growth, or the rate at which a population increases typically declines with density (a). However, if interacting negative effects (-) no longer regulate density dependence, or are masked by positive effects (+), the overall result of density on population growth might appear to be zero (b). This might occur if populations decline drastically and are no longer regulated by competition and predation, or if facilitation increases. By examining habitat conditions and community dynamics, I attempted to delineate the interacting positive and negative effects on populations of mountain chickadee and red-breasted nuthatch 14 Figure 1.2. Map of study area in Cariboo-Chilcotin region of British Columbia. Upper left map shows location within the province (51° 52'N, 122° 21'W), indicated by the shaded box, which is expanded in the larger map showing all study sites. Fifteen of 27 sites were located at Riske Creek, approximately 40 km west of Williams Lake ("Riske Creek sites" SC, RC, RP, RL, DE, SW, TO, MG, YY, ML, SD, SP, LT1. LT2, HH). Twelve sites were approximately 20 km east of Williams Lake ("Knife Creek sites" SHAC, SHCC. 7M, MM, FO, KN, CT, MC, MX, BF, PL D2). Eleven sites were cut with various prescriptions ranging from clear cuts with reserves to selection harvesting of varying proportions of pine and spruce (underlined) 15 Figure 1.3. Schematic diagram of mountain chickadee and red-breasted nuthatch relationships within the small and medium-bodied cavity-nester community: red squirrel (predator), red-naped sapsucker (facilitator), downy woodpecker (facilitator) and black-capped chickadee (competitor). Predicted effects of community members on populations of the two species are either positive (+) or negative (-) and these effects may change with the food pulse (changes highlighted with block arrows): (a) before the mountain pine beetle outbreak, and (b) during the outbreak 16 Figure 2.1. Known or suspected cavity excavators for 221 mountain chickadee nests. Numbers in brackets below each year and next to each species name indicate sample size of nests. If the cavity was freshly excavated the year of the recorded nest, the occupants were assumed to be the excavators. For example, if downy woodpeckers occupied a freshly excavated cavity in 2003 and the nest was used by mountain chickadees in 2004, the excavator was suspected to be downy woodpecker, from 1995 to 2006 across all study sites in central British Columbia 44. Figure 2.2. Mean number of standing trees (>12.5 cm DBH) on uncut and cut sites, and proportion of lodgepole pole pine trees with evidence of recent beetle attack of all trees surveyed per ha across 27 study sites in Interior British Columbia. The dip in proportion of recently mountain pine beetle-infected pine trees on the study from 2000 to 2002 was due to harvesting 7 of 11 sites between 2000 and 2002 45 Figure 2.3. Mean detections of mountain chickadees with potential (a) facilitators, and (b) competitors and predators across point count surveys, on 27 forested sites in Interior British Columbia.... 46 Figure 2.4. Predicted mean detections of mountain chickadees (at time, t+1) with increasing mountain chickadee densities (at time, t), at uncut and cut sites, from 1997-2006. Trend lines vii were calculated by substituting the model-averaged parameter estimates into the full predictive model, ln(Mt+i+0.01) = -2.00 + ln(Mt+0.01) + RegM + Rb + D + Sq + Bi + St +St2 + E + C + all interactions (Tables 2.1, 2.2) 47 Figure 2.5. Mean detections of mountain chickadees (t+1) with (a) increasing tree density from point count and vegetation surveys of 27 stands during 1997-2006, and (b) quadratic-transformed (linearized) tree density using model-averaged parameter estimates substituted into the full predictive model, ln(Mt+i+0.01).= -2.00 + ln(Mt+0.01) + RegM + Rb + D + Sq + Bi + St +S12 + E + C + all interactions (Tables, 2.1, 2.2) 48 Figure 2.6. Predicted mean detections of mountain chickadees (time, t+1) with increasing mountain chickadee densities (time, t) at high (upper 75th percentile of data) and low (lower 25th percentile) proportions of edge habitat. Trend lines were calculated using model-averaged parameter estimates substituted into the full model, ln(Mt+i + 0.01) =-2.00 + ln(Mt+0.01) + RegM + Rb + Rn + D + Sq + Bi + St + St2 + E + C + all interactions (Tables 2.1, 2.2) 49 Figure 2.7. Predicted mean detections of mountain chickadees (t+1) across chickadee densities (time, t) at high (upper 75th percentile of data) and low (lower 25th percentile) (a) Red-naped sapsucker densities, and (b) Red squirrel densities, at uncut and cut stands from 1997-2006. Trend lines were calculated using model-averaged parameter estimates substituted into the full model, ln(M,+i + 0.01) = -2.00 + ln(Mt+0.01) + RegM + Rb + Rn + D + Sq + Bi + St + St2 + E + C + all interactions (Tables 2.1, 2.2) ..' 50 Figure 2.8. Mean detections of mountain chickadees (t+1) with increasing red-breasted nuthatch densities, from point count surveys of 27 stands during 1997-2006. Trend line was calculated using model-averaged parameter estimates substituted into the full predictive model, ln(Mt+i+0.01) = -2.00 + ln(Mt+0.01) + RegM + Rb + D + Sq + Bi + St +St2 + E + C + all interactions (Tables 2.1, 2.2) : 51 Figure 3.1. Mean detections of red-breasted nuthatches, mountain chickadees, black-capped chickadees and red squirrels, across point count surveys from 1997 to 2006, on 27 forested sites in central British Columbia 76 Figure 3.2. Mean detections of red-breasted nuthatches and recent beetle-infected lodgepole pine trees across 27 forested sites using point count surveys, and vegetation surveys of 11.3 metre fixed radius plots centred around each point count station, in central British Columbia from 1997-2006. Mean densities of recent beetle-infected pine trees were proportionate to all trees surveyed in order to obtain an estimate of relative food availability on stands. I calculated a Pearson's correlation coefficient to determine how mean annual densities of nuthatches correlated with proportionate densities of recent beetle-infected pine trees, from 1997 to 2006; R=0.73, n=10, p=0.02 77 Figure 3.3. Mean detections of red-breasted nuthatches and a.) mountain chickadees, b.) black-capped chickadees, c.) red squirrels, and d.) proportion of habitat within 50m of an a forest edge for each of 27 stands, from 1997-2006. Trend lines were calculated using parameter estimates generated by the correlative model, ln(Rbt + 0.1) = M + Bc + D + Sq + Bi + At + E + St + St2 + H + Site + Year 78-79 viii Figure 3.4. Mean detections of red-breasted nuthatches (at time, t+1) and black-capped chickadees (at time, t) at 27 stands, from 1997-2006. Trend line was calculated using model-averaged parameter estimates from the top two predictive models, ln(Rbt+i + 0.1) = ln(Rbt + 0.1) + M + Be + D + Rb*D + Site + Year, and; ln(Rb,+1 + 0.1) = ln(Rbt + 0.1) + M + Be + D + Bi + At + Rb*D + Site + Year. Trend line indicates that nuthatch populations in subsequent years decreased with increasing black-capped chickadees (B = -0.86 ±0.36, ti96=-1.98, Figure 3.5. Predicted mean detections of red-breasted nuthatches (t+1) with increasing conspecific densities (t) at high (upper 75th percentile of data) and low (lower 65th percentile) densities of downy woodpeckers. Trend lines were calculated using the model-averaged parameter estimates from the predictive model, ln(Rbt+i + 0.1) = ln(Rbt + 0.1) + M + Be + D + Figure 4.1. Density (No./ha) of quaking aspen, suitable nest trees (standing dead aspen > 12.5 cm DBH), lodgepole pine and beetle-infected lodgepole pine trees in patches containing at least one suitable nest tree (Available), and in red-breasted nuthatch nest patches (Used) for pre-outbreak, outbreak, and across all years in central British Columbia, Canada. Error bars p=0.01) 80 Bi + At + Rb*D + Site + Year 81 represent standard error of the mean 96 ix ACKNOWLEDGEMENTS I thank my supervisor, Dr. Kathy Martin, for her encouragement, guidance, and support. Dr. John McLean helped me discover the fascinating world of entomology and provided valuable feedback on my thesis. I am grateful to Dr. Darren Irwin for his very insightful and thorough comments during the development of my thesis, and especially in the final writing stages. I thank Dr. Mark Drever for his assistance in data analyses and his comprehensive review of my data chapters. I thank my partner and best friend, Patrick Robinson, for his constant support, encouragement and understanding. I am grateful to my parents for their loving support and encouragement, and for their belief that I could accomplish any goal that I set for myself. My grandmother and grandfather, Margaret and Ross Norris, inspired my love for the natural world and have always supported me in all my endeavours. I thank my brothers for always standing behind me and for their friendship, and my sister for her creative inspirations. Wayne Campbell was an early inspiration in ornithology and helped me develop my birding skills. Krista DeGroot of the Canadian Wildlife Service provided insightful career advice, support and friendship. I thank several individuals who provided field assistance: Dayani Gunawardana, Meghan Newman, Alicia Newbury, Natasha Knight, Katie Aitken, Marty Mossop, Katharine Scotton, and, of course, Patrick Robinson. My fellow members of the Martin lab provided advice and valuable discussions. I am grateful for the excellent technical help from Marty Mossop. I thank Katie Aitken for the enlightening office discussions, and for her friendship, understanding and coaching through the whole process. I also thank Nicola Freeman, Alaine Camfield, Matt Tomlinson, Heather Bears, Felice Griffiths, and Diana Demarchi for their friendship. Personal funding was provided by a Natural Sciences and Engineering Research Council (NSERC) Industrial Post-graduate Scholarship, sponsored by Tolko Enterprises, Environment Canada (EC) grants through the Science Horizons program and "Space for Habitat" program funding from Brenda Morehouse, a Forest Science Program Mountain Pine Beetle initiative graduate student research fund, and a Southern Interior Bluebird Trail Society award. Fieldwork was supported by grants to Kathy Martin from the Sustainable Forest Management Network, NSERC and EC. CO-AUTHORSHIP STATEMENT This thesis contains manuscripts that were co-written. I contributed to the majority of the design of my research program, and was assisted by Dr. Kathy Martin. I performed a significant proportion of the research. My thesis incorporates data from a long-term monitoring project, which is owned by Dr. Kathy Martin. I performed all data analyses, and was assisted by Dr. Mark Drever. I wrote all three manuscripts contained in this thesis, with some assistance from Dr. Martin and Dr. Drever. xi CHAPTER 1: GENERAL INTRODUCTION Animal populations often fluctuate in size, and population growth declines as densities increase (density dependence; Elton 1924, Sutherland 1996, Newton 1998). Mechanisms driving density dependence are often attributed to community interactions and/or changes in habitat caused by natural or human-induced disturbances (Martin 1993, Dennis and Often 2000, Rodenhouse et al. 2003). The outcomes of intra- and interspecific interactions are often negative on populations, due to competition and predation, but may also be positive where one species indirectly facilitates environmental conditions for another species (Cardinale et al. 2002). The effects of environmental heterogeneity, or temporal fluxes in habitat conditions and resource availability, may be positive or negative, depending on the resources affected and the interactions with density dependent factors (Rodenhouse et al. 2003). Temporal variation in resource availability Habitat disturbances may influence populations when a limited resource changes. Broad (regional) scale, pulses of food caused by forest insect outbreaks can increase population densities of insectivorous birds through increased breeding success and/or winter survival (Sillet et al. 2000). Furthermore, positive indirect effects of a pulsed resource may occur if populations of facilitator species increase. Small scale disturbances, such as forest harvesting and increased edge effects can cause population declines in forest birds arising from low overall breeding success due to increased competition for high quality habitat and increased use of poor quality habitat (Dhondt et al. 1992, Fort and Otter 2004). The purpose of my study was to examine how temporal variability in resource availability caused by a large-scale mountain pine beetle (Dendroctonus ponderosae) outbreak and small-scale disturbances 1 caused by tree harvesting affected cavity-nester community dynamics and determine how these effects interacted to regulate populations of two species within the community. Community dynamics The cavity-nester community is structured around availability of nest cavities, analogous to a food web (Martin and Eadie 1999, Aitken et al. 2002). In the community, "facilitator" species excavate cavities, providing a critical resource for the non-excavating species (secondary cavity-nesters), which are often considered to be nest-site limited (Raphael and White 1984). Competition among cavity-nesters leads to niche partitioning and species can be classified into guilds based on nest-tree and cavity characteristics. Among over 40 cavity-nesting vertebrates in interior British Columbia, mountain chickadee (Poecile gambeli), red-breasted nuthatch (Sitta canadensis), black-capped chickadee (Poecile atricapillus) and downy woodpecker (Picoides pubescens) formed one of five guilds, namely the smallest-bodied cavity-nesters (Martin et al. 2004). These species exhibited considerable niche overlap in use of nest-trees and cavities, thus may compete for resources. In addition to nesting behaviour, all excavators are also bark insectivores, thus a mountain pine beetle outbreak may increase populations of facilitators. A "cascading effect" may occur, whereby a pulse of resources (mountain pine beetle) for one nidic level (excavators), in turn, may cause a pulse of resources (cavities) for another nidic level (secondary cavity-nesters). For my study, I focussed on one non-excavator and one excavator with similar nesting resource requirements and ecological characteristics that make them sensitive to the outbreak of mountain pine beetle and associated forest insects either directly or indirectly. Excavators are presumably more mobile in terms of nest placement due to their excavation abilities, thus should track the food pulse more accurately than non-excavators. Non-excavators should 2 respond to pulses in nest cavity availability and track changes in facilitator populations. Mountain chickadee is a secondary cavity-nester, and relies on other species to provide nest cavities, and is therefore potentially greatly influenced by facilitator populations (McCallum et al. 1999). Red-breasted nuthatch excavates its own cavities, usually in dead or decaying trees, but like mountain chickadee will also use old cavities for nesting (Ghalambor and Martin 1999). With overlap in ecological niches, these two species may be significant competitors, but the nuthatch may also be a facilitating nest cavity provider for mountain chickadee if competition for food and nest sites was low. Aside from differences in excavation ability, both species are bark and foliage insectivores, but nuthatches forage more frequently on bark than do chickadees (Ghalambor and Martin 1999, McCallum et al. 1999). Both species occupy breeding territories approximately 7-10 ha in size and are year-round residents in interior British Columbia (Ghalambor and Martin 1999, McCallum et al. 1999, Martin and Norris 2007). Many nuthatch populations exhibit irruptive and often, synchronous movements, which may be related to pulses in food availability (Enoksson and Nilsson 1983, Ghalambor and Martin 1999, Koenig 2001). Increased summer and winter food supply caused by the mountain pine beetle outbreak is expected to directly increase nuthatch populations through increased food availability and increase chickadee populations through increased populations of facilitators. Aside from the two focal species, there are four other species in this small to medium-bodied cavity nester guild (Martin et al. 2004). Black-capped chickadee is an excavating bark and foliage insectivore that occurs sympatrically with mountain chickadee and red-breasted nuthatch and may compete for food and nest-trees (Hill and Lein 1989, Smith 1993). Downy woodpecker and red-naped sapsucker (Sphyrapicus ruber) are excavators that inhabit mixed deciduous coniferous forests (Jackson and Ouellet 2002, Walters et al. 2002). The red squirrel 3 (Tamiasciurus hudsonicus) is a predator of chickadee and nuthatch nests and is also a consumer of medium-sized cavities (Banfield 1974, Martin 1993, McCallum et al. 1999). All species are year-round residents in interior British Columbia, with the exception of the red-naped sapsucker. The downy woodpecker and red-naped sapsucker provide cavities that are available in subsequent years, thus an increase in their populations could potentially facilitate populations of mountain chickadee, and possibly red-breasted nuthatch (if reuse of cavities was high), through nest cavity availability. Increased availability of bark beetles and other forest insects may temporarily improve habitat quality for chickadee and nuthatch populations. Direct effects may be through increased food availability, and indirect improvements may be through increased breeding populations of facilitating nest cavity providers. Stone (1995) found that population densities of red-breasted nuthatches, downy woodpeckers, and some secondary cavity-nesting species, including mountain chickadees and red squirrels significantly increased in mountain pine beetle-infected lodgepole pine stands in Utah. The beetle outbreak may release the food constraint at the community level and cause a shift in the roles of community members (e.g., red-breasted nuthatches may shift from being competitors to facilitators for mountain chickadees). Red squirrels can switch from preying on eggs and nestlings to bark beetle adult and larval prey, which may result in decreased nest predation for nuthatches and chickadees (Pretzlaw et al. 2006). Using point count surveys, Drever and Martin (2007) found that three resident woodpecker species and mountain chickadees showed positive trends in abundance at the stand level during 1995-2005 in the study area. In recent work, red-breasted nuthatch and mountain chickadee populations and nesting densities significantly increased with the mountain pine beetle outbreak (Martin et al. 2006, Martin and Norris 2007). While populations increased, mountain 4 chickadee population growth was negatively density dependent; and the negative effects were ameliorated with food availability and intensified with predator densities (Martin and Norris 2007). However it is unclear whether density dependence diminishes as populations collapse in the post-outbreak phase. Since density dependence may be both positively and negatively influenced by interacting factors such as facilitation and predation, the overall apparent result of density on population growth may be zero (Figure 1.1). Thus, by examining factors that interact with density, one can delineate the positive from the negative effects and determine the impacts on future populations. The main objectives of my thesis were to determine how changes in habitat composition and community interactions affect mountain chickadee and red-breasted nuthatch at the, local, regional, and nest-patch scales. Study Area The Cariboo-Chilcotin region of British Columbia supports a very rich cavity-nester community that includes eight of twelve woodpecker species found in the province (Figure 1.2; Martin and Eadie 1999). Mountain pine beetle is a natural disturbance agent in forest stands within the warm and dry Interior Douglas-fir Biogeoclimatic Zone (Finck et al. 1989). Upon attack by mountain pine beetle, other insects, such as the pine engraver (Ips pini) can attack lodgepole pine (Pinus contorta), further increasing food availability for bark insectivores (Rankin and Borden 1991). The current outbreak is the largest ever recorded for the province and for North America, and thus represents a unique opportunity for studying the ecological impacts of a large-scale natural disturbance event on forest avifauna (Martin et al. 2006). My study was part of a long-term monitoring project that examined the demography of cavity-nesting vertebrates in relation to forest stand structure, composition and condition in central British Columbia, from 1995-2006 (Table 1.1, and full description in Chapter 2). The 5 long-term data set provided a large enough sample size to examine multiple effects of competitor, facilitator and predator populations, and habitat conditions on population densities of my two focal species. Thesis predictions The mountain pine beetle outbreak is expected to influence the two species differently (Figure 1.3). Since the nuthatch is more typically a bark insectivore than the chickadee the beetle outbreak is expected to directly influence nuthatch populations through increased food availability. Since the chickadee relies on facilitators for nest cavities, the beetle outbreak should indirectly influence mountain chickadee populations through increased populations of facilitators, and have a direct effect through increased food availability. The main predictions of my thesis are: For both species: • Negative effects (density dependence, interspecific competition and predation) to be ameliorated with increased availability of mountain pine beetle For mountain chickadee (a facultative bark insectivore, non-excavator): • Populations to increase with increasing facilitators (red-naped sapsucker and downy woodpecker) and the importance of the facilitator species should shift to that with the highest response to the beetle outbreak • Populations to decrease with increasing competitors (red-breasted nuthatch and black-capped chickadee) • Red-breasted nuthatch to shift from competitor to facilitator with increasing mountain pine beetle For red-breasted nuthatch (a bark insectivore, facultative excavator): 6 • Populations to decrease with increasing competitors (downy woodpecker, black-capped chickadee and mountain chickadee) • Downy woodpecker to shift from competitor to facilitator with increasing mountain pine beetle (assuming that nuthatch nest cavity reuse increases) • Given its flexibility in excavating ability and bark insectivorous foraging nature, it should move to nest patches in closer proximity to food sources; nest patches should contain a higher proportion of beetle-infected pine trees with the progression of the outbreak Thesis overview In Chapter 2,1 used point count surveys and vegetation surveys to examine how population densities of mountain chickadees varied annually with competitors (mountain chickadees, black-capped chickadees), facilitators (red-breasted nuthatches, downy woodpeckers, red-naped sapsuckers), predators (red squirrels), tree density (including harvesting), amount of edge habitat (natural forest edge or harvested), and proportion of recently mountain pine beetle-infected pine trees at the stand level, on 27 sites from 1997-2006.1 also examined the effects of mean annual densities of mountain chickadees on predicted annual population densities to determine if density dependent processes occurred at the study area scale. I used an information-theoretic approach and a model-averaging technique to develop and generate inferences from a set of 35 mixed-effects general linear candidate models, comprised of competitor, facilitator, predator and habitat effects, in various combinations, to determine which factors contributed most to variation in predicted annual population densities of mountain chickadees. I found that the non-excavating, mountain chickadee populations were driven by negative density dependence, which was ameliorated by facilitators and intensified by predators and harvesting. 