THE INFLUENCE OF CACHE SITE AND RODENT PILFERAGE ON WHITEBARK PINE SEED GERMINATION IN THE NORTHERN AND CENTRAL ROCKY MOUNTAINS by ELIZABETH R. PANSING B.A., University of Colorado Boulder, 2010 A thesis submitted to the Faculty of the Graduate Sc hool of the University of Colorado in partial fulfillment of the requirements for the degree of Master of Science Biology 2014
ii This thesis for the Master of Science degree by Elizabeth R. Pansing has been approved for the Department of Integrative Biology by Diana F. Tomback Chair Michael B. Wunder Cheri A. Jones November 20, 2014
iii Pansing, Elizabeth R. (M.S., Biology) The Influence of Cache Site and Rodent Pilferage on Whitebark Pine Germination in the Northern and Central Rocky Mountains Thesis d irected by Professor Diana F. Tomback ABSTRACT Plant regeneration is a multistage process that includes: 1) production of viable seeds, 2) dispersal to microsite types suitable for germination, 3) persistence of seeds, 4) germination, 5) seedling survival, and 6) plant establishment. Any of these stages may act as a bottleneck to mature tree recruitment. Cache site selection by seed dispersers influences regeneration by determining both germination potential and risk of seed removal by rodent see d predators ( i.e., pilferage ) Whitebark pine ( Pinus albicaulis Engelm. ) a subalpine and treeline conifer distributed throughout the w estern United States and Canada depends on the Clark's nutcracker ( Nucifraga columbiana Wilson ) for seed dispersal Nutcrackers store seeds in many caches comprised of 1 15 seeds, which are buried 2 3 cm under substrate. In two study areas, I determined 1) germination rates of whitebark pine, 2) cache pilferage rates, 3) how germination and pilferage rates vary with elevation, microsite type, and cache size, 4 ) which rodent species are present and might be responsible for pilfering nutcracker caches, 5) whether a relationship exists between granivorous rodent density and pilferage rate In 2012, I created 735 simulated caches in six mic ro site types reported as commonly used by nutcrackers on White Calf Mountain, Glacier National Park, Montana, and Tibbs Butte, Shoshone National Forest, Wyoming. Cache sizes were randomly selected from a Poisson distribution that ranged in size from 1 to 7 seeds created with
iv cache size data from the literature In 2013, I checked each cache for seed ge rmination and pilferage. I det ermined whether the odds of germination and pilferage differ ed based on cache site characteristics by determining whether the 9 5% confidence intervals of the odds ratios included 1.0 O dds of germination in the subalpine zone of White Calf Mountain were lower than at treeline (OR = 0 .585, 95% CI = 0.360 0.952 ); differences are likely caused by the late seral successional status of the forest, which is not conducive to whitebark pine regeneration. On Tibbs Butte, odds of germination were higher near rocks and near no object relative to trees (OR = 2.26, 95% CI = 1.31 3.91; OR = 2.51, 95% CI = 1.46 4.33, respectively) O dds of pilferage were higher at treeline than in the subalpine (OR = 0.524, 95% CI = 0.389 0.706) These results suggest that the influences of cache site characteristics on whitebark pine germination vary greatly geographically and with community type ( both am ong and within Tibbs Butte and White Calf Mountain ) The form and content of this abstract are approved. I recommend its publication. Approved: Diana F. Tomback
v DEDICATION To my parents, Jay and Barbette Pansing, without whom none of this would be possible.
vi ACKNOWLEDGEMENTS My work would not have been possible without Dr. Diana Tomback. Her encouragement, support, and guidance have been a continual source of inspiration and determination. Thanks to Dr. Michael Wunder for helping me develop my q uantitative skills, and for his support and help with this project. I appreciate the time Dr. Cheri Jones spent providing feedback and especially for taking the time to visit the Denver Museum of Nature and Science to teach me mammal identification. I am p rofoundly grateful to Glacier National Park and the National Park Service Rocky Mountain Cooperative Ecosystem Services Unit for seeing promise in this study and awarding me a Jerry O'Neal fellowship for field work in the summer of 2013. The study set up wo uld not have been complete without nutcracker cache size data from Harry Hutchins I am appreciative of his willingness to contribute to this research. Thanks to Tara Carolin, Glacier National Park, and Kent Houston, Shoshone National Forest for their int erest in and coordination of this research with their agencies. Thanks to Dr. Leo Bruederle for inspiring a love of plants and for his support throughout this process. This project would not have been completed without the field help of Aaron Wagner and So l ÂŽ Diaz W ithout the constant support of Mari Majack, Chelsea Beebe, Jill Pyatt, Sarah Blakeslee, James Pembrook, Mark Visel and the other students in the IB department this project would not have been completed My thanks to each of you for continually r eminding me why I love this work.
vii TABLE OF CONTENTS CHAPTER I. WHITEBARK PINE BACKGROUND.. . ...1 Introduction...1 Taxonomic and Other Groupings..2 Morphology... 3 Distribution...4 Successional Dynamics.5 Ecosystem Function .. 6 Role in the Alpine Treeline Ecotone 7 Population Decline 7 White Pin e Blister Rust ...... . 8 M ountain Pine Beetle ... ... .10 Fire Suppression. ... ....11 Climate Change .. 11 Regeneration .... 12 Pilferag e... .13 Secondary Seed Dispersal..13 Germination... 15
viii Restoration .. 15 Figures and Tabl es18 II. INTRODUCTI ON. 21 III. METH ODS .. 26 Study Areas 26 Field Metho ds.27 Cache Simulation..... .27 Cache Assessment .......29 Rodent Trap ping. .30 Analysis ...31 Germination and Pi lferage.... 32 Rodent Abundance 34 Figures and Tab les..35 IV. RESULTS ... .....37 Germination 37 Study Area.. .. 37 Elevation Zone......37 Object38 Protection..38
ix Microsite Type ..38 Number of Seeds per Cache. 39 Pilferage ..39 Elevation Zone..40 Object40 Number of Seeds per Cache. 40 Trap ping St udy40 White Calf Mo untain40 Tibbs Butte..42 Figures and Tab les..44 V. DISCUSSION AND CONCLUSIONS .. ..59 Germination ...59 Study Area and Elevation Zone57 Object 60 Microsite...61 Number of Seeds per Cache .62 Pilferage and Seed Survival Limitations .62 Seed Dispersal by Clark's N utcrackers...64 Trapping .. 65
x Management and Res toration.66 Conclusion s.67 REFERENCE S... ... 69
xi LIST OF TABLES Table IV.1: Mean germination rates and 95% confidence intervals (CI) on White Calf Mountain and Tibbs Butte and the subalpine and treeline zones of each study area. ........44 IV.2: Mean germination rate a nd 95% confidence interval (CI) of protected (leeward of rocks, leeward of trees and among vegetation combined) and unprotected (windward of rocks, windward of trees and open combined ) cache site ..44 IV.3: Mean germination rate and correspondi ng 95% confidence interval (CI) by object in the subalpine zone of White Calf Mountain, treeline zone of White Calf Mountain, and on Tibbs Butte (elevation zones combined). ......45 IV.4: Mean germination rates and corresponding 95% confide nce intervals (CIs) by microsite type in the subalpine on White Calf Mountain, treeline on White Calf Mountain, and on Tibbs B utte (elevation zones combined)46 IV.5: Odds ratio estimate and 95% confidence interval comparing the odds of ger mination on White Calf Mountain to the odds of germination on Tibbs Butte 48 IV.6: Odds ratio estimate and 95% confidence interval comparing the odds of germination in the subalpine zone to the odds of germination in the treeline zone on both White Calf Mountain and Tibbs Butte ...48 IV.7: Odds ratio estimates and 95% confidence intervals of pairwise comparisons of the odds of germination by nurse object ..48 IV.8: Odds ratio estimate and 95% confidence interval com paring the odds of germination in unprotected sites relative to protected sites in the subalpine zone on White Calf Mountain, treeline zone on White Calf Mountain, and on Tibbs Butte .48 IV.9: Odds ratio estimates and 95% confidence intervals of pairw ise comparisons of the odds of germination by microsite type ...49 IV.10: Odds ratio estimates and 95% confidence intervals comparing odds of germination by number of seeds per cache 50 IV.11: Proportion of caches pilfered and 95% confidence intervals by study area, elevation zone, and object ..51 IV.12: Odds ratio estimate and 95% confidence interval comparing the odds of pilferage on White Calf Mountain to the odds of germination on Tibbs B utte52
xii Table IV.13: Odds ratio estimate and 95% confidence interval comparing the odds of pilferage in the subalpine zone to the odds of germination in the treeline zone ....52 Table IV.14: Odds ratio estimates and 95% confidence intervals of p airwise comparisons of the odds of pilferage by nurse object 52 Table IV.15: Odds ratio estimates and 95% confidence intervals comparing odds of pilferage by number of seeds planted per cache 53 Table IV.16: Model summari es for the subalpine trappin g web on White Calf Mountain in 2012 ...... 54 Table IV.17: Model summaries for the subalpine trapping web on White Calf Mountain in 2012 ...54 Table IV.18: Model summ aries for web 1 on White Calf Mountain in 2013 55 Table IV.19: Model summaries for web 2 on White Calf Mountain in 2013 55 Table IV.20: Model summaries for the subalpine trapping web on Tibbs Butte in 2012 ....56 Table IV.21: Model summaries for the treeline trapping web on Tibbs Butte in 2012 56 Table IV.22: Model summaries for web 1 on Tibbs Butte in 2013 ...57 Table IV.23: Model summaries for web 2 on Tibbs Butte in 2013 ...57 Table IV.24: Granivorous mammal density estimates by year and elevation zone ...58
xiii LIST OF FIGURES Figure I.1. A n old growth whitebark pine t ree.... 19 I.2. Distribution map of whitebark pine...20 I. 3. White pine bister rust on a whitebark pine....21 III.1. Study are as...35 III.2. Composition of forest trees in each elevation zone a t each study area35 III.3. Microsite types in which cach es were creat ed.36 III.4 Diagram showing triangulation method to facilitate cache re location .36 IV.1. Germination rates by study area, elevation zone, protection, object, and microsite type 47 IV.2 Proporti on of pilfered caches by cache site characteristics ...51
xiv LIST OF ABBREVIATIONS 1. WPBR White Pine Blister Rust an invasive fungal pathogen caused by Cronartium ribicola J.C. Fisher 2. USFWS United States Fish and Wildlife Service 3. OR Odd s Ratio the ratio of the odds that an event will occur relative to the odds that another event will occur 4. CI Confidence Interval 4. AIC Akaike's Information Criterion a type of model selection criterion used for model selection purposes. !"# ! ! !" ! ! where k is equal to the number of parameters, and L is the maximized likelihood.
1 CHAPTER I WHITEBARK PINE BACKGROUND Introduction Forested ecosystems cover almost one quarter of the terrestrial area of the earth and provide billions o f dollars' worth of ecosystem services each year (Costanza et al. 1997, Allen et al. 2010 ). Almost three million km 2 of the Uni ted States is forested, and these areas provide services including maintenance of biodiversity, watershed protection, carbon sequ estration and timber production (USFS 2003). Additional economic benefits are generated by direct profit, tourist activities, job creation, and labor income (Cullinane Thomas et al. 2014). Nonnative pathogens, native and nonnative pest outbreaks, changing ecological processes, and large scale wildfires, all of which have been exacerbated by climate change, have led to unprecedented tree mortality upwards of 50,000 km 2 in 2009 in the U.S. alone ( USFS 2003, Loehman and Anderson 2009, Allen et al. 2010 Man 2 010 ). Successful tree regeneration after these disturbance events is necessary for maintaining a suite of population and community dynamics, and to continued ecosystem functioning. Whereas interest in the processes affecting regeneration has increased, inf ormation about the earliest stages of tree regeneration is lacking. This problem is compounded in species with unique regeneration systems, especially those species whose seeds are dispersed by scatter hoarding animals. Whitebark pine ( Pinus albicaulis Eng elm.) is a keystone and foundation species common to upper subalpine and treeline forests of the w estern United States (Fig. I.1). Whitebark pine is currently experiencing high mortality range wide because of a n exotic
2 fungal pathogen, unprecedented outbre aks of a native pest, and long term fire exclusion practices (Tomback and Achuff 2010), and without successful regeneration range wide, whitebark pine communities are at risk. The regeneration success of whitebark pine ultimately depends on its co evolved avian mutualist, the Clark's nutcracker ( Nucifraga columbiana Wilson ; Tomback 1978, 1982, Hutchins and Lanner 1982 ). This scatter hoarding bird harvests and disperses whitebark pine seeds. The nutcracker's choice of where to store the seeds determines the environmental conditions the seeds will experience, influences whether the seeds will survive to germination, and eventually determines the spatial dynamics, population genetic structure, spatial distribution, and fitness of whitebark pine (Tomback and Lin hart, 1990 Tomback 2005 Bruederle et al., 2001). Taxonomic and Other Groupings Whitebark pine is discussed in the context of three artificial taxonomic groups: 1) the Strobus pines, a taxonomic designation based on both morphological and genomic data, 2) the stone pines ", a global grouping based on cone morphology, seed morphology and a shared seed dispersal mechanism, and 3) the "high elevation five needle white pines", a designation for high elevation white pines distributed throughout the Western Unit ed States. MatK and rbcL chloroplast DNA sequences support whitebark pine's placement in subgenus Strobus section Quinquefoliae subsection Strobus ( Liston et al. 1999, Gernandt et al. 2005 Syring et al. 2007). However, subgeneric groupings remain paraph yletic, likely because of incomplete lineage sorting (Syring et al. 2007). Whitebark pine was recommended for classification in subsection Strobus in 2005, when subsection Cembrae where whitebark pine was traditionally placed, was subsumed into subsection Strobus ( Price et al. 1998, Gernandt et al. 2005, Syring et al. 2007).
