UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Establishment and growth responses of whitebark and lodgepole pine populations in a changing climate McLane, Sierra C. 2011

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

Item Metadata

Download

Media
24-ubc_2011_spring_mclane_sierra.pdf [ 2.7MB ]
Metadata
JSON: 24-1.0071779.json
JSON-LD: 24-1.0071779-ld.json
RDF/XML (Pretty): 24-1.0071779-rdf.xml
RDF/JSON: 24-1.0071779-rdf.json
Turtle: 24-1.0071779-turtle.txt
N-Triples: 24-1.0071779-rdf-ntriples.txt
Original Record: 24-1.0071779-source.json
Full Text
24-1.0071779-fulltext.txt
Citation
24-1.0071779.ris

Full Text

Establishment and growth responses of whitebark and lodgepole pine populations in a changing climate by Sierra C. McLane  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in The Faculty of Graduate Studies (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2011 © Sierra C. McLane 2011  Abstract Climate change will affect the regeneration, growth, survival and distribution of trees. Here, I use common gardens to empirically test establishment, growth and the potential for persistence, adaptation and migration for two iconic North American trees, whitebark pine (Pinus albicaulis) and lodgepole pine (Pinus contorta ssp. latifolia). Whitebark pine is of conservation concern due to range-wide diebacks, while lodgepole pine is critical to forest productivity and carbon sequestration. Whitebark seeds were planted north of the current range in areas predicted to be climatically suitable through the 2050s; these germinated and survived in varying proportions at all locations. Establishment and growth were positively affected by moderate snow-cover durations, heavier seed weights, and warmer provenance temperatures. Whitebark pine seedlings grown from seeds sown in growth chambers spanning current and predicted-future temperatures demonstrated positive responses to warmer growing seasons. Lodgepole pine seedlings in the same chambers outgrew the whitebark pine seedlings at all but the coldest temperatures. Together, these results suggest that whitebark pine may lose its competitive advantage to other species within its narrow alpine-treeline niche as the climate warms, but that it is capable of establishing in climatically-suitable areas north of its current range. Using tree-ring data from long-term lodgepole pine common garden trials, I built universal growth-trend response functions to forecast future growth trends relative to genetics, climate and tree age. The models predict growth reductions for all populations by the end of the 21st century based on middle-ofthe-road climate models, except in far northern areas near and within Yukon, Canada. Analogous models built using summer and winter climate indices indicate that the growth declines are primarily caused by warmer summers, and may be offset by growth increases resulting from warmer winters. I found that populations are most sensitive to annual temperatures and summer aridity, but that sensitivity to climate varies due to local adaptation. Overall, my research will help forest professionals and conservationists forecast changes in forest productivity and species growth and survival under warming temperatures.  ii  Preface This thesis was written as a series of manuscripts with the intent of publication in peerreviewed journals. I took the lead in developing the ideas, collecting and analyzing the data, and producing the manuscripts, figures, and tables for all chapters of this thesis. However, I could not have succeeded without the assistance of numerous individuals. Sally Aitken helped inspire, implement, and finance all of the studies. Sally and my committee members, Valerie LeMay, Lori Daniels and Alvin Yanchuk, provided guidance on the field components and data analyses throughout, and many research assistants helped with the field work. The British Columbia Ministry of Forests, Mines and Lands, the United States Department of Agriculture Forest Service, and Whistler Blackcomb Ski Resort granted me permission to collect seeds, extract tree cores, and create common gardens on lands that they manage. Valerie LeMay advised the design of the models used in Chapters 2 and 3 and was instrumental in helping create the random coefficient model used in Chapter 4. As such, she is included as a co-author on the manuscript derived from Chapter 4. Lori Daniels provided background on the dendrochronology techniques used in Chapter 5, and is a co-author on the corresponding manuscript. Tongli Wang created the species distribution model that I used in Chapter 2 and inspired my use of universal response functions in Chapters 4 and 5. Pia Smets grew the lodgepole pine seedlings used in Chapter 3 and assisted with programming the growth chambers. Sally, Valerie, Lori, Alvin and Tongli, as well as Isla Myers-Smith, Carmen Wong, Andy Bower and anonymous individuals reviewed one or more of the manuscripts derived from the chapters in this thesis. Versions of chapters 2, 4 and 5 are submitted or in press. Chapter 2: McLane S. C., S. N. Aitken. Under Review. Whitebark pine (Pinus albicaulis) assisted migration potential: testing establishment north of the species range. Ecological Applications. Chapter 4: McLane S. C., V. M. LeMay, and S. N. Aitken. In Press. Modeling lodgepole pine radial growth relative to climate and genetics using universal growth-trend response functions. Ecological Applications. Chapter 5: McLane, S. C., L. D. Daniels, and S. N. Aitken. In Press. Climate impacts on lodgepole pine (Pinus contorta) radial growth in a provenance experiment. Forest Ecology and Management.  iii  Table of contents Abstract ......................................................................................................................................... ii Preface .......................................................................................................................................... iii Table of contents ......................................................................................................................... iv List of tables.................................................................................................................................. vi List of figures ............................................................................................................................... vii Acknowledgements ................................................................................................................... viii Chapter 1: Literature review and research objectives .............................................................. 1 1.1 Introduction ................................................................................................................. 1 1.2 Climate change in British Columbia ............................................................................. 1 1.3 Tree responses to climate change ................................................................................. 2 1.4 Phenotypic plasticity ..................................................................................................... 2 1.5 Migration....................................................................................................................... 5 1.6 Whitebark pine: A keystone species of conservation concern...................................... 8 1.7 Lodgepole pine: Important to carbon sequestration and forest productivity .............. 10 1.8 Research objectives ..................................................................................................... 11 Chapter 2: Whitebark pine’s (Pinus albicaulis) assisted migration potential: testing establishment north of the species range .................................................................................. 13 2.1 Introduction ................................................................................................................. 13 2.2 Methods....................................................................................................................... 15 2.3 Results ......................................................................................................................... 19 2.4 Discussion ................................................................................................................... 20 2.5 Conclusions ................................................................................................................. 25 Chapter 3: Whitebark pine (Pinus albicaulis) germination and growth at extreme temperatures ................................................................................................................................ 32 3.1 Introduction ................................................................................................................. 32 3.2 Methods....................................................................................................................... 33 3.3 Results ......................................................................................................................... 37 3.4 Discussion ................................................................................................................... 38 3.5 Conclusions ................................................................................................................. 41 Chapter 4: Modeling lodgepole pine (Pinus contorta) radial growth relative to climate and genetics using universal growth-trend response functions...................................................... 46 4.1 Introduction ................................................................................................................ 46 4.2 Methods....................................................................................................................... 48 4.3 Results ......................................................................................................................... 52 4.4 Discussion ................................................................................................................... 54 4.5 Conclusions ................................................................................................................. 58 Chapter 5: Climate impacts on lodgepole pine (Pinus contorta) radial growth in a provenance experiment .............................................................................................................. 70 5.1 Introduction ................................................................................................................ 70 5.2 Methods....................................................................................................................... 71 5.3 Results ......................................................................................................................... 76 iv  5.4 Discussion ................................................................................................................... 77 5.5 Conclusions ................................................................................................................. 80 Chapter 6: Conclusions .............................................................................................................. 88 6.1 Introduction ................................................................................................................. 88 6.2 Whitebark pine may require assisted migration.......................................................... 88 6.3 Lodgepole pine productivity may decline except in the far north ............................. 91 6.4 Limitations of present work ....................................................................................... 92 6.5 Future research ............................................................................................................ 94 Literature cited............................................................................................................................ 98 Appendices ................................................................................................................................. 108 Appendix 1 Methods for creating whitebark pine species distribution models .............. 108 Appendix 2 Whitebark pine germination, survival and growth data summary .............. 110 Appendix 3 Whitebark and lodgepole pine provenance data ......................................... 111 Appendix 4 Whitebark pine germination and survival data ........................................... 112 Appendix 5 Lodgepole pine chronology statistics .......................................................... 113  v  List of tables Table 2.1 Geographic and climatic variables for whitebark pine sites and populations .............. 26 Table 2.2 Regressions predicting whitebark pine quantitative traits ........................................... 27 Table 3.1 Regressions predicting whitebark pine quantitative traits by chamber ........................ 42 Table 3.2 Statistical tests for whitebark pine quantitative traits and genetics parameters ........... 43 Table 4.1 Lodgepole pine trial sites and populations .................................................................. 60 Table 4.2 Number of lodgepole pine trees per population sampled at each site .......................... 61 Table 4.3 Site and provenance temperature and precipitation data.............................................. 62 Table 4.4 Predicted maximum basal area increments for lodgepole pine .................................... 63 Table 4.5 Predicted ages at which basal area increments are maximized for lodgepole pine ..... 65 Table 5.1 Lodgepole pine trial sites, populations and bioclimatic regions .................................. 82 Table 5.2 Regressions predicting lodgepole pine ring width and sensitivity ............................... 83 Table 5.3 Correlations between residual ring widths and annual climate fluctuations ................ 84 Table A.1 Whitebark pine germination, survival and growth data summary ............................ 110 Table A.2 Whitebark and lodgepole pine provenance data ....................................................... 111 Table A.3 Whitebark pine germination and survival data ......................................................... 112 Table A.4 Lodgepole pine chronology statistics ........................................................................ 113  vi  List of figures Figure 2.1 Whitebark pine present and future species distribution models ................................. 28 Figure 2.2 Whitebark pine trial locations and populations .......................................................... 29 Figure 2.3 Whitebark pine germination, survival and mortality by site and seed treatment ....... 30 Figure 2.4 Whitebark pine germination and survival by seed weight and snowmelt timing ....... 31 Figure 3.1 Whitebark pine provenance locations......................................................................... 44 Figure 3.2 Quantitative traits for whitebark pine seedlings in growth chambers ........................ 45 Figure 4.1 Lodgepole pine trial sites and populations ................................................................. 67 Figure 4.2 Lodgepole pine basal area increments relative to tree age ........................................ 68 Figure 4.3 Universal growth-trend response functions for lodgepole pine ................................. 69 Figure 5.1 Lodgepole pine trial sites and populations ................................................................. 85 Figure 5.2 Lodgepole pine ring widths by normal site and provenance temperature .................. 86 Figure 5.3 Lodgepole pine sensitivity by normal site and provenance temperature.................... 