7 In Chapter 3,1 used point count surveys and vegetation surveys to examine how population densities of red-breasted nuthatches varied with densities of competitors and predators, and densities of aspen, recent beetle-infected lodgepole pine trees, and all trees, and proportion of edge habitat (naturally fragmented or harvested). I used an information-theoretic approach to develop a set of 18 candidate models to examine how competitor, facilitator, predator and habitat effects and their interactions influenced annual population densities of nuthatches at the stand level, on 27 sites from 1997 to 2006.1 found strong support for two competitor models and explored the ecological implications for significant correlations of nuthatch with downy woodpecker and black-capped chickadee populations. I found a significantly positive correlation of beetle-infected pine trees and nuthatch populations at the regional scale and further explored this relationship at the nest-patch scale in Chapter 4. In Chapter 4,1 addressed how the mountain pine beetle outbreak influenced habitat selection at the nest-patch scale. I used data from 27 sites and 11 years and compared vegetation patches used by nuthatches for nesting to those available to determine whether increases in food availability resulted in trade-offs in nest tree availability. Available patches were defined as those containing at least one aspen tree suitable for nesting, within the range of variation in decay and size used by nuthatches during the study. I found that nest patches had significantly higher densities of suitable nest trees than available patches before the outbreak, and had higher densities of beetle-infected pine trees during the outbreak. Since I found a dramatic change in nest-patch selection, I explored the implications for habitat selection in post-epidemic stands. In 2005 and 2006, both mountain chickadee and red-breasted nuthatch populations collapsed to below endemic (pre-outbreak) population levels. In Chapter 5,1 discuss unstable population dynamics, community interactions and fluctuations in resource availability for 8 cavity-nesters and predict the key drivers that will result in the recovery of populations in post epidemic forests. REFERENCES Aitken, K. E. H., K. L. Wiebe, and K. Martin. 2002. Nest-site reuse patterns for a cavity-nesting bird community in interior British Columbia. The Auk 119:391-402. Banfield, A. W. F. 1974. The mammals of Canada University of Toronto Press, Toronto, Ontario, CAN. Cardinale, B. J., M. A. Palmer, and S. L. Collins. 2002. Species diversity enhances ecosystem functioning through interspecific facilitation. Nature 415:426-429. Dennis, B., and M. R. M. Often. 2000. Joint effects of density dependence and rainfall on abundance of San Joaquin kit fox. Journal of Wildlife Management 64:388-400. Dhondt, A. A., B. Kempenaers, and F. Adriaensen. 1992. Density-dependent clutch size caused by habitat heterogeneity. Journal of Animal Ecology 61:643-648. Drever, M. C , and K. Martin. 2007. Spending time in the forest: responses of cavity-nesters to temporal changes in forest health and environmental conditions in interior British Columbia. Chapter 13 in Temporal explicitness in landscape ecology: Understanding wildlife responses to changes in time. J. A. Bissonette and I. Storch (Editors). Springer, New York, NY. Pp. 236-251. Elton, C. S. 1924. Periodic fluctuations in the numbers of animals: their causes and effects. Journal of Experimental Biology 2:119-163. Enoksson, B. and S. G. Nilsson. 1983. Territory size and population density in relation to food supply in the nuthatch (Sitta europaea). Journal of Animal Ecology 52:927-935. 9 Finck, D. M., P. Humphreys and G. V. Hawkins. 1989. Field guide to pests of managed forests in British Columbia, Join Publication No. 16. Canadian Forest Service/British Columbia Ministry of Forests, Victoria, BC, Canada. Fort, K. T. and K. A. Otter. 2004. Effects of habitat disturbance on reproduction in black-capped chickadees (Poecile atricapillus) in northern British Columbia. The Auk 121:1070-1080. Ghalambor, C. K., and T. E. Martin. 1999. Red-breasted nuthatch (Sitta canadensis). In A. Poole, and F. Gill, editors. The Birds of North America, no. 459. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. Hill, B. G., and M. R. Lein. 1989. Territory overlap and habitat use of sympatric chickadees. The Auk 106:259-268. Jackson, J. A., and H. R. Ouellet. 2002. Downy woodpecker (Picoidespubescens). In A. Poole, and F. Gill, editors. The Birds of North America, no. 613. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. Koenig, W. D. 2001. Synchrony and periodicity of eruptions by boreal birds. Condor 103:725-735. Martin, K., and J. Eadie. 1999. Nest webs: a community-wide approach to the management and conservation of cavity-nesting forest birds. Forest Ecology and Management 115:243-257. Martin, K., and A. R. Norris. 2007. Life in the small-bodied cavity-nester guild: Demography •of sympatric Mountain and Black-capped Chickadees within Nest Web communities under changing habitat conditions. Chapter 8, in Otter, K. (Editor). Ecology and Behavior of Chickadees and Titmice: An integrated approach, Oxford University Press. Pp. 111-130. 10 Martin, K., K. E. H. Aitken, and K. L. Wiebe. 2004. Nest sites and nest webs for cavity-nesting communities in interior British Columbia, Canada: Nest characteristics and niche partitioning. Condor 106:5-19. Martin, K., A.R. Norris, and M. Drever. 2006. Effects of bark beetle outbreaks on avian biodiversity in the British Columbia interior: Implications for critical habitat management. British Columbia Journal of Ecosystem Management 7:10-24. Martin, T. E. 1993. Nest predation and nest sites: new perspectives on old patterns. Bioscience 43:523-532. McCallum, D. A., R. Grundel, and D. L. Dahlsten. 1999. Mountain chickadee (Poecile gambeli). In A. Poole, and F. Gill, editors. The Birds of North America, no. 453. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. Newton, I. 1998. Population limitation in birds. London: Academic. Pretzlaw T., C. Trudeau, M. M. Humphries, J. M. LaMontagne and S. Boutin. 2006. Red squirrels (Tamiasciurus hudsonicus) feeding on spruce bark beetles (Dendroctonus rufipennis): energetic and ecological implications. Journal of Mammalogy 87:909-914. Rankin L. J. and J. H. Borden. 1991. Competitive interactions between the mountain pine beetle and the pine engraver in lodgepole pine. Canadian Journal of Forest Research 21:1029-1036. Raphael, M. G., and M. White. 1984. Use of snags by cavity-nesting birds in the Sierra Nevada. Wildlife Monographs 86:1-66. Rodenhouse, N. L., T. S. Sillett, P. J. Doran, and R. T. Holmes. 2003. Multiple density-dependence mechanisms regulate a migratory bird population during the breeding season. Proceedings of the Royal Society of London. B: 270:2105-2110. 11 Sillett, T. S., R. T. Holmes, and T. W. Sherry. 2000. Impacts of a global climate cycle on population dynamics of a migratory songbird. Science 288:2040-2042. Smith, S. M. 1993. Black-capped chickadee (Parus atricapillus). In Poole, A. and Gill, F. editors. The Birds of North America, no. 39. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. Stone, W. E. 1995. The impact of a mountain pine beetle epidemic on wildlife habitat and communities in post-epidemic stands of lodgepole pine forests in northern Utah. PhD thesis. Utah State University, Logan, Utah. Sutherland, W. J. 1996. From individual behaviour to population ecology. Oxford University Press. Walters, E. L., E. H. Miller, and P. E. Lowther. 2002. Red-breasted sapsucker {Sphyrapicus ruber) and red-naped sapsucker {Sphyrapicus nuchalis). In The Birds of North America, No. 663. A. Poole, and F. Gill, editors. The Birds of North America, Inc., Philadelphia, PA. 12 Table 1.1. Study sites in the Cariboo-Chilcotin area of interior British Columbia showing survey years for each site where no harvesting took place (Controls) and where a harvesting treatment was applied (Post-cut). Fixed radius point counts for breeding birds and vegetation surveys at permanent sample plots centred at each point count station were conducted annually at each of 27 study sites. Sites selected for harvesting were monitored in pre-cut conditions for two to eight years and monitored in post-cut conditions for one to nine years. Habitat variables measured for each stand: tree density, densities of mountain pine beetle-infected lodgepole pine trees (presence of mountain pine beetle was a proxy for a range of forest insects that comprised changes in food availability), and proportion of stand within 50m of an edge. Study area Site Controls (Pre-cut) Post-cut Comments Knife Creek 7M 1997-2006 SHAC 1996-2006 SHCC 1996-1997 1998-2006 MM 1997-2000 2001-2006 16 trees cut in 2000; 390 cut in 2001 FO 1996-1997 1998-2006 10 trees cut in 1997; 67 cut in 1998 KN 1996-2000 2001-2006 CT 1996-2000 2001-2006 MC 1997-2006 MX 1997-2001 2002-2006 Harvested March 2002 BF 1997-2000 D1 1997-1999 2000-2006 33 trees cut in 1999; 178 cut in 2000 D2 1997-2000 2001-2006 Riske Creek SC 1995-2006 RL 1995-2006 RP 1997-2006 RC 1995-2006 DE 1995-2000 SW 1995-2006 TO 1995-2006 MG 1995-2006 YY 1995-2006 SP 1997-2004 2005 Harvested Fall 2004 SD 1995-2005 ML 1995-2000 LT1 1998-2001 2002-2006 Harvested Spring 2002 LT2 1998-2006 HH 2001-2003 2004-2006 13 a.) b.) + Population density Population density Figure 1.1. Population growth, or the rate at which a population increases typically declines with density (a). However, if interacting negative effects (-) no longer regulate density dependence, or are masked by positive effects (+), the overall result of density on population growth might appear to be zero (b). This might occur if populations decline drastically and are no longer regulated by competition and predation, or if facilitation increases. By examining habitat conditions and community dynamics, I attempted to delineate the interacting positive and negative effects on populations of mountain chickadee and red-breasted nuthatch. 14 Figure 1.2. Map of study area in Cariboo-Chilcotin region of British Columbia. Upper left map shows location within the province (51° 52rN, 122° 21'W), indicated by the shaded box, which is expanded in the larger map showing all study sites. Fifteen of 27 sites were located at Riske Creek, approximately 40 km west of Williams Lake ("Riske Creek sites" SC, RC, RP, RL, DE, SW, TO, MG, YY, ML, SD, SP, LT_1, LT2, HH). Twelve sites were approximately 20 km east of Williams Lake ("Knife Creek sites" SHAC, SHCC. 7M, MM, FQ, KN, CT, MC, MX, BF, DI. D2). Eleven sites were cut with various prescriptions ranging from clear cuts with reserves to selection harvesting of varying proportions of pine and spruce (underlined). 15 a) Pre-outbreak conditions Red Squirrel Mountain Chickadee Red-naped Sapsucker Black-capped Chickadee Red-breasted Nuthatch Downy Woodpecker Red Squirrel Black-capped Chickadee b) Outbreak conditions Mountain Chickadee Red-breasted Nuthatch Red-naped Sapsucker, Downy Woodpecker Figure 1.3. Schematic diagram of mountain chickadee and red-breasted nuthatch relationships within the small and medium-bodied cavity-nester community: red squirrel (predator), red-naped sapsucker (facilitator), downy woodpecker (facilitator) and black-capped chickadee (competitor). Predicted effects of community members on populations of the two species are either positive (+) or negative (-) and these effects may change with the food pulse (changes highlighted with block arrows): (a) before the mountain pine beetle outbreak, and (b) during the outbreak. 16 CHAPTER 2: ECOLOGICAL SYNERGISMS IN SMALL CAVITY-NESTER COMMUNITIES: EFFECTS OF HABITAT, COMPETITION, FACILITATION, AND PREDATION ON MOUNTAIN CHICKADEE POPULATIONS 1 INTRODUCTION A central goal in population ecology is to identify mechanisms causing fluctuations in population densities. Typically, population growth declines as density increases, but the mechanisms generating density dependent processes may vary. Populations may be controlled by both 'top-down' and 'bottom-up' mechanisms where the effects are either direct or indirect (Hunter and Price 1992, Martin and Eadie 1999, Ostfeld and Keesing 2000). For birds, these mechanisms include competition, facilitation (where individuals provide resources to neighbours through biophysical interactions), predation, and availability of food and nest-sites (Cardinale et al. 2002, Sibly and Hone 2002). A temporal flux in resource availability caused by a food pulse may affect the direction and the strength of interspecific interactions, changing the nature of relationships among community members (Ostfeld and Keesing 2000). Studies examining the functional and numerical responses of birds to insect outbreaks have shown both positive and negative direct and indirect effects on population growth (Morris et al. 1958, Crawford and Jennings 1989, Petit and Petit 1996, Matsuoka et al. 2001). Pulses of food caused by insect outbreaks can increase the breeding success of insectivorous birds (Sillet et al. 2000). In addition, increased food may ameliorate the negative effects of density dependence caused by intra- and 1 A version of this chapter wil l be submitted for publication. Norris, A.R. , Drever, M.C. and Martin, K. Ecological synergisms in small cavity-nester communities: Effects of habitat, competition, facilitation and predation on mountain chickadee populations. 17 interspecific competition, causing the consumers of pulsed resources to respond numerically, with a time lag (Ostfeld and Keesing 2000). As the food supply declines, density dependent factors may interact negatively with predation pressure, limiting further population growth of the consumers of the pulsed resource, leading to a population collapse. On the other hand, positive indirect effects of a pulsed resource may occur if interspecific interactions facilitate environmental conditions for other species (Cardinale et al. 2002). The structure and function of cavity-nester communities are analogous to food webs, where excavators create cavities for which non-excavators (secondary cavity-nesters) compete (Martin and Eadie 1999, Aitken et al. 2002). Facilitation may occur among cavity-nesters as a result of two processes: increased food and nest site availability. An increase in bark beetle abundance and associated invertebrate fauna represents a major food source for woodpeckers and the feeding activity of bark scaling exposes larvae to bark gleaners, such as chickadees. As woodpecker nesting densities increase, more cavities are excavated each year, providing more holes available to secondary cavity-nesters. In Utah, a mountain pine beetle outbreak caused an increase in densities of woodpeckers and nuthatches, which increased cavity availability and potentially facilitated increases in secondary cavity-nester populations (Stone 1995). Mountain chickadee is a small-bodied, secondary cavity-nesting bird (non-excavator), common in montane forests in western North America and typically uses coniferous trees for foraging and aspen trees for nesting (Hill and Lein 1989; McCallum et al. 1999; Aitken et al. 2002). In central British Columbia, mountain chickadee is an abundant year-round resident, occupying stands of mixed coniferous-deciduous forest, especially where Douglas-fir, lodgepole pine and quaking aspen are present (Martin and Norris 2007). Since mountain chickadee is potentially limited by nest-sites, an increase in both food and nest-site availability may significantly increase population densities (Aitken et al. 2002; Martin et al. 2004). Thus it 18 is an ideal candidate species to examine the effects of density dependence, interspecific competition, facilitation, and a resource pulse caused by the mountain pine beetle outbreak on population growth. Out of over 40 vertebrate cavity-nesting vertebrate species in interior British Columbia, mountain chickadee used nest-trees and cavities most similar to those used by black-capped chickadee, red-breasted nuthatch and downy woodpecker, using the smallest diameter of trees, with the most advanced decay and the smallest cavity entrance holes (Martin et al. 2004). Black-capped chickadee and red-breasted nuthatch are about the same size as the mountain chickadee (-10-11 gm mass), and are both excavators that inhabit a range of deciduous and mixed forest habitats across North America (Smith 1993, Ghalambor and Martin 1999). Downy woodpecker (-24 gm mass) and red-naped sapsucker are exclusively excavators that inhabit mixed deciduous coniferous forests, similar to the mountain chickadee (Jackson and Ouellet 2002, Walters et al. 2002). The nuthatch, downy woodpecker and red-naped sapsucker provide cavities that are available in subsequent years, thus could potentially facilitate mountain chickadee populations through nest cavity availability. However, with overlap in resource requirements and habitat characteristics, the mountain chickadee may also compete with the black-capped chickadee, nuthatch and/or downy woodpecker. The red squirrel is a consumer of medium-sized cavities and a predator of mountain chickadee nests (McCallum et al. 1999). Populations of a European relative to the mountain chickadee, great tit (Parus major), are regulated by density-dependent mechanisms and densities increased in years of high beech mast crops (McCleery and Perrins 1985). In interior British Columbia, population densities of mountain chickadees, woodpeckers and predators varied annually across a range of habitat stands, and showed substantial increases in breeding density over the past decade concurrently 19 with an outbreak of mountain pine beetle and with small scale harvesting (Martin and Norris 2007). Previous work showed that as mountain chickadee densities increased from 1995 to 2004, population growth declined, and density dependence was ameliorated with food availability and intensified with predator densities (Martin and Norris 2007). It is unclear whether density dependence diminishes as populations collapse and community dynamics change in the post-outbreak phase. I examined how population densities of mountain chickadees at time t+1, varied with competitors, facilitators, predators, tree density (including harvesting), amount of edge habitat (natural forest edge or harvested), and proportion of recently mountain pine beetle-infected pine trees, at time t. In addition to these stand scale variables, I tested an effect o f regional densities of mountain chickadees to determine whether density dependent processes occurred at a broader spatial scale. I constructed three main types of general linear mixed models (competitor, facilitator and habitat) to determine which ecological factors were most important in predicting population densities of the mountain chickadee. I predicted that: (i) facilitators (downy woodpeckers and red-naped sapsuckers) would ameliorate the negative effects o f density dependence, and nuthatches would shift from being competitors to facilitators with the beetle outbreak; (ii) intact (unharvested), interior forests and densities of mountain pine beetle-infected pine trees would also ameliorate density dependence; and (iii) predators (red squirrels) and competitors (conspecific: mountain chickadee densities at the regional and stand levels; and heterospecific: black-capped chickadees, red-breasted nuthatches) would had a negative top-down effect on chickadee populations (Figure 1.3). 