3 Subsection Strobus is linked by morphological characters including haploxylon y, five needles per fascicle, deciduous fascicle sheaths, generally thin cone scales, terminal umbros that lack prickles, and reduced or absent seed wings ( Gernandt et al. 2005 Syring et al. 2007). However, this designation is built upon a paraphyletic section grouping that may be unstable. The stone pines include whitebark pine, Korean ( P. koriaensis Siebold & Zucc.), Japanese stone ( P. pumila (Pall.) Regel), Siberian stone ( P. sibirica Du Tour), and Swiss stone ( P. cembra L.) pines. Taxonomically, these pines are placed in subsection Strobus, but share additional morphological traits including indehiscent co nes, seeds that lack wings, and seed dispersal by nutcrackers (genus Nucifraga ; Weaver 2001 Gernandt et al. 2005). Whitebark pine is the only new world representative of the stone pines. The high elevation five needle white pines, informally referred to a s the "high five pines are a group of species that grow at high elevations in the mountains of the w estern United States and Canada The group includes whitebark, limber ( P. flexilis E. James), Rocky Mountain bristlecone ( P. aristata Engelm.), Great Basi n bristlecone ( P. longaeva D.K. Bailey), foxtail ( P. balfouriana Balf.), and southwestern white ( P. strobiformis Engelm.) pines (Tomback and Achuff 2010, Tomback et al. 2011). Taxonomically, this group includes species in both s ubsections Strobus and Balfo urianae (Gernandt et al. 2005). Whitebark pine is the only "high five" pine, indeed the only pine in North America that exhibits both indehiscent cones and fully wingless seeds. Morphology Whitebark pine grows to heights of up to 21 meters, and possesses an irregular and spreading crown. Mature trees have smooth, pale gray bark that often separates into
4 plates with age (Kral 19 93). Like other white pines in s ubsection Strobus whitebark pine is characterized by five needles per fascicle, with deciduous fas cicular sheaths. The needles are generally 3 7 cm in length and can persist for 5 8 years. The pollen cones are carmine in color, with persistent seed cones that range from deep purple black to a chocolate brown. Seed cones are arranged in whorls horizonta l to the branch axis, making access to the cones easier for the nutcrackers that disperse the seeds. The cones have thin scales, with thickened apophyses, up curved tips, and short terminal umbos. Diagnostically, their cones are indehiscent (scales remain sealed), and the ir seeds are large and wingless (Kral 1993). Distribution Whitebark pine inhabits subalpine and treeline forests of the w estern United States and Canada. Its range, which extends from 37 to 55Â¡N and 107 to 128Â¡W, is divided into two distinc t regions: a western coast al, and an interior (Fig. I.2 ; Ogilvie 1990 McCaughey and Schmidt 2001). The western distribution includes the Sierra Nevada, Klamath, Coastal, and Cascade mountains of California, Oregon, Washington, and British Columbia. The in terior distribution is comprised of the Northern Rocky Mountains of British Columbia and Alberta, continuing south throug h Idaho, Montana, and Wyoming. Disjunct populations throughout Washington and British Columbia link these portions of the range, provid ing near continuity (Little and Critchfield 1969, Arno and Hoff 1989, Ogilvie 1990). Although the geographic range of whitebark pine is large, the elevational distribution is restricted to a narrow band in the upper subalpine and treeline. Whitebark pine's elevational range spans 900 2 200 m in the coastal portions of Canada, 1 600
5 2 300 m in the Canadian Rockies, and 1 100 3 660 m in the United States (Tomback and Achuff 2010). Successional Dynamics Whitebark pine is present in a wide variety of comm unity types throughout its range and occurs in a number of successional stages. It is an early seral, or pioneering species in some regions because it is able to rapidly colonize disturbed areas (e.g., burned and clear cut sites), and is often one of the f irst conifers to successfully regenerate (Tomback 1986, Tomback et al. 1993, 2001 a Arno 2001). This is a result of two interacting factors: nutcrackers cache in open and disturbed areas, and whitebark pine germinants and seedlings survive well in exposed and harsh conditions (Tomback 1986, 2001, Arno and Hoff 1989, Maher and Germino 2006 McKinney and Tomback 2011 ). Whitebark pine is a poor competitor, and on productive sites, it is generally outcompeted by shade tolerant conifers including subalpine fir ( Aibes lasiocarpa (Hook.) Nutt.), Engelmann spruce ( Picea engelmannii Parry ex Engelm.), or mountain hemlock ( Tsuga mertensiana (Bong.) CarriÂre ; Arno and Hoff 1989, Arno 2001). In these seral communities, whitebark pine is generally replaced over time (Ar no and Weaver 1990). In the harsher areas of the upper subalpine and alpine treeline ecotone, whitebark pine persists in self replacing climax stands. These areas are less productive than the surrounding areas, with drier and colder conditions prevailing ( Arno and Weaver 1990, Arno 2001). Whitebark pine assumes a variety of growth forms. On productive sites trees are tall and straight with spreading crowns on productive sites ; on wind swept droughty sites with poor soil, whitebark pine grows more slowly, a nd assumes more wind battered
6 shapes Within the alpine treeline ecotone, harsh conditions shape the growth of the tree, and dwarfed, or krummholz, gr owth forms are common (Tomback and Achuff 2010). Whitebark pine stems often grow in multi tree clusters (F ig. I.1) as a result of Clark's nutcrackers placing more than one seed in a cache ( Tomback 1978, Tomback and Linhart 1990). Ecosystem Function Whitebark pine assumes an important ecological role in the subalpine forests and alpine treeline ecotones it inh abits, and is considered both a foundation and keystone species (Tomback et al. 2001, Ellison et al. 2005). Foundation species define ecosystem structure by stabilizing fundamental ecosystem processes, and keystone species contribute more to biodiversity t han their relative numbers would sugg est (Paine 1969, Dayton 1972). Whitebark pine initiates tree island development within the alpine treeline ecotone, facilitates the growth of other conifers, regulates runoff and slows snow melt, fosters community devel opment after fire, and reduces soil erosion (Arno and Hammerly 1984, Farnes 1990, Tomback and Linhart 1990, Tomback et al. 1993, 2001, Callaway 1998, Resler and Tomback 2008 ). Its seeds are an important component of many species' diets, including Clark' s n utcrackers grizzly bears ( Ursus arctos horribilis ), red squirrels ( Tamiasciurus hudsonicus ) and other species in 12 avian and three mammalian families (Tomback and Kendall 2001). Of particular importance from an ecological services viewpoint is whitebark pine's substantial infl uence on hydrologic processes. Whitebark pine's role in regulating snowmelt provides steady water flow to agricultural communities and cities throug hout the summer (Farnes 1990). Additionally, whitebark pine's root systems provide
7 pr otection from erosion, retaining soil by growing in areas where other species may not be able to grow (Arno and Hammerly 1984). Role in the Alpine Treeline Ecotone In many treeline ecotone s throughout the Rocky Mountains, whitebark pine is an important tre e island initiator (Tomback and Resler 2007, Resler and Tomback 2008, Tomback et al. 2014). Whitebark pine seedlings are more successful at colonizing open treeline areas than other treeline conifers, including subalpine fir and Engelmann spruce. Seedling mortality rates of greater than 90% are not uncommon for these species (Smith et al. 2009). However, Maher and Germino (2006) found that whitebark pine establish ed more successfully than other treeline species and that more whitebark pine cotyledon seedlin gs (seedlings less than 1 year old) were present in their treeline study areas. Because whitebark pine acts as a tree is land initiator in many treeline communities, it is predicted to facilitate the upward movement of treeline in response to climate change by initiating tree growth above the current tree limit (Resler and Tomback 2008). In areas where treeline movement relies on whitebark pine's role as a tree island initiator, loss of whitebark pine may diminish the ability of treeline to ascend in elevati on (Tomback and Resler 2007). Population Decline Whitebark pine p opulations are declining nearly range wide because of four interacting factors: 1) an invasive fungal pathogen ( Cronartium ribicola J.C. Fischer) that causes white pine blister rust (WPBR), 2 ) a mountain pine beetle ( Dendroctonus ponderosae Hopkins) epidemic of unprecedented severity, 3) advancing succession in some regions due to fire suppression, and 4) climate chang e (Tomback and Achuff 2010,
8 Macf arlane et al. 2013). Because of widespread p opulation decline and loss of ecosystem function in July 2011, the U.S. Fish and Wildlife Service announced that whitebark pine warranted protection under the Endangered Species Act, but precluded its listing to address other species of higher priority (U SFWS 2011). In Canada, whitebark pine was listed it as endangered under the Species at Risk Act in 2012 (Government of Canada 2012). White Pine Blister Rust. White pine blister rust a disease caused by the invasive pathogen C. ribicola is now present th roughout the ranges of all high five pines except Great Basin bristlecone (Schwandt et al. 2010 ). WPBR has been documented in nearly all areas throughout whitebark pine's range except in the interior ranges of the Great Basin (Schwandt et al. 2010, Tomback and Achuff 2010). C. ribicola was introduced to North America from Europe in the early 1900's via shipments of infected eastern white pine seedlings ( Pinus strobus L ; Hunt 2009, Geils et al. 2010). The first reports of white pine blister rust in w estern N orth America came from Washington state and British Columbia in 1921 and were attributed to a single introduction event in British Columbia in 1910 (Detwiler 1922, G ssow 1922, Mielke 1943) However, a single introduction would likely require higher rates of spread than those observed, and higher mortality rates of imported seedlings than those reported. These lines of evidence led Hunt (2009) to hypothesize that there were likely multiple introduction events, an idea supported by Spaulding (1922) and Joy (1939). White pine blister rust is a heteroecious disease that exhibits a complex life cycle comprising five spore stages. The definitive hosts are white pines, and alternate hosts are gooseberries and currants in the genus Ribes as well as herbaceous pla nts in the genera
9 Castilleja and Pedicularis. Although a long documented history exists with Ribes the role of Castilleja and Pedicularis in the pathogen life cycle was discovered only recently (Hiratsuka and Maruyama 1976, Patton and Spear 1989, Zambino et al. 2005, McDonald et al. 2006 Zambino et al. 2007 ). The life cycle begins as aeciospores are released from aecial blisters formed on woody tissue of the pine host (Fig. I.3 ). The aeciospores are transport ed by wind to an alternate host, where uredinio spores are produced on the underside of the foliage of the alternate host These urediniospores can either infect other alternate hosts, or germinate to produce telio spores on the same plant. These teliospores subsequently germinate and produce basidiospor es, which are also wind dispersed. These spores infect white pines. Once on the pine host, the basi d i ospores germinate and mycelia en ter needles through the stomata. After two to three years of fungal growth pycniospores are produced in the margin of deve loping cankers. These spores produce sweet nectar that attracts insects to facilitate cross fertilization for spore production. These fertilized pycnia go on to produce aeciospores (Ge il s et al. 2010 and references therein ). Although the impacts of WPBR ar e severe, the time from infection to death can range from a few years to a decade or more, depending on the size of the tree ( McDonald and Hoff 2001 ) Infection begins when wind dispersed b asidiospores germinate within pine needle stomata and mycelia subs equently grow into the branches and bole. The branches at the site of infection are generally killed long before the tree, resulting in loss of reproductive function prior to tree death. Generally, mortality occurs when the infection reaches the bole and t he tree is girdled (Schwandt et al. 2010).
10 Throughout the range of whitebark pine, mean infection levels of C. ribicola vary from 20 70% and appear to be highest in the Northern Continental Divide Ecosystem of the Northern Rocky Mountains (Smith et al. 200 8, 2013, Schwandt et al. 2010). These areas of elevated WPBR incidence are attributed to moisture levels and abundance of alternate hosts ( Geils et al. 2010, Smith McKenna et al. 2013). Mountain Pine Beetle. Mountain pine beetle ( Dendroctonus ponderosae Ho pkins) is a native insect that infects susceptible pine species throughout the w estern United States and Canada Its primary hosts are lodgepole ( P. contorta Douglas) and ponderosa ( Pinus ponderosa Douglas ex C. Lawson) pines, but all pine species, includi ng whitebark pine, are susceptible to attack (Logan and Powe ll 2001). Mountain pine beetle is one of the few bark beetle species that must kill its host to successfully reproduce. The beetles use a mass attack strategy to overwhelm tree def enses, culminati ng in death when sugar transport is halted ( Amman 1985 ) In the late spring or early summer, adult beetles emerge from their host trees to disperse to new pines. As a beetle makes its first attacks on a tree, it emits an aggregate pheromone to attract othe r beetles. After the host's defenses are overwhelmed, the beetles then emit an anti aggregate pheromone signaling other beetles that the tree is "full" (Amman 1982). The beetles then bore into the phloem of the tree, where females lay their eggs. Once hatc hed, the larvae fee d on phloem and mutualistic fungi brought to the trees by the adult beetles. Once mature, the beetles exit the tree to begin the cycle anew (Safranyik and Carroll 2006) Mountain pine beetle life cycles are connected to temperature. In t he elevation band in which the primary hosts grow, the life cycle occurs over the course of one year. At higher elevations with colder temperatures, life cycles have generally extended to two
11 years, and epidemic outbreaks were uncommon (Amman 1973, Safra ny ik 1978 Perkins and Swetnam n 1996 ) This was historically the case for whitebark pine, where low temperatures suppressed attack (Logan and Powell 2001, Powell and Logan 2005). But because of increasing minimum temperatures attributed to climate change, th e beetles have been able to successfully reproduce at higher rates, which have sustained large scale outbreak levels ( Logan et al. 2010 ). As of 2008, the current beetle infestation had killed over 450,000 acres of whitebark pine ( Gibson et al. 2008 ). Fire Suppression. Prior to widespread fire suppression, whitebark pine forests were characterized by mixed severity fire regimes, with stand replacing and low intensity fires also common (Arno and Hoff 1989). These fires impeded the growth of late successional subalpine fir that would eventually outcompete and replace whitebark pine (Arno and Hoff 1989). Fires additionally benefit whitebark pine because of its superior ability to regenerate in open, burned areas on poor seedbeds (Arno and Hoff 1989, Tomback et al. 1993 2001 ). Established practices of f ire exclusion in the early 20 th century contribute d to whitebark pine decline by altering the successional dynamics of those forest types in the Rocky Mountains where w hitebark pine acts largely as an early seral species (Arno 2001, Tomback et al. 2001 ). Without frequent disturbances, which historically came in the form of fire, subalpine fir and Engelmann spruce have replaced whitebark pine in many areas (Keane 2001). Climate Change. Surface temperature increases observed in the last century are projected to continue throughout the next century (IPCC 2013). The International Panel on Climate Change (IPCC, 2013) predicts that temperatures will likely increase by more
12 than 2Â¡C from 19th century averages by the end of the 21st century. These increases will also alter precipitation regimes Therefore, a species' bioclimatic niche may migrate as a result of climate change. Species may respond in three ways: 1) migration to track the ecological niche, 2) adaptation to new conditions, or 3) local exti rpation (Aitken et al. 2008). W hitebark pine 's range is projected to decrease by up to 98% by 2085 (Hamann and Wang 2006) W armer temperatures in many areas are predicted to exceed the thermal tolerance of whitebark pine, and w ill additionally favor species such as subalpine fir that are better suited for these environments Climatically suitable areas are expected north of the current range, but natural dispersal to these areas is unlikely due to slow tree growth rate, reduced seed production because of tree dama ge and mortality, leading to decreased regeneration and low rates of resistance to blister rust (Ha ma n n and Wang 2006, Schrag et al. 2007, Aitken et al. 2008). The alpine treeline ecotone inhabited by whitebark pine wil l likely change as the climate warms. Ecotones are transition areas between ecosystems, where community composition is largely determined by climatic tolerance (Holtm e ier 2009) Globally average growing season minimum temperatures and shortened growing se ason length define treeline elevations (KÂšrner 2012). Therefore, increasing temperatures are predicted to result in increasing treeline elevations (Lenoir et al. 2008). Regeneration Limited regeneration could result in a substantial overall decrease in for est area and density. In areas of high mortality where regeneration is limited by decreased seed production and dispersal, persistence of whitebark pine on the landscape is dependent upon both natur al and artificial regeneration. However, little is known about the e arly life
13 stages which have only recently become a research focus. Historically, ecologists have used mature forest structure to assess forest demography and population dynamics. This focus makes untested assumptions about early growth stages t hat might restrict regeneration, and limits our understanding of forest population dynamics. Regeneration can conceptually be divided into six stages, any of which can limit success: 1) production of viable seeds, 2) dispersal via either primary (e.g. by nutcrackers) or primary and secondary mechanisms to microsite types suitable for germination, 3) persistence of seeds on the substrate surface or within a seed bank, 4) germination, 5) survival, and 6) establishment, when mortality drops to a lower and mor e consistent rate (Grubb 1977, Germino et al. 2002, Malanson et al. 2007). I focus here on how the stages occurring after nutcracker dispersal affect germination. Pilferage. Scatter hoarding species those that create many caches each with a few seeds like ly experience cache pilferage by conspecifics and/or heterospecifics. Indeed, caching behavior itself is hypothesized to have evolved to decrease the risks of pilferage (Anderson and Krebs 1978, Smith and Reichman 1984, Vander Wall 2010). In response to pi lferage, preventio n strategies emerged, including avoidance of sites where risk of theft is high, varying cache types to include both scatter and larder hoarding, and caching in locations that thieves cannot access (Hampton and Sherry 1994, Emery et al. 2 004, Preston and Jacobs 2005, Leaver et al. 2007 MuÂ–oz and Bonal 2011). B ecause Clark's nutcrackers are sympatric with populations of granivorous rodents, some caches may be made in areas where pilferage may occur. Granivorous rodents pilfer simulated nu t cracker caches (Hutchins and Lanner 1982, McCaughey 1990 ), yet the ultimate fate of these seeds is unknown (Lorenz et al. 2008) Because many species that pilfer nutcracker
14 caches are scatter hoarders themselves (i.e. create their own caches ; Hutchins and Lanner 1982, Vander Wall et al. 2005), cache pilferage could lead to one or both of two outcomes: secondary dispersal, and/or seed loss because of predation. Secondary Seed Dispersal. Se ed pilferage does not always equate to seed predation; many plant spe cies regenerate as the result of secondary dispersal ( Vander Wall 2003 Tomback et al. 2005 Vander Wall et al. 2005 ). Secondary or Phase II dispersal is the movement of previously dispersed seeds to new locations where germination may be possible ( Vander Wall et al. 2005 ). Two types of secondary dispersal are identified The first occurs when the primary and secondary dispersal mechanisms are similar. The second, te rmed diplochory, occurs when the secondary dispersal mechanism differs from the primary disp ersal mechanism (Vander Wall et al. 2005 a ). Vander Wall and Longland (2005) consider diplochory to be advantageous because it, unlike other dispersal mechanisms, maximizes the distance from parent trees. This decreases impacts from density depe ndent seed p redators, increases th e colonization of new areas, and often leads to d irected dispersal to sites favorable for germination (Connell 1971, Janzen 1971, Howe and Smallwood 1982, Clark et al. 2007 Cain et al. 2000, Wenny 2001, Nathan et al. 2002 ). Because o f the implications of secondary dispersal, which can impact both plant spatial distribution and fitness within ecosystems, Vander Wall et al. (2005) suggests monitoring seeds to provide a more thorough understanding of seed fate and dispersal. Some researc hers have suggested that whitebark pine seed dispersal may be diplochorous, a result of small, granivorous rodents scatter hoarding seeds pilfered from nutcracker caches (Lorenz et al. 2008) However, direct evidence is lacking.
15 Germination. Seeds cached b y nutcrackers that also escape predation from other granivores may remain in the soil for one or more years before germinating, which generally occurs after two winters (Tomback et al. 2001). This soil seed bank is thought to increase reg eneration success. Because nutcrackers begin to harvest seeds in late August or early September, seeds are generally immature when dispersed. The soil seed bank is thought to allow seeds time to mature (Tillman Sutela et al. 2007). Information on whitebark pine seed germina tion in the field is limited. McLane and Aitken (2012) found that cumulative germination rates of seeds planted in an open air common garden in British Columbia were 0.7% after the first winter, 9.6% after the second, and 94% after the third. Restoration The long term survival of whitebark pine throughout much of its range depends upon both conservation and restoration (McDonald and Hoff 2001, Keane and Schoettle 2011). In areas of extreme mortality from blister rust and/or mountain pine beetles chances o f natural regeneration are slim D eclining numbers of cone producing trees and the loss of cone bearing branches lead to decreased seed cone production. Regeneration is further hampered as the probability of seed dispersal is reduced when cone production i s low (McKinney and Tomback 2007, McKinney et al. 2009, Barringer et al 2012). Therefore, in the absence of a nearby healthy seed source, planting efforts are necessary to maintain viable populations. These efforts although on small scales, are ongoing i n many locations throughout whitebark pine's range (Schwandt et al. 2010). However, because blister rust can infect young white pine seedlings (<3 years of age) which are generally more susceptible than older seedlings and mature trees rust resistant seedl ings
16 are planted to increase the likelihood of tree survival and the spread of resistance genes ( Hunt 1991, Schoettle and Sniezko 2007). In 1946, the U S Forest Service began white pine blister rust resistance screening in western white pine ( P. monticol a Douglas ex D. Don) followed by sugar pine ( P. lambertiana Douglas) both species of considerable economic value Screening has since expanded to include whitebark pine as well as limber, southwestern white, foxtail, Great Basin, and Rocky Mountain bris tlecon e pine but on more limited scales ( Schwandt et al. 2011 ). Th e screening process identifies potentially rust resistant seed source trees Cones ideally are collected from apparently healthy trees i n heavily rust infected areas but some collecting is in regions of low rust infection Seeds from the cones are used to grow s eedlings in the nursery the seedlings are exposed to high spore loads of basidiospores from Ribes leaves, and then seedlings are monitored for a suite of resistant characte ristics ( Sniezko et al. 2010). Those seedlings that show resistance are trac ked back to specific parent tree s The resistant parents are referred to as plus trees Over 650 whitebark pine plus trees have been identified, and their seeds are used to grow seedlings for planting in restoration efforts. These seedlings grow for two years in a nursery before they are out planted in restoration projects (Burr et al. 2001, Maholovich and Dickerson 2004, Burns et al. 2008 ). Planting seedlings is time consuming, labor inte nsive and incredibly costly. The time required to grow seedlings, screen for resistance, and plant the seedlings is about three years. The cost of this restoration technique ranges from $ 1980 to $2405 per hectare (Schwandt et al. 2011, Tomback et al. 2011 ) Because 84% of whitebark pine's range exists on federal land, these costs fall largely on the federal government (Keane 2000).