87  vii  Acknowledgements Foremost, I acknowledge my supervisor, Sally Aitken, for providing knowledge of forest systems, enthusiasm for my ideas, critical feedback and friendship throughout. I am also grateful for the wisdom and teaching of my committee member, Valerie LeMay. Val helped me analyze complex data sets in ways I could not have conceptualized otherwise, and persisted in contributing time and ideas until we got the models right. My committee members Lori Daniels and Alvin Yanchuk were also unflagging supporters, Lori in teaching me dendrochronology techniques and Alvin in helping me understand the fundamentals of forest genetics. I thank Tongli Wang for his inspiration in analyzing old provenance data in new ways, for creating the species distribution model that inspired my whitebark pine projects, and for contributing climate data for northwestern North America to me and countless others. Thanks also to Pia Smets, who assisted with my growth chamber experiment and coordinated logistics that enabled all of my projects, and to Norm Hodges for resuscitating my computer when necessary. I am grateful to the current and former members of the Aitken lab, with whom I had many fruitful discussions throughout the years. Particular thanks to Andy Bower, whose research on whitebark and lodgepole pine directly informed my projects. Andy, as well as Isla Myers-Smith, Carmen Wong, and anonymous reviewers provided detailed manuscript reviews that proved invaluable in shaping my final research stories. I had the pleasure of working with many wonderful research assistants and colleagues in the field and in the lab during my degree, including Sarah Braun Wildeman, Charlotte Whitney, Christine Chourmouzis, Lisa Erdle, Nicholas Ukrainetz, Liz Poulsom, Isabelle Behret, Olivia Freeman, Magdalena Moczulski, Jason “Buck” Buchwald, Tiffany Foster, Jesse Wildeman and Trevor Jones. I am also grateful for the inspiration, assistance and support of the following people and organizations: Dave Kolotelo, Alvin Yanchuk and Peter Ott of the BC Ministry of Forests, Mines and Lands, Joanne Vinnedge of the BC Ministry of Environment, Don Pigott of Yellow Point Propagation, Guido Schnelzer of Last Frontier Heli-skiing, Steve Brushey and Scotty Aitken of the BC Ministry of Transportation, Arthur DeJong of Whistler Blackcomb Ski Resort, Andy Bower, Lee Riley, Carol Aubry and Rich Sniezko of the USDA Forest Service, and the pilots and base managers of Highland Helicopters, Canadian Helicopters and Lakelse Air. Finally, many thanks to the following individuals and families for keeping me warm, fed, happy and inspired at home, at UBC and in the field: Denny Capps, Emily Gonzales, Trevor Lantz, Peter Arcese, Jamie Leathem, Ian Dallmeyer, Athena McKown, Nina Lobo, Virginie Pointeau, Tanya Stickford, the Vinnedge family, Sybille Haeussler, Dave Coates, Ben Heemskirk, Leanne Helkenberg, Bill and Betty Geier, Ron and Lila from Triple Creek Ranch, Al and Irene Whitney and Jack Woods. Funding was provided by the Forest Investment Account through the BC Forest Genetics Council to the Centre for Forest Conservation Genetics (SN Aitken), as well as by University of British Columbia Graduate Scholarships (SC McLane).  viii  Chapter 1: Literature review and research objectives 1.1 Introduction Climatic warming that began during the twentieth century is projected to continue beyond the fossil fuel era, provoking concerns regarding how forests will respond (Walther et al. 2002; Root et al. 2003). Climate is the primary factor impacting the range and distribution of tree species, and it is the principal driving force behind forest macroevolution and migration (Whitlock and Bartlein 1997). As such, warming temperatures and changing precipitation regimes will cause changes to tree germination, growth, phenology, distribution and survival. Climate change is expected to impact populations and species differently depending on genetics, growing location, and the magnitude and rate of the climatic shift. Forests affect hydrological patterns, atmospheric carbon and other gasses, animals and plants dependent on forest habitats, and human resources including timber products. Consequently, understanding how trees may be impacted by climate change is of major relevance.  1.2 Climate change in British Columbia Among the most prominent statements made by the Intergovernmental Panel on Climate Change (IPCC) in its 4th Assessment Report are that global temperatures rose 0.74 °C between 1905 and 2006, that current carbon dioxide and methane concentrations “exceed by far” the natural range of the last 650,000 years, and that 11 of the 12 years between 1995 and 2006 were among the warmest since 1850 (IPCC 2007). High northern latitudes are experiencing the most dramatic climatic changes, with Arctic temperatures increasing at twice the global average rate over the last 100 years. Climate model simulations comparing climate trends based on natural forcing (volcanoes and solar flares) alone to natural and anthropogenic forcing combined leave no question that human activities have caused the majority of the warming that we have experienced over the last century (IPCC 2007). Coastal British Columbia (BC), Canada, has warmed 0.6 °C, interior BC by 1.1 °C, and northwestern BC by 1.7 °C since the late 1800s (Ministry of Water, Land and Air Protection 2002). The 1990s were the warmest decade in the last 1,000 years, based on a variety of tree ring and ice core chronologies, and the early years of the 21st century show similar recordbreaking temperature trends. Precipitation records for the southern part of the province show increases of 2% per decade along the coast and in the central interior, 3% per decade in the 1  southern interior and 4% per decade in the southern Rocky Mountains over the last hundred years. Precipitation is estimated to have increased by almost 2% per decade in the northern part of the province as well. In western North America, annual temperatures are predicted to increase ~3.5 °C during the 21st century in southern and central areas, and ~4.5 °C in northern areas (Christensen et al. 2007). Summer temperatures are predicted to increase by ~3 °C throughout this region. However, the greatest warming is expected to occur during winter in northern locations, with predictions of up to 10 °C temperature increases in the northernmost parts of Canada and Alaska, USA by 2100. During the same time period, precipitation is predicted to increase by ~5% in southern and central parts of western North America, and by ~20% in the far northwest.  1.3 Tree responses to climate change Tree populations may respond to rapid climate change in four ways: 1) persist through phenotypic plasticity, by which populations withstand new climatic conditions in situ by altering phenotypes without adaptive changes to their genotypes; 2) adapt through microevolution, by which progeny display novel genotypes that have higher fitness under the new climate conditions; 3) migrate, by which seeds disperse to new environments with suitable climates; or 4) become extirpated (Aitken et al. 2008; Visser 2008). Phenotypic plasticity and migration are both impacted by genetic adaptation and consequently cannot be considered in its absence (Davis et al. 2005). However, adaptation occurs slowly for trees with long generation times, and thus is not expected to impact plasticity or migration substantially in the 100+ years that it will take for the climate to change markedly (Skelly et al. 2007). In contrast, trees are hallmarks of phenotypic plasticity, often weathering long periods of suboptimal climatic conditions over their century or multi-century lifetimes. Natural migration is unlikely to occur quickly for trees, but locations with suitable climates under future scenarios are readily identifiable and may be used to assist the migration of some tree species; therefore migration is considered here as well.  1.4 Phenotypic plasticity Phenotypic plasticity refers to the phenotypic expressions of a genotype under different environmental conditions. Quantifiable phenotypic traits include germination, survival, morphological features such as height, diameter and biomass, phenological cycles such as timing of needle flush and bud set, and reproductive indices such as fecundity and seed quality. In this 2  dissertation, I focused primarily on growth, due to its importance for forest productivity and carbon sequestration rates and its widespread use as a fitness indicator. Woody plants experience two distinct forms of growth within their stem and root systems: primary growth and secondary growth. The former affects elongation and is typically characterized by tree height, while the latter affects radial expansion and is generally measured as diameter at breast height. Both are impacted by climate, genetics and stand density, with height generally under stronger genetic control (Cornelius 1994), while diameter is more affected by weather during the year of and the year previous to growth (Kozlowski et al. 1991) and by tree spacing (Husch et al. 1972). In temperate, taiga and tundra biomes, tree growth is typically most responsive to climate at the latitudinal and altitudinal margins of species ranges where climate is the primary limiting factor (Fritts 1976). Individual trees or populations with high variance in growth relative to annual or seasonal climate patterns are described as sensitive, while those with low variance are considered complacent. A review of climate-driven changes in natural forest productivity during the 20st century revealed overall productivity increases (Boisvenue and Running 2006). However, patterns were more complex on local scales. Some researchers found increasing height and diameter growth and conversion of krummholz into erect growth forms at higher-latitude and elevation species margins (Peltola et al. 2002; Stromgren and Linder 2002; Daniels and Veblen 2004; Gamache and Payette 2004; Millar et al. 2004; Danby and Hik 2007), while others recorded growth declines at lower-latitude range margins and other locations where high temperatures induced heat stress (Barber et al. 2000; D‟Arrigo et al. 2004; Adams and Kolb 2005; Reich and Oleksyn 2008). Experiments in which trees were grown under elevated temperature regimes, both in situ using soil or occasionally whole-tree warming devices (Peltola et al. 2002; Stromgren and Linder 2002; Kilpelainen et al. 2006; Danby and Hik 2007) and ex situ using growth chambers or relocation studies (Hobbie and Chapin 1998; Olszyk et al. 1998; Lahti et al. 2005), have also been performed to examine growth responses under elevated temperatures. These studies reveal overwhelming evidence for earlier growth initiation under elevated temperatures, with corresponding growth increases recorded in some circumstances. Optimal conditions for growth are not necessarily the same as those necessary for germination and early survival (Baskin and Baskin 1998; Daniels and Veblen 2004). Seed germination is a challenging time for trees, as it entails transitioning from the seed stage, which is tolerant of diverse environmental conditions, to the seedling stage, which is very vulnerable. 3  Because germinating at the wrong time or in the wrong environmental conditions almost certainly leads to death of the newly germinated individual, complex mechanisms are in place to ensure that germination is more likely to occur at a time when conditions are optimal for survival (Taiz and Zeiger 2002). Exogenous forms of dormancy include physical impenetrability of the seed coat, while endogenous dormancy can consist of physiological safeguards against premature germination or morphological requirements for additional embryo maturation, such as moisture, warmth and chilling requirements (Baskin and Baskin 1998). In many species, a final moisture requirement must be met before the seed coat will rupture, allowing the radicle to extend. Adequate moisture ensures that the newly emerged seedlings will be buffered against desiccation. For new germinants, the highest mortality rates often occur during the first year of growth due water stress, with particularly high mortality rates recorded in alpine environments (Germino et al. 2002). Due to local adaptation, populations are expected to respond to rising temperatures in different ways. One fear is that maladaptation could compromise the competitive ability of some populations (O‟Neill et al. 2008; Bell and Gonzales 2009; Pautasso 2009). Understanding the differences in growth responses among populations allows us to better quantify forest productivity changes that may occur through the 21st century, and to identify populations or species that may fare worse or better as temperatures rise. One of the best research tools for examining these kinds of population differences is common garden studies.  1.4.1 Common gardens: A tool for assessing genetic responses to climate Geneticists use common gardens to observe phenotypic differences among populations planted in common environments (Turesson 1922). Many of these experiments include populations and planting locations chosen to span a wide range of climatic conditions. In forestry, these experiments are referred to as provenance trials, and they are carried out in field sites, seedling nurseries and controlled environments such as growth chambers (Savolainen et al. 2007). „Provenance‟ in this context refers to the geographic and climatic origin of a population, while „population‟ refers to the trees grown from seeds collected at the provenance and grown in multiple locations. Provenance experiments have uncovered significant geographic variation in phenotypic traits such as height and diameter growth, bud phenology and cold hardiness, with particularly strong latitudinal and altitudinal clines in cold-related traits (e.g. timing of primary growth 4  cessation and bud set) relative to provenance temperature gradients (e.g. mean annual temperature, mean coldest month temperature, number of frost free days) (Langlet 1971; Morgenstern 1996). Among populations, there are trade-offs between mean growth and timing of growth cessation. For example, populations from colder climates tend to cease growth and set bud earlier, develop cold hardiness sooner in the fall, and achieve less total growth (Saxe et al. 2001; Howe et al. 2003). This finding illustrates the balance between stabilizing hard selection eliminating phenotypes that are killed by frost in the fall and directional soft selection increasing the competitiveness of taller individuals. Whereas the original purpose of provenance experiments was to find seed sources for reforestation that would optimize wood productivity relative to climate, provenance-trial data are now being reanalyzed using the spatial-climatic variation among test sites as a proxy for temporal-climatic variation to assess the potential impacts of climate change on growth (Rehfeldt et al. 1999, 2001, 2002; Andalo et al. 2005; Chuine et al. 2006; Wang et al. 2006a; O‟Neill et al. 2008; Reich and Oleksyn 2008; Wang et al. 2010). As part of this dissertation, I analyzed new radial-growth data from a large lodgepole pine (Pinus contorta ssp. latifolia Engelm.) provenance trial in British Columbia to assess growth-rate differences among populations growing across a diverse climatic spectrum, as well as the relative influences of genetics and normal climate on the sensitivity of growth to annual and seasonal climate. I also established new common gardens to test the establishment potential of a threatened species, whitebark pine (Pinus albicaulis Engelm.), in model-predicted climatically-suitable locations north of the species range.  1.5 Migration Migration in response to shifting climates is well documented for trees based on pollen records and chloroplast DNA (cpDNA) analyses. Layers of pollen entrapped in bog and lake sediments recorded the northward migration of trees following the last glacial maximum approximately 18,000 years ago. Initial analyses using pollen records showed post-glacial migration rates of up to 200 metres per year, but cpDNA evidence of low-density refugial populations much farther north than previously thought has caused these estimates to be revised downwards (McLachlan et al. 2005). With seeds recognized to have spread both from the northern edge of the continuous species range and from discontinuous refuge populations, actual migration rates have been revised to less than 100 m per year. 5  Whether range shifts associated with climate change are already occurring is a research topic receiving considerable empirical attention, with mixed results. On a relatively local scale, some researchers have seen advancements in elevation (Danby and Hik 2007) and latitude (Gamache and Payette 2005; Caccianiga and Payette 2006), as well as increased species richness (Pauli et al. 2007). Global analyses among taxa have revealed poleward range shifts of over six kilometres per decade and altitudinal advancements of metres per decade (Parmesan and Yohe 2003). However, an analysis of vegetation in northern Canada using infrared satellite imaging showed little to no forest expansion over the last 25 years, despite regional warming of 0.6° C (Masek 2001). Hypothesized reasons for the slow migration include lack of seed sources, barriers to seed dispersal, and poor soil conditions. Regardless of predicted migration rates, there is interest in determining future-predicted ranges for tree species under climate change scenarios. For species of economic concern, such predictions allow reforestation practitioners to create seed-planting zones that maximize survival and growth despite shifting climates (T. Wang pers. comm.). For species of conservation concern, the predictions can be used to assess the adequacy of existing conservation areas for protecting biodiversity as optimal climates move polewards (Araujo et al. 2004; Hamann et al. 2005). For species that cannot migrate to new climatically-suitable locations fast enough, but also cannot persist within their current ranges, humans can consider intervening to facilitate movement to new locations. Called assisted migration or facilitated relocation, this technique is currently being assessed from both ecological and ethical perspectives (McLachlan et al. 2007; Hoegh-Guldberg et al. 2008; Mueller et al. 2008; Ricciardi and Simberloff 2009). All of these management techniques require knowledge of species‟ climatic tolerances and their potential future climatic ranges.  