20 METHODS Study area Fieldwork was conducted in the Cariboo-Chilcotin region of central interior Brit ish Columbia, Canada (51° 52TSL 122° 21'W). The area was comprised o f mixed coniferous and deciduous forest embedded in a matrix of grassland and shallow ponds within the warm and dry Interior Douglas-fir Biogeoclimatic Zone (Meidinger and Pojar 1991). Predominant tree species were quaking aspen (Populus tremuloides), Douglas-fir (Pseudotsuga menziesii), lodgepole pine (Pinus contorta), and white and hybrid spruce {Picea glauca x engelmannii). Sampling sites (mostly 15 to 32 ha in size) varied in character from continuous forest to two sites that were a series of ' forest islands' (0.2 to 5 ha) within the grassland matrix. Most sites were mature forest, nine of which were selectively cut for pine and/or spruce between 1997-2002. Eighteen of the 27 sites were located at Riske Creek, approximately 40 km west of Wil l iams Lake, and 11 sites were near 150 M i le House, approximately 20 km east of Wil l iams Lake (Table 1.1, Figure 1.2). The Riske Creek sites were a mixture of deciduous and coniferous forest, in varying matrices of grasslands, ponds, and small lakes and wetlands. The 150 M i l e House sites were predominantly dry coniferous forest with deciduous riparian zones around small streams. Recently, these interior forests of British Columbia have undergone significant changes in condition due to outbreaks of mountain pine beetle {Dendroctonus ponderosae) and western spruce budworm (Choristoneura occidentalis Freeman; Westfall 2004). The B C Ministry of Forests estimated over a 42-fold increase in area affected by mountain pine beetle in B C from 1999 to 2004 (-165,000 ha to 7 mil l ion ha; Westfall 2004). Vegetation surveys of the area also revealed significant increases in bark beetle attack on conifers, especially lodgepole pine 21 (Martin et al. 2006). Bark beetle larvae and other forest insects associated with bark beetles and decaying trees increase food availability for insectivorous cavity-nesters (Otvos 1979). Population and habitat monitoring I used field data that were collected from May through July 1995-2006 for analyses involving nest data and data from 1997-2006 for all other analyses. From 0500-0930 hours, fixed radius (50m) point counts were completed at each station for 6 min using observation and playback methods (-15-20 points/site). During the counts, playbacks were conducted for woodpeckers since they are difficult to detect using conventional survey techniques. Playbacks followed the initial 6-minute listening period at alternating stations, and the call of each woodpecker was played twice, followed by 30 s of Ustening time. I noted the species and number of birds within 50 m and recorded whether they were seen, or heard calling, singing, or dramming. Over 425 stations were surveyed four times (rounds) each year on 27 sites. Nests of all cavity-nesting species were located and monitored, and the species that excavated the cavity was recorded. Further details of population monitoring methods are provided in Martin and Eadie (1999) and Aitken et al. (2002). For density estimates of chickadees, nuthatches and squirrels, all four point count rounds in each year were included and only data collected during the initial 6 min of observation were used to ensure sampling unit consistency. For estimates of downy woodpecker and red-naped sapsucker densities, only data from the stations with playbacks in the first three rounds of counts were used in the analyses (playbacks were not conducted in the last round because of low response late in the season). For each species at the stand scale, the number of individuals observed on all point count stations was totalled across each site and divided by the number of stations surveyed and by the number of rounds to obtain estimates of mean individuals/ha for each site and year. I calculated regional densities of mountain chickadees by averaging all 22 density estimates across sites within 1.). Riske Creek and, 2.) Knife Creek study areas for each year. To determine habitat condition and characteristics, 0.04 ha circular vegetation plots were established and centered at each point count station. On continuous forest sites, transects were spaced systematically in a 100 x 100 m grid starting at a grassland or wetland edge and extending 500 m into the forest. On sampling sites with forest islands where it was not possible to establish a grid, vegetation plots were placed at least 100 m apart. Tree species, diameter at breast height (DBH), decay class (decay class 1 was a live, healthy tree, 2 a live tree with visible sign of bark boring insects, 3-8 were standing dead trees; Thomas et al. 1979), and general health (e.g., presence of boring insects) were recorded for all trees 2sl2.5 cm DBH in each plot. Mountain pine beetles were detected on lodgepole pine trees by the presence of outflows of dried resin on the outer bark, or by small holes (~2mm in diameter) in the bark. Density of recent beetle-infected pine was determined by the number of lodgepole pine trees with decay class 2 and evidence of bark boring insects/ha. I assumed that lodgepole pine trees that were alive, with evidence of beetle attack contained live beetles and represented a food source in late summer and over the following winter. Since the abundance of pine trees varied across sites, I calculated the proportion of recent beetle-infected pine trees from the density of beetle-infected pine trees divided by the density of all trees for each site. I recorded the distances from each point count station to a wet or dry forest edge using an eTrex global position system unit (GPS; Garmin Internation Inc., Olathe, KS), and calculated the proportion of edge habitat by estimating the percent of vegetation plots within 50 m of an edge, using Axe View (GIS). Timing of cutting varied across the study with sites monitored in pre-cut conditions for two to eight years and monitored in post-cut conditions for one to nine years (more details in Table 1.1). In my analyses, I classified uncut and pre-cut sites as "controls" 23 and the remaining sites where any harvesting had occurred as "cut" in all years following the harvesting. This class variable allowed me to discrirniriate potential edge effects from naturally open habitat and harvested sites. Modeling population densities I estimated population densities by taking the natural log of the average number of individuals (N) per study site at time (year), t and t+1.1 added a correction factor of 0.01 to each population density estimate because these data contained some zeroes, and the addition of a small constant to every population estimate avoids division by zeroes (Framstad et al. 1997). All models followed the general format: xt+i = B 0 + Bx *xt + fly * y,+Site + Year; where; xt+i = density of chickadees>at time t + 1; B 0 = model intercept; Bx^ estimated effect of chickadee densities ; xt = density of chickadees at time t; y = all other fixed effects at time t (e.g., densities of competitor species); By = parameter estimate for fixed effect; parameter estimates were averaged across 27 sites and 10 years for each model. Modeling approach I used general linear mixed-effects models to examine how population densities of mountain chickadees, at time t+1, were influenced by densities of competitors, facilitators, predators, and habitat effects, at time 1.1 used maximum likelihood ratios for parameter estimation and Akaike's Information Criterion (AJC) as the basis for model selection (Burnham and Anderson 1998). Variables included in the candidate models were density (at time t) of mountain chickadees at the stand (In (Nt + 0.01) and regional scales, density of red-breasted nuthatches, downy woodpeckers, red-naped sapsuckers, red squirrels, and proportion of trees recently infected by mountain pine beetle, density of all trees, and the proportion of 24 edge, and whether the site was intact (uncut) or harvested (cut). Previous iterations of these and other population growth models revealed no significant contribution of "black-capped chickadees" to explaining variation in mountain chickadee densities, so I dropped this variable from the models presented in this chapter (Martin and Norris 2007). Models chosen for analysis included all main effects and all two-way interactions of mountain chickadee densities (at both the stand and study area levels) with the remaining eight predictor variables. I examined interaction factors to determine: 1.) Whether density dependence was confounded by other variables, and; 2.) Whether density dependence was ameliorated or intensified with additional factors (Figure 1.1). I excluded data from the sites in the year immediately following harvesting to reduce a possible influence of disturbance from cutting activities during late winter and early spring when birds were establishing territories. I included site and year as random effects in all models to control for site-specific persistent effects (e.g., similar tree species composition and condition within sites' or within site variation due to harvesting). This allowed me to examine populations and habitat conditions for each site through time, while accounting for potential pseudoreplication and unbalanced data across years due to multiple measurements of plots within sites and differences in number of surveys conducted for each site (Pineiro and Bates 2000). For each model, I calculated the Akaike's information criterion, corrected for small sample sizes (where N/K < 40), AICc=AIC + 2K(K+1)/(N-K-1), where AIC is calculated from the maximum likelihood ratios, K is the number of model parameters, and N is total number of site/year units used (Burnham and Anderson 1998). Values of AICc measure the fit of the model, and penalize for each parameter in the model. AICc uses information theory to parsimoniously estimate the relative distance between predictions generated by a model and the observed data, thus smaller AICc values indicate a tighter fit of the model to the data. I 25 calculated AAICc, which equals the AICc for the model of interest minus the smallest AICc for the set of models considered. The best model has a AAICc of zero. From the AICc values I also calculated the AICc weight of each model, w,= exp (-1/2 Ai) / £ exp (-1/2 Ai), where Ai is the AAICc for model, i. AICc weights of all models sum to 1, and weights closest to 1 are the models that fit the data most accurately (Burnham and Anderson 1998). However, if there is no obvious top model then the variation is spread amongst effects in multiple models, and a model-averaging approach may be employed to extrapolate the significant effects from the suite of models (Burnham and "Anderson 1998). To assess the distributional assumptions of error within groups (e.g., within sites, across years), I examined plots of the innermost residuals of the fully parameterized model. I examined boxplots of residuals by site and year to ensure errors of both random effects were centred at zero and I examined scatterplots of standardized residuals versus fitted values to ensure homoscesdasticity within site and year (Pinheiro and Bates 2000). All statistical analyses and evaluations of model assumptions were performed using the program R (The R Foundation for Statistical Computing Version 2.1.1; Pinheiro and Bates 2000). I constructed a series of 35 a priori candidate models, comprised of potential competitor (intra- and interspecific), facilitator, predator and habitat effects, in various combinations. If density dependent processes regulated populations, then I predicted mountain chickadee densities (at time t) to be the most significant determinant of densities at time t+1. If competition limited population densities of mountain chickadees, then I predicted the competitor models to rank highest among the suite of models. If other effects influenced density dependence, then I expected strong interactions of these effects with mountain chickadee densities (at time t): positive effects, such as food and facilitators should ameliorate 26 the negative effects of density dependence, and predators and heterospecific competitors should intensify the effects. If mountain chickadees were nest-site limited, then I predicted facilitator models to perform best, and if they were limited by both nest-sites and food then I predicted the combined models of facilitators and habitat effects to rank highest. If top-down predator effects regulated mountain chickadee populations, then I predicted high ranking of models including densities of predators. If habitat condition and structure was most important in predicting chickadee densities then I expected the habitat models to perform best. Since no single model accounted for most of the variation in densities of mountain chickadees (the AICc weight of the best model was 0.33), I calculated the average parameter estimate and standard errors for each effect, based on estimates generated from each of the 35 models, accounting for AICc weights of each model. This improves the accuracy of parameter estimation when there is no obvious top model, and protects against spurious results where a single model, not well supported by the data, may predict erroneously significant effects (Burnham and Anderson 1998). Thus a significant model-averaged parameter estimate must be a highly significant predictor in more than one model and/or significant in models that fit the data closer than other models. R E S U L T S Of 221 mountain chickadee nests used from 1995 to 2006, where I determined the original excavator, 34% were excavated by red-naped sapsuckers, 27% by red-breasted nuthatches, 22% by downy woodpeckers, 9% of nests were in knot-holes formed from broken branches, and the remaining 8% were excavated by hairy woodpeckers, black-capped chickadees, three-toed woodpeckers or northern flickers (Figure 2.1). Since sapsuckers, nuthatches and downy woodpeckers excavated approximately 83% of cavities used by mountain chickadees, I 27 determined that these species would have the greatest potential to facilitate mountain chickadee population densities. However, due to significant overlapping niches among some of these species with mountain chickadees, I expected nuthatches to also compete with chickadees when resources were limited (Aitken et al. 2002, Martin et al. 2004). Changes in habitat characteristics Changes in habitat conditions over the decade included a strong increase in proportions of recently beetle-infected lodgepole pine and harvesting of 11 sites (Figure 2.2). Most harvesting took place between 2000 and 2001 when average tree density declined from 574 to 377 trees/ha on cut sites. Tree density on uncut sites remained fairly constant at ~500 trees/ha during the study. The most substantial increase in mountain pine beetle was between 2003 and 2004, when the average proportion of lodgepole pine trees with recent beetle attack rose from 5% to 8% of all trees surveyed. Changes in the community The average number of mountain chickadees and their competitors, predators and facilitators fluctuated over the recent decade (Figure 2.3). Highest mountain chickadee densities were between 2002 and 2004 with an average density of 2.2 to 2.4 individuals/10 ha. During this period, all facilitators increased, with red-breasted nuthatches reaching an unprecedented high in 2003 and red-naped sapsuckers and downy woodpeckers increasing substantially from previous years. In 2005, red-naped sapsuckers and downy woodpeckers increased from 2004, while red-breasted nuthatches declined to their lowest levels in the study. Other researchers in the area also observed recent regional declines in nuthatch and chickadee populations (Koot unpublished Christmas Bird Count data). In terms of predators and competitors, red squirrels showed an increasing trend from 2001 to 2004, and then declined along with mountain chickadees, red-breasted nuthatches and black-capped chickadees in 2005 28 then increased again in 2006. Overall, major increases in mountain chickadees coincided with increased facilitators, but populations declined with high squirrel populations in recent years. Model selection The top five models (where AAIC < 4) included facilitator and habitat predictor variables (Table 2.1). None of the models including only facilitators or predators could predict mountain chickadee population densities, suggesting that habitat effects mediated community dynamics. The strongest model describing variation in mountain chickadee population densities included conspecifics (mountain chickadee densities at the stand and study area levels), facilitators (densities of nuthatches and sapsuckers) and habitat condition (recent beetle-infected pine trees, tree density, percent edge, and harvesting) predictor variables and an interaction of mountain chickadee and sapsucker densities (AICc weight = 0.33). However, since the top model had such a low AICc weight, I made all inferences from the model-averaged parameter estimates. Using the full model (K=27, n=202), I found only three significant effects (red sapsucker: t=-5.50, df=176, p=0.02; harvesting: t=-2.88, df=176, p<0.01, and mountain chickadee x sapsucker interaction: t=-2.19, df=176, p=0.01), which were comparable to the model averaged parameter estimates for those values. Thus, I found no evidence of Freedman's paradox (where parameter estimates erroneously appeared significant when n/K < 10; Burnham and Anderson 1998). Model averaging Habitat effects Averaging across all 35 candidate models examined, I found three habitat variables significantly affected mountain chickadee populations; densities were maintained at higher levels on uncut sites across all densities, compared to cut sites (Figure 2.4); N t +, increased with tree density to approximately 500 trees/ha then declined with further increasing tree densities 29 (Figure 2.5); and; N t + 1 decreased with N, in stands with high proportions of edge habitat (Figure 2.6). Mountain chickadee density dependence When all other effects were constant, mountain chickadee densities at time, t+1, did not increase or decrease with densities at time, t, meaning that population growth appeared to be unrelated to density (Figure 2.4). Further, at any given N„ densities in the following year were higher in uncut than cut stands. Mountain Chickadee density dependence was only significant when interacting variables were examined (Table 2.2, Figures 2.7, 2.8). There were both positive and negative effects of mountain chickadees due to the three interacting variables, so the overall effect of mountain chickadee density alone appeared non-significant. Significant interacting effects were: proportion of edge habitat, and densities of red-naped sapsuckers and red squirrels (Table 2.2). At low densities mountain chickadee populations in subsequent years were higher in stands containing a high proportion of edge habitat compared to interior forest stands (low edge; Figure 2.6). But, populations were negatively density dependent in stands with high proportions of edge. In stands with low proportions of edge habitat, chickadee populations were maintained at a constant level across all densities at time t, and increased to higher levels following high chickadee densities, compared to stands with high edge. Thus edge had a positive effect,on chickadees at low densities and a negative effect at high densities. When I delineated stands with high and low densities of red-naped sapsuckers at time t, I found that at low densities, mountain chickadee populations increased to higher levels following years of high sapsucker densities compared to years following low sapsucker densities (Figure 2.7a). However, these higher populations were strongly, negatively density dependent, and the positive effect of sapsuckers declined with increasing chickadee densities. 