17 Unfortunately, the current budgets of most federal agencies render large scale restoration efforts unlikely. Additional logisti c and legal difficulties arise when Wilderness Areas are considered. In these areas, mechanized transport is prohibited, making any restoration efforts complicated, if not impossible. Interpretation of the Wilderness Act of 1964 presents additional challen ges. Wilderness areas strive to protect the "untrammeled" and "natural" conditions of the areas ( Section 2 (c) P.L. 88 577 ). Introduction of white pine blister rust and the stressors induced by anthropogenic climate change threaten the "naturalness" of the se Wilderness Areas, suggesting that restoration activities may be appropriate. However, restoration activities themselves necessitate "trammeling" these areas. Because more than 48% of whitebark pine s distribution exists on designated or proposed Wildern ess (Keane 2000), development of restoration methods consistent with Wilderness values will determine the overall success of these activities. Direct seeding, an emerging restoration method may be one solution. Direct seeding is a relatively untested resto ration technique that involves planting seeds in lieu of seedlings, eliminating the need to grow seedlings in the nursery for two years prior to planting. Direct seeding has the potential to increase the area in which restoration activities could occur by making it possible to plant in Wilderness Areas and other remote locations. It would also drastically decrease the cost, time, and labor required for restoration, reducing the time to planting after cone collection (Schwandt et al. 2011, Tomback et al. 201 1 ). Obstacles to the application of large scale direct seeding exist, yet preliminary trials have been successful (Schwandt et al. 2007, DeMastus 2013). Before implementing
18 large scale direct seeding efforts, we need a better understanding of the ecologica l processes that influence seed survival, germination, seedling survival, and establishment. Important among these include understanding the impact of cache pilferage by small rodents, and identifying suitable microsite types and optimum cache sizes for ge rmination and seedling survival (DeMastus 2013).
19 Figures and Tables Fig. I.1: A n old growth whitebark pine tree This individual exhibits a common clustered growth form that results from multi seed caches. Image: D.F. Tomback.
20 Fig. I.2: Distribution map of whitebark pine. The range of whitebark pine is shown in red ( WPEF 2014)
21 Fig. I.3: White pine blister rust on a whitebark pine. Blisters containing aec iospores on the bole of a whitebark pine These spores will go on to infect a susceptible alternate host. Image: D.F. Tomback
22 CHAPTER II INTRODUCTION P lant regeneration is integral to maintaining stable community structure and composition (Leak and Grabe r 1976). Successful plant regeneration is a multistage process that includes: 1) production of viable seeds, 2) dispersal of seeds via either primary or primary and secondary mechanisms to microsite types suitable for germination, 3) persistence of seeds o n the substrate surface or within a seed bank, 4) germination, 5) seedling survival, and 6) establishment (Grubb 1977, Germino et al. 2002, Malanson et al. 2007). Any of these stages either independently or in combination may act as a bottleneck to tree re cruitment (e.g., Andersen 1989, Crawley 1989, Eriksson and Ehrl ÂŽ n 1992, Ribbens et al. 1994) In fact, regeneration is usually discussed solely in the context of its limitations, which are grouped in two general categories: 1) seed limitation, a result of poor seed production and/or limited seed dispersal; and 2) establishment limitation, caused by seed mortality (e.g., predation, solar radiation, pests, etc.), poor seedbeds, or the scarcity of microsites favorable for germination and establishment (Eriksso n and EhrlÂŽn 1992 Ribbens et al. 1994, Clark et al. 1999 ). For species with declining populations, rates of natural regeneration influence future population demographics, recovery from disturbance, and management strategies (Guariguata and Pinard 1998, Schoettle and Sniezko 2007, Kollmann et al. 2008). For declining keystone and foundation species that stabilize ecosystem function and support biodiversity, altered recruitment patterns may impact entire ecosystems, influencing community structure and spec ies richness (Paine 1969, Dayton 1972, Ellison et al. 2005).
23 Whitebark pine ( Pinus albicaulis E ngelm.) is both a keystone and foundation species in upper subalpine and treeline forests of the Western United States and Canada (Tomback et al. 2001, Ellison et al. 2005). It is declining nearly range wide because of white pine blister rust (WPBR) caused by the exotic fungal pathogen Cronartium ribicola J.C. Fischer, unprecedented outbreaks of the native mountain pine beetle ( Dendroctonus ponderosae Hopkins), a nd successional replacement resulting from fire suppression practices (Tomback and Acuff 2010). WPBR presents a novel challenge to whitebark pine ecosystems. Since its introduction in the early 1900's, it has caused population declines of up to 90% in some regions (Smith et al. 2008). Unlike mountain pine beetle, WPBR infects all life history stages and influences recruitment by reducing seed production, decreasing the probability of seed dispersal, and killing seedlings, saplings and mature trees (Hoff and McDonald 1980, McKinney and Tomback 2007, Schoettle and Sniezko 2007, McKinney et al. 2009). Whitebark pine is a slow growing tree with cone production commencing around 20 to 30 years of age (Krugman and Jenkins on 1974). As a masting species characteri zed by synchronou s and episodic seed production, whitebark pine's reproductive output varies from year to year (Janzen 1974, Tomback 1982, McCaughey and Tomback 2001, Crone et al. 20 11 ). In mast years, increased seed supply may lead to decreased seed preda tion, but limitations to the regeneration process may result because of limitations to seed dispersal (e.g., more seeds are produced than can be dispersed), microsite availability, and germination and seedling survival (Tomback 1982, Crawley and Long 1995) In non masting years seed availability and seed predation are limiting but seed dispersal is likely not limiting.
24 Seed dispersal in whitebark depends on a co evolved mutualism with a scatter hoarding bird, the Clark's nutcracker ( Nucifraga columbiana Wilson ; Tomback 1978, 1982, Lanner 1982, Tomback and Linhart 1990). Nutcrackers harvest seeds from cones in the late summer and early fall and subsequently create many caches of 1 to 15 seeds each, burying seeds 2.0 3 cm under the substrate. The seeds ar e use d as food for themselves and their offspring during the winter and spring (Tomback 1978, 1982, Hutchins and Lanner 1982, Lorenz et al. 2011). The energy contained in the seeds of a single nutcracker's collection of caches often exceeds the amount of e nergy required to sustain the bird throughout the winter and spring by as much as 45%; these seeds may be used to feed young and compensate for pilferage of caches, but some may not be recovered (Vander Wall and Balda 1977, Tomback 1982). Seeds in unutiliz ed caches may germinate, if the seeds survive and the cache site supports germination (Tomback 1978, Mellmann Brown 2005). The cache sites selected by nutcrackers ultimately influence the microclimatic conditions experienced by the seeds and the probabilit y of seed persistence (Hulme 1996, Tomback et al. 1993, Tomback 2005). Nutcrackers cache across a large elevational range, including at the lower limits of coniferous forest as well as in the alpine treeline ecotone and alpine tundra; they cache seeds in a variety of microsite types, slope aspects, and forest communities (Vander Wall and Balda 1977, Tomback 1978, 1982, 1986, Hutchins and Lanner 1982, Baud 1993, Tomback et al. 2001). Lorenz et al. (2011) compared the microsite types of randomly selected loca tions to nutcracker cache site locations in the Cascade Mountains. They found that caches were closer to large trees (>15cm DBH and >3m in height) and had higher percent overstory and understory cover compared to
25 random locations, suggesting that nutcracke rs cache in areas closer to existing tree growth. In the Sierra Nevada and Rocky Mountains, respectively, Tomback (1978) and Hutchins and Lanner (1982) frequently observed nutcrackers caching near the base of trees, the base of rocks, in open pumice, in pi ne needle litter, near fallen logs or branches, in trees, and among plants. To date, Tomback (1978) is the only study to report the relative frequency of microsite types selected by nutcrackers, in which the three most commonly utilized cache sites were ne ar the base of trees, in open pumice, and near the base of rocks. Other less frequently observed microsite types (needle litter, near or among logs/branches/roots) are associated with microsite types near the base of trees. Once cached in the soil white bark pine seeds become part of a persistent soil seed bank, and germination may be delayed two or more years (Tomback et al. 2001, Tillman Sutela et al. 2008). Whitebark pine seeds exhibit seed characteristics and life history traits suggesting that becaus e nutcrackers may cache some seeds prior to full maturity seeds cached in soil complete maturation over one or more years (Tomback et al. 2001, Tillman Sutela et al. 2008). This delayed germination may increase risks of pilferage, predation, and disea se. S eed predation rates of 10 90% have been documented for experimenter created whitebark pine caches ( Tomback 1980, Hutchins and Lanner 1982, McCaughey 1990). Predation was attributed to chipmunks ( Tamias spp. ), deer mice ( Peromyscus maniculatus ) and other rodents. However, these studies cannot equate pilferage with predation, because seed fate remains unknown. R odents have emerged as importan t secondary dispersal agents vectors that move previously dispersed seeds to new locations where germination may be p ossible ( Vander Wall 2003 Tomback et al. 2005 Vander Wall et al. 2005 ). Primary and secondary dispersal by
26 rodents has been documented in a variety of tree species (Vander Wall 2002, Vander Wall et al. 2005 ), but the role of rodents in whitebark pine reg eneration has not been studied Additionally, pilferage rates are likely influenced by predator densities, with increased densities leading to higher rates of pilferage and predation (McCaughey and Tomback 2001). Additional approaches are needed to assess the cache site characteristics in which whitebark pine regeneration occurs, lending insight into whether nutcrackers effectively disperse whitebark pine seeds. Previously conducted regeneration assessments have used existing regeneration to reconstruct the microsite types that support regeneration (Tomback et al. 1993, 2001, Maher and Germino 2006). However, few studies document germination rates under natural conditions, leaving the question of what microsite types selected by nutcrackers support germinati on and seedling establishment. Additionally, the suitability of commonly used nutcracker cache site types for germination success across the elevational gradient of whitebark pine remains poorly investigated. Lastly, the effect of pilferage and predation o n whitebark pine germination is currently unknown, as is the impact of variation in rodent density on pilferage. To address the gaps in knowledge presented above, my objectives were to determine 1) germination rates of whitebark pine, 2) cache pilferage r ates, 3) how germination and pilferage rates vary with elevation, microsite type, and cache size, 4 ) which rodent species are present and might be responsible for pilfering nutcracker caches 5) whether a relationship exists between granivorous rodent dens ity and pilferage rate
27 CHAPTER III METHODS Study Areas My study was conducted in both the subalpine forest and alpine treeline ecotone (hereafter, "treeline") of two study areas on the Eastern Front of the Northern and Central Rocky Mountains. The norther n study area was located on White Calf Mountain, Glacier National Park, Montana (48Â¡ 38' 20. 9 5" N, 113Â¡ 24' 08. 72" W; 1 920 2,272 m elevation) and encompassed an area of 0.51 km 2 The southern study area was on Tibbs Butte, Shoshone National Forest, Wy oming (44Â¡ 56' 28.33" N, 109Â¡ 26' 39.69" W; 2,983 3238 m elevation) and encompassed 0.74 km 2 (Fig. III.1). The study areas were separated by approximately 510 km and 3Â¡ of latitude. The subalpine zone of White Calf Mountain is comprised of dense, closed canopy, late seral forest, dominated by subalpine fir ( Abies lasiocarpa (Hook.) Nutt.). Whitebark pine, Engelmann spruce ( Picea engelmannii Parry ex Engelm.), aspen ( Populus tremuloides Michx.), and lodgepole pine ( Pinus contorta Douglas ex Loudon) are min or components. Predominant understory vegetation includes Arnica cordifolia Hook., Arctostaphylos uva ursi (L.) Spreng., Pedicularis spp., and Thalictrum occidentale A. Gray. Treeline is the zone of transition from upright, closed canopy forest to treeles s alpine tundra. In this zone, trees assume a dwarfed or krummholz form and occur as solitary trees or multi tree groups called tree islands, which are surrounded by a matrix of herbaceous and/or shrubby vegetation (Marr 1977, Habeck 1969, Holtmeier and Br oll 1992). At treeline on White Calf Mountain, tree islands comprised of whitebark pine and
28 subalpine fir dominate the steep terrain. A. uva ursi, Achillea millefolium L. Hedysarum sulphurescens Rydb. Potentilla diversifolia Lehm. and grasses are the mo st common species within the understory matrix. The subalpine forest of Tibbs Butte comprises late seral whitebark pine forest on south facing slopes and a late seral mixed forest dominated by whitebark pine and Engelmann spruce on all other aspects. Anten naria spp., Carex spp., Vaccinium scoparium W.W. Sm., and grasses comprise the understory vegetation. The treeline community includes mostly solitary krummholz trees with whitebark pine and Engelmann spruce co dominating. The understory matrix consists pri marily of Geum rossii (R.Br.) Ser., Potentilla diversifolia Lehm., and Saxifrage spp. Field Methods Cache Simulation. In September 2011, I collected cones from Divide Mountain, Blackfeet Nation, Montana, and the Line Creek Research Natural Area, Custer Nat ional Forest, Montana as seed sources for our experiments on White Calf Mountain and Tibbs Butte, respectively. Both collection locations are within the seed transfer zone of the corresponding study area (Mahalovich and Hipkins 2011). Cones were transporte d to the University of Colorado Denver for cold storage. In October 2011, I removed seeds from cones and discarded unfilled moldy, and/or pest infested seeds from the seed pool, simulating nutcracker seed selection behavior (Vander Wall and Balda 1977, T omback 1978, 1982). See ds weighing more than 0.128 g the average weight of viable seeds repo rted in Tomback et al. (2011) were considered full. I then stored the seeds selected until caching at ~ 1.5Â¡ C from November through June. Resistance status of the parent trees to WPBR was unknown.
29 Prior to entering the field in 2012, I used ArcGIS (version 10.1) to generate random points representing caching sites. I first delineated the study areas and stratified each area into a subalpine and treeline "zone". Z one s were separated by visually evaluating an ESRI World Imagery satellite map to locate the end of contiguous closed canopy forest. Caching sites were randomly generated from within each zone at each study area. The number of seeds planted per cache was dete rmined by randomly generating numbers from a Poisson d istribution (" = 3, range = 1 15), created using cache size data from Tomback (1978) and Hutchins and Lanner (1982). The microsite types used as caching locations were based on the three most frequent ly observed cache locations at the base of trees, at the base of rocks, and in open areas (Tomback 1978, Hutchins and Lanner 1982, Balda and Kamil 1990, Lorenz et al. 2011). Whereas these locations in part dictate the environmental conditions to which a seed is exposed, the orientation of the cache with respect to the object may also impact regeneration. This is especially true at treeline, where spatial relationships shape community structure through facilitation and other biotic/abiotic interactions. Th ese interactions are most pronounced on the leeward/down wind side of objects, where protection from harsh winds is greatest (Callaway 1998, B r ooker 2008, Blakeslee 2012, Pyatt 2013). Thus, for cache locations near trees and rocks, microsite types were des ignated as being windward or leeward of the object, and vegetated microsites were considered protected versions of open microsites. Microsite types were therefore defined by two characteristics: 1) "object" with the levels "rock", "tree", or "none", and 2) "protection" with levels "protected" and "unprotected". These two characteristics combined to for m 6 microsite types: protected l eeward of a rock, leeward of a tree,
30 among herb aceous vegetation; unprotected windward of a rock windward of a tree open (Fi gure III.3 ). Each caching location was systematically assigned to one of these six microsite types. At treeline and in the subalpine, I determined the leeward and windward side of objects at treeline using wind flagging patterns of nearby conifers. In the subalpine, I used the prevailing wind directions determined from treeline flagging to assign leeward vs. windward designations. In July 2012, I created 735 caches as follows: 362 caches (1,273 seeds) on White Calf Mountain and 372 caches (1,288 seeds) on T ibbs Butte. I placed one cache in the nearest assigned microsite type to each randomly selected location (point). I buried caches under ~2.5 cm of substrate the average depth of nutcracker caches (Tomback 1978, Hutchins and Lanner 1982). To ensure I could locate cache sites the following year, I triangulated cache position by placing seeds in one vertex of a Daubenmire frame, and two nail spikes in adjacent vertices (Fig. III.4 ). I then geo referenced cache locations using a sub meter accuracy GPS unit (Tr imble GeoXT GeoExplorer 2008 series). Data recorded at each cache included the number of seeds planted (range = 1 7), percent range of vegetation cover ( 0 25%, 26 50%, 51 75%, 76 100%), dominant vegetation type (grass, forb, woody, herbaceous, litter), a nd direct measures of slope steepness, slope aspect, and object height. In the case of no object, the height of vegetation within the Daubenmire frame was recorded. Cache Assessment. In July through early August 2013, I returned to both study areas to asse ss each cache. After determining the number of cotyledon seedlings (hereafter "germinants") present at each cache site I documented the number of seed coats from rodent predation found in/near the microsite. These seed coats corresponded
31 to the loss of pa rt or all the seed cache. Seed coats shed from germinants display characteristic longitudinal fissuring, and were easily distinguished from seed coats resulting from seed predation by rodents (Mirov 1967, McCaughey 1994). I only counted seed coats showing evidence of seed predation. To estimate the number of seeds pilfered, I excavated each cache without disturbing the seedlings and determined the number of seeds still present in the cache to compare with the number cached in 2012. I attributed missing seed s to pilferage by small granivorous rodents, which rely on olfactory cues to discover caches (Vander Wall 1998). I assumed no pilferage by other animals because: 1) nutcrackers rely on spatial memory to locate caches, 2) red squirrels harvest cones from tr ees and cache them in middens, and 3) bears raid squirrel middens for cones but have not been observed pilfering nutcracker caches (Balda and Kamil 1992, Mattson and Blanchard 1994, Vander Wall 1998). After excavating caches, all intact seeds were re burie d in the cache. Rodent Trapping. In July 2012 and 2013, I live trapped small mammals to estimate the abundance of granivorous rodents that might be responsible for cache pilferage and seed predation. In 2012, I created two trapping webs at each study area, one in the subalpine and one in the treeline zone. Webs were constructed following Buckland et al. (2001), with the central point of each web randomly selected using ArcGIS (version 10.1). Each web comprised a total of 55 Sherman Live Traps (22.9x 8.9 x 7 .6 cm), arranged as follows: 10 trap lines with five traps placed at five meter intervals radiated from a central point, with 5 traps arranged around the central point. I intended to use Distance Sampling to estimate density, but capture probability did n ot decline with distance from the center of the web. In stead the probability of
32 capture increased with distance suggesting that a larger web was needed (Buckland et al. 2001). For the 2013 season, I intended to create one web comprised of 190 Sherman Liv e Traps in the treeline zone, but rough, steep terrain precluded this option, and I created two webs in each study area. Each web comprised 95 traps with 9 traps placed along each trap line. Spacing between traps and the number of trap lines were consisten t across years. The center point for each web was randomly selected following 2012 methods. Because the target species are both diurnal and nocturnal (e.g., Tamias spp., Peromyscus maniculatus. ), each trapping bout occurred from early evening to early morn ing. I baited traps with black oil sunflower seeds at approximately 18:00h each evening and checked for capture at approximately 7:00h each morning. I processed newly captured individuals by 1) marking them uniquely with nail polish, 2) measuring ear notch left hind foot, and tail length, and 3) recording weight and sex. These measurements assisted in species identification. I released animals at the trapsite of capture following processing. In 2012, the trapping effort consisted of 8 nights at each study location. In 2013, I again trapped for 8 nights at White Calf Mountain, but inclement weather at Tibbs Butte precluded trapping for one night, and data are based on 7 nights. Analysis Germination and Pilferage. Germination was analyzed as a binomial varia ble, where germination refers to one or more seeds within a cache producing a germinant. Odds of g ermination were compared among cache site characteristics by calculating t he odds ratio (OR) and corresponding 95% confidence interval of all pairwise compari sons for all levels of each cache site characteristic variable (e.g., study area, elevation zone
33 object (tree, rock, none), protection (protected, unprotected), microsite type (lee rock, windward rock, lee tree, windward tree, vegetation, open) and numbe r of seeds per cache ) I concluded that the odds of germination differed between levels of cache site characteristic variables if the 95% confidence interval of the OR did not include 1.0 All analyses were conducted using R (R core team 2014) Because exp lanatory variables were nested I initially estimated the O R comparing study area, followed by elevation zone, and microsite type. If the confidence interval of the OR for any comparison included 1.0 I concluded there was not sufficient evidence that the odds of germination differed between levels and I pooled data for subsequent analyses. If the confi dence interval did not include 1.0 indicating the od ds of germination differ ed between levels data were segregated in subsequent comparisons Additionally, o dds ratio estimates and confidence intervals were used to determine whether the odds of germination differed among object s and protection to determine whether these explained any differences that might not be observed based on microsite type alone (e.g., if microsite types differed, was it due to the object, the fact that it was protected, or both). Lastly, I tested whether the odds of germination varied with the number of seeds planted per cache to determine whether there is an optimal number of seeds per cache for germination. Pilferage a binomial variable where "pilferage" refers to the loss of one or more seeds per cache was analyzed similarly to germination data. I estimated the odds ratios and 95% confidence intervals of all pairwise comparisons for all levels of a given explanatory variable In contrast to germination data, I analyzed study area, elevation zone, object, and number of seeds per cache as explanatory variables. Because I found no
34 evidence in the literature suggesting that rodent movemen t var ies based on object orientation in relation to prevailing wind protection and microsite type were not analyzed. Rodent Abundance. Rodent abundance methods were designed for use with distance sampling However, when analyzing capture data in Program D istance ( Thomas et al. 2010 ) no models including the uniform key function converge d I therefore used mark recapture methods to estimated rodent abundance using the RCapture package in R ( R core team 2014 Baillargeon and Rivest 2012). This estimation met hod assumes a closed population (i.e., the population size is not variable over the sampling effort). RCapture estimates abundance by fitting both log linear models and generalized linear models to mark recapture data (Cormack 1989). The package allows for four model variants: 1) temporal variation of capture probabilities ( Mt ), 2) variation of capture probability among individuals ( Mh ), 3) behavioral variation ( Mb ), where capture results in changes in capture probability, and 4) constant capture probabilit y ( M0 ). Darroch et al. (1993), Chao's (1987), and Poisson (Rivest and Baillargeon 2007) variants of Mh and Mth models were also fit. I selected the single best model for each web, using the model with the lowest #AIC value. Abundance was estimated by maxim izing the multinomial likelihood following Cormack (1992). I subsequently estimated density by dividing each abundance estimate by the area of the web.