1.5.1 Species distribution models: A tool for predicting potential migration responses to climate change The close association between climate and plant species ranges is behind the growing array of species distribution models (SDMs) that are being used in combination with global climate models (GCMs) to predict future climate change-driven range shifts for a wide variety of taxa. The most common form of SDM is called a bioclimatic envelope model (BEM). BEMs map the current distribution of a species based on the climatic parameters of its existing range, and then predict the species‟ future climatic range distribution under modified temperature 6  regimes (Pearson and Dawson 2003; Guissan and Thuiller 2005; Hamann and Wang 2006). By comparing the extents of present vs. future bioclimatic envelopes for any particular species, one can predict the direction and rate of migration that will have to occur in order for the species to inhabit the same bioclimatic niche in the future that it inhabits at present. Mechanistic models that rely on species‟ or populations‟ physiological responses to climate rather than their range distributions are also being created (Pearson and Dawson 2003). However, data for mechanistic models are not available for many species, and mechanistic models fail to account for species interactions. Thus, most current-generation SDMs are derived using presence/absence data for species distributions and interpolated weather station data for climate. The estimated migration rates of < 100 m per year following the last glacial maximum are troubling in light of SDM predictions of how fast trees will need to migrate during the 21st century. Test of 14 combinations of GCMs and SDMs show up to 100% of the models predicting plant migration rates of 1000 m per year or higher under 2X CO2 climate forcing (Malcolm et al. 2002). Similar estimates have been derived for tree species using regional data from western (Hamann and Wang 2006) and eastern (McKenney et al. 2007) North America. Species in high-latitude biomes will need to migrate fastest, due to more extreme warming occurring towards the poles. High-elevation habitats are also predicted to be more sensitive to climate change than lowland areas (Berry et al. 2002), although a smaller distance must be covered to gain equivalent cooling in alpine areas (Loarie et al. 2009). Given that they provide habitat for forest animals and nucleate new tree islands for some plant species, trees will likely influence the migration rates of associated plant and animal communities. There are many shortcomings to most of the SDMs currently available for predicting species migration rates and future distributions (Araujo and Guisan 2006; Dormann 2007). Few SDMs to date have the capacity to account for non-climatic factors that determine the difference between the fundamental (habitable) and realized (inhabited) niche of a species, including biotic interactions, limiting resources, genetics, barriers to seed dispersal and range fragmentation (Pearson and Dawson 2003; Guisan and Thuiller 2005). This level of ecological detail is unavailable for many species, and yet non-climatic factors can strongly impact the migration capacity of a species, potentially causing predicted and realized future species distributions to differ considerably. In addition, novel climates comprised of new combinations of temperature and moisture variables may occur in the future, for which species presence cannot be accurately predicted. 7  Most current SDMs also do not account for population genetics and life history (e.g. reproduction) traits. Due to lack of data, most models do not account for central-peripheral structure among populations, genetic diversity levels and degree of local adaptation (Mimura and Aitken 2007a; O‟Neill et al. 2008), differences between leading and trailing edge populations (Hampe and Petit 2005), or varying levels of gene flow received by populations in different positions relative to the species range (Mimura and Aitken 2007b). Given that trees often demonstrate moderate population differentiation for traits relating to adaptation to local climate (Howe et al. 2003; Savolainen et al. 2007), individual populations‟ climatic envelopes are likely substantially smaller that the climate envelope of the species as a whole, as inferred from species distributions and climate modeling. The need for accuracy in range shift predictions is particularly vital for threatened species due to extinction being a possible consequence of using inaccurate data in conservation policy decisions (Midgley et al. 2002). While climate is often a dominant limiting factor at broad geographic scales, as the spatial resolution becomes finer, topography, distribution of suitable microsites, landscape connectivity and biotic interactions weigh in as critical organizing features. Bioclimatic models are well recognized for being useful at macro-scales, where influences of biotic factors are superseded by influences of climate (Pearson et al. 2002; Pearson and Dawson 2003). Unfortunately, species in need of conservation attention tend to have dramatically reduced range and population sizes, most often due to extirpation or habitat fragmentation caused by humans (Hartl and Clark 1997). As a consequence of their small effective population sizes, threatened species are often genetically impoverished and at risk of inbreeding depression. These demographic and genetic factors have the potential to reduce the migration potential and competitive ability of threatened populations faced with climate change. Species distribution models should be used with caution in conservation planning for threatened species in particular, and never without field tests and full transparency regarding the models‟ limitations. In this dissertation, I test the effectiveness of a species distribution model that predicts the climatic range of a threatened species, whitebark pine.  1.6 Whitebark pine: A keystone species of conservation concern Whitebark pine is a high-elevation species native to western North America. It is part of a small group of stone pines (Pinus subsection cembrae) characterized by nondehiscent cones closed at seed maturity, and large, wingless seeds that are released and dispersed almost 8  exclusively by birds in the nutcracker genus (Nucifraga) of the corvid family (Tomback 2001). The species is being considered for endangered status in the United States and Canada due to extensive die-offs associated with the introduced disease white pine blister rust (caused by the fungus Cronartium ribicola) and projected impacts of mountain pine beetle (Dendroctonus ponderosae) (Kendall and Keane 2001; Zeglen 2002; Smith et al. 2008; Logan et al. 2010). The species is also losing habitat due to vegetative exclusion by more shade-tolerant tree species caused by fire suppression in the more southern portions of the species range (Arno 2001). The longer-term question of how the species‟ climatic range may shift given the projected warming of the 21st century is also a concern (Hamann and Wang 2006; Warwell et al. 2007). Whitebark pine grows from 37° N latitude in the Sierra Nevada to 56° N in the coast ranges of British Columbia, and from 43° to 53° N in the Rocky Mountains (Ogilvie 1990). The range of temperatures associated with whitebark pine do not vary dramatically throughout its range (normal growing-season (May through August) temperatures for the species: 9.7 °C mean, 7.2 °C 10th percentile, 11.9 °C 90th percentile (Wang et al. in prep.)); rather, the elevations at which it grows drops from 3,600 metres above sea level in the Sierra Nevada to 900 metres in central British Columbia. Like the other stone pines, whitebark pine tends to inhabit areas at timberline with extensive snow packs that keep soils saturated during spring, and with cool enough summers that soil drought is rarely experienced (Weaver 2001). The bioclimatic envelope of whitebark pine is bracketed by -14° C average January (-34° C minimum) and 18° C average July (29° C maximum) temperatures. The northern range of the species is not primarily limited by cold, as shown by the species‟ high cold tolerance (Bower and Aitken 2006) and the more northerly ranges of less cold tolerant subalpine fir (Abies lasiocarpa) and lodgepole pine. Nor are its southern populations necessarily limited by heat. Two-year-old seedlings photosynthesized optimally at 20° C and optimal root growth occurred at 30° C (Jacobs and Weaver 1990), and stands have been observed to grow well in 28° C average (39° C maximum) July temperatures (Weaver 1994). However, Bower and Aitken (2008) found that whitebark pine seedlings grown in Vancouver, BC grew better with cooled than with ambient soil temperatures. Most of the precipitation in whitebark pine environments falls as snow, accumulating as snow pack that rarely melts before mid-May (Weaver 2001). Between summer rains and melting snow fields, whitebark pine rarely experiences drought, although it has been hypothesized that strong winds limit the altitudinal range of whitebark pine by causing foliar desiccation. While 9  snow cover is vital for protecting whitebark pine seeds from winter cold (Mellmann-Brown 2005), too short of a growing season due to late snow melt is most likely the principal high latitude and elevation limiter of the whitebark pine species range. This was demonstrated by Weaver (1994), who found that growing season lengths shorter than three months was the critical threshold determining whitebark pine timberline in the Rockies. It is also possible that wind conditions, which are not accounted for in the species distribution model, cause soil freezing due to snow scouring in alpine areas (Weaver 2001).  1.7 Lodgepole pine: Important to forest productivity and carbon sequestration Lodgepole pine is a valuable timber species that dominates early-successional forest ecosystems throughout much of temperate western North America. Its four subspecies inhabit a wide variety of environmental conditions across it range, with annual precipitation levels ranging from 250 to 5000 mm, absolute minimum temperatures from -57 to 7 °C, absolute maximum temperatures from 27 °C to 38 °C, elevations from just above sea level over 3,000 m in the Sierra Nevada, soil conditions from very dry to very wet, and nutrient condition from very poor to very rich (Lotan and Critchfield 1990). The subspecies that is the focus of this study, ssp. latifolia occurs from northern Oregon, USA, through the southern part of Yukon, Canada, and from the rain shadow of Pacific coastal mountain ranges to the Rocky Mountains in both countries. The mountain pine beetle epidemic that is currently sweeping through British Columbia is expected to kill 80% of lodgepole pine trees in British Columbia by 2013 (Safranyik and Wilson 2006). The beetle is endemic to British Columbia, but the abundance of even-aged stands created by fire suppression and reforestation efforts combined with a series of warmer than usual winters and dry summers has resulted in pine beetle populations increasing by orders of magnitude (Axelson et al. 2009). Lodgepole pine is not expected to go extinct within its current range, but the die-backs could profoundly impact fire dynamics, competition, and population genetics in some areas. The Illingworth lodgepole pine provenance trials provide a unique opportunity to examine the growth of lodgepole pine populations relative to climate. The trials were established in 1974 by the British Columbia Ministry of Forests, Mines and Lands in order to identify populations with high growth and yield characteristics, as well as to determine seedtransfer guidelines for post-harvest restock (Illingworth 1978). The trials revealed that, on 10  average, local populations tend to grow better than non-local populations in local climates (Rehfeldt et al. 1999). However, a latitudinal discrepancy exists between the climate of provenance origin and that of optimum growth, with populations from southern provenances occurring in climates ~0.5 °C below their optima, while populations from northern provenances are in locations ~7 °C colder than optimal. This is thought to be the result of intraspecific density-dependent selection, wherein central populations (central corresponding to southeastern British Columbia) genetically swamp more peripheral populations. The Illingworth populations were also found to have higher adaptability when moved upward (2 °) than downward (1 °) in latitude with no impact on growth. This may be due to incomplete range expansion following the last major glacial period approximately ten thousand years ago (Johnstone and Chapin 2003) or recent climate change. The Illingworth trial trees are ideal for studying radial-growth responses to climate relative to genetics because of the homogeneity of site conditions. In natural forests, age, intertree distances, competition with tree and non-tree vegetation, nutrients related to parent material and soil processes, and genetics often function as confounding factors when looking at growth relative to climate. In contrast, the Illingworth trial controlled for many of these confounding variables, allowing the examination of growth responses attributed almost entirely to genetics and climate.  1.8 Research objectives In the following chapters, I examine the impacts of climate and genetics on the establishment and growth of two iconic North American trees, whitebark and lodgepole pine, and use the results to predict how climate change may affect these species. In Chapter 2, I planted whitebark pine seeds in common gardens north of the current species range in areas that are predicted to be climatically suitable under both present and future climates, and monitored establishment and growth relative to climate, site features and genetics for four years. In Chapter 3, I grew whitebark and lodgepole pine seedlings in growth chambers to test the species‟ fundamental climatic niches and their relative competitive potential across an extreme climatic spectrum. In Chapter 4, I used lodgepole pine radial-growth data from the Illingworth trial to build universal growth-trend response functions that predict diameter growth over time relative to present and future normal annual and seasonal climate and genetics. In Chapter 5, I assessed  11  the responsiveness and sensitivity of annual radial growth to annual and seasonal climate fluctuations on a regional basis using the same Illingworth-trial radial-growth data.  12  Chapter 2: Whitebark pine (Pinus albicaulis) assisted migration potential: Testing establishment north of the species range1 2.1 Introduction Species distribution models (SDMs) predict that climatic niches of species will shift towards the poles and to higher elevations with climate warming. However, there is considerable uncertainty as to whether trees can migrate fast enough to stay within their moving niches, given the rates at which temperatures are expected to rise (Christensen et al. 2007). Fossil pollen records and molecular data indicate that many tree species migrated 10 to 20 km per century following the last glacial maximum (Davis and Shaw 2001; McLachlan et al. 2005). However, migration rates up to an order of magnitude higher may be necessary for some trees to stay within their climatic tolerances (Malcolm et al. 2002; Iverson et al. 2004; Hamann and Wang 2006; Loarie et al. 2009). The failure of species to migrate fast enough could lead to population collapses and extinctions, with those with limited ranges, small population sizes, and major barriers to dispersal considered most vulnerable (Parmesan 2006). One way to avert species losses may be to assist the migration of vulnerable organisms in situations where natural migration is implausible (McLachlan et al. 2007; Hoegh-Guldberg et al. 2008; Mueller et al. 2008; Ricciardi and Simberloff 2009). Those that favour assisted migration believe that, under well-researched and ecologically appropriate circumstances, relocating a threatened species can protect it from extinction while minimally impacting the recipient ecological community. Opponents argue that the probability of the new species negatively affecting the biota within its new environment is not worth the risks. Both sides agree that decision-making frameworks based on rigorous ecological risk assessments and economic costbenefit analyses must be developed prior to taking action. Whitebark pine (Pinus albicaulis Engelm.) is a threatened keystone species that inhabits mountainous regions in western North America. Its nutritious seeds are distributed by Clark‟s nutcrackers (Nucifraga columbiana) and serve as a critical food source for grizzly bears (Ursus arctos) (Mattson et al. 1992). The slow-growing tree also creates microhabitats for recruitment  1  McLane S. C., S. N. Aitken. Under review. Whitebark pine (Pinus albicaulis) assisted migration potential: testing establishment north of the species range. Ecological Applications. 