30 Where sapsucker densities were low, chickadee populations in the following year were maintained at a constant level (not density dependent), but populations were not able to increase to the high levels found where sapsucker densities were high. Thus red-naped sapsuckers had a positive effect on chickadee growth rate at low densities but this positive effect diminished as densities increased. At any given regional density at time t, densities in the following year were maintained at constant and higher levels on stands where squirrel densities were low (Figure 2.7b). Where squirrel densities were high, populations on those stands decreased with increases in regional densities. Thus squirrel densities had a negative effect on subsequent chickadee populations, and may have caused negative density dependence. Other facilitator effects Mountain chickadee densities significantly increased with increasing red-breasted nuthatch densities O = 1.18 ± 0.55, t23 = 2.15; Figure 2.8). I found no effect of downy woodpecker densities on mountain chickadee populations. In summary, I found strong evidence for a positive effect of facilitating red-breasted nuthatches, a negative effect of cutting and high squirrel densities, and a non-linear effect of tree density on mountain chickadee population density. Also, edge and red-naped sapsucker densities interacted with chickadee densities to have both negative and positive effects on populations. DISCUSSION Populations are regulated by temporal variation in strengths of top-down and bottom-up effects, both of which are significantly affected by pulsed resources (Jedrzejewska and Jedrzejewski 1998; McShea 2000; Ostfeld and Keesing 2000). In Europe, mast production of 31 oak (Quercus robur) and hornbeam (Carpinus betulus) trees caused functional and numerical responses of rodents, followed by subsequent numerical responses of predators, which had detrimental effects on local songbird reproductive success (Jedrzejewska and Jedrzejewski 1998). I observed similar cascading effects at the community level whereby the pulse of mountain pine beetle availability caused functional responses of woodpeckers and nuthatches, followed by subsequent increases in mountain chickadees and red squirrels. Overall, I found that density dependence, facilitators, predators and habitat effects were the major determinants of population change. I suggest that mechanisms affecting population densities of mountain chickadees were dependent on annual variability in resources and community interactions. Habitat effects Standing live and dead trees provide cavity-nesters with foraging and nesting substrates, consequently, areas of high densities of these resources are often preferred nest-sites (Mannan et al. 1980, Raphael and White 1984). However, since red squirrels compete for cavities, predation risk may be high in these areas of high resource availability (Li and Martin 1991). Tree density, edge effects, harvesting and densities of red squirrels influenced chickadee population densities. In my system, low tree density can indicate either a greater amount of edge habitat with fewer trees, typical of a recently cut or regenerating young forest, or fewer but larger trees, typical of an interior, mature forest. My result that chickadee populations increased with tree density to a maximum of 500 trees/ha and then declined, and that population growth was lower in harvested stands suggests that mountain chickadees prefer interior, intact forest types (Figures 2.4, 2.5). Upon examining the negative interaction of chickadee density with edge, I found that edge was only detrimental at high chickadee densities, suggesting that edge may compound the negative effects of density dependence. Densities of red squirrels were also negatively correlated with forest edge (K. Martin 32 unpublished data), so while negative density dependence was ameliorated in interior forests, predation may have dampened this positive effect. Predation may have limited chickadee populations in interior forests, displacing populations to edge habitat, which then declined due to negative, density dependent effects, such as competition with other cavity-nesters. Intra- and Interspecific Competition Competitors may limit populations if resources are limited or distributed unevenly as a consequence of habitat heterogeneity (Dhondt et al. 1992). As predicted, mountain chickadee populations were strongly regulated by density dependence through high proportions of edge habitat, and high densities of sapsuckers and squirrels. My results match those of tit (Parus spp.) studies in Europe that found lower fecundity of populations during periods of higher densities of two competitor species because a greater proportion of poor quality territories were occupied (Dhondt et al. 1992). Contrary to this European study, my population models did not predict any significant negative effects of interspecific competition on mountain chickadee growth rate. Red-breasted nuthatches may be the closest competitor to mountain chickadees in the area because they occur at similar densities, had very similar foraging niches and used old cavities (Aitken et al. 2002, Martin and Norris 2007). Yet, I found a positive correlation between nuthatch densities and subsequent chickadee densities, suggesting that nuthatches positively affected mountain chickadee populations. Christmas bird count data in the region also showed a significant positive correlation between the number of mountain chickadees and red-breasted nuthatches observed from 1995 to 2006 (r = 0.56, df = 12, p = 0.04; Cathy Koot unpublished data). Thus I found the strongest evidence for population regulation by intra- rather than interspecific competition, which was driven by interacting habitat, facilitator and predator effects. 33 Facilitation Species that indirectly facilitate populations of other species do not tend be limited by the same resources and this facilitation relationship may change with resource availability (Cardinale et al. 2002). Pulsed resources have directly positive effects on consumers and their predators and indirect positive effects on other community members (Ostfeld and Keesing 2000). In addition to nuthatches, red-naped sapsuckers also positively affected mountain chickadee population densities, but this positive effect diminished as chickadee densities increased, suggesting that high sapsucker populations caused density dependent population growth for mountain chickadees. Cavities excavated by red-naped sapsuckers were preferred by chickadees and high sapsucker densities in one year, may have increased cavity availability for chickadees in following years. Mountain chickadee populations may have been nest-site limited because nest-box addition experiments in the area resulted in significant increases in nesting densities (Aitken 2007). As a result of the increased nest-site availability by sapsuckers, populations on these stands might have increased above carrying capacity, and then other forces, such as predation may have negatively impacted density dependence. Red squirrels used cavities most similar to those excavated by red-naped sapsuckers out of 40 other vertebrate cavity-nesters, suggesting that squirrels may have also increased with sapsucker densities and functioned as both competitors and predators to mountain chickadees (Martin et al. 2004). Red-naped sapsucker densities varied considerably over the study period, but not in synchrony with local events (Figure 2.3a). Red-naped sapsuckers typically excavated cavities in live trees, so holes were available for several years after they are used by sapsuckers, and as sapsucker populations increased, cavity resources for chickadees accumulated (Martin et al. 2004). However, sapsuckers tended to nest within 30m of forested edges, where chickadee 34 populations were negatively density dependent, whereas nuthatches nested closer to the forest interior with the outbreak (Chapter 4, Martin et al. 2004). Before the beetle outbreak, I found that mountain chickadees used the more abundant sapsucker cavities, possibly allowing chickadee growth rate to increase with high sapsucker densities, and as mountain pine beetle increased and subsequently, densities of red-breasted nuthatches, use of nuthatch cavities increased (Figure 2.1). In a previous paper, we found that as the beetle outbreak progressed, mountain chickadees used significantly smaller cavities, which were more characteristic of nuthatch cavities (Martin and Norris 2007). Nuthatch cavities, smaller in size compared to red-naped sapsucker cavities, provide greater protection from medium-sized cavity competitors, such as bluebirds, swallows and starlings, and predators because use of these cavities is limited to smallest-bodied cavity-nesters (Aitken et al. 2002). Thus with an increase in smaller cavities that are closer in proximity to potential food sources and situated in interior forests where density dependence is low, I suggest that the dominant facilitator changed from red-naped sapsucker to red-breasted nuthatch with the beetle outbreak. I suspect that mountain pine beetle abundance significantly increased over-winter survival of woodpeckers and nuthatches, increasing their feeding activities on lodgepole pine trees and their reproductive output in following years. In turn, this created increased feeding opportunities and nest cavities for mountain chickadees, causing a release in nest-site resource limitation, allowing dramatic increases in population growth. Predation Typical predator-prey models predict that predators have little impact on prey populations at very low prey densities, but when prey populations increase, predators limit population growth and may exhibit a numerical response to increased prey (Rosenzweig and MacArthur 1963). Since squirrels generally increased with mountain chickadees, predators exhibited some 35 numerical response to resource availability during this time, which may have resulted in the negative impacts on chickadee population densities. Red Squirrels are generalist consumers, capable of switching between conifer seeds and bird eggs and nestlings, thus an increase in squirrels does not always influence chickadee demography nor necessarily indicate a numerical response to prey populations (Banfield 1974; Mahon and Martin 2006). Since squirrel densities were often at the same levels or higher than mountain chickadees, it is most likely that squirrels increased due to some factor besides increased availability of chickadee prey. Squirrels may respond to increased conifer seed supply, especially since avian prey availability is seasonally limited. A masting event in lodgepole pine trees in the area might have occurred synchronously with the beetle outbreak because masting and beetle outbreaks are often linked to the same abiotic factor, drought (Raffa and Berryman 1983). Furthermore, squirrels can switch from preying on eggs and nestlings to bark beetle prey during outbreaks (Pretzlaw et al. 2006). Regardless, given that the model-averaged parameter estimate of the squirrel*mountain chickadee interaction was significantly negative; I conclude that high squirrel densities had a significant negative top-down effect on mountain chickadee population densities. Understanding population dynamics allows us to examine mechanisms regulating population trends and predict more accurately the effects of environmental variability. In my study, annual variability in resource availability influenced the specific roles of community members, and these had varying impacts on mountain chickadee populations. I suggest temporal fluctuations in both community interactions and habitat structure be examined together in evaluating populations. 36 R E F E R E N C E S Aitken, K. E. H. 2007. Resource availability and lirnitation in a cavity-nesting bird and mammal community in mature conifer forests and aspen groves of interior British Columbia. PhD thesis. University of British Columbia, Vancouver, British Columbia. Aitken, K. E. H., K. L. Wiebe, and K. Martin. 2002. Nest-site reuse patterns for a cavity-nesting bird community in interior British Columbia. The Auk 119:391-402. Banfield, A. W. F. 1974. The mammals of Canada University of Toronto Press, Toronto, Ontario, CAN. Burnham, K. P., and D. R. Anderson. 1998. Model selection and inference: a practical information-theoretic approach. Springer-Verlag, New York, USA. Cardinale, B. J., M. A. Palmer, and S. L. Collins. 2002. Species diversity enhances ecosystem functioning through interspecific facilitation. Nature 415:426-429. Crawford, H. S., and D. T. Jennings. 1989. Predation by birds on spruce budworm Choristoneura fumiferana: functional, numerical, and total responses. Ecology 70:152-163. Dhondt, A. A., B. Kempenaers, and F. Adriaensen. 1992. Density-cependent clutch size caused by habitat heterogeneity. Journal of Animal Ecology 61:643-648. Framstad, E., N. C. Stenseth, O. N. Bjornstad, and W. Falck. 1997. Limit cycles in Norwegian lemmings: tension between phase-dependent and density dependence. Proceedings of the Royal Society London Series B 246:31-38. Ghalambor, C. K., and T. E. Martin. 1999. Red-breasted nuthatch (Sitta canadensis). In A. Poole, and F. Gill, editors. The Birds of North America, no. 459. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. 37 Hill, B. G., and M. R. Lein. 1989. Territory overlap and habitat use of sympatric chickadees. Auk 106: 259-268. Hunter, M. D., and P. W. Price. 1992. Playing chutes and ladders: Heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology 73:724—732. Jackson, J. A., and H. R. Ouellet. 2002. Downy woodpecker {Picoidespubescens). In A. Poole, and F. Gill, editors. The Birds of North America, no. 613. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. Jedrzejewska, B. and W. Jedrzejewski. 1998. Predation in vertebrate communities. The Bialowieza Primeval Forest as a case study. Springer Verlag, Berlin. Li, P., and T. E. Martin. 1991. Nest-site selection and nesting success of cavity-nesting birds in high elevation forest drainages. The Auk 108:405-418. Mahon, C.L., and K. Martin. 2006. Nest survival of chickadees in managed forests: habitat, predator, and year effects. Journal of Wildlife Management 70:1257-1265. Mannan, R. W. E. C. Meslow, and H. M. Wight. 1980. Use of snags by birds in Douglas-fir forests, western Oregon. Journal of Wildlife Management 44:787-797. Martin, K., and J. Eadie. 1999. Nest webs: a community-wide approach to the management and conservation of cavity-nesting forest birds. Forest Ecology and Management 115:243-257. Martin, K., and A. R. Norris. 2007. Life in the small-bodied cavity-nester guild: Demography of sympatric mountain and black-capped chickadees within Nest Web communities under changing habitat conditions. Chapter 8, in Otter, K. (Ed.). Ecology and Behavior of Chickadees and Titmice: An Integrated Approach, Oxford University Press, UK. Pp. 111-130. 38 Martin, K., K. E. H. Aitken, and K. L. Wiebe. 2004. Nest sites and nest webs for cavity-nesting communities in interior British Columbia, Canada: Nest characteristics and niche partitioning. Condor 106:5-19. Martin, K., A.R. Norris, and M. Drever. 2006. Beetles in the Nest Web: Impacts of bark beetle populations on wildlife communities in Interior British Columbia Forests 1^1 years of wildlife-habitat data. British Columbia Journal of Ecosystems Management 7:10-24. Matsuoka, S. M., C. M. Handel, and D. R. Ruthrauff. 2G01. Densities of breeding birds and changes in vegetation in an Alaskan boreal forest following a massive disturbance by spruce beetles. Canadian Journal of Zoology 79:1678-1690. McCallum, D. A., R. Grundel, and D. L. Dahlsten. 1999. Mountain chickadee (Poecile gambeli). In A. Poole, and F. Gill, editors. The Birds of North America, no. 453. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. McCleery, R. H., and C. M. Perrins. 1985. Territory size, reproductive success and population dynamics in the great tit, Parus major. Pages 353-373 in R. M. Sibley and R. H. Smith, editors. Behavioural ecology. Blackwell Scientific, Oxford, England. McShea, W. J. 2000. The influence of acorn crops on annual variation in rodent and bird populations. Ecology 81:228-238. Meidinger, D., and J. Pojar, J. 1991. Ecosystems of British Columbia. British Columbia Ministry of Forests Special Report Series, no. 6. Victoria, British Columbia. Morris, R. F., W. F. Cheshire, C. A. Miller, and D. G. Mott. 1958. The numerical response of avian and mammalian predators during a gradation of the spruce budworm. Ecology 39:487^ 194. Ostfeld, R. S. and F. Keesing. 2000. Pulsed resources and community dynamics of consumers in terrestrial ecosystems. Trends in Ecology and Evolution 15: 232-237. 39 Otvos, I. S. 1979. The effects of insectivorous bird activities in forest ecosystems: An evaluation. In J. G. Dickson, R. N. Connor, R. R. Fleet, J. C. Kroll, and J. A. Jackson, (Editors). The Role of Insectivorous Birds in Forest Ecosystems, Academic Press, New York. Petit, L. J., and D. R. Petit. 1996. Factors governing habitat selection by prothonotary warblers: Field tests of the Fretwell-Lucas models. Ecological Monographs 66:367-387. Pinheiro, J. C, and D. M. Bates. 2000. Mixed-effects models in S and S-Plus. Springer-Verlag, New York, NY. Pretzlaw T., C. Trudeau, M. M. Humphries, J. M. LaMontagne and S. Boutin. 2006. Red squirrels (Tamiasciurus hudsonicus) feeding on spruce bark beetles (Dendroctonus rufipennis): energetic and ecological implications. Journal of Mammalogy 87:909-914. Raffa K. F., and A. A. Berryman. 1983. The role of host plant resistance in the colonization behaviour and ecology of bark beetles (Coleoptera: Scolytidae). Ecological Monographs 53:27^9. Raphael, M. G., and M. White. 1984. Use of snags by cavity-nesting birds in the Sierra Nevada. Wildlife Monographs 86:1-66. Rosenzweig, M. L., and R. H. MacArthur. 1963. Graphical representation and stability conditions of predator-prey interactions. American Naturalist 97:209-223. Sibly, R. M., and J. Hone. 2002. Population growth rate and its determinants: an overview. Philosophical transactions of the Royal Society B: Biological Sciences 357:1153-1170. Sillett, T. S., R. T. Holmes, and T. W. Sherry. 2000. Impacts of a global climate cycle on population dynamics of a migratory songbird. Science 288:2040-2042. 40 Smith, S. M. 1993. Black-capped Chickadee {Parus atricapillus). In Poole, A. and Gill, F. editors. The Birds of North America, no. 39. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. Stone, W. E. 1995. The impact of a mountain pine beetle epidemic on wildlife habitat and communities in post-epidemic stands of lodgepole pine forests in northern Utah. PhD thesis. Utah State University, Logan, Utah. The R Foundation for Statistical Computing Version 2.1.1 Walters, E. L., E. H. Miller, and P. E. Lowther. 2002. Red-breasted Sapsucker {Sphyrapicus ruber) and Red-naped Sapsucker {Sphyrapicus nuchalis). In The Birds of North America, No. 663. A. Poole, and F. Gill, editors. The Birds of North America, Inc., Philadelphia, PA. Westfall, J. 2004. 2004 Summary of Forest Health Conditions in British Columbia. Pest Management Report Number 15. Forest Practices Branch, British Columbia Ministry of Forests. 41 Table 2.1. General linear mixed models predicting population densities of mountain chickadees in Interior British Columbia from 1997 to 2006. Variables examined were mountain chickadee (at the local scale: M; and at the regional scale: RegM), proportion of beetle-infected pine trees (Bi), red-breasted nuthatch (Rb), downy woodpecker (D), red-naped sapsucker (Rn), red squirrel (Sq) abundance, tree density (St) and its quadratic form (St_2) and percentage edge (E), and whether the site was harvested (C). Site and year were included as random variables. For each model, the number of parameters (K), the number of observations (N: all site/year combinations for 27 sites and 10 years), the maximum loglikelihood ratio (logLik), the Akaike's Information Criterion corrected for small samples (AICc), the difference in AICc compared to the model with the least AICc value (AAICc), and its weight (AICc wt). Shown are 14 of 35 models examined. Parameters Model description K N logLik AICc AAICc weight M Rb Rn St St2 Bi C E RegM M*Rn Facilitators + Habitat 14 202 189.33 408.91 0.00 0.33 M Rb Rn St St2 Bi C E RegM M*E Facilitators + Habitat 14 202 190.59 411.43 .2.52 0.09 M St St2 Bi E C RegM M*E Habitat 12 202 192.94 411.53 2.62 0.09 M Rb Rn St St2 Bi C E RegM Facilitators + Habitat 13 202 191.98 41L89 2.98 0.07 M St St2 Bi E C RegM Habitat 11 202 194.58 412.55 3.64 0.05 M Sq St St2 Bi E C RegM RegM*Sq Predators + Habitat 13 202 192.57 413.09 4.18 0.04 M Rb Rn St St2 Bi E C RegM RegM*Rn Facilitators + Habitat 14 202 191.70 413.65 4.74 0.03 -M Rb Rn St St2 Bi C E RegM M*Bi Facilitators + Habitat 14 202 •191.7.7 413.80 4.89 0.03 M Sq St St2 Bi E C RegM Predators + Habitat 12 202 194.08 413.80 4.89 0.03 M Sq St St2 C Bi E RegM Predators + Habitat 12 202 194.08 413.80 4.89 0.03 M Rb Rn St St2 Bi E C RegM RegM*Bi Facilitators + Habitat 14 202 191.83 413.90 4.99 0.03 M Rb Rn St St2 Bi E C RegM RegM*E Facilitators + Habitat 14 202 191.85 413.95 5.04 0.03 M Rb D Rn Sq Bi E C St St2 RegM All main effects 14 202 194.22 418.70 9.79 0.00 M Rb D Rn Sq Bi E St St2 C RegM M*Rb M*Rn M*Sq M*Bi M*Stem M*RegM RegM*Rb RegM*Rn RegM*Sq RegM*Bi RegM*E RegM*St Full model 27 202 181.48 425.64 16.73 0.00 42 Table 2.2. Model-averaged parameter estimates for the explanatory variables of mountain chickadee population densities at time, t+1: mountain chickadee (at the local scale: M; and at the regional scale: RegM), mountain pine beetle (Bi), red-breasted nuthatch (Rb), downy woodpecker (D), red-naped sapsucker (Rn), red squirrel (Sq) abundance, tree density (St) and its quadratic transformed variable (St2) and percentage edge (E), at time, t and all two-way interaction terms. Standard errors (SE) and the number of models of 35 examined, in which each of the effects appear (N) are given. Parameters have a significant effect on mountain chickadee growth rates at the a = 0.05 level where t-value > /1.98/, and at the a = 0.1 level where t-value > / 1.65/. Sign of estimate indicates a positive or negative effect on mountain chickadee densities in subsequent years. Parameter Estimate SE N t Intercept -2.00 0.58 35 -3.466 C -0.60 0.23 23 -2.600 St2 -4.06 1.63 24 -2.488 M*Rn -1.51 0.65 4 -2.344 Rb 1.18 0.55 23 2.146 Sq*RegM -6.59 3.76 3 -1.751 M*E -0.29 0.17 2 -1.750 Rn -1.45 0.92 23 -1.590 St*RegM 8.38 5.57 1 1.505 Sq 0.63 0.46 16 1.366 D -0.92 .0.74 7 -1.246 St 1.79 1.58 24 1.135 Bi 1.37 1.28 24 1.073 M*RegM -0.85 0.88 1 -0.960 E -0.34 0.36 24 -0.956 M*Bi -0.86 1.14 4 -0.759 Rn*RegM -4.23 5.74 4 -0.738 RegM -0.59 1.18 35 -0.502 M 0.04 0.09 35 0.473 Bi*RegM 6.09 13.58 4 0.449 E*RegM -0.79 1.91 4 -0.414 M*St -0.08 0.32 1 -0.242 M*Sq -0.09 0.38 3 -0.233 M*Rb 0.12 0.57 4 0.209 Rb*RegM 0.41 5.84 4 0.070 43 • Red-naped Sapsucker (76) Red-breasted Nuthatch (59) Downy Woodpecker (49) Northern Flicker (3) • Black-capped Chickadee (4) fi Natural hole (20) • Hairy Woodpecker (7) W,. Three-toed Woodpecker (3) 1.2 i 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 (3) (2) (10) (16) (9) (8) (29) (27) (29) (28) (45) (15) Figure 2.1. Known or suspected cavity excavators for 221 mountain chickadee nests. Numbers in brackets below each year and next to each species name indicate sample size of nests. If the cavity was freshly excavated the year of the recorded nest, the occupants were assumed to be the excavators. For example, if downy woodpeckers occupied a freshly excavated cavity in 2003 and the nest was used by mountain chickadees in 2004, the excavator was suspected to be downy woodpecker, from 1995 to 2006 across all study sites in central British Columbia. 