35 Figures and Tables Figure III.1: Study areas : WCM: White Calf Mountain Glacier National Park, M ontana; TB: Tibbs Butte, Shoshone National Forest, Wyoming. Shaded areas represent Glacier National Park (GNP) and Yellowstone National Park (YNP). Figure III.2: Composition of forest trees in each elevation zone at each study area Res ults of an informal survey showing tree species composition and relative abundance by study area and elevation zone. tbsa = Tibbs Butte subalpine, tbtl = Tibbs Butte treeline, wcsa = White Calf Mountain subalpine, and wctl = White Calf Mountain treeline.
36 Figure III.3: Microsite types in which caches were created. Breakdown of planting site characteristics by object, whether the microsite type was protected (lee) or unprotected (windward), and the microsite type that corresponds to each object and protection type. Fig. III. 4: Diagram showing triangulation method to facilitate cache re location Seeds were placed in one vertex of a Daubenmire frame and nail spikes were placed in two adjacent vertices. Upon returning to t he cache location, the frame was placed around the nail spikes, indic ating the location of the cache Wind Wind Protected Unprotected Leeward of Tree Leeward of Rock Vegetation Windward of Tree Windward of Rock Open Object Tree Rock None
37 Chapter IV RESULTS Germination In the summer of 2013, I located 717 of the 735 caches created in 2012 351 on White Calf Mountain and 366 on Tibbs Butte o f which 0.319 ( 95% CI: 0.285 0.354) germinated. Study Area. On White Calf Mountain, 0.256 ( 95% CI: 0.211 0.302) of the caches germinat ed, and on Tibbs Butte, 0.380 ( 95% CI: 0.330 0.430) germinated (Table IV.1, Figure IV.1) To determine whether germ ination ra tes differed between study area s I calculated the odds ratio and 95% confidence interval of the odds of germination on White Calf Mountain relative to the odds of germination on Tibbs Butte. The odds of germination on White Calf Mountain were 0. 5 63 times the odds of germination on Tibbs Butte (OR = 0. 563 95% CI: 0. 409 0. 775 ; Table IV.5 ). These results suggest that the odds of germination differ by study area; therefore all future analyses were segregated by study area. Elevation Zone. On White Calf Mountain, 0.206 ( 95% CI: 0.146 0.266) and 0.307 ( 95% CI: 0.239 0.375) of the caches germinated in the subalpine and treeline zones, respectively ( Table IV. 1, Figure IV.1 ). Odds of germination in the subalpine were 0.588 times those at treeline ( OR = 0.585 95% CI: 0. 360 0. 952 ; Table IV.6) The 95% confidence interval did not include 1.0 suggesting that germination rates differed between elevation zones on White Calf Mountain. All further analyses of caches located on White Calf Mountain were s egregated by elevation zone.
38 On Tibbs Butte, 0.412 ( 95% CI: 0.340 0.485) of the caches germinated in the treeline zone, and 0.349 ( 95% CI: 0.281 0.417) germinated in the subalpine zone (Table IV.1, Figure IV.1 ). In contrast to the differences in the o dds of germination on White Calf Mountain, the odds ratio confidence interval of germination on Tibbs Butte indicate d that the odds of germination did not differ by elevation zone (OR = 1.31 95% CI: 0. 857 2. 00 ; Table IV.6 ). Therefore, all further analys es for Tibbs Butte were pooled Object To determine whether the odds of germination differed depending on the object near which caches were created, I calculated the odds ratio and 95% CI of each pairwise comparison (no object: tree, rock: tree, rock: no object ). Ninety five percent confidence intervals included 1.0 for all object comparisons i n both the subalpine and treeline zones on White Calf Mountain suggesting that the odds of germination did not differ by object (Table IV.7 ). However, on Tibbs Butt e, odds of ger mination near no object were 2.5 1 times the odds of germination near trees (OR = 2.5 1, 95% CI: 1.46 4.33 ; Table IV.7 ), and the odds of germination near rocks were 2.26 times the odds of g ermination near trees (OR = 2.26 95% CI: 1.31 3.91 ; Table IV.7 ). The odds of germination near rocks did not differ relative to the odds of germination near no object (OR = 0. 899 95% CI: 0. 543 1. 49 ; Table IV.7 ). Protection Odds of germination did not differ by p rotection type (protected vs. unprotected ) in the subalpine or treeline of White Calf Mounta in, or on Tibbs Butte (Table IV.8 ). Microsite Type. Odds of germination did not differ among microsite types in the subalpine or treeline zones on White Calf Mountain (Table IV.9) However, odds of germin ation did differ among certain microsite types on Tibbs Butte. Odds of
39 germinati on in the lee of rocks were 3.12 ( OR = 3.12, 95% CI: 1.40 6.92 ) times the odds of ger mination windward of trees, 2.32 ( OR = 2.32, 95% CI: 1. 09 4. 91 ) times higher in the ope n than leeward of trees, 2. 77 ( OR = 2.77, 95% CI: 1. 26 6.10 ) times greater in vegetated microsites than windward of trees, 2. 49 ( OR = 2.49, 95% CI: 1.13 5.53 ) times higher windward of rocks than windward of trees, and 3.46 ( OR = 3.46, 95% CI: 1.57 7 62 ) times higher in the open than windward of trees (Table IV. 9 ) All other confidence intervals for odds ratios did not include 1.0 indicating that the odds of germination did not differ among other microsite types (Table IV .9 ) Number of Seeds Per Ca che On Tibbs Butte, the odds of germination of one seed caches were 0.441 times the odds of pilferage of three seed caches (OR = 0.441, 95% CI = 0.195 0.995 ). However, the odds of germination of all other cache sizes on Tibbs Butte and in both elevation zones on White Calf Mountain did not differ (i.e., the 95% confidence interval of the odds ratio estimate included 1.0; Table IV.10) Pilferage Differences in odds of pilferage by cache site characteristics were analyzed similarly to odds of germination, with pilferage represented as a binomial variable where caches were considered pilfered if one or more seeds were missing from the cache. Of the 717 caches found in 2013, rodents pilfered 0.543 ( 95% CI: 0.492 0.593 ) Of those pilfered caches, 0. 749 lost all seeds from the cache and the remainder lost only a fraction of the seeds contained within the cache On White Calf Mountain, rodents pilfered 0.544 of caches ( 95% CI: 0.492 0.596 ; Table IV.11 Figure IV.2 ), and on Tibbs Butte, rodents pilfered 0.541 ( 95% CI: 0.490 0.592 ; Table IV.11 Figure IV.2 ). Odds of pilferage did
40 not diffe r between study areas (OR = 1.00 95% CI: 0. 747 1. 34 ; Table IV.12 ), and study areas were pooled for subsequent analyses. Elevation Zone. Rodents pilfered seeds from 0.461 ( 95% CI: 0.409 0.514 ; Table IV.11 Figure IV.2 ) of the caches in the subalpine zone and 0.621 ( 95% CI: 0.571 0.671 ; Table IV.11 Figure IV.2 ) of the caches at treeline. The odds ratio estimate and 95% confidence interval suggest that the odds of pilfer age in the subalpine were approximately 0.544 the odds of pilferage at treeline (OR = 0. 524 95% CI: 0. 389 0. 706 ; Table IV.13 ) Object. In the subalpine zone, the odds of pilferage near no object were 0. 588 times the odds of pilferage near trees ( OR = 0. 588, 95% CI: 0. 350 0. 988 ; Table IV.14 ) Odds of pilferage were not different between rocks and trees, or rocks and no object. At treeline, there were no differences in the odds of pilferage by object (Table IV .14 ). N umber of Seeds Per Cache In the subal pine zone, the odds of pilferage of o ne seed caches were lower relative to the odds of pilferage of six and seven seed caches ( Table IV.15) Odds of pilferage of two seed caches were lower relative to six seed caches (OR = 0.256 95% CI = 0.099 0.663 ), a nd the odds of pilferage of four seed caches were lower relative to six seed caches (OR = 0.316 95% CI = 0.120 0.829 ; Table IV.15). At treeline, odds of pilferage of two seed caches were lower relative to the odds of pilferage of three, four, five, and seven seed caches (Table IV.15). Trapping Study White Calf Mountain. On White Calf Mountain, I captured Zapus princeps Tamias spp ., and P. maniculatus in the subalpine web, and Tamias spp ., P. maniculatus
41 and Microtus sp in the treeline web. In 2012, I recorded 70 captures of 38 individuals from the subalpine web and 55 captures of 39 individuals from the treeline web on White Calf Mountain (Table IV.24 ) The lowest $AIC values for the treeline web were the Mb and Mbh models (Table IV.16, IV.17) ; howev er, abundance estimates using these models were negative. I therefore used the model with the lowest $AIC value with a non negative abundance estimate. For the subalpine web on White Calf Mountain, the best model was a Poisson variant accounting for heter ogeneity in both time and individual capture probabilities ( Mth Table IV.16 ), which estimated 29.5 individuals ( 95 % CI = 29 .0 56.9 ). The model with the lowest $AIC value and positive abundance estimate for the treeline web was the Chao variant of the Mth model for heterogeneity in time and individual capture probabilities (Table IV.17) which estimated 53.2 individuals ( 95% CI: 41.2 71.9). In 2013, I recorded 84 captures of 35 individuals for the first web and 25 captures of 11 individuals for the second web on White Calf Mountain (Table IV.24 ) The best model for both White Calf trapping webs was a Darroch variant of the Mh model, which accounted for heterogeneity in capture probabilities (Table IV.18) Abundance was estimated to be 44 individuals ( 95% CI: 36.0 67.9) for the first web, and 18.2 individuals ( 95% CI: 11 .0 52.4) for the second web ( Table IV.24 ). I calculated density as the abundance estimate divided by web area. For 2012, I calculated rodent density for both elevation zones. For 2 013 data, I calculated rodent density using the abundance estimate of one randomly selected web in the treeline zone. Rodent density in 2012 was 168.6 (95% CI = 147.7 289.8) individuals per hecta re in
42 the subalpine zone and 25 7 .2 ( 95% CI = 209.8 366.2) individuals per hectare in the treeline zone. In 2013, rodent density was estimated using the abundance estimate from the first web. Density was estimated to be 5 6 .0 (95% CI = 45.8 86.5) individuals per hectare in 2013 at treeline (Table IV.24) Tibbs B utte. On Tibbs Butte, I captured Tamias spp in the subalpine web, and P. maniculatus, Tamias spp ., Spermophilus lateralis and a Microtus sp. in the treeline web. In 2012, I recorded 16 captures of 6 individuals at the treeline web and 18 captures of 9 in dividuals in the subalpine web (Table IV.24 ) In 2012, $AIC values for both the subalpine and treeline web were lowest using Mb and Mbh models; however, abundance estimates using these models were negative. For the subalpine web, I used the model with the lowest $AIC and positive abundance estimate the Mt model, which accounts for heterogeneity in time (Table IV.20 ) Abundance in the subalpine web was estimated to be 9 .0 individuals ( 95% CI: 9 .0 48.8). The best model for treeline ($AIC = 0 ) accounted for heterogeneity in capture probabilities over time ( Mt ; Table IV.21 ), which estimated abundance at treeline to be 7 .0 individuals ( 95% CI: 7 .0 9.2). In 2013, I recorded 117 captures of 55 individuals; this comprised 32 captures of 27 individuals at the first web and 61 captures of 33 individuals at the second. The lowest $AIC values for both webs were Mb and Mbh models; however, abundance estimates using these models were negative. Therefore, for the first web I used the Darroch et al. (1993) model accounting for both time and trap heterogeneity ( Mt h Darroch ; Table IV.22 ), which estimated an abundance of 38 .0 individuals (95% CI= 28. 6 73.3 ).