13  by other tree species in the high-elevation environments where it grows (Callaway 1998). Logan et al. (2010; J. Logan pers. comm.) postulate that nearly all cone-bearing whitebark pine trees in the greater Yellowstone ecosystem will be dead by 2015. In addition, over 50% of all whitebark pine in the northern United States and in Canada are dead or dying due to the combined impacts of mountain pine beetle (Dendroctonus ponderosae) and white pine blister rust (Cronartium ribicola) (Kendall and Keane 2001; Zeglen 2002, Smith et al. 2008). Due to these population declines, the federal governments of the United States and Canada (COSEWIC 2010) are evaluating the species for endangered status as of 2010. Forecasts from SDMs indicate that whitebark pine will need to migrate hundreds of kilometres over the next century to stay within its climatic niche (Hamann and Wang 2006). The species is projected to lose 73% of its current climatic range within British Columbia (BC), Canada, by 2085, while gaining an equivalent-sized new climatic range to the northwest of its current northern range limit (Figure 2.1). Simultaneously, the species is projected to lose over 97% of its current climatic niche within the United States (Warwell et al. 2007). Interestingly, SDMs show that much of the area in northwestern BC that is predicted to be habitable by whitebark pine in the future is already climatically suitable for the species under current conditions, implying that whitebark pine does not inhabit its full climatic niche. Either nonclimatic factors are restricting recruitment outside of the current range, or SDMs are not accurately predicting the present and future climatic envelopes of the species. The former has important implications for the ability of the species to inhabit its predicted climatic niche in the future, while the latter points to the need for validating and improving SDMs. Whitebark pine seeds are typically immature at harvest and require a warm, moist maturation period followed by extended chilling to break dormancy in order to germinate. This process occurs naturally during the summer and winter following planting by nutcrackers, resulting in most germination occurring two summers after productive cone crops are observed (Tomback et al. 2001). Tree nurseries have established seed pre-treatment protocols to speed up and promote germination (Berdeen et al. 2007; Riley et al. 2007), but the long-term establishment potential of treated versus untreated whitebark pine seeds under field conditions has not previously been tested. Likewise, whitebark pine is known to be limited by habitat conditions, particularly those related to snowmelt timing and sun and wind protection (Weaver 2004; Mellman-Brown 2005; Maher and Germino 2006; McCaughey et al. 2009). It also has low to moderate genetic differences among populations, with genetic clines following 14  provenance (place of origin) climatic gradients (Mahalovich et al. 2006; Bower and Aitken 2008). However, these limiting factors have not been tested for trees planted north of the current species range, and must be better understood before managers can consider assisting the migration of whitebark pine. Whitebark pine is ideal for such trials because of its threatened status, promising future range predictions, restricted ecological niche, and negligible risk of unwanted spread due to slow (30-50+ years) reproductive maturation (McCaughey and Tomback 2001). In this study, we examined the impacts of seed maturity, habitat quality, and genetics on whitebark pine establishment throughout the species‟ realized and predicted climatic range within British Columbia. Seeds from multiple populations were planted in areas predicted by SDMs to be habitable for whitebark pine under current and 2055 climate conditions, focusing particularly on areas northwest of the current species range. We tested the overall hypothesis that whitebark pine can establish in these model-predicted areas, and furthermore, hypothesized that a) whitebark pine seeds subjected to screening and induced maturation treatments prior to planting have greater establishment potential than untreated seeds; b) establishment is affected by site climate and microsite conditions, particularly those relating to snow duration; and c) quantitative-trait differences among populations correspond to clines in provenance climate. Our results allowed us to assess the accuracy of SDMs for predicting current range limits for whitebark pine, and inform the creation of assisted migration guidelines for the species. This information is key to understanding the potential for this important threatened species to survive 21st-century climate change.  2.2 Materials and Methods 2.2.1 Seed collection and treatment We collected cones from seven whitebark pine provenances representing a wide geographic and climatic gradient within the northwestern extent of the species range (Figure 2.2, Table 2.1). Cones from 10 trees per provenance were caged using wire exclosures in June and July 2007 to prevent harvesting by Clark‟s nutcrackers and rodents. The cones were collected, seeds extracted, and mean family seed weights recorded in August and September 2007. The seeds from each parent tree were considered an open-pollinated family. To test techniques for promoting germination, we treated half of the seeds in each family using variations on the protocols of Berdeen et al. (2007) and Riley et al. (2007). The treatment 15  protocol comprised numerous steps. In January 2008, the seeds were x-rayed using a Faxitron XRay machine (2 minute exposure, 15 kVp). Seeds in which the embryo filled less than 20% of the corrosion cavity were classified as nonviable and discarded. Starting in February 2008, the x-rayed seeds were soaked in warm water for two days to promote imbibation, matured at 15 °C for one month, and then cold stratified at 2 °C for three months. Upon termination of stratification in June, the treated seeds were transported to the test sites at 3 °C. Prior to planting, 1-2 mm of tissue was clipped from the radical end of each seed coat using a razor blade to promote radical emergence. Nonviable seeds were tallied prior to being discarded at all stages of the treatment process.  2.2.2 Site selection and establishment We established common gardens in eight locations representing the current observed and 2055 projected climatic ranges of whitebark pine. Common gardens are experiments where multiple genetic lineages are grown in common environments to examine the impact of genetics and site conditions on quantitative-trait differences among individuals. We established the gardens in areas broadly predicted to be climatically suitable according to SDMs created by T. Wang (University of British Columbia) using the protocols of Hamann and Wang (2006; Appendix 1). Two of the gardens are located within current species range, while six are located north of the range (Figure 2.2, Table 2.1). Collectively, the trial locations span nearly 10 ° latitude, from 600 km southeast to 800 km northwest of the current northern range margin. A ninth location near Tweedsmuir Park that was snow-covered at the time of planting turned out to have inappropriate substrate and was abandoned. Within each trial location, we established two sites with the intention of replicating conditions experienced by whitebark pine within the subalpine extent of its current ecological niche. Sites were located 50 to 500 m apart, 100 to 200 m below the highest tree islands above continuous treeline, on south-facing, 5 to 20 ° slopes with coarse, well-drained soils. Areas with evidence of human or animal disturbance were avoided. Thermachron iButton (Dallas Semiconductor) temperature data loggers were installed at each site in September 2007, one at ground level and one at 10 cm depth underground, and programmed to record temperatures four times daily. We planted the seeds in experimental units of two to mimic the multi-seed caching behaviour of Clark‟s nutcrackers (Tomback 2001). Four two-seed caches per family were 16  planted in two blocks per site for replication. Two of the four caches per family and block were planted using untreated seeds in September 2007 and the other two using treated seeds in June 2008. The seeds were planted 2 cm deep in a 0.25 by 0.5 m grid alternating by seed treatment, with vegetation cleared within 5 cm of each cache. A total of 8,960 untreated and 6,992 treated seeds were planted. Families lacking adequate numbers of treated seeds were not represented in every block. Late snowmelt in 2008 prevented planting the treated seeds in all of Bell II site 2 and half of Blackcomb site 2, so a third site was created at Bell II very near site 1, and additional rows were added at Blackcomb site 2 (Table 2.1).  2.2.3 Data collection To track seedling establishment over time, we recorded germination, survival, health, height and needle fascicle data in 2008, 2009 and 2010. Germination and survival were recorded as binary variables, health was ranked as poor, moderate or good based on foliage and stem appearance, height was measured to the nearest half cm, and needle fascicles were counted for all buds displaying needle tissue. We collected microsite data, including slope, convexity, soil type, soil depth and vegetation height for every seed-cache location in 2008. Slope was estimated by 10 ° increments, while convexity was classified as flat, concave or convex and soil type as pure mineral, mineral topped with organic, or very (> 80%) rocky; all of these values represent means for the 10 cm radius surrounding each cache. Soil depth to impenetrable rock was measured to the nearest cm adjacent to each cache, while vegetation height represents mean height outside of the 5 cm cleared area but inside the 10 cm radius around each cache. We averaged the soil-surface temperature data to create mean growing season (June – September) and winter (October – May) temperatures for the duration of the experiment (2007 – 2010). Most sites were covered in snow, and therefore registered surface and sub-surface temperatures near 0 °C, for the majority of the winter months. Snowmelt dates, determined as the first day of the year that temperatures exceeded 5 °C for seven days in a row following the winter period of near-freezing temperatures, were also determined from the temperature-sensor data. Snowmelt dates could not be determined for the Atlin sites because a snowpack never developed, presumably due to wind scouring, as indicated by winter temperatures fluctuating well below 0 °C. A small number of temperature sensors lost functionality or disappeared; when this occurred, we estimated seasonal temperatures and snowmelt dates from below-ground 17  sensors at the same site or from above and below-ground sensors at the other site within the same location. We also obtained normal (1971 – 2000) mean annual temperature (MAT) and precipitation as snow (PAS) data for each provenance and site using ClimateWNA (an extension of ClimateBC (Wang et al. 2006b)) (Table 2.1).  2.2.4 Statistical analyses We conducted all statistical analyses using SAS software, version 9.2 (SAS Institute 2008) with seed cache as the experimental unit. Data from the following site and population were excluded from all analyses due to near-zero germination rates: 1) Whistler site 2, treated seeds, attributed to seed-quality depletion (last site planted); and 2) Smithers provenance, all seeds, attributed to very low seed viability (Figure 2.2). Of the 280 seed caches at Smithers site 2, 137 were excluded from the analysis due to seed herbivory by rodents in June 2008 evidenced by digging and seed-coat remnants. Surprisingly, no signs of seed or seedling herbivory were noted at other sites. We examined the effects of seed maturity, site conditions and provenance climate on germination, survival, health, height and number of needle fascicles using predictive models. Initial regressions were performed for each dependent variable with the explanatory variables separated into four categories: 1) cache microsite (vegetation height, soil type, soil depth, slope, convexity); 2) site climate (normal MAT and PAS, summer temperature, winter temperature, snowmelt date); 3) provenance climate (normal MAT and PAS) and seed weight; and 4) seed treatment. Significant variables from these submodels were pooled and interactions added to build a full model for each response variable. Stepwise procedures were used to determine variable significance for all models. Germination was analyzed using a multinomial logistic model based on cumulative logistic models fitted by SAS PROC LOGISTIC: ( ) ( )    Eq. 2.1  where Yi = 1 or 0 for each cumulative model; F(x) is a linear function of the explanatory variables; and ε represents error. The first model gives the probability of number of germinants = 0 vs. 1 or 2 and the next model gives the probability of germinants = 0 or 1 vs. 2. The probability of each level is then obtained by subtraction. Model fit was assessed using maxrescaled R2 values (SAS 9.2 Documentation, SAS Institute 2008). Germination was modeled separately for treated seeds that germinated in 2008 versus untreated seeds that germinated in 18  2009. Total treated-seed germination was also estimated with discarded non-viable seeds accounted for in the calculation. This allowed a direct evaluation of the impact of seed treatment on germination potential. As with germination, survival and health were analyzed using multinomial logistic models described by Eq. 2.1. The first survival model gives the probability of survival = 0 if no germinants survived vs. 0.5 (i.e. if 1/2 germinants survived) or 1 (i.e. if 1/1 or 2/2 germinants survived). The second model gives the probability of survival = 0 or 0.5 vs. 1. Survival was modeled separately for the treated versus untreated seedlings using the 2010 data set. For health, the first model gives the probability of health = 1 if health was poor vs. 2 or 3 if health was moderate or good, respectively, and the second model gives the probability of health = 1 or 2 vs. 3. The 2010 dataset for live seedlings that germinated in 2008 was used for the health analysis. If there were two germinants in a cache, only the health of the larger seedling was analyzed. Height and number of needle fascicles were analyzed using a general linear model fitted using PROC GLM: ( )    Eq. 2.2  where Yi is the height or the number of fascicles of the larger seedling in each cache, respectively; F(x) is a linear function of the explanatory variables; and ε represents error. Both variables were analyzed using the 2010 data set for live seedlings that germinated in 2008. Model fit was assessed using variance explained.  2.3 Results Germination occurred and seedlings survived and grew in all 16 common garden sites (Figure 2.3a). By August 2008, 28.5% of the treated and 0.7% of the untreated seeds had germinated (Figure 2.3b, Appendix 2). By July 2009, these numbers had risen to 29.9% of the treated and 9.6% of the untreated seeds. As of 2010, an additional 0.1% of the treated and 0.6% of the untreated seeds had germinated. This translates to 95% of the total treated-seed seedlings germinating in 2008 (2 months after planting), and 94% of the total untreated-seed seedlings germinating in 2009 (second year after planting). Total treated-seed germination fell to 20.4%, when the proportion of seeds discarded during the seed-treatment process was taken into account. Explained variance was higher for untreated seeds that germinated in 2009 than treated seeds that germinated in 2008, the primary difference being that site climate influenced the former but not the latter (Table 2.2). Seed weight was a major explanatory variable for 19  germination of both treated (Figure 2.4a), and untreated seeds. Germination of treated seeds was higher in microsites with organic soils. Almost all germinants survived until assessed during the summer that they germinated (Figure 2.3b, Appendix 2). Survival rates were lower for treated-seed germinants (63.9% in 2009) than untreated-seed germinants (79.2% in 2010) by their respective second summers. Mortality rates levelled off for the treated-seed seedlings after 2009, with 53.9% of the total original germinants still alive in 2010. Calculated in terms of total original seeds (i.e. including seeds discarded during the seed-treatment process), 8.1% of the total untreated and 11.7% of the total treated seeds germinated and survived through summer 2010 (Figure 2.3b). Survival of the untreated-seed seedlings was primarily negatively associated with colder winters. Survival of treated-seed seedlings was primarily positively associated with seed weight (Figure 2.4a) and warmer summers, and negatively associated with later snowmelt dates (Table 2.2, Figure 2.4b). Seedlings that germinated in 2008 averaged 3.2 cm tall with 6.0 flushed needle fascicles as of 2010. Height, fascicles and health were primarily positively associated with both provenance and site temperature, and negatively associated with later snowmelt (Table 2.2). Other than fewer needle fascicles developing on seedlings surrounded by taller vegetation, microsite factors negligibly influenced height, fascicles and health.  