44 700 600 500 ro w 400 CD CO CD 200 100 - •— Uncut sites -•- - Cut sites o • Recent beetle attack 9 » % * * % I \ \ (5 - 9 - -1 1 1 1 1 1 1 1 r-, — 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 0.09 co 0.08 w 0 CD 0.07 | Q-0.06 | & a. CD CD 0.05 0.04 0.03 § 0.02 c g •0.01 | Figure 2.2. Mean number of standing trees (>12.5 cm DBH) on uncut and cut sites, and proportion of lodgepole pole pine trees with evidence of recent beetle attack of all trees surveyed per ha across 27 study sites in Interior British Columbia. The dip in proportion of recently mountain pine beetle-infected pine trees on the study from 2000 to 2002 was due to harvesting 7 of 11 sites between 2000 and 2002. 45 a. Facilitators 0.3 0.25 -\ ^ 0.2 -I tn c o S 0.15 a> T3 (D OJ to S 0.1 3 0.05 - •— Mountain Chickadee o - - Red-breasted Nuthatch -o— Red-naped Sapsucker -o- • • Downy Woodpecker :0' - c -r- —l 1 1 r - i r " 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 0.35 0.3 0.25 tn o 0.2 o £ 0) T> (1) cn £5 CD -5 0.1 b. Competitors & Predators 0.15 4 0.05 4 - Mountain Chickadee Recj squirrel - o - - Red-breasted Nuthatch • ••©••• Black-capped Chickadee o o- o - I r -1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Figure 2.3. Mean detections of mountain chickadees with potential (a) facilitators, and (b) competitors and predators across point count surveys, on 27 forested sites in Interior British Columbia. 46 0 -0.2 H CD — Uncut — Cut $ - 1 . 6 -D_ -1.8 A -2 -5 -4 -3 -2 -1 0 Mean # mountain chickadees/ha (ln(Nt+0.01)) Figure 2.4. Predicted mean detections of mountain chickadees (at time, t+1) with increasing mountain chickadee densities (at time, t), at uncut and cut sites, from 1997-2006. Trend lines were calculated by substituting the model-averaged parameter estimates into the full predictive model, ln(Mt+]+0.01) - -2.00 + ln(Mt+0.01) + RegM + Rb + D + Sq + Bi + St +St2 + E + C + all interactions (Tables 2.1,2.2). 47 c co o> -4.5 -5 200 400 600 Mean # trees/ha 800 1000 0 -0.5 o + 1 -1 z — -1.5 co cu CD T3 CO o E % c co CD -2\ -2.5 \ -3.5 A b.) 0.2 0.4 0.6 Mean proportion of trees/ha((N/1000)A2) 0.8 Figure 2.5. Mean detections of mountain chickadees (t+1) with (a) increasing tree density from point count and vegetation surveys of 27 stands during 1997-2006, and (b) quadratic-transformed (linearized) tree density using model-averaged parameter estimates substituted into the full predictive model, ln(Mt+i+0.01) = -2.00 + ln(Mt+0.01) + RegM + Rb + D + Sq + Bi + St +8^  + E + C + all interactions (Tables, 2.1,2.2). 48 0 CD O O + + -0.5 4 -1 — High E d g e Low E d g e E 1.5 ' -o CD -2-H -2.5 -5 -4 -3 -2 -1 0 M e a n # mountain ch i ckadees /ha (ln(Nt+0.01)) Figure 2.6. Predicted mean detections of mountain chickadees (time, t+1) with increasing mountain chickadee densities (time, t) at high (upper 75th percentile of data) and low (lower 25th percentile) proportions of edge habitat. Trend lines were calculated using model-averaged parameter estimates substituted into the full model, ln(Mt+i + 0.01) = -2.00 + ln(Mt+0.01) + RegM + Rb + Rn + D + Sq + Bi + St + St2 + E + C + all interactions (Tables 2.1, 2.2). 49 a.) M*Rn Interaction -0.2 --0.4 --0.6 i ro JC cu CO TJ ro J£ 1 5-0.8 _ d I* - n | | -1 .2- I -1.4 TJ 0) Tj '•5 £> CL -1.6 A -1.8 H - High Sapsucker density L- Low Sapsucker density Uncut Cut -5 -4 -3 -2 -1 Mean # mountain chickadees/ha (ln(Nt+0.01) -0.5 ro x: ~&> co cu TJ ro O 55 O O E ? S .5 E ~ TJ TJ TJ e 0 . -2.5 b.) RegM*Sq Interaction L - Low Squirrel density H - High Squirrel density 0.1 0.2 0.3 Mean # mountain chickadees/region/ha 0.4 Figure 2.7. Predicted mean detections of mountain chickadees (t+1) across chickadee densities (time, t) at high (upper 75th percentile of data) and low (lower 25th percentile) (a) Red-naped sapsucker densities, and (b) Red squirrel densities, at uncut and cut stands from 1997-2006. Trend lines were calculated using model-averaged parameter estimates substituted into the full model, ln(Mt+, + 0.01) = -2.00 + ln(M,+0.01) + RegM + Rb + Rn + D + Sq + Bi + St + St2 + E + C + all interactions (Tables 2.1,2.2). 50 u -3.5 4 o o o o % -4 H o c CO °> -4.5 -o o o o o -5 + 0 0.1 0.2 0.3 0.4 0.5 0.6 Mean # red-breasted nuthatches/ha Figure 2.8. Mean detections of mountain chickadees (t+1) with increasing red-breasted nuthatch densities, from point count surveys of 27 stands during 1997-2006. Trend line was calculated using model-averaged parameter estimates substituted into the full predictive model, ln(Mt+i+0.01) = -2.00 + ln(Mt+0.01) + RegM + Rb + D + Sq + Bi + St +St2 + E + C + all interactions (Tables 2.1,2.2). 51 CHAPTER 3: CASCADING EFFECTS OF A MOUNTAIN PINE BEETLE OUTBREAK ON THE SMALL CAVITY-NESTER COMMUNITY INFLUENCES REGULATION OF RED-BREASTED NUTHATCH POPULATIONS 1 INTRODUCTION Fluctuations in populations may be caused by intraspecific competition (density dependence), interspecific interactions, predation and environmental change, and interactions between these drivers (Martin 1993, Dennis and Otten 2000, Rodenhouse et al. 2003). Elucidating the mechanisms generating changes in populations is essential for predicting how populations will respond to environmental disturbances, such as habitat fragmentation or climate change (Saether et al. 2000). The effects of competition and predation often depress population densities, while environmental change may have variable effects, depending on the resources affected and the interactions with density dependent factors (Rodenhouse et al. 2003). Negative effects of habitat change may arise if abundance of a limiting resource, such as nesting habitat, decreases. Forest harvesting and increased edge effects can cause population declines due to increased competition for high quality habitat and/or increased use of poor quality habitat (Dhondt et al. 1992, Fort and Otter 2004). An increase in a limiting resource, such as food, may minimize the negative effects of competition and/or predation, causing increases in population densities. For example, a seed crop in British Columbia increased an alternate food supply for predators and decreased predation on chestnut-backed chickadee (Poecile rufescens) nests (Mahon and Martin 2006). The effects on populations of habitat 1 A version of this chapter will be submitted for publication. Norris, A.R. and Martin, K. Cascading effects of a mountain pine beetle outbreak on the small cavity-nester community influences regulation of red-breasted nuthatch populations. 52 change caused by natural and human-induced disturbance events and their interactions with community dynamics require further investigation. The red-breasted nuthatch is a small-bodied, weak primary cavity-nesting bird, (~ 11 g mass), that will either excavate a new nest cavity in dead or decaying trees or re-use an existing cavity (Ghalambor and Martin 1999). It excavates new cavities at approximately the same frequency as it uses old cavities (K. Martin, unpublished data). The nuthatch is a common bark insectivore in montane and boreal forests in North America and typically uses aspen trees for nesting and large diameter coniferous trees for foraging (Ghalambor and Martin 1999, Aitken et al. 2002). In central British Columbia, red-breasted nuthatch is an abundant year-round resident, occupying stands of mixed coniferous-deciduous forest, especially where Douglas-fir, lodgepole pine and quaking aspen are present (Martin and Norris 2007). Nuthatch populations in North America and Europe often exhibit high temporal variability in densities (Enoksson andNilsson 1983, Matthysen 1998, Ghalambor and Martin 1999). Fluctuations in population sizes of nuthatches are often attributed to density dependence, winter food availability and winter weather in Europe (Matthysen 1998). Furthermore, increases in European nuthatches (Sitta europaea) corresponded to increases in populations of great tits (Parus major) and blue tits {Parus caeruleus), which are close competitors to nuthatches and also close relatives of North American chickadees {Poecile spp.). Irruptions in North American nuthatch populations show strong correlations with cone production (Ghalambor and Martin 1999). It remains unclear, however, how changes in resource availability and community dynamics, such as increasing food supply, competitors and predators, interact to influence observed temporal variability typically observed in North American nuthatch populations. Among all cavity-nesting species in interior British Columbia, red-breasted nuthatch used nest-trees and cavities most similar to those used by mountain chickadee, black-capped 53 chickadee and downy woodpecker (Martin et al. 2004). Downy woodpecker and black-capped chickadee are exclusively excavators, and the former species provides nest cavities that are available in subsequent years for use by other species within the cavity-nester community. This resource may facilitate population growth of nest-site limited populations, such as non-excavating red-breasted nuthatches or the mountain chickadee which is a non-excavating secondary cavity nester that relies solely on previously excavated cavities for nesting (Martin et al. 2004). A l l four species occupy mixed coniferous-deciduous forest habitats across North America and occur sympatrically in British Columbia, however, black-capped chickadee generally occupies a higher deciduous component than the others (H i l l and Le in 1989, Smith 1993, McCa l lum et al. 1999, Jackson and Ouellet 2002, Martin and Norris 2007). The chickadees are approximately the same size as the nuthatch (~11 g mass), whereas the downy woodpecker is larger (-24 g mass). Wi th overlap in resource use and habitat preferences, I predicted that the two chickadee species would compete with nuthatches, as would downy woodpeckers. However, since downy woodpeckers provide nest cavities, I predicted that they would have a facilitating effect when competition for resources was low. The red squirrel is a primary predator of nuthatch and chickadee nests and is also a consumer o f medium-sized cavities, using cavities for rearing pups, roosting and storing seed cones (Banfield 1974, Ghalambor and Martin 1999, M c C a l l u m et al. 1999, Mahon and Martin 2006). Hence, I predicted that squirrel densities would have a negative effect on nuthatch populations. Recently, the condition of interior forests of British Columbia has changed significantly due to outbreaks of mountain pine beetle and western spruce budworm (Choristoneura occidentalism Westfall 2004). The B C Ministry of Forests estimated a 42-fold increase in area affected by mountain pine beetle in B C from 1999 to 2004 (~165,000 ha to seven mill ion ha; Westfall 2004). The vegetation surveys also revealed significant increases in bark beetle 54 infection of conifers, particularly lodgepole pine (Martin et al . 2006). Forest insect outbreaks result in increases in year-round food availability for insectivorous resident birds, such as nuthatches, chickadees and woodpeckers (Otvos 1979). Adult beetles increase summer food availability that may enhance breeding success and bark beetle larvae provide a valuable food source in winter, which could increase winter survival and subsequent population densities (Boutin 1990). Over winter, nuthatches often forage in mixed species flocks with chickadees and downy woodpeckers, ranging over areas that are much larger than breeding territories (Ghalambor and Martin 1999). Many nuthatch populations exhibit irruptive and often, synchronous movements, but whether these irruptions are directly related to food supply remains unclear (Enoksson and Nilsson 1983, Ghalambor and Martin 1999, Koenig 2001). The objectives of this paper were to: 1.) determine how changes in nuthatch populations correlated with changes in habitat conditions and populations o f competitors and predators and; 2.) examine how these community and habitat effects influenced nuthatch populations in subsequent years. Since the mountain pine beetle outbreak increased food availability for all bird species examined, I predicted positive correlations among densities of nuthatches and mountain chickadees, black-capped chickadees, and downy woodpeckers, and densities of beetle-infected trees at the stand level. Since the outbreak increased winter food availability for nuthatches, and they ranged over areas larger than the stand level during winter I also predicted that nuthatch populations would increase with beetle at the study area level. However, at high populations of nuthatches, I predicted intra- and interspecific competition and predation to limit nuthatch populations and facilitation and increased food supply to ameliorate density dependence. Thus mountain and black-capped chickadees and red squirrels should have a negative effect and downy woodpeckers should have a positive effect on nuthatch populations in subsequent years. 55 M E T H O D S Study area Fieldwork was conducted in central interior British Columbia, Canada (51° '52*N, 122° 21'W). The area was comprised of mixed coniferous and deciduous forest embedded in a matrix of grassland and shallow ponds wiftun the warm arid dry Interior Douglas-fir Biogeoclimatic Zone (Meidinger and Pojar 1991). Predominant tree species were lodgepole pine {Pinus contorta), Douglas-fir (Pseudotsuga menziesii), quaking aspen (Populus tremuloides), and white and hybrid spruce (Picea glauca x engelmannii). Sampling sites (mostly 15 to 32 ha in size) varied in character from continuous forest to two sites that were a series of 'forest islands' (0.2 to 5 ha) within the grassland matrix. Most sites were mature forest, and eleven were selectively cut for pine and/or spruce in 1997-2002. Additional study area details are given in Chapter 1 (General Introduction) and in Aitken et al. (2002) and Martin et al. (2004). Population and habitat monitoring Field data were collected from May through July 1997-2006. From 0500T0930 hours, fixed radius (50m) point counts were completed at each station for 6 min using observation and playback methods (-15-20 points/site). Playbacks for downy woodpeckers followed the initial 6-minute listening period at alternating stations, and the call was played twice, each followed by 30 s of listening time. I recorded the number of nuthatches, chickadees, downy woodpeckers and squirrels within 50 m and recorded whether they were seen, or heard calling, singing, or drumming. Over 425 stations 27 sites were surveyed four times (rounds) each year, except in 2006 where the last round of counts was omitted. Further details of population monitoring methodology are provided in Martin and Eadie (1999) and Aitken et al. (2002). 56 For density estimates of chickadees, nuthatches and squirrels, all point count rounds in each year were included and only data collected during the initial 6 min o f observation were used to ensure sampling unit consistency. For estimates o f downy woodpecker densities, only the first three rounds of counts where playbacks were conducted were included in the analyses (playbacks were not conducted in the last round because of low response late in the season). For each species at the stand scale, the number of individuals observed on all point count stations surveyed was totalled across each site and divided by the number of point count stations surveyed and by the number o f rounds to obtain estimates of mean individuals/ha for each site and year. To determine habitat condition and characteristics, 0.04 ha circular vegetation plots were established and centered at each point count station. O n continuous forest sites, transects were spaced systematically in a 100 x 100 m grid starting at a grassland or wetland edge and extending 500 m into the forest. On sampling sites with forest islands where it was not possible to establish a grid, I placed vegetation plots at least 100 m apart. Tree species, diameter at breast height ( D B H ) , decay class (decay class 1 was a live, healthy tree, 2 a live tree with visible sign o f bark boring insects, 3-8 were standing dead trees; Thomas et al. 1979), and general health (e.g., presence o f boring insects) were recorded for al l trees ;>12.5 cm diameter within each plot. Mountain pine beetles were detected on trunks of lodgepole pine trees by the presence of outflows of dried resin on the outer bark, or by small holes (~2mm in diameter) in the bark. Density of recent beetle-infected pine was determined by the number o f lodgepole pine trees with decay class 2 and evidence of bark boring insects/ha. I assumed that lodgepole pine trees that were alive, with evidence o f beetle attack contained live beetles and represented a food source for nuthatches in late summer and over the following winter. Since the abundance of pine trees varied across sites, I calculated the proportion o f recent beetle-57 infected pine trees from the density of beetle-infected pine trees divided by the density of all tree species for each site. I recorded the locations of all point count stations and distances from each point count station to a grassland or wetland edge using an eTrex global position system unit (GPS; Garmin Internation Inc., Olathe, KS). I calculated the proportion of edge habitat by estimating the percent of vegetation plots within 50 m of an edge, using aerial photographs and Arc View (GIS). Eleven of the 27 sites were cut with various prescriptions ranging from clear cuts with reserves to selection harvesting of varying proportions of pine and spruce (beetle hazard reduction). Cut sites were paired with nearby and distant uncut controls with comparable stand composition, age arid landscape context. Timing of cutting also varied across the study with sites monitored in pre-cut conditions for one to eight years and monitored in post-cut conditions for one to nine years (Table 1.1). In my analyses, I classified uncut and pre-cut sites as "controls" and the remaining sites where any harvesting had occurred as "cut" in all years following the harvesting. This class variable allowed me to discriminate potential edge effects from naturally open habitat and harvested sites. Modeling population densities I estimated population densities by taking the natural log of the average number of individuals (N) per study site at time, t (year), and t+1. Since some data contained zeroes, I added a correction factor of 0.1 to each population density estimate of nuthatches to avoid the division by zeroes (Framstad et al. 1997). My models followed the general format: x t +i = fi0 + Bx *x t + fiy * yt + Site + Year; where; xt+i = density of nuthatches at time t+1; B 0 = model intercept; Bx= estimated effect of nuthatch densities ; xt = density of nuthatches at time t; y = all other fixed effects at time t (e.g., densities of competitor species); By = parameter estimate for fixed effect; parameter estimates were averaged across 27 sites and 10 years for each model. 