43 For the second web, lowest $AIC values closely ranked three models with positive abundance estimates: Mt Mth Darroch (Table IV.23) I selected the top model by determining which model best predicted the number of new captures on each occasion (Rivest and Baillergeon 2007). This was the Mt model, which estimate d 42 individuals (95% CI= 33.94 64.19). On Tibbs Butte in 2012, rodent density was estimated to be 45.8 (95% CI = 45.8 244.5) individuals per hectare in the subal pine and 35.7 (95% CI = 35.7 46.9) individuals per hectare at treeline. The de nsity at treeline in 2013 was 4 8 .4 (95% CI = 36.4 93.3) individuals per hectare (Table IV.24 )
44 Figures and Tables Table IV.1: Mean seed cache germination rates and 95% confidence intervals ( CI ) on White Calf Mountain and Tibbs Butte in the subalpine and treeline zones of each study area Table IV.2: Mean germination rate and 95% confidence interval (CI) for caches placed in 2012 in protected (leeward of rocks, leeward of trees and among vegetation combined ) and unprotec ted (windward of rocks, windward of trees and open combined ) cache sites in the subalpine and treeline zones on White Calf Mountain and elevation zones combined on Tibbs Butte. Study Area Proportion Germinated 95% CI Elevation Zone Proportion Germinated 95% CI White Calf Mountain 0.256 0.211 0.320 Subalpine 0.206 0. 146 0.266 Treeline 0.307 0.239 0.375 Tibbs Butte 0.380 0.330 0.430 Subalpine 0.412 0.340 0.485 Treeline 0.349 0.281 0.417 Study Area El evation Zone Protection Proportion Germinated 95% CI White Calf Mountain Subalpine Protected 0.217 0.133 0.302 Unprotected 0.193 0.108 0.278 Treeline Protected 0.333 0.234 0.432 Unprotected 0. 2 8 1 0.188 0.374 Tibbs Butte Subalpine & Treeli ne Protected 0.412 0.340 0.463 Unprotected 0.349 0.281 0.437
45 Table IV.3: Mean germination rate and corresponding 95% confidence interval (CI) by object in the subalpine zone of White Calf Mountain, treeline zone of White Calf Mountain, and on Tibbs Butte (eleva tion zones combined). Study Area Elevation Zone Object Proportion Germinated 95% CI White Calf Mountain Subalpine None 0.172 0.075 0.270 Rock 0.237 0.129 0.346 Tree 0.207 0.103 0.311 Treeline None 0.353 0.222 0.484 Rock 0.328 0.213 0.441 Tree 0.241 0.131 0.0.351 Tibbs Butte Subalpine & Treeline None 0.456 0.369 0.543 Rock 0.429 0.342 0.518 Tree 0.250 0.173 0.327
46 Table IV.4: Mean germination rates and corresponding 95% confidence intervals (CIs) by microsite type in the subalpine on White Calf Mountain, treeline on White Calf Mountain, and on T ibbs Butte (elevation zones combined). Study Area Elevation Zone Microsite Proportion Germinated 95% CI White Calf Mountain Subalpine Leeward of Rock 0.267 0.108 0.424 Leeward of Tree 0.250 0.0896 0.410 Vegetation 0.147 0.0280 0.266 Windward of Rock 0.207 0.0595 0.354 Windward of tree 0.167 0.0333 0.300 Open 0.208 0.0459 0.371 Treeline Leeward of Rock 0.412 0.246 0.577 Leeward of Tree 0.250 0.0896 0.410 Vegetation 0.320 0.137 0.503 Windward of Rock 0.242 0.096 0. 389 Windward of tree 0.233 0.0820 0.385 Open 0.385 0.198 0.572 Tibbs Butte Subalpine & Treeline Leeward of Rock 0.458 0.331 0.585 Leeward of Tree 0.288 0.173 0.404 Vegetation 0.429 0.306 0.551 Windward of Rock 0.403 0.281 0.525 Windward of tree 0.213 0.110 0.316 Open 0.484 0.359 0.608
47 Fig. IV.1: Germination rates by A: study area, B: elevation zone, C: protection, D: object, E: microsite type and F: number of seeds per cache Elevation zones on Tibbs Butte are not broken out because germination rates did not differ between zones. Light gray: subalpine zone on White Calf Mtn., dark gray: treeline zone on White Calf Mtn., black: Tibbs Butte. Points represent the mean germination rate for each cache site type, and bars represent the 95% confidence i nterval. Tree Rock None Tree Rock None Tree Rock None 0.0 0.2 0.4 0.6 Proportion Germinated Unprotected Protected Unprotected Protected Unprotected Protected 0.0 0.2 0.4 0.6 Proportion Germinated Wind Tree Wind Rock Vegetation Open Lee Tree Lee Rock Wind Tree Wind Rock Vegetation Open Lee Tree Lee Rock Wind Tree Wind Rock Vegetation Open Lee Tree Lee Rock 0.0 0.2 0.4 0.6 Proportion Germinated 7 6 5 4 3 2 1 7 6 5 4 3 2 1 7 6 5 4 3 2 1 0.00 0.25 0.50 0.75 1.00 Proportion Germinated Tibbs Butte White Calf Mtn. 0.0 0.3 0.6 0.9 Proportion Germinated Treeline Subalpine Treeline Subalpine 0.0 0.2 0.4 0.6 Proportion Germinated A B C D E F
48 Table IV. 5 : Odds ratio estimate and 95% confidence interval comparing the odds of germination on White Calf Mountain to the odds of germination on Tibbs Butte. 95% confidence interval does no t include 1. Table IV. 6 : Odds ratio estimate and 95% confidence interval comparing the odds of germination in the subalpine zone to the odds of ge rmination in the treeline zone on both White Calf Mountain and Tibbs Butte SA: subalpine, TL: treeline. 95% confidence interval does not include 1. Table IV .7 : Odds ratio estimates and 95% confidence intervals of pairw ise comparisons of the odds of germination by nurse object. R: rock, N: no object, T: tree. 95% confidence interval does not include 1. Table IV.8: Odds ratio estimate and 95% confidence interval comparing the odds of germination in unprotected sites relative to protected sites in the subalpine zone on White Calf Mountain, treeline z one on White Calf Mountain, and on Tibbs Butte U: unprotected, P: protected. 95% confidence interval does not include 1. Comparison OR 95% CI WC/TB 0.563* 0.409 0.775 White Calf Mtn. Tibbs Butte Comparison OR 95% CI OR 95% CI SA/TL 0.585* 0.360 0.952 1.31 0.857 2.00 White Calf Mtn: Subalpine White Calf Mtn.: Treeline Tibbs Butte Comparison OR 95% CI OR 95% CI OR 95% CI N/T 0.799 0.315 2.03 1.71 0.746 3.94 2.51 1.46 4.33 R/T 1.19 0.498 2.86 1.54 0.698 3.38 2.26 1.31 3.91 R/N 1.49 0.602 3.70 0.896 0.416 1.93 0.899 0.543 1.49 White Calf Mtn.: Subalpine White Calf Mtn.: Treeline Tibbs Butte Comparison OR 95% CI OR 95% CI OR 95% CI U/P 0. 860 0.411 1.80 0.781 0.411 1.48 0.900 0.590 1.37
49 Table IV .9 : Odds ratio estimates and 95% confidence intervals of pairwise comparisons of the odds of germination by micros ite type. LR: leeward of rock, LT: leeward of tree, V: vegetation, O: open, WR: windward of rock, WT: windward of tree. 95% confidence interval does not include 1. White Calf Mtn.: Subalpine White Calf Mtn.: Treeline Tibbs Butte Comparison OR 95% CI OR 95% CI OR 95% CI LR/LT 1.09 0.336 3.54 2.10 0. 703 6.27 2.08 0.973 4.46 LR/V 2.11 0.606 7.3 4 1.49 0.504 4.39 1.23 0.550 2.30 LR/WR 1.39 0.416 4.67 2.19 0.766 6.24 1.25 0.607 2.57 LR/WT 1.82 0. 518 6.38 2.30 0. 775 6. 82 3.1 2 1.40 6.92 LR/O 1.38 0.386 4.94 1.12 0. 3 94 3.18 0.900 0.441 1.84 V/LT 0.517 0.144 1. 86 1. 41 0.426 4.68 1.85 0.873 3.93 WR/LT 0.783 0.226 2.71 0. 960 0. 298 3.09 1.67 0.782 3.56 LT/WT 1.67 0.461 6. 03 1.10 0. 329 3.65 1.49 0.650 3.44 O/LT 0. 789 0.214 2.91 1.88 0. 585 6. 01 2.32 1.09 4.91 V/ WR 0.661 0.179 2.44 1. 47 0. 462 4.68 1.11 0.545 2.26 V/WT 0.862 0.223 3.33 1. 55 0. 469 5.10 2.77 1.26 6.10 V/O 0.655 0.167 2.57 0.753 0. 238 2.39 0.800 0.395 1.62 WR/WT 1.30 0.350 4.86 1.05 0. 329 3. 36 2.49 1 .13 5.53 O/WR 1.01 0.266 3.83 1.95 0. 636 6.00 1.39 0.681 2.83 O/WT 1.32 0.332 5.21 2. 05 0. 645 6.54 3.46 1.57 7.62
50 Table IV.10: Odds ratio estimates and 95% confidence intervals comparing odds of germination by number of seeds per cache. 95% confidence interval does not include 1. 0 Comparison White Calf Mtn. Subalpine White Calf Mtn. Treeline Tibbs Butte 1/2 1 .44 0. 345 5.98 0. 727 0. 255 2. 07 0.673 0. 291 1.56 1/3 0.892 0. 236 3.38 1. 05 0. 372 2.96 0. 441* 0. 195 0.995 1/4 0.808 0. 202 3.23 1.74 0. 553 5.45 0.695 0. 298 1.62 1/5 1.0 5 0. 233 4.69 1.24 0. 309 4.96 0.838 0. 331 2. 12 1/6 2.77 0. 439 17.5 0. 788 0. 183 3.39 0. 884 0. 298 2.62 1/7 4.00 0. 392 40.8 0. 720 0. 184 2.82 0. 354 0. 113 1.10 2/3 0. 621 0. 199 1.94 1. 44 0. 549 3.80 0. 654 0. 349 1. 23 2/4 0. 563 0. 169 1. 87 2.39 0.810 7.04 1.03 0. 528 2.02 2/5 0. 729 0. 192 2.76 1.70 0 447 6.48 1.24 0. 576 2.69 2/6 1. 93 0. 350 10.6 1.08 0. 265 4.43 1.31 0. 505 3.41 2/7 2.79 0. 304 25.6 0. 990 0. 266 3.69 0. 525 0. 190 1.45 3/4 0. 905 0. 306 2.68 1. 65 0. 566 4. 83 1. 58 0. 832 2.99 3/5 1.17 0. 343 4.01 1.18 0. 312 4.45 1. 90 0. 904 4.00 3/6 3.10 0. 609 15.8 0. 75 0. 185 3.05 2.01 0. 789 5.10 3/7 4.48 0. 518 38.8 0. 686 0. 186 2.53 0.802 0. 297 2.17 4/5 1.30 0. 357 4.69 0. 713 0. 173 2.93 1.21 0. 553 2.63 4/6 3.43 0. 644 18.3 0. 454 0. 103 2.00 1.27 0. 486 3.33 4/ 7 4.95 0. 554 44.3 0. 415 0. 103 1.67 0. 509 0. 183 1.41 5/6 2. 65 0. 451 15.5 0. 636 0. 119 3.41 1.05 0. 375 2.97 5/7 3.82 0. 397 36.8 0. 582 0. 118 2. 88 0. 422 0. 142 1.25 6/7 1. 44 0. 118 17.7 0.914 0. 174 4.81 0. 400 0. 117 1.36
51 Tibbs Butte White Calf Mtn. 0.0 0.2 0.4 0.6 Proportion Pilfered Subalpine Treeline 0.0 0.2 0.4 0.6 Proportion Pilfered 7 6 5 4 3 2 1 7 6 5 4 3 2 1 0.0 0.3 0.6 0.9 Proportion Pilfered Tree Rock None Tree Rock None 0.0 0.2 0.4 0.6 Proportion Pilfered B A C D Table IV.11 : Proportion of caches pilfered and 95% confidence intervals by study area, elevation zone, and object. Fig. IV.2: Proportion of pilfered caches by cache site characteristics. Mean proportion pilfered and 95% confidence interval by A: study area, B: elevation zone, C : object, and D: number of seeds per cache Gray: treeline zone, black: subalpine zone. Study areas are not broken out because pilferage rates did not differ between areas. Study Area Proportion Pilfered 95% CI Elevation Zone Proportion Pilfered 95% CI Object Proportion P ilfered 95% CI White Calf Mountain 0.544 0.492 0.596 Subalpine 0.463 0.411 0.515 Rock 0.487 0.397 0.578 None 0.387 0.299 0.474 Tree 0.517 0.426 0.608 Tibbs Butte 0.544 0.493 0.595 Treeline 0.622 0.572 0.672 Rock 0.615 0.532 0.69 9 None 0.591 0.501 0.681 Tree 0.653 0.573 0.743
52 Table IV .12 : Odds ratio estimate and 95% confidence interval comparing the odds of pilferage on White Calf Mountain to the odds of germination on Tibbs Butte. Table IV .13 : Odd s ratio estimate and 95% confidence interval comparing the odds of pilferage in the subalpine zone to the odds of germination in the treeline zone. SA: subalpine, TL: treeline. 95% confidence interval does not include 1. Table IV .14 : Odds ratio estimates and 95% confidence intervals of pairwise comparisons of the odds of pilferage by nurse object. R: rock, N: no object, T: tree. 95% confidence interval does not include 1. Comparison OR 95% CI WC/TB 1.00 0.747 1.34 Comparison OR 95% CI SA/TL 0.52 4* 0.389 0.706 Subalpine Treeline Comparis on OR 95% CI OR 95% CI N/T 0.588* 0.350 0.988 0.751 0.442 1.28 R/T 0.887 0.530 1.48 0.830 0.495 1.39 R/N 1.51 0.899 2.53 1.11 0.662 1.85
53 Table IV.15 : Odds ratio estimates and 95% confidence intervals comparing odds of pilferag e by number of seeds planted per cache. *95% confidence interval does not include 1. Comparison Subalpine Treeline 1/2 0.821 0.377 1.79 1.01 0.466 2.21 1/3 0.513 0.239 1.10 0.484 0.227 1.03 1/4 0.667 0.301 1.48 0.348* 0.156 0.780 1/5 0.477 0.202 1.13 0.332* 0.131 0.840 1/6 0.211* 0.075 0.591 0.524 0.199 1.38 1/7 0.321* 0.113 0.911 0.396 0.134 1.17 2/3 0.625 0.327 1.19 0.477* 0.251 0.905 2/4 0.812 0.409 1.61 0.343* 0.171 0.690 2/5 0.581 0.272 1.24 0.327* 0.142 0. 754 2/6 0.256* 0.099 0.663 0.516 0.215 1.24 2/7 0.391 0.150 1.02 0.390* 0.143 1.07 3/4 1.30 0.665 2.54 0.720 0.366 1.41 3/5 0.930 0.441 1.96 0.686 0.303 1.55 3/6 0.410 0.160 1.05 1.08 0.458 2.56 3/7 0.626 0.242 1.62 0.819 0.304 2.20 4/5 0.716 0.329 1.56 0.953 0.403 2.26 4/6 0.316* 0.120 0.829 1.51 0.610 3.72 4/7 0.482 0.182 1.28 1.14 0.407 1.14 5/6 0.441 0.159 1.22 1.58 0.573 4.35 5/7 0.673 0.241 1.88 1.19 0.387 3.68 6/7 1.53 0.471 4.95 0.756 0.237 2.41
54 Table IV.16: Model summaries for the subalpine trapping web on White Calf Mountain in 2012. Table IV.17: Model summaries for the subalpine trapping web on White Calf Mountain in 2012. Model Deviance DF AIC AIC M t 15 246 66.8 0 M th(Poisson) 13.9 245 67.7 0.9 M th(Darroch) 14 245 67.8 1 M th(Chao) 15 245 68.8 2 M b 30.6 252 70.4 3.6 M bh 30.6 251 72.4 5.6 M o 61.4 253 99.2 32.4 M h(Chao) 61.4 253 99.2 32.4 M h(Poisson) 59.8 25 2 99.6 32.8 M h(Darroch) 59.8 252 99.6 32.8 Model Deviance DF AIC AIC M bh 72.2 251 128 0 M b 74.5 252 129 1 M th(Chao) 64 245 132 4 M th(Darroch) 66.6 245 135 7 M t 70.7 246 137 9 M th(Poisson) 69 245 137 9 M h(Chao) 98.9 252 153 25 M h(Darroch) 101 252 155 27 M o 104 253 156 28 M h(Poisson) 103 252 157 29
55 Table IV.18: Model summaries for web 1 on White Calf Mountain in 2013. Table IV.19: Model summaries for web 2 on White Calf Mountain in 2013. Model D eviance DF AIC AIC M b 89.4 252 147.2 0 M bh 88.3 251 148.1 0.9 M th(Darroch) 78.1 245 150 2.8 M th(Chao) 74.7 243 150.5 3.3 M th(Poisson) 89.4 245 161.2 14 M h(Chao) 108 251 167.8 20.6 M h(Darroch) 111.7 252 169.5 22.3 M t 101.4 246 171.2 24 M h(Poisson) 120.7 252 178 .5 31.3 M o 129.6 253 185.4 38.2 Model Deviance DF AIC AIC M b 52.6 252 74.5 0 M bh 52.6 251 76.5 2 M h(Darroch) 56 252 77.9 3.4 M h(Chao) 55.4 251 79.3 4.8 M th(Darroch) 43.5 245 79.3 4.8 M h(Poisson) 58.1 252 80 5.5 M th(Poisson) 46.3 245 82.2 7.7 M th(Chao) 42.8 243 82.7 8.2 M o 64.3 253 84.2 9.7 M t 53.9 246 87.8 13.3
56 Table IV.20: Model summaries for the subalpine trapping web on Tibbs Butte in 2012. Table IV.21: Model summaries for the treeline trapping web on Tibbs Butte in 2012. Model Deviance DF AIC AIC M b 2 2.2 252 43.5 0 M bh 21.7 251 44.9 1.4 M t 23.3 246 56.6 13.1 M th(Darroch) 22.7 245 57.9 14.4 M th(Poisson) 23.1 245 58.3 14.8 M th(Chao) 22.2 242 63.4 19.9 M o 44.7 253 63.9 20.4 M h(Chao) 44.2 252 65.4 21.9 M h(Darroch) 44.6 252 65.8 22.3 M h(Poisson) 44 .7 252 65.9 22.4 Model Deviance DF AIC AIC M b 16.9 252 35.5 0 M bh 16.5 251 37.1 1.6 M t 21.5 246 52.1 16.6 M th(Darroch) 19.5 245 52.1 16.6 M th(Poisson) 20.5 245 53.1 17.6 M th(Chao) 18.2 242 56.9 21.4 M o 41.4 253 58 22.5 M h(Chao) 39.7 252 58.4 22.9 M h(Darroch) 41 252 59.6 24.1 M h(Pois son) 41.4 252 60 24.5
57 Table IV.22: Model summaries for web 1 on Tibbs Butte in 2013. Table IV.23 : Model summaries for web 2 on Tibbs Butte in 2013. Model Deviance DF AIC AIC M b 46.9 124 91.5 0 M bh 45.7 123 92.3 0.8 M th(Darroch) 44.3 118 100.9 9.4 M th(Poisson) 46.2 118 102.8 11.3 M th( Chao) 43.9 115 106.4 14.9 M t 52.9 119 107.5 16 M h(Darroch) 80.6 124 125.2 33.7 M h(Poisson) 81.8 124 126.4 34.9 M o 85.5 125 128.1 36.6 M h(Chao) 80.3 121 130.9 39.4 Model Deviance DF AIC AIC M b 47.7 124 93 0 M bh 46.7 123 94.1 1.1 M t 48.8 119 104.1 11.1 M th(Darroch) 47.1 118 104.4 11.4 M th(Poisson) 48.1 118 105.5 12.5 M th(Chao) 46.3 117 105.6 12.6 M o 89.5 125 132.8 39.8 M h(Chao) 88.3 124 133.6 40.6 M h(Darroch) 88.9 124 134.2 41.2 M h(Poisson) 89.3 124 134.7 41.7
58 Table IV.24 : Granivorous mammal density estimates by year and elevation zone. Mammal density estimates by year and elevation zone. 2013 density estimates were based on the abundance estimate of one randomly selected web. Area Year Elevation Zone Abundance (95% CI) no. individuals Web Area hectares Density (95% CI) individuals/hectare White Calf Mountain 2012 Subalpin e 33.1 29 .0 56.9 0.1963 5 16 8 .6 147.7 289.8 Treeline 50.5 41.2 71.9 0.1963 5 25 7 .2 209.8 366.2 2013 Treeline 44 36.0 67.9 0.7854 0 5 6 .0 45.8 86.5 Tibbs Butte 2012 Subalpine 9 .0 9.0 48.8 0.1963 5 4 5.8 45.8 244.5 Treeline 7 .0 7.0 9.2 0 .1963 5 35.7 35.7 46.9 2013 Treeline 38 .0 28.6 73.3 0.7854 0 4 8 .4 36.4 93.3
59 CHAPTER V DISCUSSION AND CONCLUSIONS Germination Study Area and Elevation Z one. Mean overall germination rate, which in cludes both study areas, was 0.319 ( 95% CI: 0.28 5 0.354% and ranged from 0.2 06 ( 95% CI: 0.14 6 0.266 ) to 0.412 (95% CI = 0.34 0 0.485 ) across study area s and elevation zone s Mean germination rates appear to be consistent with other studies, which have ranged from 0. 07 to 0.58 0 depending on study location (Schwandt et al. 2007, Schwandt et al. 2011, McLane and Aitken 2012, DeMastus 2013). Differences in the odds of germination were observed between study areas and, on White Calf Mountain, between elevation zones. The odds of germination on Tibbs Butte were nearly twice as great as on White Calf Mountain (OR (WC: TB) = 0.587 95% CI: 0.425 0.811 ), suggesting that successful whitebark pine germination may depend upon the region in which seeds are cached These findings are support ed by direct seeding trials conducted in the North ern Rocky Mountains and Cascade s which reported a range of germinations among study areas (range = 5.0 58.7% ); however these results include second and third year germination, which is common in whitebark pine (Schwandt et al. 2007, Schwandt et al. 2011 DeMastus 2013). The low germination rate (0.206, 95% CI: 0.146 0.266) in the subalpine forest on White Calf Mountain could be attribu ted to its dense, late successional forest structure. These forest types decrease whitebark pine regeneration by limit ing local caching opportunities for nutcrackers and more importantly for discussion here, the availability
60 of sunlight for seedlings (Tomback 1978, 2001 Baud 1993, Tomback e t al. 1993, Bansal et al. 2011) Results from Tomback et al. (2001) show that whi tebark pine germinates across a range of forest types and community structures. However, the v ariation in odds of germination by study area and elevation zone on White Calf Mountain could be attributed to several mechanisms including seed source, climatic vari ation, soil type, or topography Seed viability does appear to vary across whitebark pine's range in nursery settings, but these differences have often been attributed to variation in seed maturity and embryo development and not inherent variation amon g seed sources (Mahalovich et al. 2006). However, because seeds were collected at the same time at both study areas, and seeds were weighed to select those more likely to contain a full embryo, seed maturity did not likely contribute to the differences in germination rates. In many other plant species from a range of families, d ifferences in seed viability and subsequent germination based on seed source are common (Krugman and Jenkinson 1974, Ginwal et al. 2005). More research is needed to assess the role o f regional differences such as climate, seed source, soil type, and topography play in whitebark pine germination. Object. Comparisons of the odds of germination by elevation zone revealed no difference between the subalpine and treeline zone on Tibbs Butt e (Table IV.6) but the odds of germination in the subalpine of White Calf Mountain were 0.588 ( 95% CI: 0.362 0.958) times the odds of germination at treeline. This suggests that, on White Calf Mountain, the treeline environment was more conducive to ger mination than the subalpine environment.