2.4 Discussion Our study demonstrates that whitebark pine is able to germinate and survive in locations hundreds of kilometres north of its current northern range limit. Further monitoring will be necessary to determine if these planted populations will persist. However, our results broadly confirm the accuracy of species distribution models indicating that whitebark pine can be successfully relocated to climatically-suitable environments outside of the current species range. The major factors influencing establishment were seed maturity, growing season length as determined by snow, and provenance temperature.  2.4.1 Seed maturity: the combined effects of treatment and weight Treating the whitebark pine seeds effectively promoted seed maturation and broke dormancy, causing the majority of germination to occur a year earlier and inducing higher germination rates relative to leaving seeds untreated. However, when nonviable seeds discarded during the treatment process were accounted for, only twice as many treated seeds germinated 20  than untreated seeds overall. Furthermore, percent survival was lower for treated-seed than untreated-seed seedlings, such that the number of survivors relative to the number of original seeds did not differ majorly between seed treatments. We cannot be sure that the mortality-rate differences were due to seed treatment rather than weather or other factors, since the major mortality pulses occurred in different years for the two seed treatments. Likewise, the proportion of seeds discarded during the treatment process may have been unusually high due to poor seed development. However, our results suggest that it is worth considering whether treating whitebark pine seeds is worth the effort in contexts where terrain and germination speed are not limiting factors, such as restoration-planting initiatives. Seed weight was highly variable among the whitebark pine families and ended up being a primary predictor of establishment potential. We found that heavier seeds had better-developed embryos, germinated and survived in greater numbers and developed into larger, healthier seedlings. Whitebark pine is an exception to the global trend that seed size decreases with latitude (Moles and Westoby 2003), likely reflecting the co-evolution between the species and its primary disperser, the Clark‟s nutcracker. Whitebark pine seeds must be large in order to provide a net energetic gain for the nutcracker, and yet the short growing-season length in the subalpine environments inhabited by the tree prohibits full seed maturation prior to harvest (Tomback 2001). Many families from three of our seven initial populations – Apex, Whistler and Smithers – had low seed weights and poorly developed embryos and correspondingly demonstrated particularly low germination rates (Appendix 2). Based on weights and x-rays of seeds collected from the same provenances in earlier years (S.C. McLane pers. obs.; D. Pigott of Yellow Point Propagation pers. comm.), we believe the poor seed quality of these populations results primarily from maternal effects caused by unusually cold conditions during cone development, rather than genetic or normal-climate factors. These sorts of weather-driven maternal effects are common for plant species inhabiting harsh environments (Moles and Westoby 2003).  2.4.2 Site conditions: the paramount influence of snow cover Most precipitation in whitebark pine environments falls as snow, accumulating as a snow pack that rarely melts before mid-May (Weaver 2001). This was validated by our ClimateWNA data, where precipitation as snow was highly correlated with mean annual precipitation for the sites (r = 0.98) and for the provenances (r = 0.92). Between summer rains and melting snow 21  fields, mature whitebark pine trees rarely experience drought, although it has been hypothesized that strong winds partially limit the species‟ altitudinal range by causing foliar desiccation. Previous researchers have found that snow cover is vital for protecting whitebark pine seeds from cold (Mellmann-Brown 2005), but that survival is poor in areas where the growing season is too short due to snow persistence (Weaver 1994). We found snow to play this classic, paradoxical role in our study system. Earlier snowmelt was associated with greater survival rates and better health and growth, and lower amounts of normal precipitation as snow were positively associated with germination. However, extremely low germination of untreated seeds and high mortality of treated-seed seedlings occurred in the Atlin sites where insulating snowpacks never developed, presumably due to wind scouring (Figure 2.4). At the same time, mean annual and summer temperatures were positively associated with germination, survival and growth. Together, these results confirm that a balance between sufficient snow duration and a longenough growing season are critical for whitebark pine establishment and persistence. Microsite influenced establishment less than we expected. Whitebark pine has previously been found to establish best in proximity to landscape features including trees, herbs, logs, rocks and stumps, due to the protection these features provide against desiccation by sun and wind (Mellman-Brown 2002; Maher and Germino 2006; McCaughey et al. 2009). We may have inadvertently decreased seedling establishment by systematically locating our sites away from these sorts of landscape features and by removing vegetation within 5 cm of the seed caches to minimize confounding effects.  2.4.3 Genetic effects follow provenance-temperature clines Few studies have examined quantitative-trait differences among whitebark pine populations, reflecting the expense and difficulty of conducting common-garden experiments for such a slow-growing species with difficult-to-procure seeds. However, two such studies in relatively benign nursery environments indicate that seedlings from milder provenances grow more but have lower cold tolerance than those from harsher locations (Mahalovich et al. 2006; Bower and Aitken 2008). Our data partially corroborated these trends, with increased germination and survival capacity, growth and height recorded for populations from warmer provenances. This effect was not an artefact of seed weight, which was not correlated with provenance temperature (r = 0.02). Cold tolerance and phenology will be monitored during  22  future site visits to better determine which populations would be optimal to plant in northern locations should assisted migration be deemed necessary or desirable.  2.4.4 Using SDMs to predict the climatic range for whitebark pine Our demonstration that whitebark pine seedlings can establish in model-predicted areas north of the species range, and that establishment is partially predicted by modeled climate variables including normal precipitation as snow, is preliminary evidence that species distribution models can predict climatically-suitable habitat for this species. However, annual snowmelt timing, as recorded by iButtons, was an important predictor that was not captured by our species distribution model, and that is not included in other SDMs to our knowledge. Despite being temperature and precipitation-driven, snow persistence is highly influenced by slope, aspect, wind, local topography and freeze-thaw cycles. Adding a snow-duration variable to predictive habitat models for whitebark pine and other cold-adapted species could greatly improve their accuracy. Some researchers are generating regional snowmelt-timing models (e.g. Beniston et al. 2003), and satellite-derived snowpack data are available for some portions of the globe (Canadian Meteorological Centre 2010).  2.4.5 Should we assist the migration of whitebark pine? Whitebark is declining precipitously within its current range, and is not expected to adapt nor migrate fast enough to keep pace with climate change. Whitebark pine generations are 50 to 100 years long, making it highly unlikely that the species can adapt to the ~3 °C temperature increases (Christensen et al. 2007) predicted for northwestern North America by the 22nd century. Natural migration is also unlikely for the species. Numerous biotic and abiotic factors interplay with climate to determine migration potential, including reproductive strategy, recruitment potential, geographic barriers to dispersal and interactions between species (Davis et al. 1998). For whitebark pine, a major and unusual migration constraint is the species‟ dependency on the Clark‟s nutcracker for seed distribution. Long-distance dispersal could be facilitated if nutcrackers fly seeds to previously uninhabited areas as they become climatically suitable for both species. However, the SDM prediction that much of whitebark pine‟s potential future range is climatically suitable at present calls into question why the nutcracker has not moved the species northward already.  23  Whitebark pine and other montane species have the advantage that small uphill migrations yield large temperature reductions relative to migrating across flat terrain (Loarie et al. 2009). However, whitebark pine often lives just below alpine areas in which soils tend to be poorly formed or absent. Given the predicted rate of climate change, it will take centuries for adequate soils to develop in such environments. The most likely scenario is therefore that whitebark pine will be outcompeted by faster-growing competitors encroaching from lower elevations (SCM and SNA unpublished data) while remaining unable to migrate far uphill and slow in migrating northward. The probability of whitebark pine becoming invasive in novel environments is extremely low. Plants in general are at low risk for intracontinental invasions (Mueller and Hellman 2008), and whitebark pine‟ slow (30-50+ years) reproductive maturation, infrequent cone crop, poor competitive ability relative to other trees, and habitat-specialist life-history strategy make it particularly unlikely to demonstrate uncontrolled population growth (MaCaughey and Tomback 2001; Richardson and Rejmánek 2004). However, other ecological and economic factors should be accounted for in assessing whitebark pine‟ case for assisted migration. Most critical is the continued seeking in nature, or creation, of white pine blister rust-resistant genetic strains, without which translocating the species could be futile. Fortunately, provenances with higher levels of resistance are beginning to be identified (Mahalovich et al. 2006). While concern for whitebark pine is high, the species is not yet federally legislated as endangered, and as such its migration should not be facilitated at present. Our common gardens are for research only; we will monitor the surviving trees and evaluate removing them before they reach reproductive maturity. Current efforts should go instead towards a) continuing to determine natural levels of resistance to blister rust; b) improving SDMs to more accurately predict the species‟ future range extent, and d) evaluating Clark‟s nutcracker and whitebark pine dynamics at the current northern edge of the species range to determine the pine‟s natural dispersal potential, and d) restoration planting within the current species range using rustresistant provenances that are pre-adapted for warmer temperatures (Keane et al. in press). Concurrently, ecological and ethical decision-making frameworks for assisted migration should continue to be developed (Richardson et al. 2009) using whitebark pine as a test case because of its threatened status and non-invasive life-history attributes.  24  2.5 Conclusions We found that whitebark pine can establish in model-predicted climate zones north of the current range limit. The major factors influencing establishment were seed maturity and site conditions, particularly the duration of snow cover. Treating seeds caused germination to occur earlier and boosted germination potential, but percent survival of seedlings grown from treated seeds was lower, indicating that treating seeds is a questionable use of resources for restoration planting. The species distribution model that we used was broadly accurate for predicting climatically-suitable growing locations for whitebark pine, although it could be improved by adding a snow-duration variable. Further monitoring will be necessary to determine the longterm establishment potential of our whitebark pine trees in the new locations. We believe that whitebark pine is a species that could eventually benefit from a program of assisted migration.  25  Table 2.1: Geographic and climatic variables for the two sites within each of the eight trial locations and the seven sampled whitebark pine populations (see Figure 2.2). Closest Town Provenance Tatla Lake, BC Fort St. James, BC Smithers, BC Penticton, BC Whistler, BC John Day, OR Entiat, WA  Latitude Longitude Elevation MAT PAS (°N) (°W) (m) (°C) (cm)  Other explanatory variables Seed weight (g per 100 seeds) 16.6 52.54 125.81 1541 0.1 524 16.9 54.88 125.37 1490 0.2 396 10.2 54.77 127.28 1500 0.5 473 10.2 49.37 119.92 2148 0.7 336 9.7 50.10 122.90 1882 0.8 1290 12.2 44.28 118.70 2438 3.7 572 12.7 47.99 120.41 1998 6.5 463 Summer Winter Day of Trial locations and sites (all in BC) temp (°C) temp (°C) snowmelt Atlin 2 59.7302 133.5177 1368 -2.5 287 11.1 -4.3 n/a Atlin 1 59.7292 133.5182 1357 -2.4 287 11.4 -3.8 n/a Bell II 2 & 3 56.7627 129.6864 1494 -1.7 628 6.4 -0.1 173 Bell II 1 56.7627 129.6902 1455 -1.6 628 m.d. -0.4 181 Smithers 2 54.7771 127.3034 1676 0.3 599 9.8 -2.0 149 Smithers 1 54.7763 127.2957 1650 0.4 594 9.6 -1.2 153 Whistler 1 50.0899 122.8959 1970 0.5 1430 10.3 0.2 177 Whistler 2 50.0890 122.8957 1952 0.6 1422 9.5 0.2 186 Haines Jct. 2 59.5669 136.4616 852 0.7 1024 12.1 -0.4 121 Terrace 2 54.8312 128.7073 1316 0.7 915 13.1 1.2 98 Haines Jct. 1 59.5668 136.4630 842 0.8 1020 12.1 -0.7 135 Stewart 1 56.1701 130.0434 1274 0.8 1985 10.7 0.4 139 Stewart 2 56.1692 130.0427 1278 0.8 1989 10.6 0.4 150 New Hazelton 1 55.3235 127.5239 1543 0.9 427 9.6 -1.8 144 New Hazelton 2 55.3223 127.5247 1527 0.9 424 10.6 -1.9 143 Terrace 1 54.8320 128.7047 1319 0.9 943 12.5 0.8 123 Notes : Provenances and sites are ordered from smallest to largest normal mean annual temperature (MAT). MAT and PAS (precipitation as snow) were generated using ClimateWNA (Wang et al. 2006) and represent 1971-2000 normals. Temp = temperature. Summer and winter temp and day of snowmelt are averages for the study period (2007 - 2010), derived from iButton temperature sensors. BC = British Columbia, Canada; WA = Washington, USA; OR = Oregon, USA.  26  Table 2.2: Models predicting whitebark pine germination, survival, health, height and needle fascicles relative to 1) microsite factors, 2) test site climate, and 3) provenance climate and seed weight. Full models were built using the significant (p < 0.05) variables from the three submodels. Variables with negative slopes are bolded and italicized. Dependent variable  Model  Significant variables (p < 0.05; deduced using stepwise method, listed in decreasing order of significance) soil type (organic) sPAS seed weight, pMAT seed weight, soil type (organic), pMAT, sPAS soil type (organic), soil depth, slope sPAS , sMAT, 2009 snowmelt date , 2009 winter temp pMAT, seed weight pMAT, sPAS , seed weight, sMAT, 2009 snowmelt date , 2009 winter temp vegetation height ave summer temp, ave snowmelt date , sMAT seed weight, pMAT seed weight, ave summer temp, ave snowmelt date , sMAT, pMAT, vegetation height soil depth, vegetation height ave winter temp pMAT ave winter temp , vegetation height vegetation height ave summer temp, sMAT pMAT, seed weight pMAT, sMAT, vegetation height , seed weight  K  N  R²  Microsite 1 3152 0.03 Germination of Test site 1 3199 0.02 treated seeds in Provenance 2 3199 0.13 2008 Full model 4 3152 0.16 Microsite 3 3679 0.07 Germination of Test site 4 2973 0.14 untreated seeds Provenance 2 3690 0.08 in 2009 Full model 6 3439 0.22 Microsite 1 1425 0.01 Survival of Test site 3 1224 0.11 treated-seed Provenance 2 1428 0.02 seedlings as of 2010 Full model 6 1224 0.14 Survival of Microsite 2 582 0.05 untreated-seed Test site 1 584 0.10 seedlings as of Provenance 1 623 0.03 2010 Full model 2 623 0.09 Fascicles of Microsite 1 799 0.05 summer1 Test site 2 758 0.06 germinants in Provenance 2 799 0.07 2010 Full model 4 758 0.17 Height of Microsite 0 799 n/a summer1 Test site ave snowmelt date , sMAT, ave summer temp 3 758 0.11 germinants in Provenance seed weight, pMAT 2 799 0.05 2010 Full model ave snowmelt date , pMAT, sMAT, seed weight, ave summer temp 5 758 0.16 Microsite vegetation height 1 799 0.03 Health of Test site ave snowmelt date , ave winter temp , sPAS , sMAT, ave summer temp 5 758 0.12 summer1 Provenance pPAS , seed weight, pMAT 3 799 0.07 germinants in ave snowmelt date , ave winter temp , sPAS , pMAT, pPAS , ave summer 2010 Full model temp, seed weight, sMAT 8 758 0.16 Notes: Seed treatment was included for needle fascicles, height and health, but was not significant and therefore is not listed. K = number of significant variables; N = number of seed caches observed; R² = max-rescaled R² for germination, survival and health. The prefix "p" denotes Provenance while "s" denotes site. MAT = normal mean annual temperature; PAS = normal precipitation as snow; ave = average; yearly winter temperatures reflect the winter ending in the listed year. Height and seed weight are log-transformed.  