58 Modeling approaches I used two approaches (current year and lag effect) to examine the effects of competitor, facilitator and predator populations and habitat conditions on nuthatch populations. In the first approach, I used a general linear mixed-effect model to examine how nuthatch densities (at time, t) correlated with (1) densities of mountain chickadees, black-capped chickadees, downy woodpeckers and red squirrels; (2) proportions of recent beetle-infected pine trees, and edge habitat (including harvesting), and; (3) densities of aspen trees and all trees (> 12.5 cm diameter). This approach allowed me to examine how nuthatch populations changed in response to changing habitat conditions and assess whether interspecific competitors and potential predators may have increased along with nuthatches at the local, site level. In my second approach I used a suite of general linear mixed-effects models to examine how population densities of red-breasted nuthatches at time, t+1, were influenced by densities of conspecifics (density dependence), all other effects examined in the first approach at time t, and interactions between nuthatch densities and all other effects. For both approaches I used maximum likelihood ratios for parameter estimation and Akaike's Information Criterion (AIC) as the basis for model selection in the second approach (Burnham and Anderson 1998). Models chosen for analysis included all main effects and all two-way interactions of nuthatch densities with the remaining eight predictor variables. I excluded data from the sites in the year immediately following harvesting to reduce a possible influence of disturbance from cutting activities during late winter and early spring when birds were establishing territories. I included site and year as random effects in all models to control for site-specific persistent effects (e.g., similar tree species composition and condition within sites or within site variation due to harvesting). This allowed me to examine populations and habitat conditions for each site through time, while accounting for potential pseudoreplication and unbalanced data across 59 years due to multiple measurements of plots within sites and differences in number of surveys conducted for each site (Pineiro and Bates 2000). For each model, I calculated the Akaike's information criterion, corrected for small sample size (where N/K < 40), AICc=AIC + 2K(K+1)/(N-K-1), where AIC is calculated from the maximum likelihood ratios, K is the number of model parameters, and N is the sample size (Burnham and Anderson 1998). Values of AICc measure the fit of the model, penalized for each parameter in the model, and corrected for small sample size. AICc uses information theory to parsimoniously estimate the relative distance between predictions generated by a model and the observed data, thus smaller AICc values indicate a tighter fit of the model to the data. I also calculated AAICc, which equals the AICc for the model of interest minus the smallest AICc for the set of models being considered. The best model has a AAICc of zero. From the AICc values I also calculated the AICc weight of each model, Wj= exp (-1/2 Ai) / £ exp (- 1/2 Ai), where Ai is the AAICc for model, i. AICc weights closest to 1 are best models (Burnham and Anderson 1998). To assess distributional assumptions of within group error, I examined plots of the innermost residuals of the fully parameterized models for each of the two approaches. Also, I examined boxplots of residuals by site and year to ensure errors of both random effects were centred at zero. Finally, I examined scatterplots of standardized residuals versus fitted values to ensure homoscesdasticity within site and year (Pinheiro and Bates 2000). All statistical analyses and evaluations of model assumptions were performed using the program R (The R Foundation for Statistical Computing Version 2.1.1; Pinheiro and Bates 2000). I constructed a series of 18 candidate models, comprised of competitor, predator and habitat effects, in various combinations with interaction effects. If competition limited red-60 breasted nuthatch densities, then I predicted the competitor models (mountain chickadee, black-capped chickadee, downy woodpecker) to rank highest among the suite of models. If nuthatch populations were driven by overall changes in the habitat then I predicted the habitat effects (beetle-infected pine, aspen, tree density, harvesting, edge) to rank highest. If irruptions in nuthatch populations were caused by changes in food availability, then I predicted beetle-infected pine trees to be the most significant habitat effect. If top-down, predator effects regulated nuthatch populations, then I predicted high ranking of predator models. Even though the top model was well supported (AICc wt = 0.69), I used the model-averaging approach to improve the accuracy of my predictions and calculated the average parameter estimate and standard error for each effect, using estimates generated from the top two models (where the sum of the AICc weights was 0.82), accounting for AICc weights of each model (Burnham and Anderson 1998). This approach protects against spurious results where a single model, not well supported by the data, may predict erroneously significant effects (Burnham and Anderson 1998). RESULTS Overall changes in populations at the study (regional) level Across all study sites, the average densities of red-breasted nuthatches and their competitors and predators fluctuated over the recent decade (Figure 3.1). In the years prior to the peak of the mountain pine beetle outbreak, densities of nuthatches remained at a fairly constant level, ranging from 0.08 to 0.13 individuals / ha from 1997 to 2002. Then, the densities of nuthatches doubled between 2002 and 2003. This doubling in densities occurred one year prior to the highest proportion of recently beetle-infected lodgepole pine trees observed during the study, in 2004 (Figure 3.2). During this time, there were similar densities 61 of mountain chickadees, which were also at their highest levels, and almost doubled between 2001 and 2002. In 2004, the year of the peak in recent beetle attack, black-capped chickadees were at their highest levels observed during the study and red squirrels increased to their second highest levels observed, following a steady increase since 2001. In 2005, however, nuthatches declined by 76% to an unprecedented low of 0.06 individuals / ha, coinciding with a drastic decline in recent beetle-infected pine. Mountain chickadees and black-capped chickadees also declined in 2005 by 26% and 80%, respectively. Densities of downy woodpeckers doubled from 2004 to their highest levels in 2005 and then declined by 50% in 2006. In 2006, mountain chickadees declined by another 29%, to the lowest density observed during the study. Nuthatches and other birds remained low in 2006 with the only increase shown by red squirrels. Concurrent correlations with nuthatch populations at time t at the stand level Nuthatches increased significantly with increased densities of mountain chickadees, black-capped chickadees, and red squirrels and harvested habitat, but decreased with natural edge habitat (Table 3.1). There were no significant correlations with densities of recent beetle-infected pine trees at the stand level, suggesting that nuthatches did not respond to summer food availability provided by adult mountain pine beetles. Predictors of nuthatch populations at time t+1 at the stand level Among the suite of 18 models, the best predictor of densities of nuthatches was a competitor model that was comprised of densities of red-breasted nuthatches, mountain chickadees, black-capped chickadees, and downy woodpeckers and an interaction of densities of nuthatches and downy woodpeckers, in the previous year (Table 3.2). Using this model, I 62 found that densities of nuthatches significantly decreased following years with increased densities of black-capped chickadees (8 = -1.04 ± 0.43 SE, p=0.02) and with decreased densities of downy woodpeckers (B = 1.24 ± 0.46 SE, p<0.01). Densities of nuthatches significantly increased following years with high densities of downy woodpeckers (B = 1.24 ± 0.46 SE, p<0.01). The next strongest model contained competitor and habitat effects, and like the top model, predicted a significantly negative effect of black-capped chickadees (B = -1.08 ± 0.44 SE, p=0.01) and a positive interaction of densities of nuthatches and downy woodpeckers on nuthatch densities (B = 1.22 ± 0.46 SE, p<0.01). This model was comprised of the same effects as the top model, but also included densities of beetle-infected pine and aspen trees. However, these habitat effects did not explain miich variation in nuthatch densities (beetle-infected pine: B = -0.39 ± 0.50 SE, p=0.44; aspen: B = -0.21 ±0.24 SE, p=0.40). When I averaged the parameter estimates across the top two models, my findings matched the predictions of the top model (Table 3.3). Densities of nuthatches decreased following years with increased black-capped chickadees (Figure 3.4). Furthermore, I observed negative density dependence in nuthatches following years of low downy woodpecker densities, but positive population growth with high downy woodpecker densities (Figures 3.5). Densities of nuthatches increased moderately following years of high mountain chickadees densities, though not significantly. I found no effect of proportionate densities of recent beetle-infected pine or densities of aspen trees at the stand scale. 63 D I S C U S S I O N Density dependence affected by interspecific interactions The strong interaction effect of nuthatches and downy woodpeckers in both top models suggested that it was a strong driver of nuthatch populations. I found evidence for negative density dependence of nuthatch populations, but only at low densities of downy woodpeckers (Figure 3.5). Overall densities of downy woodpeckers were low prior to the mountain pine beetle outbreak. During this time food probably limited nuthatch populations, providing a mechanism for density dependence, as predicted. The negative effects of density dependence were eliminated with high densities of downy woodpeckers, suggesting either covariation or facilitation of nuthatch populations by downy woodpeckers. The superabundance of food caused by the outbreak may have released the constraints of negative density dependence and interspecific competition, allowing populations to increase concurrently. Similarly, in southern Sweden, increases in European nuthatch densities and decreases in territory sizes correlated with increases in winter food supply (Enoksson and Nilsson 1983). Another explanation for my result is that downy woodpeckers facilitated nuthatch populations by increasing nest cavity availability. In outbreak years, nuthatches used cavities that were higher, deeper, and larger in entrance area (which were more characteristic of downy woodpecker than nuthatch cavities), than those used in pre-oufbreak years (Martin and Norris 2007). Also, I have some evidence that nuthatches used a greater proportion of old cavities during the outbreak (43% of 97 nests in 1997-2002 compared to 65% of 87 nests in 2003-2006, K. Martin, unpublished data). My results suggest that as nuthatches and downy woodpecker increased the food pulse, nuthatches may have benefited from this covariation because they potentially became limited by cavity availability, as they increased their use of old cavities. 64 Interspecific interactions My findings that the highest mean densities of nuthatches and both chickadee species occurred during the peak mountain pine beetle outbreak years at the study level and that nuthatches significantly increased with both chickadees at the site level suggests that all three species responded to similar habitat conditions, increasing the potential for interspecific competition. My result that densities of nuthatches decreased following years of high densities of black-capped chickadees provided some evidence that interspecific competition negatively affected nuthatch populations. Black-capped chickadees and red-breasted nuthatches compete for food and nest trees and will occasionally engage in aggressive interactions with each other during the breeding season, potentially causing nuthatches to be driven out of territories occupied by black-capped chickadees (Ghalambor and Martin 1999). However, competition would probably not lead to exclusion of one species because of the superabundance of food during this time. An alternative explanation for the negative association is that nuthatches changed breeding habitat selection criterion as the mountain pine beetle outbreak progressed, moving from areas of high aspen density, typically inhabited by black-capped chickadees, to areas of high beetle-infected pine density (Norris and Martin 2007). While the mountain pine beetle outbreak increased food availability and subsequent populations for nuthatches and chickadees, it may have influenced the breeding habitat of nuthatches more than the others, or nuthatches were more flexible in habitat selection, and moved from sympatric habitat providing more available habitat to black-capped chickadees, resulting in an apparently negative association. As nuthatches moved to more pine dominated habitats, both species may have been released from nest site and territorial competition with greater niche partitioning. 65 Habitat and predator effects Nuthatches may assess food availability during incipient stages of insect outbreaks (Crawford et al. 1990). Territory size often decreases with increased winter food supply, allowing populations to increase (Enoksson and Nilsson 1983). If winter food availability were expected to be high, then it would be advantageous for breeding pairs to maximize the number of young in the summer preceding the outbreak. Since bird surveys were conducted during the breeding season and vegetation surveys were conducted in late summer the correlation between nuthatch densities and recent beetle-infected pine provides evidence that nuthatches assessed winter food availability in the breeding season preceding the outbreak. Mean annual nesting densities increased and clutch sizes increased by approximately 30% with the outbreak, which suggests that nuthatches may have anticipated the winter food pulse and adjusted their reproductive output accordingly (Martin and Norris 2007). I did not examine territory size and movement among stands, but further research should test whether winter food supply decreases territory size in the breeding season, ameliorating density dependence and causing increases in nesting densities, maximizing overall reproductive effort. My result that nuthatches were negatively associated with edge habitat provides further evidence that nuthatches moved into the coniferous forest interior with the progression of the mountain pine beetle outbreak. Also, since nuthatches apparently increased with harvesting, this negative result is specific to naturally fragmented habitat, rather than due to harvesting. The naturally fragmented edge habitat was comprised of high proportions of aspen, areas typically inhabited by black-capped chickadees. Thus, I found that by avoiding naturally fragmented edges, nuthatches avoided competition with black-capped chickadees as nuthatch populations increased. In the next chapter, I examined how nest-patch selection of nuthatches changed with the progression of the beetle outbreak and found that in outbreak years, 66 nuthatches chose nest-patches with higher densities of beetle-infected pine trees and lower densities o f aspen trees, compared to pre-outbreak years. The benefits nuthatches accrued by moving closer to food sources and away from competitors may have been counterbalanced by costs associated with increased predation pressure. Densities of nuthatches significantly increased with densities o f squirrels in interior forests, which may have increased the potential for predation on nests. In chapter 2,1 found evidence that densities o f squirrels facilitated negative density dependence in mountain chickadee populations. However, red squirrels are generalist consumers and may switch from preying on bird nests to consuming seeds in years of crop masting, thus an increase in squirrel densities may not always influence demography of small-bodied cavity nesters (Mahon and Mart in 2006). Similarly, in this chapter, I found no evidence that squirrel populations negatively affected nuthatch populations. Predicted changes in climate are l ikely to increase the frequency and intensity o f natural disturbance events (Overpeck et al. 1990). The mountain pine beetle outbreak appeared to positively influence nuthatch populations, directly at the study level and indirectly through increases in downy woodpeckers at the.stand level. However, as food availability diminishes in post-epidemic stands, it is unknown how communities wi l l respond. Stone (1995) observed a similar peak and decline o f red-breasted nuthatch populations in relation to mountain pine beetle-killed lodgepole pine stands. The strong declines in 2005 and 2006 suggest that the food pulse is gone and populations are below endemic levels. It is possible that nuthatches have moved to other forest types, possibly to stands dominated by Douglas-fir and/or spruce. Their ability to assess food availability in incipient stages of insect outbreaks and their irruptive population dynamics might facilitate flexibility with respect to habitat type and condition. Since the mountain pine beetle outbreak is the largest ever in North America and expanding 67 eastward and southward, my results have a general application for small-bodied cavity-nester communities in North America. R E F E R E N C E S Aitken, K. E. H., K. L. Wiebe, and K. Martin. 2002. Nest-site reuse patterns for a cavity-nesting bird community in interior British Columbia. The Auk 119:391-402. Banfield, A. W. F. 1974. The mammals of Canada University of Toronto Press, Toronto, Ontario, CAN. Boutin, S. 1990. Food supplementation experiments with terrestrial vertebrates: patterns, problems, and the future. Canadian Journal of Zoology 68:203-220. Burnham, K. P., and D. R. Anderson. 1998. Model selection and inference: a practical information-theoretic approach. Springer-Verlag, New York, USA. Carroll, A. J., S. W. Taylor, and J. Regniere. 2004. Effects of climate change on range expansion by the mountain pine beetle in British Columbia. Pp. 223-232 In: Proceedings of "Mountain Pine Beetle Symposium: Challenges and Solutions", Kelowna, BC, Oct. 30-31, 2003. Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, Victoria, British Columbia, Information Report BC-X-399. 298 pages. Crawford, H. S., D. T. Jennings, and T. L. Stone. 1990. Red-breasted nuthatches detect early increases in spruce budworm populations. Northern Journal of Applied Forestry 7:81-83. Dennis, B., and M. R. M. Otten. 2000. Joint effects of density dependence and rainfall on abundance of San Joaquin kit fox. Journal of Wildlife Management 64:388-400. Dhondt, A. A., B. Kempenaers, and F. Adriaensen. 1992. Density-dependent clutch size caused by habitat heterogeneity. Journal of Animal Ecology 61:643-648. 68 Enoksson, B . and S. G . Nilsson. 1983. Territory size and population density in relation to food supply in the nuthatch (Sitta europaea). Journal of Animal Ecology 52:927-935. Fort, K . T. and K . A . Otter. 2004. Effects of habitat disturbance on reproduction in black-capped chickadees (Poecile atricapillus) in northern British Columbia. The A u k 121:1070-1080. Framstad, E . , N . C . Stenseth, O. N . Bjornstad, and W . Falck. 1997. Limi t cycles in Norwegian lemmings: tension between phase-dependent and density dependence. Proceedings o f the Royal Society London Series B 246:31-38. Ghalambor, C . K . , and T. E . Martin. 1999. Red-breasted nuthatch (Sitta canadensis). In A . Poole, and F. G i l l , editors. The Birds of North America, no. 459. Academy o f Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D . C . H i l l , B . G . , and M . R. Lein. 1989. Territory overlap and habitat use of sympatric chickadees. The A u k 106: 259-268. Jackson, J. A . , and H . R. Ouellet. 2002. Downy Woodpecker (Picoidespubescens). In A . Poole, and F. G i l l , editors. The Birds of North America, no. 613. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D . C . Koenig, W . D . 2001. Synchrony and periodicity of eruptions by boreal birds. Condor 103:725-735. Mahon, C. L . 2006. Temporal and spatial variation in habitat quality: breeding ecology of the chestnut-backed chickadee in uncut and partial cut forests in British Columbia PhD thesis. University of British Columbia, Vancouver, British Columbia. Mahon, C . L . and K . Martin. 2006. Nest survival o f chickadees in managed forests: Habitat, predator, and year effects. Journal of Wildlife Management 70:1257-1265. 69 Martin, K., K. E. H. Aitken, and K. L. Wiebe. 2004. Nest sites and nest webs for cavity-nesting communities in interior British Columbia, Canada: Nest characteristics and niche partitioning. Condor 106:5-19. Martin, K., and J. Eadie. 1999. Nest webs: a community-wide approach to the management and conservation of cavity-nesting forest birds. Forest Ecology and Management 115:243-257. Martin, K., and A R. Norris. 2007. Life in the small-bodied cavity-nester guild: Demography of sympatric mountain and black-capped chickadees within Nest Web communities under changing habitat conditions. Chapter 8, in Otter, K. (Ed.). Ecology and Behavior of Chickadees and Titmice: An integrated approach, Oxford University Press. Pp. 111-130. Martin, K., A.R. Norris, and M. Drever. 2006. Effects of bark beetle outbreaks on avian biodiversity in the British Columbia interior: Implications for critical habitat management. British Columbia Journal of Ecosystem Management 7:10-24. Martin, T. E. 1993. Nest predation and nest sites: new perspectives on old patterns. Bioscience 43:523-532. Matthysen, E. 1998. The nuthatches. T & A D Poyser, London. McCallum, D. A., R. Grundel, and D. L. Dahlsten. 1999. Mountain chickadee (Poecile gambeli). In A. Poole, and F. Gill, editors. The Birds of North America, no. 453. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. Meidinger, D., and J. Pojar, J. 1991. Ecosystems of British Columbia British Columbia Ministry of Forests Special Report Series, no. 6. Victoria, British Columbia. Norris, A. R. and K. Martin. 2007. Mountain pine beetle presence affects nest-patch choice of red-breasted nuthatches. Journal of Wildlife Management. In Press. 70 Otvos, I. S. 1979. The effects of insectivorous bird activities in forest ecosystems: An evaluation. In J. G. Dickson, R. N. Connor, R. R. Fleet, J. C. Kroll, and J. A. Jackson, (Editors). The Role of Insectivorous Birds in Forest Ecosystems, Academic Press, New York. Overpeck, J. T., D. Rind, and R. Goldberg. 1990. Climate-induced changes in forest disturbance and vegetation. Nature 343:51-53. Pinheiro, J. C , and D. M. Bates. 2000. Mixed-effects models in S and S-Plus. Springer-Verlag, New York, NY. Rodenhouse, N. L., T. S. Sillett, P. J. Doran, and R. T. Holmes. 2003. Multiple density-dependence mechanisms regulate a migratory bird population during the breeding season. Proceedings of the Royal Society of London. B: 270:2105-2110. Saether, B.-E., J. Tufto, S. Engen, K. Jerstad, O.W. Rostad and J. E. Skatan. 2000. Population dynamical consequences of climate change for a small temperate songbird. Science 287:854-856. Smith, S. M. 1993. Black-capped Chickadee (Parus atricapillus). In Poole, A. and Gill, F. editors. The Birds of North America, no. 39. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. Steen, H., and D. Haydon. 2000. Can population growth rates vary with the spatial scale at which they are measured? Journal of Animal Ecology 69:659-671. Stone, W. E. 1995. The impact of a mountain pine beetle epidemic on wildlife habitat and communities in post-epidemic stands of lodgepole pine forests in northern Utah. PhD thesis. Utah State University, Logan, Utah. Taylor, S. W., and A. L. Carroll. 2004. Disturbance, forest age, and mountain pine beetle outbreak dynamics in BC: A historical perspective. In T. L. Shore, J. E. Brooks, and J. E. 71 Stone, editors. Pages 41-51 in Mountain Pine Beetle Symposium: Challenges and Solutions. October 30-31, 2003, Kelowna, British Columbia, Canada. Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, Victoria, British Columbia, Information Report BC-X-399. The R Foundation for Statistical Computing Version 2.1.1 Westfall, J. 2004. 