61 The success of early life stages (i.e. germinants and seedlings) initiates the spatial dynamics of mature trees ( Rey and Alcantara 2000, Rochefort and Peterson 1996 ). Patterns of germination on Tibbs Butte indicate that the odds of germination near rocks and sites with no object were increased when compared to trees (no object/tree OR = 2.5 1, 95% CI: 1.46 4.33; rock/tree OR = 2.26 95% CI: 1.31 3.91 ). Similar patterns have been observed in the spatial distributi on of mature whitebark pine at treeline throughout the Northern Rocky Mountains, where mature trees are more likely found growing as solitary individuals than associated with other trees (Tomback et al. 2014). Additionally Tomback et al. (2014) observed t hat solitary trees were most often associated with rocks. I found no difference in the odds of germination near rocks relative to sites with no object (OR = 0.899 95% CI: 0.543 1.49 ), suggesting that differential survival may account for the observed di screpancy between the spatial distribution of seedlings and mature trees For this reason, information on the relative survival of seedlings associated with each object (rock, none, or tree) is necessary for elucidating the bottlenecks associated with esta blishment to mature tree spatial patterns. T hese results appear to conflict with the spatial distribution of germinants in a regeneration assessment conducted on Tibbs Butte and nearby Wyoming Creek in 1993 1994 which found germinants at a higher density leeward of trees (Mellmann Brown 2005). However, the results presented in Mellmann Brown (2005) are based on a small sample size and less rigorous sampling strategy. Microsite. Odds of germination by microsite type did not differ on White Calf Mountain bu t did on Tibbs Butte Odds of germination were increased leeward and windward of rocks, in vegetation, and in open microsites relative to the odds of
62 germination windward of trees (Table IV.9). Odds of germination were additionally increased in open micros ites relative to the odds of germination leeward of trees (Table IV.9). These results produce an unclear signal in the variation of odds of germination by microsite type, suggesting that the odds of germination are better explained by cache site object. An important consideration is that these germination rates are confounded by pilferage. All germination rates are tied to the odds of pilferage at each cache. Therefore, these results should not be interpreted without understanding pilferage dynamics. Numbe r of Seeds Per Cache In general, the odds of germination did not vary with cache size (Table IV.10). The only exception was on Tibbs Butte, where the odds of germination of one seed caches were lower relative to 3 seed caches (OR = 0.441, 95% CI = 0.195 0.995 ). These results suggest that cache size does not have a strong influence on the odds of germination. Pilferage and Seed Survival Limitations A striking result is the pilferage rate observed: 0.541 of all caches assessed lost one or more seeds. Odd s of pilferage in the subalpine were 0.52 4 ( 95% CI: 0.389 0.706 ) those at treeline indicating that pilferage rates increased at treeline In the subalpine, the odds of pilferage for one seed caches were lower relative to six and seven seed caches (Tabl e IV.15) odds of pilferage for two seed caches were lower relative to six seed caches, and odds of pilferage of four seed caches were lower relative to six seed cac hes. At treeline, the o dds of pilferage were lower for two seed caches relative to three, f ou r, five, and seven seed caches. All other cache sizes had similar odds of pilferage. These results suggest that general trends in pilferage by cache
63 size were either not observed, or that seed pilferage is a function of random discovery by rodents that p ass by. B ecause rodents rely on olfactory cues to locate caches, elevated olfactory stimulation caused by larger cache size should increase ability to loc ate seeds more information is needed to understand the characteristics of caches that increase the pr obability of cache pilferage (Vander Wall 1998). Vander Wall et al. (2005) caution against associating seed loss caused by pilferage with seed predation. Secondary seed dispersal the movement of seeds to new locations where germination may be possible ha s been documented in other pine species (Vander Wall et al. 2005). Although it has been hypothesized that rodents are secondary dispersers of whitebark pine seeds, seed coats showing evidence of predation were found near 65% and 22% of pilfered caches on T ibbs Butte and White Calf Mountain, respectively. The proportion of seed coats showing signs of predation on Tibbs Butte suggest that, in a majority of cases, pilferage results in seed predation and the probability of re caching and secondary dispersal is low. Preliminary seed tracking results on White Calf Mountain with seeds covered in fluorescent pigment suggest that seeds pilfered from caches are not likely to contribute to regeneration via secondary dispersal. In 2013, I tracked fluorescent pigment co ated seeds over three nights on White Calf Mountain using the methodology of Tomback et al. (2005) and Longland and Clements (1998). I detected a total of 19 track lines, of which 7 resulted in caches and 8 ended in rodent holes. The 7 caches were all surf ace caches comprised of 1 3 seeds. These results indicate that pilfered seeds are either larder hoarded or surface cached. Neither fate is likely to lead to germination (Levitt 1972, Tomback et al. 2005).
64 Seed pilferage appears to result largely in seed loss caused by either seed predation or deposition of seeds in locations unfavorable for germination, suggesting that pilferage decreases seed availability for regeneration and recruitment. Previous studies have shown that seed predation generally does no t constrain regeneration when microsites are limiting. However, when seed availability is low, seed predation can act as a limiting factor ( Archibald et al. 2012, Calvi Â– o Cancela 2007, Nathan and Ne'eman 2004 Crawley 2000 Ims 1990, Andersen 1989 ). Seed l imitations might negatively impact regeneration in non mast years, but limitations might be overcome in mast years when seeds are not limiting. Seed Dispersal by Clark's Nutcrackers The number of seeds cached by an individual nutcracker has been estimat ed from data gathered both in the eastern Sierra Nevada mountain range and the Rocky Mountains in the Greater Yellowstone Area (Hutchins and Lanner 1982, Tomback 1982). Based on field observations i n the Sierra Nevada, Tomback (1982) estimated that an indi vidual nutcracker caches 35,000 seeds and Hutchins and Lanner (1982) working in the Rocky Mountains, estimated a higher value of 98,000 seeds annually These figures both assume seed supply is abundant. Of those seeds cached Tomback (1982) estimated 45% or 15,750 44,100 seeds are not retrieved (Tomback 1982). Using the low end of the number of seeds above (15,750), the average number of seeds per cache ( ! ) and the overall cache germination rate of 0.319 observed on White Calf Mountain and Tibbs Butte (Tomback 1982), I calculated that, on average, an individual bird would create at least 1,674 caches that would germinate. This is despite an average pil ferage rate of 0.54
65 This calculation assumes that caching behavior (number of caches created, microsite types used, number of seeds per cache) is constant across whitebark pine's range. However, both Tomback (1982) and Hutchins and Lanner (1982) observed nutcrackers cache most frequently at the base of trees, which on Tibbs Butte supported the lowest average germination rate of any nurse object (25.0%). Even assuming a 25.0% germination rate for all caches created (i.e., every cache is created in tree micr osites), I estimate that 1,076 3,675 caches would germinate each year as a result of one nutcracker's efforts However, the timing of pilferage could drastically alter these results. The figures calculated above assume that pilferage occurs after nutcrac kers retrieve their seeds, whereas the timing of pilferage likely overlaps with nutcracker retrieval and therefore will decrease the number of seeds available to the nutcracker, potentially increasing the number of caches a bird must revisit prior to obtai ning sufficient energy stores. The true number of caches that germinate as a result of a single nutcracker s efforts is likely less than estimated and would be clarified by a better understanding of pilferage dynamics, especially the timing of pilferage. T rapping Trapping efforts resulted in the identification of both diurnal and nocturnal rodents that might be responsible for cache pilferage. These species display life history tr aits of both scatter and larder hoarding, indicating that the possibility of s econdary dispersal will likely vary depending upon the relative contribution of each species to cache pilferage. Scatter hoarding P. maniculatus and Tam i a s species have been implicated in cache pilferage and seed dispersal of a variety of pine species incl uding Jeffr e y ( P. jeffreyi A.Murray bis ), ponderosa ( P. ponderosa Douglas ex C. Lawson)
66 sugar, and limber pines (Tomback et al. 2005, Vander Wall 2003, 1993). Spermophilus lateralis is reported to scatter hoard pine seeds, but generally buries seeds too d eep for successful seedling emergence (Vander Wall 2003, 2002). Similarly, any seeds pilfered by the larder hoarding Z. princeps are unlikely to contribute to regeneration. Small mammal d ensity estimates in 2013 were similar between study areas (Table IV.2 0), whereas rodent density in the treeline zone on White Calf Mountain appeared to decrease by 4.5 times from 2012 to 2013. Similar pilferage rates at both study areas suggest that similar densities may lead to similar pilferage rates. However, this assume s that pilferage occurs during the spring and summer rather than during the fall. M oisture plays an integral role in successful location of cached lodgepole and Jeffrey pine seeds suggesting that t he period after snowmelt when the soil is saturated might allow rodents to locate cached seeds more easily (Vander Wall 1995, 1998) Management and Restoration DeMastus (2013) suggests that direct seeding efforts could be considered successful with combined germination and survival rates of 10%. Although surviva l rates are still unknown, whitebark pine survival rates have been reported to be approximately 50% on the Beartooth Mountains, where Tibbs Butte is located (Maher and Germino 2006). Assuming this survival rate, second year survival on Tibbs Butte would ap proach 18%, suggesting that observed germination rates may be sufficient to support direct seeding at this area. Additionally, inf ormation herein provides valuable insight into the feasibility of restoring whitebark pine communities in a changing climate ( Keane et al. 2012). As temperatures continue to warm, treeline is predicted ascend and appropriate management strategies are necessary to accommodate these changes (Schrag et al. 2007).
67 Conclusions I examined germinat ion rates of whitebark pine, cache pil ferage rates, how germination and pilferage rates vary with elevation, microsite type, and cache size, which rodent species are present and might be responsible for pilfering nutcracker caches, and whether a relationship exists between granivorous rodent d ensity and pilferage rate at White Calf Mountain Glacier National Park, Montana and Tibbs Butte Shoshone National Forest, Wyoming Results suggest that o ne major barrier to seed germination and seedling e stablishment faced by whitebark pine r e sults from high rates of seed pilferage However, the timing of cache pilferage relative to nutcracker retrieval has the potential to significantly influence seed survival limitations. Pilferage occurring prior to nutcracker retrieval could require nutcrackers to vis it more caches to obtain the seeds they require energetically, resulting in fewer seeds contributing to regeneration. Additionally, e vidence from seed coats and seed tracking suggests that seed pilferage likely results in seed loss due to predation and s econdary dispersal by these rodents appear s infrequent or unlikely. Many regions with heavy whitebark pine mortality may become more seed limited as time progresses because of increasing damage from white pine blister rust and continued mortality from mou ntain pine beetle. This will further reduce seed production and the probability of seed dispersal (McKinney et al. 2009, Barringer et al. 2012, Scott 2013). As seed availability decreases in these areas, seed survival limitations may increase, further redu cing the possibility of regeneration in the most severely affected areas (Andersen 1989). These results suggest that, in regions of high mortality,
68 restoration via planting or direct seeding may become more important as populations decline. Germination doe s not appear to act as a bottleneck to regeneration when seeds are available (e.g., in mast years). However, seedling survival might influence tree recruitment more than germination. More information on the effects of microsite types and other cache site c haracteristics on seedling survival and establishment will help future restoration planning and the possible implementation of direct seeding techniques.
69 REFERENCES Aitken, S.N., S. Yeaman, J.A. Holliday, T. Wang, and S. Curtis McLane. 2008. Adaptation, migration or extirpation: climate change outcomes for tree populations. Evolutionary Applications 1: 95 111. Allen, C.D., A.K. Macalady, H. Chenchouni, D. Bachelet, N. McDowell, M. Vennetier, T. Kitzberger, A. Rigling, D.D. Breshears, E.H. Hogg, P. Gonzal ez, R. Fensham, Z. Zhang, J. Castro, N. Demidova, J. Lim, G. Allard, S. W. Running, A. Semerci, and N. Cobb. 2010. A global overview of drought and heat induced tree mortality reveals emerging climate change risks for forests. Forest Ecology and Management 259: 660 684. Anderson, M. and J. Krebs. 1978. On the evolution of hoarding behavior. Animal Behavior 26: 707 711. Amman, G.D. 1973. Population changes of the mountain pine beetle in relation to elevation. Environmental Entomology, 2: 541 547. Amman, G.D. 1982. Characteristics of mountain pine beetles reared in four pine hosts. Environmental Entomology 11:590 593. Amman, G.D. 1985. A test of lodgepole pine hazard rating methods for mountain pine beetle infestation in southeastern Idaho. In Proceeding s of the IUFRO conference on the role of the host in the population dynamics of forest insects, joint meeting of IUFRO Working Parties S2 07 05 & S2 07 06. Banff Centre, September 4 7, 1983. L. Safranyik (editor). Banff, AB. Canadian Forest Service, Pacifi c Forestry Centre. pp.186 200. Andersen, A.N. 1989. How important is seed predation to recruitment in stable populations in long lived perennials? Oecologia 81: 310 315. Arno, S.F. and R.P. Hammerly. 1984. Timberline: Mountain and arctic forest frontiers Mountaineers Books, Seattle, Washington, USA Arno, S.F. and R.J. Hoff. 1989. Silvics of whitebark pine ( Pinus albicaulis ). Gen. Tech. Rep. IMRS GTR 253. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station. 14 p. Arno S.F. and T. Weaver. 1990. Whitebark pine community types and their patterns on the landscape. In: W.C. Schmidt and K.J. McDonald, compilers. Whitebark pine ecosystems Ecology and management of a high mountain resource. USA Forest Service, Intermountain Research Station, General Technical Report INT 270. Pages 97 105. Arno, S.F. 2001. Community types and natural disturbance processes: In: Whitebark Pine Communities: Ecology and Restoration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 74 88. Islan d Press, Washington DC, USA
70 Baillargeon, S. and L. Rivest. 2012. Rcapture: loglinear models for capture recapture Experiments. R package version 1.3 1. http://CRAN.R project.org/package=Rcapture Balda, R.P. and A.C. Kamil. 1992. Long term spatia l memory in C lark s nutcracker, Nucifraga columbiana Animal Behavior, 44(4): 761 769. Bansal, S., K. Reinhardt, and M.J. Germino. 2011. Linking carbon balance to establishment patterns: comparison of whitebark pine and Engelmann spruce seedlings along an herb cover exposure gradient at treeline. Plant Ecology 212: 219 228. Barringer, L.E., D.F. Tomback, M.B. Wunder, and S.T. McKinney. 2012. Whitebark pine stand condition, tree abundance, and cone production as predictors of visitation by Clark's nutcracker. PLoSONE 7(5): e37663. doi:10.1371/journal.pone.0037663 Baud, K.S. 1993. Simulating Clark's nutcracker caching behavior : germination and predation of seed caches. M.A. thesis. University of Colorado Denver ,Denver, Colorado. Blakeslee, S.C. 2012. Assessing whiteb ark pine vigor and facilitation roles in the alpine treeline ecotone. M.S. thesis, University of Colorado Denver, Denver, Colorado. Brooker, R.W., F.T. Maestre, R.M. Callaway, C.L. Lortie, L.A. Cavieres, G. Kunstler, P. Liancourt, K. TielbÂšrger, J.M.J. Tra vis, F. Anthelme, Z. Kikvidze, J. Olofssson, F. Pugnaire, C.L. Quiroz, P. Saccone, K. Schiffers, M. Seifan, B. Touzard, and R. Michalet. 2008. Facilitation in plant communities: the past, the present, and the future. Journal of Ecology 96: 18 34. Bruederl e, L.P, D.L. Rogers, K.V. Krutovskii, and D.V. Politov. 2001. Population genetics and evolutionary implications. In: Whitebark Pine Communities: Ecology and Restoration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 137 153. Island Press, Washington DC, USA. Buckland, S.T., D.R. Anderson, K.P Burnham, J.L Laake, D.L. Borchers, and L.T. Thomas. 2001. Introduction to distance sampling: estimating abundance of biological populations. Oxford University Press, New York, New York, USA. Burns, K. S., A.W. Sc hoettle, W.R. Jacobi, and M.F. Mahalovich. 2008. Options for the management of white pine blister rust in the Rocky Mountain region. Gen. Tech. Rep. RMRS GTR 206. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Sta tion. 26 p. Burr, K. E., E. Eramian, K. Eggleston. 2001. Growing whitebark pine seedlings for restoration. In: Tomback, D. F.; Arno, S. F.; Keane, R. E., eds. Whitebark pine communities: ecology and restoration. Washington, DC: Island Press: 325 345. Cain, M.L., B.G. Milligan, and A.E. Strand. 2000. Long distance seed dispersal in plant populations. American Journal of Botany 87(9): 1217 1227.