27  Figure 2.1: Species distribution models depicting whitebark pine‟ a) current observed range in British Columbia (BC), Canada; b) current predicted range in BC based on 1961-1990 climate normal; and c) 2025 and d) 2085 future predicted ranges in BC based on IS92a CGCM1 GAX future-climate scenarios. The models were created by T. Wang (University of British Columbia, unpublished, using methods from Hamann and Wang 2006). See Figure 2.2 for 2055 predicted range, scale and geographic location and Appendix 1 for the model-creation methods.  28  Figure 2.2: Trial locations and provenances relative to the 1990s observed and 2055 predicted whitebark pine species range within British Columbia, Canada. Of the eight trial locations, two are within and six are north of the current species range; all are in areas predicted to be habitable under both present and 2055 climate regimes. The two bolded locations, Whistler and Smithers, are both trial locations and provenances. The predicted species range was created by T. Wang (University of British Columbia, unpublished, using methods from Hamann and Wang 2006) using the IS92a CGCM1 GAX future-climate scenario. The map scale is accurate in the map centre but approximate at the boundaries due to projection skew. The current species range within Canada is accurate with respect to latitude and longitude but not elevation.  29  Figure 2.3: Whitebark pine germination, survival and mortality by a) site (treated-seed data only) and b) seed treatment and year.  30  Figure 2.4: Whitebark pine treated-seed seedlings that a) germinated and survived relative to seed weight, and b) survived relative to snowmelt timing.  31  Chapter 3: Whitebark pine (Pinus albicaulis) germination and growth at extreme temperatures 3.1 Introduction Climate change will challenge the ability of tree species to persist within their current ranges. Persistence requires either phenotypic plasticity, i.e. the genetic flexibility to deal with the new climatic conditions in situ, or adaptation, i.e. the genetic capacity to respond to natural selection (Aitken et al. 2008). Adaptation occurs slowly for trees with long generation times, and consequently is considered an improbable option for such species given the predicted rate of climate change (Christensen et al. 2007). In contrast, trees are hallmarks of phenotypic plasticity, often weathering long periods of suboptimal climatic conditions over their century or multi-century lifespans. An important phenotypic change already noted in mid and high-latitude environments is overall increased tree growth (Boisvenue and Running 2006) caused by higher photosynthetic rates (Saxe et al. 2001; Way and Oren 2010) and elongated growing seasons (Bergh and Linder 1999). Realized ecological niches are almost always smaller than fundamental climatic niches due to species interactions, resource limitations and physical barriers to dispersal (Sexton et al. 2009). Competition by more heat-adapted species or populations, in particular, is known to restrict some populations to colder portions of their fundamental niches (e.g. lodgepole pine (Pinus contorta, Rehfeldt et al. 1999)). These restricted populations possess the latent genetic capacity to experience growth increases as temperatures rise. However, if competitors gain proportionately greater growth advantages, competition-limited populations may experience net growth and habitat losses due to climatic warming. Whitebark pine (Pinus albicaulis Engelm.) is a keystone species that inhabits a narrow treeline niche in western North America. The species is being considered for endangered status by the United States (www.fws.gov/mountain-prairie/species/plants/whitebarkpine) and Canadian governments (www.sararegistry.gc.ca) due to population declines caused primarily by white pine blister rust (Cronartium ribicola) and mountain pine beetle (Dendroctonus ponderosae) (Tomback and Achuff 2010). However, climate change is predicted to cause even broader population declines for whitebark pine (Hamann and Wang 2006; Warwell et al. 2007; Wang et al. in prep.). Global circulation models indicate that average air temperatures will rise by ~ 3.4 to 4.5 °C within whitebark pine‟s current range during the 21st century (Christensen et 32  al. 2007; Nogués-Bravo et al. 2007). Since normal growing-season (May through August) temperatures for the species currently average 9.7 °C (10th percentile = 7.2 °C; 90th percentile = 11.9 °C; Wang et al. in prep.), this corresponds to mean normal growing-season temperatures experienced by the species increasing to as high as 14.2 °C. In order to evaluate the potential for whitebark pine to persist within its current range through phenotypic plasticity as the climate changes, we tested the extent of whitebark pine‟s fundamental climatic niche, as well as the temperature at which a lower-elevation competitor begins to dominate. To assess the species‟ adaptation capacity, we calculated the degree of genetic differentiation among populations and the heritabilities of traits of interest for multiple populations. We conducted these assessments by germinating seed and growing seedlings from multiple whitebark pine populations in growth chambers programmed to simulate growingseason temperatures ranging from levels currently observed within the species range to levels ~4 °C above those predicted by the end of the 21st century. Such high temperatures were used because growth-chamber experiments identify growth limits at higher temperatures than those observed in nature due to the absence of compounding stressors such as vegetative competition, desiccation, insect and disease damage, and disturbance by wind, snow or animals (P. Smets and S.N. Aitken unpublished data). Likewise, seedlings were used because seed germination and early seedling survival are the life stages that may be most strongly impacted by climate change for trees (Smith et al. 2003). Seeds and mature trees are relatively tolerant of diverse environmental conditions, whereas seedlings are highly vulnerable to heat, desiccation and cold. Our objectives were to: 1) compare the growth of whitebark vs. lodgepole pine (Pinus contorta var. latifolia Engelm.) seedlings growing across an extreme temperature gradient; 2) predict whitebark pine germination, survival and growth using growing-season temperature, provenance (i.e. seed source) climate variables, and seed weight; and 3) assess trait heritability and genetic differentiation among populations.  3.2 Methods 3.2.1 Sample materials and experimental design We collected open-pollinated whitebark pine seeds from ten seed parents in each of six provenances during August and September 2007 (McLane and Aitken under review; Figure 3.1). The provenances were selected to represent a wide geographic gradient within the northwestern portion of the species range (Appendix 3). The seeds were manually extracted and treated using 33  the protocol of McLane and Aitken (under review). Treatment consisted of 1) x-raying the seeds to assess embryo development, 2) maturing the seeds under warm, moist conditions for one month, 3) cold stratifying the seeds for three months, and 3) nicking seed-coat tissue off of the radicle end of each seed immediately prior to planting. Numbers of non-viable seeds were recorded prior to being discarded at all stages of the treatment process. In October 2008, we sowed the seeds in Superblock 60 cell 220mL capacity Styroblocks (Beaver Plastics Ltd., Acheson, Alberta). Styrofoam containers were used to provide some root insulation because whitebark pine had previously been found to be sensitive to root warming (Bower and Aitken 2008). The seeds were sown in a completely randomized design with replicate cells representing each family within each of five chambers. Available seed for each family were distributed equally among chambers, but total seeds varied by family. The number of seeds per family sown per cell was determined individually for each family based on previous germination results, with the intention of maximizing the probability of one seed germinating per cell. To compare growth rates between whitebark pine and a competitor, we sowed one separate Styroblock of lodgepole pine seeds per chamber. Equal numbers of lodgepole pine seeds from six bulked seedlots representing whitebark pine‟s geographic span within British Columbia were sown in each chamber (Appendix 3). For both species, when multiple seeds germinated per cell, the seedling farthest from the centre of the cell was transplanted into a new cell within the same chamber or removed. This occurred at the beginning of the first growing season for lodgepole and at the end of the first growing season for whitebark, before duplicate seedlings would have been in competition. Immediately following sowing, we placed the Styroblocks in the five growth chambers (Conviron, Winnipeg, Manitoba, models E-15 and PGV-35) and began the programmed temperature regimes. We designed the regimes to mimic daily and weekly cycles for the 17week period spanning from mid-May to early-September. Average temperatures ranged from 7.9 in the coldest chamber to 17.9 ºC in the hottest chamber, with the cold chamber representing temperatures within the current whitebark pine species range and the hot chamber representing an extreme warming scenario (Table 3.1). We subdivided each week of the growing season into two phases: three days with greater temperature variance representing sunny conditions, followed by four days with lesser temperature variance representing cloudy conditions. Each day, temperatures were ramped on an 34  hourly basis from minimum temperatures at 1:00 am to maximum temperatures at 1:00 pm. Temperatures were regulated by internal thermometers and verified by two DS1921G Thermachron iButtons (Dallas Semiconductor) per chamber, one on the soil surface and one buried 3 cm deep. A seasonally fluctuating photoperiod representing the average weekly daylight available at 52.5 ºN, which is approximately the middle of the current whitebark pine species range within British Columbia, was maintained in all chambers. In January 2009, the first growing season ended and we set the chambers to 4 ºC with lights off for 5 weeks to mimic winter dormancy (A.D. Bower, unpublished data). In March 2009, we initiated the second growing season using the same chambers, temperature regimes and protocols as the first growing season. Throughout the experiment, seedlings were well watered in all chambers, and Styroblocks were rotated bi-weekly to minimize intra-chamber environmental variation. In some chambers, nighttime temperatures were extended during the first third of each growing season to maintain desired weekly temperature averages while accommodating minimum allowable temperatures of 4 ºC in the E-15 chambers. Due to a mechanical problem with the 10.8 ºC chamber, the Styroblocks from that chamber were placed in the 13.0 ºC chamber for the first 11 days of the first growing season; no seedlings emerged during that time. We measured quantitative traits for the whitebark pine seedlings throughout the experiment, including germination capacity, days to germination, survival of germinants, days to second-year needle flush, final height increment, final number of needle fascicles, and dry biomass traits including root mass, shoot mass (stem plus needle mass), root:shoot ratio and total mass. Germination capacity was calculated based on the total number of seeds sown using data from the end of the first growing season, prior to the elimination of all but one seedling per cell. Very few (<5) additional seedlings germinated during the second growing season. Final height and survival were also recorded for the lodgepole pine seedlings. Explanatory variables included: normal (1971 – 2000) provenance mean annual temperature (MAT) and precipitation as snow (PAS), estimated using ClimateWNA v4 (an extension of ClimateBC (Wang et al. 2006b)); average chamber growing-season temperature; and average family seed weight. PAS was used instead of mean annual precipitation (MAP) because most precipitation in whitebark pine environments falls as snow (Weaver 2001). This is reflected in the high correlation (r = 0.92) between PAS and MAP for the six provenances used.  35  3.2.2 Data analysis 3.2.2.1 Germination, survival and growth We performed all analyses using SAS software, version 9.3 (SAS Institute Inc. 2010). Final height data were compared between the whitebark and lodgepole pine seedlings using PROC GLM. Additional regressions were performed to assess the effects of seed weight and provenance temperature and precipitation as snow on the whitebark pine quantitative traits. One set of regressions was performed among chambers with chamber temperature and temperature squared included as explanatory variables. The other set of regressions was performed by chamber. For both sets of analyses, multinomial logistic models based on cumulative logistic models fitted using PROC LOGISTIC were used for germination capacity and survival, while general linear models fitted using PROC GLM were used for the continuous traits, as per the methods described in Chapter 2 of this dissertation.  3.2.2.2 Heritability and population differentiation The proportion of phenotypic variation attributable to additive genetic variation (heritability or h2) and the proportion of genetic variation due to among-population differences (QST) were estimated from variance components derived using a mixed model: Yijklm = µ + ci + b(c)ij + pk + f(p)kl + cpik + eijklm  Eq. 3.1  where Yijklm is the observed value for seedling m in family l within population k in block j within chamber i, ci is the effect of chamber i, b(c)ij is the effect of block j within chamber i, pk is the effect of population k, f(p)kl is the effect of family l within population k, cpik is the interaction between chamber i and population k, and eijklm is the error. Chamber was considered a fixed effect while population, family within population, and block within chamber were considered random effects. Germination potential was analyzed using a logistic model fitted using PROC GLIMMIX with a negative binomial distribution and a logit link function. Error variance was standardized to π2/3 for this model (Yanchuk et al. 2008). The continuous traits were analyzed using general linear mixed models fitted using PROC MIXED. Heritability and QST were estimated using the methods of Bower and Aitken (2006, 2008). Within-population additive genetic variance was estimated as three, rather than four times the family variance, because whitebark pine openpollinated progeny are more closely related than true half-sibs due to inbreeding and correlated paternity (Krakowski et al. 2003; Bower and Aitken 2006). Genetic parameters were not 36  calculated for days to germination and survival, due to insufficient variance given the unbalanced data set.  3.3 Results The lodgepole pine seedlings grew significantly taller than the whitebark pine seedlings at the three highest temperatures (p < 0.001 for all) and significantly shorter at the coldest temperature (p < 0.0001; Figure 3.2a). Height was equivalent for the two species at the secondcoldest temperature. Average chamber growing-season temperature accounted for the majority of the explained variance for all whitebark pine traits except germination capacity when data were analyzed with chambers combined (Table 3.2a). Days to germination and needle flush decreased with increasing growing season temperature, while survival, height, root mass, shoot mass, root:shoot ratio and total biomass increased, and number of fascicles demonstrated a parabolic trend, increasing at lower temperatures but decreasing at the highest temperatures (Figure 3.2). Levels of explained variance (R2) were low to moderate for these traits (0.03 ≤ R2 ≤ 0.41). Germination capacity was greater among populations from warmer provenances. Heavier seed weight lead to greater root, shoot and total mass (Table 3.2a). When population variation was analyzed within chambers, the largest number of traits varied significantly among populations growing at the coldest growing-season temperature, while the fewest varied for populations growing at the middle temperature (Table 3.1). Of the traits that varied significantly, variance was poorly to moderately explained by the models in all chambers (0.06 ≤ R2 ≤ 0.21). Greater germination capacity and survival, earlier flushing and greater growth were primarily associated with warmer provenance temperatures and lower provenance precipitation as snow. Raw germination and survival data are summarized by population and chamber for whitebark pine and survival data are summarized by chamber for lodgepole pine in Appendix 4. Heritability was moderate for all analyzed traits (0.20 < h2 < 0.41; Table 3.2b). Population differentiation was moderate for germination capacity (QST = 0.15) and low for all other traits (0.03 < QST < 0.09).  