2004 Summary of Forest Health Conditions in British Columbia. Pest Management Report Number 15. Forest Practices Branch, British Columbia Ministry of Forests. 72 Table 3.1. Parameter estimates for the explanatory variables for red-breasted nuthatch densities at time, t: mountain chickadee (M), black-capped chickadee (Be), downy woodpecker (D), red squirrel (Sq), recent beetle-infected pine (Bi), aspen trees (At), proportion of edge habitat (E), tree density (St) and its quadratic form (St2), and harvested sites (H), with significant correlations with red-breasted nuthatch densities at the 0.05 level of confidence, bolded. Effect Estimate Std. E r r o r D F t value P(>|t|) (Intercept) 0.08 0.05 226 1.58 0.116 M 0.14 0.06 226 2.48 0.014 Be 0.34 0.11 226 3.03 0.003 D 0.12 0.08 226 1.39 0.166 Sq 0.14 0.04 226 3.71 <0.01 Bi -0.19 0.13 226 -1.46 0.147 At 0.06 0.05 226 1.28 0.201 E -0.07 0.03 226 -2.13 0.034 St -0.09 0.16 226 -0.58 0.566 St2 0.16 0.16 226 0.99 0.322 H 0.04 0.02 226 2.18 0.030 73 Table 3.2. General linear mixed models predicting densities of red-breasted nuthatches, at time, t+1, in central British Columbia from 1997 to 2006. Variables examined were densities of red-breasted nuthatches (Rb), mountain chickadees (M), black-capped chickadees (Be), downy woodpeckers (D), red squirrels (Sq), recent beetle-infected pine trees (Bi), aspen trees (At), all trees (St, and its quadratic form, St2), proportion of edge habitat (E), and instances of harvesting (H), at time, t. Site and year were included as random effects. For each model, the number of parameters (K), the number of observations (n), the maximum loglikelihood ratio (loglik), the Akaike's Information Criterion corrected for small samples (AICc), the difference in AICc compared to the model with the least AICc value (AAICc), and its weight (wt). Parameters Model description K N loglik AICc AAICc wt Rb M Be D Rb*D Competitor 9 202 -48.47 115.88 0.00 0.69 Rb M Be D Bi At Rb*D Competitor + Habitat 11 202 -47.92 119.23 3.35 0.13 Rb M Be D Rb*M Competitor 9 202 -50.865 120.67 4.79 0.06 Rb M Be D Rb*Bc Competitor 9 202 -51.812 122.56 6.68 0.02 Rb M Be D Bi At Rb*M Competitor + Habitat 11 202 -50.209 123.81 7.93 0.01 Rb M Be D Bi At Rb*At Competitor + Habitat 11 202 -50.333 124.06 8.18 0.01 Rb M Be D Bi At D*Bi Competitor + Habitat 11 202 -50.351 124.09 8.21 0.01 Rb M Be D Sq Bi At E St St2 H Full model (no interactions) 16 202 -44.741 124.42 8.54 0.01 Rb M Be D Bi At Rb*Bi Competitor + Habitat 11 202 -50.603 124.60 8.72 0.01 Rb M Be D Bi At Bc*Bi Competitor + Habitat 11 202 -50.635 124.66 8.78 0.01 Rb Bi At St St2 E Rb*E Habitat 11 202 -50.663 124.72 8.84 0.01 Rb Bi At St St2 E Rb*At Habitat 11 202 -50.732 124.85 8.98 0.01 Rb M Be D Bi At Rb*Bc Competitor + Habitat 11 202 -50.836 125.06 9.18 0.01 Rb Bi At St St2 E Rb*St2 Habitat 11 202 -51.266 125.92 10.04 0.00 Rb Bi At St St2 E Rb*Bi Habitat 11 202 -52.398 128.18 12.31 0.00 Rb M Be D Sq Bi AT E St St2 H Rb*M Rb*Bc Rb*D Rb*Sq Rb*Bi Rb*At Rb*E Rb*Sf2 Full model + interactions 23 202 -39.048 130.30 14.42 0.00 Rb Sq Bi At St St2 E Rb*Bi Predator + Habitat 12 202 -52.372 130.40 14.52 0.00 RbSq BiAtSt St2ERb*Sq Predator + Habitat 12 202 -52.847 131.34 15.47 0.00 Table 3.3. Model-averaged parameter estimates for the explanatory variables for red-breasted nuthatch densities (at time t+1) from top two models, where AAICc < 4. Variables examined were densities of red-breasted nuthatches (Rb; at time t), mountain chickadees (M), black-capped chickadees (Be), downy woodpeckers (D), recent beetle-infected pine trees (Bi), aspen trees (At) and an interaction of red-breasted nuthatches with downy woodpeckers (Rb*D). Significant effects on red-breasted nuthatch densities at the 0.05 level of confidence are bolded. Effect Estimate Std. E r r o r df t P Intercept -1.37 0.13 196 -8.38 <0.01 Rb -0.04 0.06 196 -0.50 0.44 M 0.25 0.19 196 1.09 0.15 Be -0.86 0.36 196 -1.98 0.01 D 1.86 0.63 196 2.41 <0.01 Bi -0.-39 0.50 194 -0.78 0.44 At -0.21 0.24 194 -0.85 0.40 Rb*D 1.01 0.37 196 2.22 0.01 75 0.35 0.3 0.25 -\ — • — red-breasted nuthatch - -Q- - mountain chickadee - - * - - black-capped chickadee • - -O- • • downy woodpecker • - -x- - - red squirrel —0 1 1 1 1 1 1 f-1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Figure 3.1. Mean detections of red-breasted nuthatches, mountain chickadees, black-capped chickadees and red squirrels, across point count surveys from 1997 to 2006, on 27 forested sites in central British Columbia. 76 0.3 0.25 CD O 13 =3 C o I 0.1 0.2 0.15 0.05 -•— Red-breasted nuthatch -a- - Beetle-infected pine 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 _ C 'a. T J t3 JP S 15 c o "•c o CL o 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Figure 3.2. Mean detections of red-breasted nuthatches and recent beetle-infected lodgepole pine trees across 27 forested sites using point count surveys, and vegetation surveys of 11.3 metre fixed radius plots centred around each point count station, in central British Columbia from 1997-2006. Mean densities of recent beetle-infected pine trees were proportionate to all trees surveyed in order to obtain an estimate of relative food availability on stands. I calculated a Pearson's correlation coefficient to determine how mean annual densities of nuthatches correlated with proportionate densities of recent beetle-rinfected pine trees, from 1997 to 2006; R=0.73,n=10,p=0.02. 77 Figure 3.3a o o <D o «3 _: l o Cvl I o g J * ^ 0 'oV**yo o ® o J ° ° ° o o ?) _ o o o X J (V o f t o _ o ° 0 o O {BOCDOO o o 0.0 T T 0.1 0.2 0.3 0.4 M e a n # of mounta in c h i c k a d e e s / h a Figure 3.3b o + o Id C o s o I o 7 I I o o .oo o 8 ° % 0 GkxO ^ 5 o c cr <D 0 < O O O O GD'O OO o. J 0.0 0.1 0.2 0.3 0.5 o Mean # of b lack-capped chickarJees/ha 78 Figure 3.3c o o T IT) T I o o o <2 8 o o "b o ^ % <V o o " o O O O o -I o. o nJ „<fc oc-<s» o Q> © o o o c 1 i 0.2 o o 0.0 0.4 0.6 0.8 M e a n # of red squ i r re ls /ha Figure 3.3d c o O 1 1 1 1 T~ . 0.2 0,4 0.6 0.8 1.0 M e a n proport ion of s tand within 50m of a n edge Figure 3.3. Mean detections of red-breasted nuthatches and a.) mountain chickadees, b.) black-capped chickadees, c.) red squirrels, and d.) proportion of habitat within 50m of an a forest edge for each of 27 stands, from 1997-2006. Trend lines were calculated using parameter estimates generated by the correlative model, ln(Rbt + 0.1) = M + Bc + D + Sq + Bi + At + E + St + St2 + H + Site + Year. 79 0 P -0.5 -d + + -2.5 "1 1 •—i 1 1 1 1 1 1 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 Mean # of black-capped chickadees/ha, at Nt Figure 3.4. Mean detections of red-bfeasted nuthatches (at time, t+1) and black-capped chickadees (at time, t) at 27 stands, from 1997-2006. Trend line was calculated using model-averaged parameter estimates from the top two predictive models, ln(Rbt+i + 0.1) = ln(Rbt + 0.1) + M + Be + D + Rb*D + Site + Year, and; ln(Rb,+i + 0.1) = ln(Rb, + 0.1) + M + Be + D + Bi + At + Rb*D + Site + Year. Trend line indicates that nuthatch populations in subsequent years decreased with increasing black-capped chickadees (15 = -0.86 ±0.36, ti96=-1.98, p=0.01). 80 o + CD J= "w CD J Z o CD £ C " 5 *t c CD CD E •o . CD O Vj CD i— CL -1.15 -1.2 H -1.25 -1.3 -1.35 -1.4 -1.45 high downy woodpecker density — - low downy woodpecker density -2.5 0 -2 -1.5 -1 -0.5 Mean # of nuthatches/ha, ln(Nt + 0.1) Figure 3.5. Predicted mean detections of red-breasted nuthatches (t+1) with increasing conspecific densities (t) at high (upper 75th percentile of data) and low (lower 65th percentile) densities of downy woodpeckers. Trend lines were calculated using the model-averaged parameter estimates from the predictive model, ln(Rbt+i + 0.1) = ln(Rbt + 0.1) + M + Bc + D + Bi + At + Rb*D + Site + Year. 81 C H A P T E R 4 : M O U N T A I N P I N E B E E T L E P R E S E N C E A F F E C T S N E S T P A T C H C H O I C E O F R E D - B R E A S T E D N U T H A T C H E S 1 I N T R O D U C T I O N Red-breasted nuthatches (Sitta canadensis) are small-bodied cavity-nesting birds common in old forest stands in western North America (Ghalambor and Martin 1999). In central British Columbia, nuthatches are year-round residents, and typically occupy mature mixed stands of coniferous-deciduous forest, especially where spruce (Abies spp.), Douglas-fir (Pseudotsuga menziesii) and lodgepole pine (Pinus contortd) are present (Martin and Norris 2007). During the breeding season, nuthatches are primarily bark and foliage-gleaning insectivores, but will take advantage of the most abundant food supply available, including seeds and berries (Ghalambor and Martin 1999). They are classified as facultative weak primary cavity-nesters and will either excavate new nest cavities in dead or decaying trees or re-use existing cavities. Choosing a patch of habitat within a landscape in which to situate a nest, or, "nest patch selection", by red-breasted nuthatches can be influenced by nest tree condition and availability (Li and Martin 1991, Steeger and Hitchcock 1998, Bunnell et al. 2002, Martin et al. 2004). Rotting hardwood trees are often selected over healthy or coniferous trees because they offer a soft substrate for excavation while retaining a firm shell of sapwood, for stable and secure cavities (Ghalambor and Martin 1999). Red-breasted nuthatches in central British Columbia selected 90% of nests in decaying or dead aspen trees even though these comprised only 10-1 A version of this chapter is in review for publication. Norris, A. R., and K. Martin. 2007. Mountain pine beetle presence affects nest-patch choice of red-breasted nuthatches. Journal of Wildlife Management. 82. 15% of trees available (Martin et al. 2004). These trees occur at a low frequency in old forests and are further reduced in managed forests (Steeger and Hitchcock 1998). Food abundance and proximity to foraging habitat can also influence nest patch selection (Li and Martin 1991, Stauss et al. 2005). Nests close to abundant food resources allow adults to make more frequent foraging trips, minimizing the time away from the nest and increasing the number of successful fledglings (Eeva etal. 1989, Stauss et al. 2005). Dead and decaying trees offer greater food availability because they tend to harbor a greater abundance of insects than healthy trees (Allen et al. 1996). In south-eastern British Columbia, the best predictors of nuthatch density were densities of standing dead trees, particularly those killed by mountain pine beetle (Steeger and Hitchcock 1998). In northern Utah, red-breasted nuthatch abundance increased with degree of mountain pine beetle infection in stands of lodgepole pine up to a maximum of 2 nuthatches/ha in stands with 70% tree mortality (Stone 1995). Thus, rotting hardwood trees represent nest trees and dead lodgepole pine trees represent food availability for nuthatches. Within the last decade, the condition of central interior forests of British Columbia has changed drastically due to a mountain pine beetle outbreak (Martin et al. 2006). The outbreak caused a high degree of spatial variability in beetle-attacked trees, causing the foraging substrate to be unevenly distributed across the landscape (Martin et al. 2006). Furthermore, aspen trees may also be distributed unevenly, thus imposing a potential trade-off between greater hest-tree or food availability on nest patches. My objective was to determine whether nuthatch nest patch selection changed during this outbreak, possibly due to the trade-off between patches with abundant nest trees versus foraging trees. If nuthatch nest patch selection criteria shifted from abundance of nest trees to foraging trees during the beetle outbreak, then I expected nest patch vegetation to have a greater density of live and dead aspen trees than 83 "available" patches (those containing at least one suitable-sized dead aspen nest tree) during pre-outbreak years, and for nest patches to have a greater density of beetle-infected pine trees than available patches during outbreak years. While multiple studies found significant associations of nuthatches with diseased trees (Steeger and Hitchcock 1998, Stone 1995), my study is the first to document nest and foraging substrate availability and use before and during a major mountain pine beetle outbreak and examine how trade-offs in these two resources change with the outbreak. METHODS Study area The study area was located within the warm and dry Interior Douglas-fir biogeoclimatic zone, near Williams Lake, in central interior British Columbia (51°52'N, 122°21'W). Predominant tree species were interior Douglas-fir and lodgepole pine interspersed with patches of grassland and stands of quaking aspen (Populus tremuloides; Martin and Eadie 1999). The 27 sites were located within 40 km of Williams Lake, and ranged from deciduous and coniferous forest stands (>30 ha) to small, isolated, natural forest fragments (0.1 to 5 ha), in a matrix of grassland, ponds, and wetlands. Nest patch selection Between May and July 1995-2005,1 conducted systematic nest searches for all active nests on all sites, as well as checking old nest trees for re-use. A nest was considered active if at least one egg or chick was observed upon visual inspection of a cavity, or an attending pair was observed feeding nestlings. Nests were monitored until the nestlings fledged from the cavity. Each year, vegetation data were collected at all active nest patches and at systematically placed plots (representing availability), 100m apart throughout all sites. Tree 84 species, diameter at breast height (DBH), decay class (decay class 1 was a live, healthy tree, 2 a live tree with visible sign of disease or decay such as bark boring insects, 3-8 were standing dead trees; Thomas et al. 1979), and general health (e.g., presence of boring insects) were recorded for all trees 1^2.5 cm DBH within an 11.3 m radius (0.04 ha) circular plot centered around each nest tree or "available" plot. Mountain pine beetles were detected on lodgepole pine trees by the presence of outflows of dried resin on the outer bark, or by small holes (~2mm in diameter) in the bark. Density of beetle-infected pine was determined by the number of lodgepole pine trees with evidence of bark boring insects/0.04 ha. Additional nest monitoring and vegetation survey methodology are given in Aitken et al. (2002) and Martin et al. (2004). I assumed that nuthatches had equal access to all habitat within each site, but were probably limited by nest tree availability so I chose all systematic vegetation plots in my study with at least one standing dead aspen tree that was within the observed range of variation in size of nest trees used by nuthatches on my sites (x = 23.2 ±7.6 cm DBH; Martin et al. 2004), hereafter "suitable nest trees," and used these as my "available" patches (n = 1136 plots). Densities of aspen and suitable nest trees were determined from the number of total aspen (live and dead) and standing dead aspen trees/0.04 ha, respectively. Statistical analyses I used generalized linear mixed-effects models to determine whether nuthatch nest patches differed in tree species composition and condition from those available before and during a mountain pine beetle outbreak. I converted the binary dependent data (used or available) into a logistic distribution using a logistic regression (logit) link function and used penalized quasi-likelihood ratios for parameter estimation (Breslow and Clayton 1993). Mixed-effects models allow the use of unbalanced and potentially pseudoreplicated data, by splitting the within-group variation due to random effects, from the between-group variation due to fixed effects 85 (Pinheiro and Bates 2000). Fixed effects were density of aspen, density of suitable nest trees (standing dead aspen), and density of beetle-infected pine trees for each patch. I included site and year as random effects in all models to control for site-specific persistent effects (e.g., similar tree species composition and condition within sites or within site variation due to harvesting) and potential pseudoreplication and unbalanced data across years due to multiple measurements of systematic plots and nest patches (e.g., some nest patches were used by nuthatches for multiple years). All two-way interaction terms were examined for each model constructed but none were significant, so I present only the most parsimonious univariate effects models. I constructed three separate models, based on the mountain pine beetle outbreak time period, to compare the relative contribution of each effect on nest patch selection between periods. The "pre-outbreak" model included data from 1995-2002; "outbreak" from 2003-2005; and "overall" included all years. All statistical analyses were conducted using the program R (R Development Core Team 2006). RESULTS I constructed three generalized linear mixed models to compare densities of aspen trees, suitable nest trees and beetle-infected pine trees among patches selected for nesting (n=231) to those available, unused patches (n=l 136). Of the 231 nests examined in this study, 90% were located in aspen trees. As predicted, nuthatches selected patches with significantly higher mean densities of aspen trees in pre-outbreak years (used = 316 trees/ha ± 24 SE; available = 219 trees/ha ± 7 SE), during the outbreak (used = 248 trees/ha ± 22 SE; available = 199 trees/ha ± 11 SE), and across all years (used = 288 trees/ha ± 17 SE; available = 213 trees/ha ± 6 SE; Table 4.1). However, in outbreak years, nuthatches chose patches with lower mean densities of aspen compared to pre-outbreak years (Figure 4.1). 86 Nest patches had significantly higher mean densities of suitable nest trees than available patches in pre-outbreak years (used =103 trees/ha ± 14 SE; available = 63 trees/ha ± 2 SE) and across all years (used = 90 trees/ha ± 9 SE; available = 63 trees/ha ± 2 SE), but not in outbreak years (used = 69 trees/ha ± 6 SE; available = 62 trees/ha ± 2 SE; Table 1, Figure 4.1). The mean density of beetle-infected pine trees among used (29 trees/ha ± 5 SE) and available patches (24 trees/ha ± 2 SE) did not differ prior to the outbreak (Figure 4.1, Table 4.1). However, nuthatches selected patches with significantly higher mean densities of beetle-infected pine trees in outbreak years (used = 63 trees/ha ±11 SE; available = 46 trees/ha ± 5 SE) and across all years (used = 43 trees/ha ± 6 SE; available = 30 trees/ha ± 2 SE). DISCUSSION Nest patch selection in pre-outbreak years Nuthatches showed a strong preference for nest patches with high aspen densities, probably as a result of greater nest tree availability in these patches (Martin et al. 2004). Quaking aspen was the dominant deciduous tree species in the study area and exhibited both clumped and single-tree distribution, depending on the mode of reproduction (sexual vs. asexual; Callan 1998). In asexually reproducing aspen, the presence of one dead or decaying aspen tree often indicates that other trees in the aggregation are in the same condition providing multiple nest trees within a single patch. Nest patch selection varied temporally as I found that nest patches contained higher densities of aspen and suitable nest trees, but only during pre-outbreak years (Figure 4.1). I provide three possible explanations for my result. First, nuthatches may have cued in to aggregated aspen because these patches would likely contain a suitable nest tree, thus minimizing the time spent searching for a nest site. Or, nuthatches may have selected nest trees in patches because these were preferable to nesting in 87 dispersed trees or low density aspen patches for foraging or other activities. Lastly, nesting in patches with high densities of aspen and suitable nest trees may increase the probability of other cavity-nesters nesting nearby, increasing vigilance for predators in the area (Brown and Brown 1987, Soler and Soler 1996). Nest patch selection in outbreak years During outbreak years, nuthatches switched to nest patches with higher densities of beetle-infected pine trees. Nest patches in outbreak years had lower densities of aspen than those in pre-outbreak years, probably due to shifting to mixed forest patches with beetle-infected pine. Also, this shift probably resulted in fewer suitable nest trees in patches. Thus, preference for both aspen and beetle-infected pine trees may have constrained nest patch selection. My data support the conclusions of previous studies that forest insect outbreaks contribute to habitat preference and nest patch choice of cavity nesting birds (Conner et al. 1999, Crawford and Jennings 1989, Morris et al. 1958, Steeger and Hitchcock 1998). Patches containing high densities of aspen and beetle-infected pine were preferred over other suitable patches, possibly due to greater food availability. Although I did not examine the entire arthropod community, mountain pine beetle were likely the most prevalent insect available for bark gleaning birds in my forest stands during the outbreak and may have constituted a significant portion of their diets, especially after 2001 (A. Norris, personal observation). Nuthatches rely on adult beetles and larvae as a primary food source during the breeding season and over winter; beetles comprised 80% of nuthatch diets in ponderosa pine (Pinus ponderosa) forests, and 64% in Douglas-fir forests during the breeding season in Oregon (Anderson 1976). Food availability in aspen trees and other conifers may have been higher than in pine trees prior to the outbreak, but the relative abundance of insects in those tree species declined with the beetle outbreak, precipitating stronger selection for pine patches. However, my observation that nest patches 88 prior to 2003 contained higher densities of pine than available indicates a general preference for patches with pine trees even before the beetle outbreak. Implications for populations As year-round residents and bark insectivores, red-breasted nuthatches could be expected to show both functional and numerical responses to mountain pine beetle outbreaks (Morris et al. 1958, Crawford et al. 1990, Stone 1995). Mountain pine beetles provide nuthatches with increases in both winter food supply (developing larva underneath bark) and summer food supply (emerging adults) potentially increasing survival and subsequent population densities. I observed increased densities of nuthatch nests with the onset of the mountain pine beetle outbreak (Martin et al. 2006), and an increase in average clutch size from 5.0 ± 0.25 SE eggs in 1999-2001 (n=38 nests) to 5.6 ± 0.20 SE in 2002-2005 (n=79 nests; Martin and Norris, unpublished data). This suggests that nuthatches may be able to predict summer food availability and select territories accordingly, resulting in increased reproductive effort. While nuthatches showed immediate positive responses to mountain pine beetle outbreaks, populations in the post-outbreak phase may decline due to increased mortality of pine and depletion of food resources. Stone (1995) reported declines in red-breasted nuthatch abundance after 70% mortality of lodgepole pine stands caused by mountain pine beetle. Forest insect outbreaks are often accompanied by increased harvesting or targeted cutting of dead trees, thus imposing further constraints on nesting and foraging habitat and exacerbating the detrimental effects of diminishing habitat quality in post-epidemic stands (Martin et al. 2006). Weak cavity-nesters are limited by availability of nest trees and food, which are often linked because decayed trees offer both nest sites and foraging opportunities. Preference for aspen trees for nesting and conifer bark beetle-attacked trees for feeding presents a potential 89 trade-off between selecting patches with better nesting habitat or foraging habitat. In mixed forests red-breasted nuthatches can capitalize on bark beetle outbreaks because aspen trees are still available for nesting (albeit fewer), but foraging conditions are improved. MANAGEMENT IMPLICATIONS It is important to retain a range of habitat stands and conditions to sustain red-breasted nuthatch and other small-bodied cavity-nester populations during post-epidemic conditions. My result that nuthatches selected diseased and decaying aspen and pine trees suggests that managers should maintain an array of tree species and conditions in mixed forest stands to maintain cavity-nester breeding habitat. Diseased trees offer both nesting and foraging substrates, and healthy trees offer potential future resources. REFERENCES Aitken, K. E .H., K. L. Wiebe, and K. Martin. 