71 Calcagno, V. 2013. glmulti: Model selection and multimodel inference made easy. R package version 1.0.7. http:// CRAN.R project.org/package=glmulti Callaway, R.M. 1998. Competition and facilitation on elevation gradients in subalpine forests of the Northern Rocky Mountains. Oikos 82: 561 573. Calvi Â– o Cancela, M. 2007. Seed and microsite limitations of recruitment an d the impacts of post dispersal seed predation at the within population level. Plant Ecology 192(1): 35 44. Chao, A. 1987. Estimating the population size for capture recapture data with unequal catchability. Biometrics 43: 783 791. Clark, J.S., C. Fas tie, G. Hurtt, S.T. Jackson, C. Johnson, G.A. King, M. Lewis, J. Lynch, S. Pacala, C. Prentice, E.W. Schupp, T. Webb III, and P. Wyckoff. 1998. Reid's paradox of rapid plant migration: dispersal theory and interpretation of paleoecological records. Bioscie nce 48(1): 13 24. Clark, J.S., B. Beckage, P. Camill, B. Cleavland, J. HilleRisLambers, J. Licter, J. McLachlan, J. Mohan, and P. Wyckoff. 1999. Interpreting Recruitment Limitation in Forests. American Journal of Botany 86: 1 16. Clark, C.J., J.R. Poul sen, D.J. Levey, and C.W. Osenberg. 2007. Are plant populations seed limited? A critique and meta analysis of seed addition experiments. The American Naturalist 170: 128 142. Connell, J. H. 1971. On the role of natural enemies in preventing competitive e xclusion in some marine animals and in rain forest trees. In: Den Boer, P. J. and Gradwell, G. (eds), Dynamics of populations. PUDOC, pp. 298 312. Cormack, R.M. 1989. Loglinear models for capture recapture. Biometrics 45: 395 413. Cormack, R.M. 1992. Int erval estimation for Mark Recapture studies of closed populations. Biometrics 48: 567 576. Costanza, R., R. d'Arge, R. deGroot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem, R.V. O'Neill, J. Paruelo, R.G. Raskin, P. Sutton, and M. van den Belt. 1 997. The value of the world's ecosystem services and natural capital. Nature 387: 253 260. Crawley, M.J. 1989. Insect herbivores and plant population dynamics Annual Review of Entomology 34: 531 564. Crawley, M.J. and C.R. Long. 1995. Alternate bearin g, predator satiation and seedling recruitment in Quercus robur L. Journal of Ecology 83: 683 696. Crone, E.E., E.J.B. McIntire, and J. Brodie. 2011. What defines mast seeding? Spatio temporal patterns of cone production by whitebark pine. Journal of Ec ology 99: 438 444.
72 Cullinane Thomas, C., C. Huber, and L. Koontz. 2014. 2014 National Park visitor spending effects: economic contributions to local communities, states, and the nation. Natural Rsource Report NPS/NRSS/EQD/NRR 2014/765. National Park Se rvice, Fort Collins, Colorado. Darroch, J.N., S.E. Fienberg, G.F.V. Glonek, and B.W. Junker. 1993. A three sample multiple recapture approach to sensus population estimation with heterogeneous catchability. Journal of the American Statistical Association 88: 1137 1148. Dayton, P. K. 1972. Toward an understanding of community resilience and the potential effects of enrichments to the benthos at McMurdo Sound, Antarctica. pp. 81 96 in Proceedings of the Colloquium on Conservation Problems Allen Press, Lawren ce, Kansas. DeMastus, C.R. 2013. Effective methods of regenerating whitebark pine ( Pinus albicaulis ) through direct seeding. M.S. thesis, Montana State University, Bozeman, Montana. Detwiler, S.B. 1922. Blister rust appears in the Puget Sound region. Ameri can Forestry 28: 97 98. Ellison, A.M., Bank, M.S., Clinton, B.D., Colburn, E.A., Elliot, K., Ford, C.R., Foster, D.R., Kloeppel, B.D., Knoepp, J.D., Lovett, G.M., Mohan, J., Orwig, D.A., Rodenhouse, N.L., Sobczak, W.V., Stinson, K.A., Stone, J.K., Swan, C .M., Thompson, J. Von Holle, B. and Webster, J.R. 2005. Loss of foundation species: consequences for the structure and dynamics of forested ecosystems. Frontiers in Ecology 3: 479 486. Emery, N.J., J.M. Dally, and N.S. Clayton. 2004. Western scrub jays ( A phelocoma californica ) use cognitive strategies to protect their caches from thieving conspecifics. Animal Cognition 7: 37 43. Eriksson, O. and J. EhrlÂŽn. 1992. Seed and microsite limitation of recruitment in plant populations. Oecologia 91: 360 364. Fa rnes, P.E. 1990. SNOTEL and snow course data describing the hydrology of whitebark pine ecosystems. In Schmidt, W.C., McDonald, K.J., comps. Proceedings Symposium on whitebark pine ecosystems: Ecology and management of a high mountain resource. Gen. Tech. Rep. INT 270. Ogden, UT: U.S. Department of Agriculture, Forest Service, Intermountain Research Station: 302 304. Germino, M.J., W.K. Smith, and A. C. Resor. 2002. Conifer seedling distribution and survival in an alpine treeline ecotone. Plant Ecology 16 2: 157 168. Geils, B.W., K.E. Hummer, and R.S. Hunt. 2010. White pines, Ribes and blister rust: a review and synthesis. Forest Pathology 40: 147 185. Gernandt, D.S., G.G. LÂ—pez, S. O. GarcÂ’a, and A. Liston. 2005. Phylogeny and classification of Pinus Ta xon, 54: 29 42
73 Gibson, K., Skov, K., Kegley, S., Jorgensen, C., Smith, S. and Witcosky, J. 2008. Mountain pine beetle impacts in high elevation five needle pines: current trends and challenges. R1 08 020. Missoula, MT: U.S. Department of Agriculture Fores t Service, Forest Health Protection. Ginwal, H.S., S.S. Phartyal, P.S. Rawat, and R.L. Srivastava. 2005. Seed source variation in morphology, germination and seedling growth of Jatropha curcas Linn. in Central India. Silvae Genetica 54: 76 80. Government of Canada, 2012: Order amending Schedule 1 to the Species at Risk Act Canada Gazette Part II, Vol 146, No. 14, SOR/2012 113, June 20, 2012 [online]. Available from htt p://www.sararegistry.gc.ca/virtual_sara/files/orders/g2 14614i_e.pdf [accessed 22 February 2014]. Grubb, P.J. 1977. The maintenance of species richness in plant communities: the importance of the regeneration niche. Biological Reviews 52: 107 145. Guarig uata, M. R. and M.A. Pinard. 1998. Ecological knowledge of regeneration from seed in neotropical forest trees: Implications for natural forest manage ment. Forest Ecology and Management, 112, 87 99. G ssow, H.T. 1922.In British Columbia. Report of the Dominion Botanist. Ottawa, ON. Canadian Department of Agriculture, Dominion Experimental Farms. pp 7 9. Habeck, J R. 1969. A gradient analysis o f a timberline zone at Logan Pass, Glacier Park, Montana. Northwest Science. 43(2): 65 73. Hamann, A. and T. Wang. 2006. Potential effects of climate change on ecosystem and tree species distribution in British Columbia. Ecology 87: 2773 2786. Hampton, R.R., and Sherry, D.F. 1994. The effects of cache loss on ch oice of cache sites in the black capped chickadee. Behavioral Ecology 5: 44 50. Hiratsuka, Y.; Maruyama, P. J., 1976: Castilleja miniata: a new alternate host of Cronartium ribicola. Plant Disease Reports 60: 241. Hoff, R.J. and G.I. McDonald. 1980. Impr oving rust resistant strains of inland western white pine. Holtmeier, F., and G. Broll. 1992. The influence of tree islands and microtopography on pedoecological conditions in the forests alpine tundra ecotone on niwot ridge, Colorado front range, USA. Ar ctic and Alpine Research 24:216 228. Holtmeier, F K. 2009. Mountain Timberlines: Ecology, Patchiness, and Dynamics. 2nd edn. Springer Science + Business Media B.V. Howe, H.F. and J. Smallwood. 1982. Ecology of seed dispersal. Annual Review of Ecology and Systematics 13: 210 228.
74 Hulme, P.E. 1996. Natural regeneration of Yew ( Taxus baccata L.): microsite, seed or herbivore limitation. Journal of Ecology 84(6): 853 861. Hunt, R.S. 2009. History of western white pine and blister rust in British Columbi a. The Forestry Chronicle 85(4): 516 520. Hutchins, H.E. and R.M. Lanner. 1982. The central role of Clark s nutcracker in the dispersal and establishment of whitebark pine. Oecologia 55: 192 201. (IPCC) International Panel on Climate Change. 2013. Summ ary for Policymakers. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Stocker, T.F., D. Qin, G. K. Plattner, M. Tignor, S.K. Allen, J. Bos chung, A. Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA. Janzen, D.H. 1971. Seed predation by animals. Annual Review of Ecology and Systematics 2:465 492. Joy, E.L. 1939. His tory of blister rust infections in the west. Type written, USDA, Forest Service, Spokane, Washington Keane, R.E. 2000. The importance of Wilderness to whitebark pine research and management. In: McCool, Stephen F.; Cole, David N.; Borrie, William T.; O Lo ughlin, Jennifer, comps. 2000. Wilderness science in a time of change conference Volume 3: Wilderness as a place for scientific inquiry; 1999 May 23 27; Missoula, MT. Proceedings RMRS P 15 VOL 3. Ogden, UT: U.S. Depart ment of Agriculture, Forest Service Rocky Mountain Research Station. Keane, R.E. 2001. Successional dynamics: modeling an anthropogenic threat. In: Whitebark Pine Communities: Ecology and Restoration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 159 192. Island Press, Washington DC, USA. Keane, R.E. and A.W. Schoettle. 2011. Strategies, tools, and challenges for sustaining and restoring high elevation five needle white pine forests in western North America. In: Keane, R.E., D.F.Tomback, M.P. Murray, and C.M. Smith, eds. 2011. The fut ure of high elevation five needle white pines in Western North America: Proceedings of the High Five Symposium. 28 30 June 2010; Missoula, Montana. Proceedings RMRS P 63. Fort Collins, Colorado: U.S. Department of Agriculture, Forest Service, Rocky Mountai n Research Station. 376 p. Online at http://fs.fed/us/rm/pubs/rmrs_p063.html Keane, R.E., D.F. Tomback, C.A. Aubry, A.D. Bower, E.M. Campbell, C.L. Cripps, M.B. Jenkins, M.F. Mahalovich, M. Manning, S .T. McKinney, M.P. Murray, D.L. Perkins, D.P. Reinhart, C. Ryan, A.W. Schoettle, C.M. Smith. 2012. A range wide restoration strategy for whitebark pine ( Pinus albicaulis ). Gen. Tech. Rep. RMRS GTR 279. Fort Collins, CO: U.S. Department of Agriculture, Fore st Service, Rocky Mountain Research Station. 108 p.
75 Kollmann, J., J.P. Prado Cordova, and R.M. Andersen. 2008. Factors limiting regeneration of an endangered conifer in the highlands of Guatemala. Journal for Nature Conservation 16: 146 156. Korner, C. 20 12. Alpine Treelines: Functional ecology of the global high elevation tree limits. Sp ringer, Philadelphia, Pennsylva nia, USA. Krugman, S.L. and J.L. Jenkinson. 1974. Pinus L. pine. Pages 598 638 in: Schopmeyer, C.S., technical coordinator. Seeds of woody p lants in the United States. USDA Forest Service Agriculture Handbook No. 450, Washington, DC, USA. Kutner, M., C. Nachtsheim, J. Neter, and W. Li. 2005. Applied Linear Statistical Models. 5 th Ed. McGraw Hill, New York, New York, USA. Kral, R. 1993. Pinus albicaulis. Flora of North America north of Mexico. Vol. 22. Oxford University Press, New York, New York, USA. Also at website http://www.efloras.org/florataxon.aspx?flora_id=1 &taxon_id=10726 [accessed 6 February 2014]. Lanner, R.M. 1982. Adaptations of whitebark pine for seed dispersal by Clark's nutcracker. Canadian Journal of Forest Research 12: 391 402. Leak, W.B. and R.E. Graber. 1976. Seedling input, death and growth i n uneven aged northern hardwoods. Canadian Journal of Forest Research 6: 368 374. Leaver, L.A., L. Hopewell, C. Caldwell, and L. Mallarky. 2007. Audience effects on food caching in grey squirrels ( Sciurus carolinensis ): evidence of pilferage avoidance st rategies. Animal Cognition 10: 23 27. Levitt, J. 1972. Responses of Plants to Environmental Stresses. Academic, New York, New York. Lenoir, J., J.C. GÂŽgout, P.A. Marquet, P. de Ruffray, and H. Brisse. 2008. A significant upward shift in plant species opti mum elevation during the 20th century. Science 320: 1768:1771. Liston, A., W.A. Robinson, D. Pi Â– ero, and E.R. Alvarez Buylla. 1999. Phylogenetics of Pinus (Pinaceae) based on nuclear ribosomal DNA internal transcribed spacer region sequences. Molecular Ph yogenetics and Evolution 11: 95 109. Little, E. L., Jr. and W. B. Critchfield. 1969. Subdivisions of the genus Pinus (pines). USDA. Misc. Publ. 1144. 51 pp. Loehman, R., and G. Anderson. 2009. Understanding the science of climate change: talking points impacts to Western Mountains and Forests. Natural Resource Report NPS/NRPC/NRR 2009/090. National Park Service, Fort Collins, Colorado.
76 Logan, J.A. and J.A. Powell. 2001. Ghost forests, global warming and the mountain pine beetle (Coleoptera: Scolytidae). American Entomologist 47: 160 173. Logan, J.A., W.W. MacFarlane, and L. Wilcox. 2010. Whitebark pine vulnerability to climate driven mountain pine beetle disturbance in the Greater Yellowstone Ecosystem. Ecological Applications, 20: 895 902. Longland, W.S and Clements, C. 1995. Use of fluorescent pigments in studies of seed caching by rodents. Journal of Mammology 76(4): 1260 1266. Lorenz, T. J., C. Aubry, and S. Robin. 2008. A review of the literature on seed fate in whitebark pine and the life history traits of Clark's nutcracker and pine squirrels. Gen. Tech. Rep. PNW GTR 742. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 62 p. Lorenz, T.J., K.A. Sullivan, A.V. Bakian, and C.A. Aubry. 2011. Cache site selection in Clark's nutcracker ( Nucifraga columbiana ). The Auk 128: 237 247. Mahalovich, M.F. and G.A. Dickerson. 2004. Whitebark pine genetic restoration program for the Intermountain West (United States). In: Sniezko, R.A., S. Samman, S.E. Schlarbaum and H.B. Kriebel, eds. 2004. Breeding and genetic resources of five needle pines: growth, adaptability and pest resistance; 23 27 July 2001; Medford, OR, USA. IUFRO Working Party 2.02.15. Proceedings RMRS P 32, USDA Foreset Service, Rocky Mountain Resear ch Station, Fort Collins, CO, USA. Mahalovich, M.F., K.E. Burr, and D.L. Foushee. 2006. Resistance, and cold hardiness among seed sources in the inland Northwest: planting strategies for restoration. In: Riley, L.E.; Dumroese, R.K.; Landis, T.D., tech. coo rds. 2006. National Proceedings: Forest and Conservation Nursery Associations 2005 Proc. RMRS P 43. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 160 p. Mahalovich, M.F. and V.D. Hipkins. 2011. Molecu lar genetic variation in whitebark pine ( Pinus albicaulis Engelm.) in the Inland West. In: Keane, R.E., Tomback, D.F., Murray, M.P. and Smith, C. M., eds. 2011. The future of high elevation, five needle white pines in Western North America: Proceedings of the High Five Symposium. 28 30 June 2010; Missoula, MT. Proceedings RMRS P 63. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 376 p. Online at http://www.fs.fed.us/rm/pubs/rmrs_p063.html. Maher, E.L. and M.J. Germino. 2006. Microsite differentiation among conifer species during seedling establishment at alpine treeline. Ecoscience 13: 334 341. Malanson, G.P., D.R. Butler, D.B. Fagre, S.J. Walsh, D.F. Tomback, L.D. Daniels, L.M. Resler, W.K. Smith, D.J. W eiss, D.L. Peterson, A.G. Bunn, C.A. Hiemstra, D. Liptzin, P.S. Bourgeron, Z. Shen, and C.I. Millar. 2007. Alpine treeline of Western North America: linking organism to landscape dynamics. Physical Geography 28: 378 396.