37  3.4 Discussion Our data demonstrate that whitebark pine has a broad fundamental niche, with the genetic capacity to germinate and survive across a wide range of temperatures. Furthermore, the species has the capacity to germinate earlier and grow faster at warmer growing-season temperatures up to very high temperature levels. However, lodgepole pine was able to outgrow whitebark pine at all but the coldest growing-season temperatures. This bodes poorly for whitebark pine‟s ability to persist within its current ecological niche as the climate warms. Our results corroborate previous findings that whitebark pine is limited by competition rather than heat at its lower-elevation range margin. Natural stands of whitebark pine have been observed to grow well in 28 °C average (39 °C maximum) July temperatures (Weaver 1994). In laboratory settings, two-year old seedlings have been shown to photosynthesize optimally at 20 °C, and to optimize root growth at 30 °C (Jacobs and Weaver 1990; but see Bower and Aitken 2008). Whitebark pine is also not limited by cold temperatures at its upper-elevation range margin, but rather is thought to be limited by short growing seasons (< 100 days) caused by extended snow persistence (Ogilvie 1990). Ecosystems uphill from whitebark pine‟s niche are typically dominated by alpine plants living in marginal soils at the lower margin, and rocky ridges with little to no soil development at the upper margin. The fear is that whitebark pine may get “squeezed” out of its narrow ecological niche as temperatures rise. Lodgepole pine, subalpine fir (Abies lasiocarpa), Engelmann spruce (Picea engelmannii), and mountain hemlock (Tsuga mertensiana) all presently co-occur with whitebark pine at lower elevations (Ogilvie 1990, Weaver 2001). Since we demonstrated that lodgepole pine seedlings have the genetic capacity to grow faster than whitebark pine seedlings under equivalent increased-temperature levels, it is likely that lodgepole pine will increasingly outcompete whitebark pine at their ecotonal boundary as warming proceeds. We expect similar competitive dominance by other trees that co-occur with whitebark pine at its downhill range margin (Peterson and Peterson 2001; Ettl and Peterson 2006). At higher elevations, the duration of snow cover will likely decline with climatic warming (Christensen et al. 2007). However, true alpine areas are not expected to become habitable for whitebark pine for centuries, due to the extremely slow pace of soil development in such locations (Körner 2003). The realized niche of whitebark pine will therefore diminish in size as competitors encroach from lower elevations while terrain remains uninhabitable at higher elevations.  38  In our study, populations from warmer provenances germinated earlier, had a better chance of surviving, and grew larger than those from colder provenances. This result is in keeping with past reports of clinal population differentiation by Mahalovich et al. (2006), Bower and Aitken (2008), and M. Warwell (USDA Forest Service, pers. comm.). However, the proportion of total phenotypic variance explained by provenance climate was low. Likewise, population differentiation values were low for our growth traits (0.03 < QST < 0.08), similar to the findings of Bower and Aitken (2008) who grew whitebark pine seedlings from 48 provenances in common gardens in Vancouver, British Columbia. Previous estimates of neutral genetic variation in whitebark pine are virtually identical to these values (0.02 < FST < 0.09; Bower et al. in press). Bower and Aitken found the timing of bud break to be under strong genetic control (QST = 0.47), while we did not (QST = 0.09). Thus, our results indicate that whitebark pine has no more population differentiation for growth traits and needle flush than for selectively-neutral genetic markers. Our moderate heritability estimates (0.20 < h2 < 0.41) demonstrate that the traits we recorded have the genetic potential to respond to selection, but our QST values suggest that there is little divergent selection acting on the traits to drive local adaptation (Skelly et al. 2007). The reason for this could be that most whitebark pine populations are under stabilizing selective pressure regardless of geographic location, because virtually all populations are confined to a narrow elevational and functional niche between competitors downslope and inhospitable alpine environments upslope. This is reflected in the relatively small mean growing-season (May through August) temperature (7.2 °C 10th percentile to 11.9 °C 90th percentile) and total growing-season precipitation (153 mm 10th percentile to 380 mm 90th percentile) range recorded for the species (Wang et al. in prep.). However, our results should be interpreted cautiously since we used few populations. Our population differentiation estimate for germination potential (QST = 0.15) was higher than the estimates for the growth traits. This estimate was impacted by the aforementioned trend that populations from warmer provenances have greater germination potential than those from colder provenances. However, it was likely influenced by maternal effects as well. Maternal effects are characteristics of an offspring‟s phenotype that are attributed to the mother‟s phenotype or environment rather than the offspring‟s genotype or environment. Seed weight was an imperfect proxy for maternal effects in this study, only partially capturing the seeddevelopment differences that we saw in the seed x-rays. We postulate that the seed-development differences were largely a factor of poor weather conditions in the year of seed maturation, i.e. 39  that they are better classified as maternal than genetic effects (McLane and Aitken under review). The magnitude of the influence of maternal effects for whitebark pine could be examined by assessing germination potential for families from multiple provenances collected over multiple years. Quantitative traits were better explained by the predictive variables at the coldest growing-season temperature than at middle and high temperatures, based on the within-chamber predictive models. The strong impact of provenance climate in the coldest chamber compliments a tree-ring study in a well-established lodgepole pine provenance trial showing that differences in climatic sensitivity are strongest among populations growing in cold trial locations (McLane and Aitken under review). The whitebark pine seedlings in the coldest chamber had delayed germination and needle flush compared with the seedlings in other chambers, as well. These trends were likely influenced by the sub-zero night time temperatures in the coldest chamber (2.0 °C), which were well below the minimum temperatures achieved in the other chambers (4 °C). Longer cold periods were imposed at night in the latter chambers to create the overall average among-chamber temperature gradient, but it appears that the sub-zero temperatures in the coldest chamber caused phenological effects beyond those induced by average temperatures alone. Major advances in germination and bud-burst timing could therefore occur in natural whitebark stands, given that night-time temperatures are predicted to increase more than daytime temperatures under climate change (Christensen et al. 2007). Also notable was that precipitation as snow had a negative influence on root mass and root:shoot ratio in some chambers. This meets expectations that populations from drier locations are adapted to allocate more resources towards root growth than those from wetter locations (Mokany et al. 2006). We were surprised that growth and survival did not decline for the whitebark pine seedlings in the warmest chamber, given that temperatures of 41.9 °C were reached. Withstanding such temperatures may be a physiological adaptation to the high solar radiation in the subalpine and alpine environments most commonly inhabited by whitebark pine, where soilsurface temperatures can reach 40 - 50 °C (Körner 2003; S.C. McLane pers. obs.). However, lodgepole pine withstands high soil-surface temperatures in its native range, too, and yet the lodgepole pine seedlings experienced comparatively poor survival and height growth in the hottest chamber. The lodgepole pine seedlings accrued far more biomass, and consequently desiccated their soil water reserves faster than the whitebark pine seedlings, in that chamber. Had the whitebark pine seedlings been equivalent in size to the lodgepole pine seedlings, we 40  believe they also would have experienced survival and growth declines due to heat-induced drought stress in the hottest chamber.  3.5 Conclusions It appears that whitebark pine will fare poorly as global temperatures rise, but not directly due to heat stress. The species is capable of responding positively to average temperatures ~4 °C higher than those seen within its current range, by germinating and breaking bud earlier and growing larger shoots and roots than it does at colder temperatures. However, lodgepole pine seedlings are able to outgrow whitebark pine seedlings under increased temperatures. If whitebark pine is able to persist within its current range despite increased competition, then it may fare better than species distribution models predict. However, phenotypic plasticity may fail the species in the long run given that temperatures are predicted to continue rising beyond the 21st century, regardless of emissions reductions. This leaves migration and adaptation as the remaining options for the species. Our finding of little genetic differentiation among populations for the traits we measured, coupled with the predicted speed of warming relative to the reproductive cycle of whitebark pine, largely negates natural or human-assisted adaptation options. Natural migration is not expected to occur quickly either, leaving assisted migration as a potential option of last resort (Chapter 2; McLane and Aitken under review). This subject needs to be more broadly discussed, and taxon-specific decision-making frameworks established, in the near future.  41  Table 3.1: Regressions predicting whitebark pine germination, survival, health, height and fascicles relative to provenance climate and seed development by chamber. Trait  7.9 (-2.0 to 31.9) F Sig vars R²  Chamber temperature (mean and range; °C) 10.8 (4.0 to 34.4) 13.0 (4.0 to 36.9) 15.4 (4.0 to 39.4) F Sig vars R² F Sig vars R² F Sig vars R²  Germination (%) a 19.90 pMAT, pPAS 0.09 20.18 Germination (days) Survival (%) a Needle flush (days) 13.59 Height (mm) 9.28  pMAT 0.11 pMAT 0.08  Fascicles (#) b  pMAT 0.09  Root mass (g)  10.79 c  pMAT 0.08 11.81  17.9 (4.0 to 41.9) F Sig vars R²  pMAT 0.05 27.22 pMAT, pPAS 0.09 34.20  8.89  pMAT 0.14  pMAT 0.06 6.61  pMAT 0.10  7.01  pMAT 0.06  8.78 pPAS , pMAT 0.14  Shoot mass (g) c Root:shoot ratioc  33.65  pPAS 0.21  9.37 pMAT , pPAS 0.08  Total mass (g) c Notes: Only significant (p < 0.05) variables (Sig vars) are listed. Variables with negative slopes are bolded and italicized. F = F-value; pMAT = provenance mean annual temperature; pPAS = provenance precipitation as snow; weight = seed weight; b  c  Wald chi-square and R² values are max-rescaled; square root transformed; log transformed.  42  a  test statistic is  Table 3.2: Statistical tests for whitebark pine seedlings a) predicting quantitative traits relative to growing-season temperature, provenance climate and seed weight; and b) estimating heritability (h²) and genetic differentiation (QST). b) Genetics tests h² QST  a) Prediction models  Trait  F-value Significant variables (p < 0.05)  R²  Germination (%) a Germination (days)  99.08 pMAT, pPAS 153.72 ctemp , ctemp², pMAT  0.08 0.40  Survival (%) a Needle flush (days) Height (mm)  6.67 ctemp ² , ctemp 117.51 ctemp , ctemp², pMAP 43.33 ctemp², pMAT  0.03 0.41 0.13  0.20 0.27  0.09 0.07  Fascicles (#) b Root mass (g)  25.56 ctemp, pMAT, ctemp² c  Shoot mass (g)  c  Root:shoot ratio  c  N/A  0.15  N/A  0.11  0.32  0.06  87.17 ctemp, ctemp² , pMAT, weight  c  0.32  0.38  0.04  22.14 ctemp, pMAT, ctemp² , weight  c  0.13  0.29  0.08  0.36  0.41  0.03  111.49 ctemp, pPAS  c  c  Total mass (g) 41.85 ctemp, ctemp² , pMAT, weight 0.22 0.33 0.03 Notes : . Only significant (p < 0.05) variables are listed, those with negative slopes bolded and italicized. ctemp = average chamber growing-season temperature; pMAT = provenance normal mean annual temperature; pPAS = normal provenance precipitation as snow; weight = seed weight; b  a  test statistic is  Wald chi-square and R² values are max-rescaled; square root transformed; c log transformed.  43  Figure 3.1: Whitebark pine provenance locations.  44  Figure 3.2: Overall mean quantitative trait values for whitebark pine seedlings growing in five controlled-environment growth chambers representing growing-season temperatures ranging from average observed temperatures for the species to extreme climate change scenarios. Error bars depict standard errors.  45  Chapter 4: Modeling lodgepole pine (Pinus contorta) radial growth relative to climate and genetics using universal growth-trend response functions2 4.1 Introduction Climate change will profoundly impact the role of forests as carbon sinks by inducing changes in growth, phenology and mortality (Saxe et al. 2001; Aitken et al. 2008; Allen et al. 2010). As temperatures rise, secondary growth in trees is predicted to increase at the upper latitudinal and elevational boundaries of species ranges where cold temperatures tend to limit growth (Peltola et al. 2002; Stromgren and Linder 2002), but to decline at southern range margins and other locations where high temperatures induce heat stress (Barber et al. 2000; Adams and Kolb 2005; Reich and Oleksyn 2008). Overall, net increases in forest productivity are predicted (Boisvenue and Running 2006); however, due to local adaptation, populations are expected to respond to temperature increases in different ways. Models that account for population differentiation warn that maladaptation, or lack of genetic variation in individual populations, could compromise the competitive ability of some populations and species (O‟Neill et al. 2008; Bell and Gonzales 2009; Pautasso 2009). Forest geneticists initially established long-term provenance trials to find populations for reforestation that would optimize wood productivity (Savolainen et al. 2007). „Provenance‟ in this context refers to the geographic and climatic origin of a population, while „population‟ refers to the trees grown from seeds sampled at the provenance and grown in multiple sites. In recent years, data from some older provenance experiments have been reanalyzed to predict how populations may respond differentially to climate change (Rehfeldt et al. 1999, 2001, 2002; Andalo et al. 2005; Chuine et al. 2006; Wang et al. 2006a; O‟Neill et al. 2008; Reich and Oleksyn 2008; Wang et al. 2010). Most of these studies use the spectrum of climatic conditions available across multiple sites to model yields – usually cumulative height or diameter growth and their derivatives such as stem volume – relative to normal site and provenance climate variables. A common output from such studies is a quantitative genetic model called a population response function, which is a type of reaction norm illustrating how the yields of 2  McLane S. C., V. M. LeMay, and S. N. Aitken. In Press. Modeling lodgepole pine radial growth relative to climate and genetics using universal growth-trend response functions. Ecological Applications. 