2002. Nest-site reuse patterns for a cavity-nesting bird community in interior British Columbia. The Auk 119:391-402. Allen, E. A., D. J. Morrison, and G. W. Wallis. 1996. Common tree diseases of British Columbia. Canadian Forest Service, Pacific Forestry Centre, Victoria B.C. Anderson, S. H. 1976. Comparative food habits of Oregon nuthatches. Northwest Science 50: 213-221. Breslow, N. E., and D. G. Clayton. 1993. Approximate inference in generalized linear mixed models. Journal of the American Statistical Association 88:9-25. Brown, C.R., and M.B. Brown. 1987. Group-living in cliff swallows as an advantage in avoiding predators. Behavioral Ecology and Sociobiology 21:97-108. Bunnell, F. L., M. Boyland and E. Wind. 2002. How should I spatially distribute dying and 90 dead wood? United States Department of Agriculture Forest Service General Technical Report PSW-GTR-181. Callan, B. E. 1998. Diseases of Populus in British Columbia: a diagnostic manual. Natural Resources Canada, Canadian Forest Service. Conner, E. F., J. M. Yoder, and J. A. May. 1999. Density-related predation by the Carolina chickadee, Poecile carolinensis, on the leaf-mining moth, Cameraria hamadryadella at three spatial scales. Oikos 87:105-112. Crawford, H. S., and D. T. Jennings. 1989. Predation by birds on spruce budworm Choristoneura fumiferana: Functional, numerical, and total responses. Ecology 70:152-163. Crawford, H. S., D. T. Jennings, and T. L. Stone. 1990. Red-breasted nuthatches detect early increases in spruce budworm populations. Northern Journal of Applied Forestry 7:81-83. Eeva, T., E. Lehikoinen, and S. Veistola. 1989. Habitat preference and breeding performance in four hole-nesting passerines at the northern limit of their range. Ornis Fenn 66:142-150. Ghalambor, C. K., and T. E. Martin. 1999. Red-breasted Nuthatch (Sitta canadensis). Account 459 in A. Poole and F. Gill, editors/The birds of North America, The Academy of Natural Sciences, Philadelphia, Pennsylvania, and The American Ornithologists' Union, Washington, D.C, USA. Li, P., and T. E. Martin. 1991. Nest-site selection and nesting success of cavity-nesting birds in high elevation forest drainages. The Auk 108:405-418. Martin, K., and J. Eadie. 1999. Nest webs: a community-wide approach to the management and conservation of cavity-nesting forest birds. Forest Ecology and Management 91 115:243-257. Martin, K., and A. R. Norris. 2007. Life in the small-bodied cavity nester guild: Demography of sympatric Mountain and Black-capped Chickadees within Nest Web communities under changing habitat conditions. Pages 111-130 in K. Otter, editor. Ecology and Behavior of Chickadees and Titmice: An integrated approach, Oxford University Press. Martin, K., K. E. H. Aitken, and K. Wiebe. 2004. Nest sites and nest webs for cavity-nesting communities in interior British Columbia, Canada: Nest characteristics and niche partitioning. Condor 106:5-19. Martin, K., A. Norris, and M. Drever. 2006. Beetles in the Nest Web: Impacts of bark beetle populations on wildlife communities in Interior British Columbia Forests -11 years of wildlife-habitat data British Columbia Journal of Ecosystem Management 7:10-24. Morris, R. F., W. F. Cheshire, C. A. Miller, and D. G. Mott. 1958. The numerical response of avian and mammalian predators during a gradation of the spruce budworm. Ecology 39:487-494. Pinheiro, J. C , and D. M. Bates. 2000. Mixed-Effects Models in S and S-PLUS, Statistics and Computing Series, Springer-Verlag, New York, NY. R Development Core Team. 2006. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Soler, M., and J.J. Soler. 1996. Effects of experimental food provisioning on reproduction in the Jackdaw, Corvus monedula, a semi-colonial species. Ibis, 138:377-383. Stauss, M. J., J. F. Burkhardt, and J. Tomiuk. 2005. Foraging flight distances as a measure of parental effort in blue tits Parus caeruleus differ with environmental conditions. Journal of Avian Biology 36:47-56. Steeger, C, and C. L. Hitchcock. 1998. Influence of forest structure and diseases on nesting 92 habitat selection by red-breasted nuthatches. Journal of Wildlife Management 62:1349-1358. Stone, W. E. 1995. The impact of a mountain pine beetle epidemic on wildlife habitat and communities in post-epidemic stands of lodgepole pine forests in northern Utah. PhD thesis. Utah State University, Logan, Utah. Thomas, J., R. Anderson, C. Masser, E. Bull. 1979. Snags: Wildlife habitats in managed forests, the Blue Mountains of Oregon and Washington. United States Department of Agriculture, Forest Service, Agriculture Handbook Number 553, Washington, D.C. 93 Table 4.1. Generalized linear mixed effects model parameter estimates of predicted red-breasted nuthatch nest patch use with the following fixed effects: Total aspen density, beetle-infected lodgepole pine density, and suitable nest tree density in central British Columbia, Canada. Time period Parameter Estimate SE DF t P Pre-outbreak Intercept -2.491 0.254 765 -9.809 <0.001 (1995-2002) Total aspen 0.035 0.012 765 2.873 0.004 Beetle-infected pine 0.029 0.050 765 0.569 0.569 Suitable nest trees 0.090 0.032 765 2.776 0.006 Outbreak Intercept -1.792 0.282 349 -6.352 <0.001 (2003-2005) Total aspen 0.038 0.018 349 2.181 0.030 Beetle-infected pine 0.090 0.032 349 2.756 0.006 Suitable nest trees 0.010 0.060 349 0.166 0.868 All years Intercept -2.321 0.207 1117 11.218 <0.001 (1995-2005) Total aspen 0.037 0.010 1117 3.702 <0.001 Beetle-infected pine 0.085 0.027 1117 3.144 0.002 Suitable nest trees 0.067 0.028 1117 2.410 0.016 94 Figure 4.1. Density (No./ha) of quaking aspen, suitable nest trees (standing dead aspen > 12.5 cm DBH), lodgepole pine and beetle-infected lodgepole pine trees in patches containing at least one suitable nest tree (Available), and in red-breasted nuthatch nest patches (Used) for pre-outbreak, outbreak, and across all years in central British Columbia, Canada. Error bars represent standard error of the mean. 95 • Pre-outbreak (1995-2002) • Outbreak (2003-2005) • All years (1995-2005) Quaking aspen Suitable nest trees 400 350 -I 300 | 250 CO a> £ 200 c CO | 150 100 50 0 140 120 100 80 60 H 40 20 0 160 i 140 120 Lodgepole pine CO 100 CO CD 2 80 -I c CO ^ 60 40 20 -I 0 5 Available n = 1136 Used n = 231 90 80 70 60 50 40 -30 -20 10 • 0 Beetle-infected pine 0 0 Available n = 1136 T Used n = 231 96 CHAPTER 5: GENERAL DISCUSSION In this study, temporal variation in resource availability influenced the specific roles of community members, which had varying impacts on mountain chickadee and red-breasted nuthatch populations. For mountain chickadees, the mountain pine beetle outbreak elicited bottom-up effects on densities of facilitators (woodpeckers and nuthatches), which contributed to a secondary resource pulse of increased nest-site availability, which may have allowed chickadee populations to increase. I was not able to test whether mountain chickadees were limited by nest-site availability, however, nest-site addition experiments in the area resulted in a significant increase in nesting densities of mountain chickadees on treatment sites, indicating nest-site limitation (Aitken 2007). My findings that densities of chickadees increased with nuthatches and that population growth was higher following high sapsucker densities, suggested that nuthatches and sapsuckers facilitated chickadee populations. My findings that population growth was higher in uncut than cut stands and that populations increased with tree density to a maximum of 500 trees/ha and then declined, suggested that mountain chickadees prefer the forest interior, as was found in other studies of mountain chickadee habitat (McCallum et al. 1999). Higher proportions of edge were only detrimental to chickadee populations following years of high densities on these stands, suggesting that edge intensified the negative effects of density dependence. In addition, high densities of red squirrels had a significantly negative top-down effect on mountain chickadee populations. Densities of red squirrels were also negatively correlated with forest edge, so while density dependence was ameliorated for chickadees in interior forests, predation risk may have counteracted this positive effect. Increasing squirrel populations and the risk of nest predation may have displaced mountain chickadees from interior forests into stands with high proportions of edge habitat, where populations were then regulated by density dependence. 97 For red-breasted nuthatches, the mountain pine beetle outbreak increased food availability, ameliorating the negative effects of density dependence. My result that nuthatch densities increased with high densities of downy woodpeckers provided some evidence for facilitation. My result that nuthatch densities decreased with increasing black-capped chickadees provided some evidence for interspecific competition limiting populations. As the outbreak progressed nuthatches moved from aspen-dominated forest edges into beetle-infected pine dominated interior forests. This shift may have caused nuthatches to spend less time searching for food and more time searching for a suitable nest tree, thus the positive effect of downy woodpeckers might have been due to increased nest-site availability. Nesting densities of nuthatches increased on stands where nest-site availability was experimentally increased, suggesting that nuthatches may assess nest-site availability (Aitken 2007). I assumed that red-breasted nuthatches were not limited by nest-site availability due to their ability to excavate, however, with the food pulse, they may have had fewer options for nesting as they moved into conifer-dominated interior forests, potentially causing nest-site limitation to increase. This would then cause increased competition with other secondary cavity-nesters and more reliance on facilitator species, causing further niche overlap with the mountain chickadee. But the advantages accrued from nesting in areas of high food availability may have outweighed the costs associated with competing for cavities. The pulse of food may have allowed reproductive output to increase, and nests in old cavities often have larger clutches than those in freshly excavated cavities because clutches may be initiated earlier (Wiebe et al. 2007). Martin and Norris (2007) found that mean clutch size of nuthatches was higher during outbreak years, than in pre-outbreak years. In summary, I found some evidence for population regulation of nuthatches by competitors, however these effects were confounded by changes in habitat conditions and subsequent changes in settling patterns. 98 I found some evidence of density dependent regulation of both mountain chickadee and red-breasted nuthatch populations. Although I examined changes in population densities and local habitat conditions to explain my observed changes in population densities of mountain chickadees and red-bfeasted nuthatches, these are correlations that may not be causally linked. Recent work showed increases in annual fecundity of mountain chickadees and red-breasted nuthatches with the progression of the mountain pine beetle outbreak (Martin and Norris 2007). However these findings may be confounded by concurrent increases in another summer food supply in the area, western spruce budworm (Choristoneura occidentalis). Other studies where population densities strongly regulate populations found direct links between fecundity and population size (Enoksson and Nilssori 1983, Rodenhouse et al. 2003). As my findings were correlations of population densities with potential food supply, further research is needed to determine how fecundity and recruitment directly interact with population density to affect mechanisms causing regulation of populations. Due to their sympatric patterns of co-occurrence, similar body size, and niche overlap within the cavity-nester community, chickadees and nuthatches are often assumed to be close competitors (Martin et al. 2004, Hill and Lein 1989). Some studies have reported instances of interspecific aggression in relation to foraging and territory establishment between the two species, while others have shown that nuthatches may assess predation risk using chickadee alarm calls, and therefore may benefit from their presence (Minock 1972, Ghalambor and Martin 1999, McCallum et al. 1999, Templeton and Greene 2007). Using call playback experiments of nuthatches and chickadees, I found instances where both species exhibited high levels aggression in response territorial calls by their heterospecific counterpart (Norris, unpublished data). Furthermore, I reported a rare case of nest-sharing between red-breasted nuthatch and mountain chickadee in my study area, indicating competition for nest cavities 99 (Robinson et al. 2005). I found both positive and negative effects of interspecific interactions on the two species, which were seemingly confounded by habitat heterogeneity. A more in-depth study disentangling habitat heterogeneity from interspecific interactions is needed to determine how species interact at the behavioural level and how direct competition might influence settling patterns in different habitats. Two main questions arise from this research: 1.) How does food availability affect fecundity and survival and what are the consequences of increased fecundity and recruitment on population regulation? 2.) How do interspecific interactions of individuals influence settling patterns? Future research should also include tests of the correlative relationships found in this study to examine how unstable resource constraints (nest sites, food and predation) imposed by the mountain pine beetle outbreak directly affect breeding success and survival of populations. Finally studies should examine the persistence of these bark insectivores in post-beetle epidemic forests to determine how populations adjust to drastic declines in food availability. Conservation and management implications Chickadees and nuthatches are ideal umbrella species for management of forest songbirds. Revealing critical drivers in populations and determining how these drivers change with the beetle outbreak at the population and community level will aid in beetle management activities. Management criteria developed for these species may provide conservative guidelines for the management of other landbirds in these conditions. Predicted changes in climate are likely to increase the frequency and intensity of natural disturbance events (Overpeck et al. 1990). My results indicate that the large-scale natural disturbance event of the mountain pine beetle outbreak had cascading effects from communities to populations to nest-patch selection of individuals. Furthermore, community 100 dynamics determined whether density dependent process regulated chickadee and nuthatch populations. Thus management of chickadees and nuthatches needs to be within the greater context of the cavity-nester community rather than at the scale of the species. Since temporal variation in food supply affected facilitators, competitors and predators, managers need to examine annual changes in the community in relation to habitat conditions. Long-term monitoring will help predict annual variation in chickadee and nuthatch populations. Dead or decaying aspen trees and beetle-infected pine trees are critical habitat components for cavity-nesting birds in mature coniferous and mixed forests. Diseased trees offer both nesting and foraging substrates, and healthy trees offer food and potential future resources. The increased annual allowable cut and other beetle-salvage management activities that involve removal of these resources impose further constraints on already declining populations (Lindenmayer and Noss 2006). It is important to retain a range of habitat stands and conditions in mixed forests to sustain mountain chickadee and red-breasted nuthatch populations during post-epidemic conditions (Martin et al. 2006). REFERENCES Aitken, K. E. H. 2007. Resource availability and limitation in a cavity-nesting bird and mammal community in mature conifer forests and aspen groves of interior British Columbia. PhD thesis. University of British Columbia, Vancouver, British Columbia. Enoksson, B. and S. G. Nilsson. 1983. Territory size and population density in relation to food supply in the nuthatch (Sitta europaea). Journal of Animal Ecology 52:927-935. Ghalambor, C. K., and T. E. Martin. 1999. Red-breasted nuthatch (Sitta canadensis). In A. Poole, and F. Gill, editors. The Birds of North America, no. 459. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. 101 Hill, B. G., and M. R. Lein. 1989. Territory overlap and habitat use of sympatric chickadees. The Auk 106:259-268. Lindenmayer, D. B., and R. F. Noss. 2006. Salvage Logging, Ecosystem Processes, and Biodiversity Conservation. Conservation Biology 20:949-958. Martin, K., and A R. Norris. 2007. Life in the small-bodied cavity-nester guild: Demography of sympatric Mountain and Black-capped Chickadees within Nest Web communities under changing habitat conditions. Chapter 8, in Otter, K. (Ed.). Ecology and Behavior of Chickadees and Titmice: An integrated approach, Oxford University Press. Invited book chapter. Pp. 111-130. Martin, K., K. E. H. Aitken, and K. L. Wiebe. 2004. Nest sites and nest webs for cavity-nesting communities in interior British Columbia, Canada: Nest characteristics and niche partitioning. Condor 106:5-19. Martin, K., A.R. Norris, and M. Drever. 2006. Effects of bark beetle outbreaks on avian biodiversity in the British Columbia interior: Implications for critical habitat management. British Columbia Journal of Ecosystem Management 7:10-24. McCallum, D. A., R. Grundel, and D. L. Dahlsten. 1999. Mountain chickadee (Poecile gambeli). In A. Poole, and F. Gill, editors. The Birds of North America, no. 453. Academy of Natural Sciences, Philadelphia, and American Ornithologists' Union, Washington, D.C. Minock, M. E. 1972. Interspecific aggression between Black-capped and Mountain Chickadees at winter feeding stations. Condor 74:454-461. Overpeck, J. T., D. Rind, and R. Goldberg. 1990. Climate-induced changes in forest disturbance and vegetation. Nature 343:51-53. Robinson, P. A., A. R. Norris and K. Martin. 2005. Interspecific nest sharing by Red-breasted Nuthatch and Mountain Chickadee. Wilson Bulletin 117:400-402. Rodenhouse, N. L., T. S. Sillett, P. J. Doran, and R. T. Holmes. 2003. Multiple density-dependence mechanisms regulate a migratory bird population during the breeding season. Proceedings of the Royal Society of London. B: 270:2105-2110. Templeton, C. N. and E. Greene. 2007. Nuthatches eavesdrop on variations in heterospecific chickadee mobbing alarm calls. Proceedings of the National Academy qf Sciences 104:5479-5482. Wiebe, K. L., W. D. Koenig, and K. Martin. 2006. Evolution of clutch size in cavity-excavating birds: The nest site limitation hypothesis revisited. The American Naturalist 167:343-353. 103 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0074949/manifest

Comment

Related Items