77 Man, G. 2010. Major forest insect and disease conditions in the United States: 2009 update. U.S. Department of Agriculture, Forest Health Protection: FS 952. Macfarlane, W.W., J.A. Logan, and W.R. Kern. 2013. An aerial survey assessment of mountain pine beetle caused whitebark pine mortali ty in the Greater Yellowstone Ecosystem. Ecological Applications .23: 421 437. Mattson, D.J., and B.M. Blanchard, B.M.1994. Bear use of whitebark pine seeds in North America. U.S. Forest Service General Technical Report. INT GTR 309. pp. 212 220. Marr, J. W 1977. The development and movement of tree islands near the upper limit of tree growth in the southern Rocky Mountains. Ecology 58:1159 1164. McCaughey, W. W. 1990. Biotic and microsite factors affecting Pinus albicaulis establishment and survival. Doct oral dissertation, Montana State University, Bozeman, Montana. McCaughey, W.W. 1994. The regeneration process of whitebark pine. In: W.C. Schmidt and F. K. Holtmeier, compilers. Proceedings International workshop on subalpine stone pines and their environ ment: the status of our knowledge. USDA Forest Service, Intermountain Research Station, General Technical Report INT 209, Ogden, Utah. McCaughey, W.W. and W. C. Schmidt. 2001. Taxonomy, Distribution, and History. Whitebark pine communities: ecology and res toration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 243 262. Island Press, Washington DC. McCaughey, W.W. and D.F. Tomback. 2001. The natural regeneration process. In: Whitebark Pine Communities: Ecology and Restoration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 105 120. Island Press, Washington DC, USA. McCaughey, W.W. and W.C. Schmidt. 2001. Taxonomy, distribution, and history. In: Whitebark Pine Communities: Ecology and Restoration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 41 73. Island Press, Washington DC, USA. McDonald, G.I., R.J. Hoff. 2001. Whitebark pine communities: ecology and restoration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 3 25. Island Press, Washington DC. McDonald, G.I., B.A. Richardson, P.J. Zamb ino, N.B. Klopfenstein, and M.S. Kim. 2006. Pedicularis and Castilleja are natural hosts of Cronartium ribicola in North America: a first report. Forest Pathology 36: 73 82. McKinney, S.T. and D.F. Tomback. 2007. The influence of white pine blister rust o n seed dispersal in whitebark pine. Canadian Journal of Forest Research 37: 1044 1057. McKinney, S.T. C.E. Fiedler, and D.F. Tomback. 2009. Invasive pathogen threatens bird pine mutualism: implications for sustaining a high elevation ecosystem. Ecological Applications 19: 597 607.
78 McKinney, S. T., and D. F. Tomback. 2011. Altered community dynamics in Rocky Mountain whitebark pine forests and the potential for accelerating declines. Pages 45 78 in: Mountain Ecosystems: Dynamics, Management and Conservati on K.E. Richards, ed. Nova Science Publishers, Inc., Hauppauge, N.Y. McLane, S.C., and S.N. Aitken. 2012. Whitebark pine ( Pinus albicaulis ) assisted migration potential: testin establishment north of the species range. Ecological Applications 22:142 153. Mellmann Brown, S., 2005: Regeneration of whitebark pine in the timberline ecotone of the Beartooth Plateau, U.S.A.: Spatial distribution and responsible agents In Broll, G., and Keplin, B. (eds.), Mountain Ecosystems: Studies in Treeline Ecology. Berlin and Heidelberg: Springer Verlag, 97 115. Mielke, J.L. 1943. White pine blister rust in western North America. Bulletin 52. New Haven, CT. Yale University, School of Forestry. 155p. Mirov, N.T. The genus Pinus The Ronald Press. New York, New York, USA. Mu Â– oz, A. and R. Bonal. 2011. Linking seed dispersal to cache protection strategies. Journal of Ecology 99: 1016 1025. Nathan, R., G.G. Katul, H.S. Horn, S.M. Thomas, R. Oren, R. Avissar, S. W. Pacala, and S. A. Levin. 2002. Nature 418: 409 413. Ogilvie, R. T. 1990. Distribution and ecology of whitebark pine in western Canada. Pages 54 60 in: Schmidt, W. C.; McDonald, K. J., compilers. Proceedings symposium on whitebark pine ecosystems: ecology and management of a high mountain resource. General Technical Report INT 270, USDA Forest Service, Intermountain Research Station, Ogden, UT, USA. Paine, R.T. 1969. A note on trophic compleity and community stability. The American Naturalist 103(929): 91 93. Patton, R. F.; Spear, R. N., 1989: Histopathology of colon ization in leaf tissue of Castilleja Pedicularis Phaseolus and Ribes species by Cronartium ribicola Phytopathology, 79: 539 547. Perkins, D.L., Swetnam, T.W., 1 996. A dendroecological assess ment of whitebark pine in the Sawtooth Salmon River region Ida ho. Canadian Journal of Forest Research 26: 2123 2133. Powell, J.A. and J.A. Logan. 2005. Insect seasonality: circle map analysis of temperature driven life cycles. Theoretical Population Biology 67: 161 179. Preston, S.D. and L.F. Jacobs. 2005. Cache d ecision making: the effects of competition on cache decisions in Merriam s kangaroo rat ( Dipodomys merriami ). Journal of Comparative Psychology 119:187 196.
79 Price, R. A., A. Liston, and S. H. Strauss. 1998. Phylogeny and systematics of Pinus Ecology and biogeography of Pinus. (ed. by D. M. Richardson). pp. 49 68. Cambridge University Press, New York, USA. Pyatt, J.C., D.F. Tomback, S.C. Blakeslee, M.B. Wunder, and L.M. Resler. in prep. Whitebark pine ( Pinus albicaulis ) as a tree island initiator in the a lpine treeline ecotone: comparing microclimate and microsite quality among common nurse objects. Ray, P. J. and J.M. Alcantara. 2000. Recruitment dynamics of a fleshy fruited plant ( Olea europaea ): connecting patterns of seed dispersal to seedling establis hment. Journal of Ecology 88(4): 622 633. Resler, L.M. and D.F. Tomback. 2008. Blister rust prevalence in krummholz whitebark pine: implications for treeline dynamics, Northern Rocky Mountains, Montana, U.S.A. Arctic Antarctic, and Alpine Research 40: 161 170. Ribbens, E., J.A. Silander, Jr., and S.W. Pacala. 1994. Seedling Recruitment in Forests: Calibrating Models to Predict Patterns of Tree Seedling Dispersion. Ecology 75: 1794 1806. Rivest, L. and S. Baillargeon. 2007. Rcapture: loglinear models fo r capture recapture in R. Journal of Statistical Software 19: 1 31. Rochefort, R.M. and D.L. Peterson. 1996. Temporal and spatial distribution of trees in subalpine meadows of Mount Rainier National Park, Washington, U.S.A. Arctic and Alpine Research 28( 1): 52 59. Safranyik, L. and A.L. Carroll. 2006. The biology and epidemiology of the mountain pine beetle in lodgepole pine forests. Pages 3 66 (Chapter 1) in L. Safranyik and W.R. Wilson, editors. The mountain pine beetle: a synthesis of biology, manag ement, and impacts on lodgepole pine. Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, Victoria, British Columbia. 304 p. Schoettle, A.W. and R.A. Sniezko. 2007. Proactive intervention to sustain high elevation pine ecosystems th reatened by white pine blister rust. Journal of Forest Research 12: 327 336. Schrag, A.M., A.G. Bunn, and L.J. Graumlich. 2007. Influence of bioclimatic variables on treeline conifer distribution in the Greater Yellowstone Ecosystem: implications for spec ies of conservation concern. Journal of Biogeography 35: 698 710. Scott, J. 2013. Clark's nutcracker occurrence, whitebark pine stand health, and cone production in the Waterton Glacier International Peace Park. M.S. thesis, University of Colorado Denver, Denver, Colorado. Schwandt, J.W., D.F. Tomback, R.E. Keane, W.W. McCaughey, and S.J. Kearns. 2007. First year results of a whitebark pine seed planting trial near Baker City, Or. USDA Forest Service R6 NR FHP 2007 01
80 Schwandt, J.W., I.B. Lockman, J.T. Kli ejunas, and J.A. Muir. 2010. Current health issues and management strategies for white pines in the Western United States and Canada. Forest Pathology 40: 226 250. Schwandt, J., K. Chadwick, H. Kearns, and C. Jensen. 2011. Whitebark pine direct seeding tr ials in the Pacific Northwest. In: Keane, R.E., D.F.Tomback, M.P. Murray, and C.M. Smith, eds. 2011. The future of high elevation five needle white pines in Western North America: Proceedings of the High Five Symposium. 28 30 June 2010; Missoula, Montana. Proceedings RMRS P 63. Fort Collins, Colorado: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 376 p. Online at http://fs.fed/us/rm/pubs/rmrs_p063.html. Schwandt, J. 2011. Highlights of the Forest Health Protection whitebar k pine restoration program. In: Keane, R.E., D.F.Tomback, M.P. Murray, and C.M. Smith, eds. 2011. The future of high elevation five needle white pines in Western North America: Proceedings of the High Five Symposium. 28 30 June 2010; Missoula, Montana. Pro ceedings RMRS P 63. Fort Collins, Colorado: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. 376 p. Online at http://fs.fed/us/rm/pubs/rmrs_p063.html Smith, C.C. and O .J. Reichman. 1984. The evolution of food caching by birds and mammals. Annual Review of Ecology Systematics 15: 329 351. Smith, C.M., B. Wilson, S. Rasheed, R.C. Walker, T. Carolin, and B. Shepard. 2008. Whitebark pine and white pine blister rust in th e Rocky Mountains of Canada and northern Montana. Canadian Journal of Forest Research, 38: 982 995. Smith, W.K., M.J. Germino, D.M. Johso, and K. Reinhardt. 2009. The altitude of alpine treeline: a bellwether of climate change effects. Botanical Review 75 : 163 190. Smith Mckenna, E.K., L.M. Resler, D.F. Tomback, H. Zhang, and G.P. Malanson. 2013. Topographic influences on the distribution of white pine blister rust in Pinus albicaulis treeline communitites. Bioscience 20: 215 229. Spaulding, P. 1929. Whit e pine blister rust: a comparison of European with North American conditions. USDA Technical Bulletin 87. Syring, J., K. Farrell, R. Businsk% R. Cronn, and A. Liston. 2007. Widespread genealogical nonmonophyly in species of Pinus subgenus Strobus Systema tic Biology 56: 163 181. Tillman Sutel, E., A. Kauppi, K. Karppinen, D.F. Tomback. 2007. Variant maturity in seed structures of Pinus albicaulis (Engelm.) and Pinus sibirica (DuTour): key to a soil seed bank, unusual among conifers? Trees 22: 225 236. Tomback, D.F. 1978. Foraging strategies of Clark s nutcracker. Living Bird 16: 123 161.
81 Tomback, D.F. 1980. How nutcrackers find their seed stores. The Condor 82(1): 10 19. Tomback, D.F. 1982. Dispersal of whitebark pine seeds by Clark s nutcracker: a m utualism hypothesis. Journal of Animal Ecology 51(2): 451 467. Tomback, D.F. 1986. Post fire regeneration of krummholz whitebark pine: a consequence of nutcracker seed caching. Madro Â– o 33: 100 110. Tomback, D.F. and Y.B. Linhart. 1990. The evolution of bird dispersed pines. Evolutionary Ecology 4(3): 185 219. Tomback, D.F., S.K. Sund, and L.A. Hoffmann. 1993. Post fire regeneration of Pinus albicaulis : height age relationships, age structure, and microsite characteristics. Canadian Journal of Forest Re search 23(2): 113 119. Tomback, D.F., A.J. Anderies, K.S. Carsey, M.L. Powell, and S. Mellmann Brown. 2001a. Delayed seed germination patterns following the Yellowstone fires. Ecology 82(9): 2587 2600. Tomback, D.F., Arno, S.F. and Keane, R.E. 2001b. The compelling case for management intervention. In: Whitebark Pine Communities: Ecology and Restoration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 3 25. Island Press, Washington DC, USA. Tomback, D.F. and K.K. Kendal. 2001c. Biodiversity losses: th e downward spiral. In: Whitebark Pine Communities: Ecology and Restoration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 243 262. Island Press, Washington DC, USA. Tomback, D.F. 2005. The impact of seed dispersal by Clark's nutcracker on whitebark p ine: multi scale perspective on a high mountain mutualism. In: Broll, G. and B. Keplin, eds. 2005. Mountain Timberlines: studies in treeline ecology. Springer, New York, New York, USA. Tomback, D.F., A.W. Schoettle, K.E. Chevalier, and C.A. Jones. 2005 Lif e on the edge for limber pine: seed dispersal within a peripheral population. Ecoscience 12: 519 529. Tomback, D.F. and L.M. Resler. 2007. Invasive pathogens at alpine treeline: consequences for treeline dynamics. Physical Geography 28: 397 418. Tomback, D.F. and P. Achuff. 2010. Blister rust and western forest biodiversity: ecology, values and outlook for white pines. Forest Pathology 40: 186 225. Tomback, D.F., P. Achuff, A.W. Schoettle, J.W. Schwandt, and R.J. Mastrogiuseppe. 2011. The magnificent hig h elevation five needle white pines: ecological roles and future outlook. In: Keane, R.E., D.F.Tomback, M.P. Murray, and C.M. Smith, eds. 2011. The future of high elevation five needle white pines in Western North America: Proceedings of the High Five Symp osium. 28 30 June 2010; Missoula, Montana. Proceedings RMRS P 63. Fort Collins, Colorado: U.S. Department of Agriculture,
82 Forest Service, Rocky Mountain Research Station. 376 p. Online at http://fs.fed /us/rm/pubs/rmrs_p063.html Tomback, D.F., K.G. Chipman, L.M. Resler, E.K. Smith McKenna, and C.M. Smith. 2014. Relative abundance and functional role of whitebark pine at treeline in the Northern Rocky Mountains. Arctic, Antarctic, and Alpine Research 4 6: 407 418. (USFS) United States Forest Service. 2003. America's forests: 2003 health update. US Department of Agriculture. AIB 773. (USFWS) United States Fish and Wildlife Service. 2011. Endangered and threatened wildlife and plants; 12 month finding on a petition to list Pinus albicaulis as endangered or threatened with critical habitat. Department of the Interior (US). 50 CFR Part 17. Vander Wall, S.B. and R.P. Balda. 1977. Coadaptations of the Clark's nutcracker and the Pinon pine for efficient seed har vest and dispersal. Ecological Monographs 47: 89 111. Vander Wall, S.B. 1988. Foraging of Clark's nutcrackers on rapidly changing pine seed resources. The Condor 90: 621 631. Vander Wall, S.B. 1995. Influence of Substrate Water on the Ability of Rodents to Find Buried Seeds. Journal of Mammalogy 76: 851 856. Vander Wall, S.B. 1998. Foraging success of granivorous rodents: effects of variation in seed and soil water on olfaction. Ecology 79 (3) : 233 241. Vander Wall, S.B. 2002. Masting in animal disperse d pines facilitates seed dispersal. Ecology 83(12): 3509 3516. Vander Wall, S.B. 2003. Effects of seed size of wind dispersed pines ( Pinus ) on secondary seed dispersal and the caching behavior of rodents. Oikos 100: 25 34. Vander Wall, S.B. and W.S. Lo ngland. 2005. Diplochory and the evolution of seed dispersal. In: Forget, P. M., J.E. Lambert, P.E. Hulme, and S.B. Vander Wall. CABI Publishing, Cambridge, MA, USA. pp 297 314. Vander Wall, S.B., P.M. Forget, J.E. Lambert, and P.E. Hulme. 2005. Seed fate pathways: filling the gap between parent and offspring. In: Forget, P. M., J.E. Lambert, P.E. Hulme, and S.B. Vander Wall. CABI Publishing, Cambridge, MA, USA. pp 1 8. Vander Wall, S.B., K.M. Kuhn, and M.J. Beck. 2005. Seed removal, seed predation, and sec ondary dispersal. Ecology 86: 801 806. Vander Wall. 2010. How plants manipulate the scatter hoarding behavior of seed dispersing animals. Philosophical Transactions of the Royal Society of Biological Sciences 365: 989 997.
83 Weaver T 2001. Whitebark pine and its environment. In: Whitebark Pine Communities: Ecology and Restoration (ed. by D.F. Tomback, S.F. Arno and R.E. Keane), pp. 41 73. Island Press, Washington DC, USA. (WPEF) Whitebark Pine Ecosystem Foundation. 2014. < http://whitebarkfound.org/wp co ntent/uploads/2014/05/Whitebark pine range colour 2014.jpg > Wenny, D.G. 1999. Two staged dispersal of Guarea glabra and G. kunthiana (Meliaceae) in Monteverde, Costa Rica. Journal of Tropical Ecology 15: 481 496. Zambino, P.J., B.A. Richardson, G.I. Mc Donald, N.B. Klopfenstein, and M. S. Kim. 2005. A paradigm shift for white pine blister rust: Non Ribes alternate hosts of Cronartium ribicola in North America. In: Guyon, J.C. comp. 2006. Proceedings of the 53rd Western International Forest Disease Work C onference; 2005 August 26 29; Jackson, WY. USDA Forest Service, Intermountain Region Ogden, Utah. Zambino, P.J., B.A. Richardson, and G.I. McDonald. 2007. First report of the white pine blister rust fungus, Cronartium ribicola on Pedicularis bracteosa. P lant Disease : 91:467.