46  individual populations at a particular age vary relative to the spectrum of climatic conditions represented among sites. Wang et al. (2010) recently developed a complimentary model, called a universal response function (URF), which integrates the climatic effects of both sites and provenances into a single, multi-dimensional model. URFs generate smoothed growth-climate relationships using the sampled populations and sites, thereby generating estimated performances for all hypothetical population and site combinations. Population and universal response functions created using cumulative height and volume growth provide a picture of how yields vary among populations growing across a spectrum of climatic conditions. However, the picture is only a snapshot representing accumulated growth at the time of measurement. Tree growth is strongly age-dependent, generally consisting of rapid, early elongation and biomass accumulation followed by a slower-growth period during maturation (Husch et al. 1972). At the junction of these two stages is a growth maximum, the magnitude and timing of which is influenced by climatic and genetic factors in addition to age. Cumulative growth measurements mask these growth-rate trends, which can be more indicative of long-term productivity than yields. Recent evaluations of lodgepole pine (Pinus contorta ssp. latifolia Engelm.) stand declines in British Columbia (BC), for example, have called into question long-standing assumptions that height and health at a given age provide an adequate index of long-term health and productivity (Heineman et al. 2010; Mather et al. 2010). In this study, we create universal “growth-trend” response functions (URFs) using radial increment data from trees growing in the extensive Illingworth lodgepole pine provenance trial located in BC and Yukon, Canada (Illingworth 1978). The sampled trees, 34 years old at the time of coring, represent 12 populations growing at 16 sites, with the increment data converted to basal area increments (BAIs) prior to modeling. We create the URFs using random-coefficient models (Littell et al. 2006; also called „parameter prediction models‟ in some texts), which allow site and provenance climate variables associated with the sampled trees to predict changes in coefficients of non-linear models that generate smoothed estimates of annual radial growth over time. The models output radial growth estimates for trees from any provenance at any age, growing in any present or future time period within reason, relative to growth-year (previous August through current July), summer, or winter climate normal measures. Previous studies evaluating the influence of genetics on radial growth using provenance-trial trees observed multiple populations at only one site and focused on average (Savva et al. 2007) or residual  47  (Oleksyn et al. 1998; Savva et al. 2002, 2008) radial growth indices rather than growth trends over time. We use the URFs to derive the estimated ages and BAIs at which radial growth is maximized for a spectrum of populations and sites relative to growth-year, summer, and winter climate periods under present and future-climate scenarios. We then use the predictions to forecast regionalized radial-growth responses to climate change throughout the 21st century. Over half of the Illingworth sites have been severely affected by mountain pine beetle (Dendroctonus ponderosae) and it is expected that most will be decimated within the next few years. Our study is therefore timely both in the context of climate change and in terms of taking full advantage of the valuable Illingworth experiment.  4.2 Methods 4.2.1 Plant materials The Illingworth provenance trial is a BC Ministry of Forests, Mines and Lands experiment comprising 153 lodgepole pine populations reciprocally planted and maintained at 60 sites throughout the interior of BC and two sites in Yukon (Illingworth 1978). The trees were planted in 1974 as three-year-old seedlings from bulked seed collections in which numerous parent trees represented each provenance. Not all populations were planted at all sites, because the original primary research objective was finding populations that would maximize yield under current climates, and it was known that populations planted far from their place of origin were unlikely to be optimal performers. Two randomized complete blocks were established at each site. Within each block, nine seedlings per population were planted in three-by-three tree plots. All trees were planted 2.5 m apart. The experimental design of the Illingworth trial was integral to our study. Uniform tree spacing minimized confounding growth trends associated with uneven stand density. Even-aged trees allowed for growth-trend comparisons over a standardized time period, and rendered intraspecific competition for light and other resources relatively consistent for all populations and sites through time. Competing vegetation was controlled in the early years of the study, but no stand thinning was performed. Principle components analyses of temperature and precipitation data and geographic information system maps were used to select 16 sites and 12 populations representing broad climatic and geographic gradients (Figure 4.1, Table 4.1). Survey data from 2005 were used to 48  verify that damage by mountain pine beetle and other pests and pathogens was minimal, and that there was adequate survivorship of the selected populations, at the selected sites. In 2006, increment cores were sampled from up to six of the nine trees per population in each block, depending on survival, with a preference for sampling trees with six live neighbours to maintain standardized stand density. Thus, up to 12 trees per population were cored at each of the 16 selected sites (Table 4.2). Trees were cored cross-slope between 30 and 40 cm above ground using a 5 mm increment borer.  4.2.2 Sample preparation Cores were dried, mounted and sanded following Stokes and Smiley (1968) prior to being measured using a WinDENDRO digital (1200 dpi resolution) tree-ring image processor (Régent Instruments, Quebec). Ring widths were recorded at a resolution of 0.001 mm. The ring-width series for every tree was visually cross-dated using the graphical output provided by WinDENDRO and then statistically cross-dated and quality checked using the International Tree-Ring Data Bank Library (ITRDBL) software program COFECHA (Holmes 1986). COFECHA was programmed to correlate each ring-width series against the other series from the same population and site using 20-year segments with a five-year lag. Annual basal area increments (BAIs) were calculated from each ring-width series as: BAI 2  = π(R n - R2n-1), where R is the tree radius obtained by summing ring widths from the innermost to the nth ring. BAIs account for tree circumference increasing over time by representing the two-dimensional cross-sectional area of each tree ring at the height of coring. BAIs thereby provide a more accurate index of growth rate, represented by structural carbohydrate allocation, than can be attained using ring widths. Although some of the cores had rings dating into the mid-1970s, all BAIs prior to 1981, corresponding to age 10, were excluded from the final data set to remove establishment-related growth effects. Site-population combinations with fewer than six live trees were also eliminated from the analyses. The final data set consisted of data from 1,246 trees representing 110 site-population combinations (Table 4.2).  4.2.3 Climate data Climate data for the 16 sites and 12 provenances were estimated using ClimateWNA (an extended version of ClimateBC (Wang et al. 2006b) for western North America, accessible at www.genetics.forestry.ubc.ca/cfcg/ClimateWNA/ClimateWNA.html) (Table 4.1). ClimateWNA 49  interpolates weather station data using high-resolution digital-elevation models that accurately capture climatic variance in western North America‟s mountainous terrain. Current normal (1971-2000) temperature and precipitation values were calculated for each of the provenances. For each site, monthly data were averaged to create biologically-relevant growth-year (previous August through current July), summer (current June, July and August) and winter (previous December through current February) climate variables. These data were averaged for the years 1981 – 2005 to create normal growth-year, summer and winter climate variables for the study period, hereafter referred to as the „1990s‟ time period. Future-predicted growth-year, summer and winter climate measures were also estimated for each site using the A1B (integrated world, balanced energy sources) ensemble mean scenario of the third-generation Coupled Global Climate Model, developed by the Canadian Centre for Climate Modeling and Analysis (Flato et al. 2000) (Table 4.3). The future scenarios represent 2011-2040, 2041-2070, and 2071-2100 climate-normals, referred to by their midpoints as the „2020s‟, „2050s‟ and „2080s‟ time periods. To pictorially and numerically display the impacts of diverse site and provenance climates on growth trends, the site and provenance temperature and precipitation data were averaged to create 10th, 50th, and 90th percentile temperature and precipitation levels for use in some graphs and tables (Table 4.3). The temperature levels are referred to as „cold‟, „cool‟, and „warm‟, with cold roughly encompassing sites and provenances in northern BC and Yukon, cool higher-elevation and north-central BC sites and provenances, and warm lower-elevation and southern BC sites and provenances. The words cool and warm were used instead of „moderate‟ and „hot‟ because the sampled sites represent only the central and northern portions of lodgepole pine‟s range, and thus do not represent sites and provenances from truly hot locations. Precipitation levels are referred to as „dry‟, „moist‟ and „wet‟.  4.2.4 Universal growth-trend response functions Universal growth-trend response functions (URFs) were created using a randomcoefficient model with normal site and provenance climate and tree age as input variables. Three separate models were created using growth-year, summer and winter site climate variables, respectively. Each random-coefficient model was fitted using three steps. First, least-squares parameter estimates for the age-related BAI growth trends were obtained for each of the 112 site i – provenance j combinations using a base model of growth over time: 50  Ykt  a  X kt  c X kt   kt b  Eq. 4.1  where Ykt is the BAI for tree k at age t; Xkt is tree age; a, b and c are parameters to be estimated, with a ≥ 0, b > 0 and 0 < c ≤ 1; and kt is the residual. A Gaussian search method was used to minimize the sums of squared errors. Second, the parameter estimates from Eq. 4.1 for each site and provenance were regressed against site and provenance temperature and precipitation values to generate starting coefficients for the random-coefficient model:  aij  0  1  tempi  2  precipi  3  tempj  4  precip j   ij  f1   ij  Eq. 4.2  bij  0  1  tempi  2  precipi  3  tempj  4  precip j   ij  f 2   ij  Eq.  cij  0  1  tempi  2  precipi  3  tempj  4  precip j   ij  f3   ij  Eq. 4.4  4.3  where aij, bij and cij are the estimated parameters from Eq. 4.1 for site i and provenance j; s, βs and s are parameters to be estimated; tempi , precipi, tempj , and precipj are normal temperature and precipitation for site i and provenance j; and ij are the residuals. Six of the 112 siteprovenance combinations for which Eq. 4.1 did not converge were excluded from fitting Eq. 4.2 through 4.4, but were included in the final fit of the random-coefficient model. Finally, the parameters of Eq. 4.1 were replaced by Eq. 4.2 through 4.4 in the randomcoefficient model to obtain predicted BAIs for trees representing all site and provenance climate values:  Yijkt  f1  X ijkt 2  f 3 f  X ijkt    ijkt  Eq. 4.5  where Yijkt is the BAI for tree k in population j at site i at age t; Xijkt is age of tree k in population j at site i at age t; f1 to f3 correspond to Eq. 4.2 through 4.4, respectively; and ijkt is the residual. First-order autocorrelation and heteroskedasticity were accounted for in the model fit. A Marquardt search method was used to minimize the sums of squared errors. Residuals plots were used to check for any remaining autocorrelation and heteroskedasticity in the final models and were found to be acceptable.  4.2.5 Impacts of present and future climate on radial growth The URFs were used to estimate the impacts of site and provenance climates on the growth trend (i.e. BAI over age), as well as the value (maxBAI) and age at which BAIs 51  maximize for each population. MaxBAIs and the ages at which they occur were estimated for each population in each time period (1990s, 2020s, 2050s, 2080s) within each site-climate level (cold, cool, warm) using each set of climate variables (growth-year, summer, winter). All statistical analyses were conducted using SAS software version 9.2 (SAS Institute Inc. 2008).  4.3 Results Universal growth-trend response functions were obtained using the following input variables: age; normal provenance temperature and precipitation; and present and future normal growth-year, summer and winter site temperature and precipitation. The functions predict annual radial growth for trees representing any chosen subset of the input variables. The models explained 64%, 64% and 63% of the total observed variation in BAI (n = 29,312) using growthyear, summer, and winter site climate variables, respectively. Each model is six-dimensional, making it impossible to visually illustrate all dimensions simultaneously. To accommodate this, an array of graphics was created, each highlighting particular climatic and genetic impacts on growth trends while holding other factors constant. Temporal radial-growth patterns vary relative to provenance temperature, site temperature, and site precipitation (Figure 4.2). Populations growing at warm sites have faster initial growth rates and faster growth declines following maxBAI than populations growing at cold sites, regardless of provenance. Cold and warm provenances growing in warm-wet sites have larger maxBAIs than their counterparts growing in warm-dry sites, with the warm provenances exceeding the cold provenances in overall maxBAI. In contrast, maxBAIs are larger at cold-dry than at cold-moist sites, with warm provenances dominating cold provenances at dry sites, but the opposite trend occurring at moist sites. In these assessments, moist rather than wet precipitation levels were used for the cold sites because a lack of data from cold-wet sites made BAI predictions for the cold-wet climate combination illogical. Additionally, provenance precipitation was held at the moist level because it was shown to negligibly impact the BAI predictions. MaxBAIs and the ages at which they occur for all of the sampled populations growing at cold, cool, and warm sites under present and future climate scenarios using growth-year, summer, and winter climate variables are summarized in Tables 4.3 and 4.4. Site precipitation was held at the moist level across the three site-temperature levels for consistency, but was allowed to vary by provenance and climate scenario. 52  Using the growth-year model, maxBAIs are predicted to be largest overall at cool sites in the 1990s (Table 4.4a). However, maxBAIs decrease 14% (from 7.7 to 6.6 cm2) at cool sites and 27% (from 6.7 to 4.9 cm2) at warm sites between the 1990s and the 2080s, while increasing 108% (from 3.9 to 8.1 cm2) at cold sites over the same time period. By the 2080s, predicted maxBAIs are largest overall at cold sites (2.0 °C normal temperature) with no evidence of declines within the 21st century. In contrast, maxBAIs begin declining by the 2050s (3.9 °C normal temperature) and the 2020s (5.1 °C normal temperature) at cool and warm sites, respectively. Using the summer model, maxBAIs are predicted to be largest at all sites in the 1990s, with the overall largest BAIs occurring at cold sites (Table 4.4b). Average maxBAIs decrease at all sites between the 1990s and the 2080s, with the severity of the decline increasing with site temperature from 17% (7.5 to 6.2 cm2) at cold sites to 30% (7.0 to 4.0 cm2) at cool sites and 33% (5.8 to 3.9 cm2) at warm sites. In the 2080s, the biggest predicted maxBAIs continue to occur at cold sites, but the size differential between warm and cold sites is larger (2.3 cm2) than in the 1990s (1.7 cm2). Thus, trees at warm sites are predicted to experience the greatest productivity losses relative to summer temperatures rising throughout the 21st century. Results from the winter model largely oppose those from the summer model, with the largest maxBAIs predicted to occur at all sites in the 2080s, and the overall largest maxBAIs occurring at warm sites (Table 4.3c). MaxBAIs increase at all sites between the 1990s and the 2080s, with the magnitude of the increase rising from 23% at cold (5.6 to 6.9 cm2) and cool (7.7 to 9.5 cm2) sites to 39% (8.3 to 11.5 cm2) at warm sites. Thus, trees at warm sites are predicted to experience the greatest productivity gains due to winter temperatures rising throughout the 21st century. Averaged among populations, the biggest maxBAI predicted in the winter model is 11.5 cm2, while the biggest in the summer model is 7.5 cm2, similar to the biggest growth-yearmodel maxBAI of 8.1 cm2 (Table 4.4). Trees from warmer provenances have larger predicted maxBAIs than trees from colder provenances at all present and future site temperatures, with the exception of cold sites in the 1990s under growth-year model scenarios (Table 4.4). Correspondingly, the variance in maxBAI size increases with provenance temperature at all site-climate levels. The ages at which BAIs maximize range from 18 to 28 years in the growth-year model, 17 to 26 years in the summer model, and 18 to 32 years in the winter model (Table 4.5). MaxBAI ages increase uniformly with provenance temperature, irrespective of site temperature 53  or time period, save for slight variations at higher provenance temperatures in the winter model. This trend corresponds with maxBAI sizes, which are consistently larger at higher provenance temperature levels. However, the differences between the ages at which BAIs maximize among populations are small (≈2 to 3 years between the coldest and warmest provenances, on average) compared to the differences in the sizes of maxBAIs (≈100% increase between the coldest and warmest provenances, on average) within any given time period. The ages at which maxBAIs occur decrease with increasing site temperatures among sites and time periods in all three models (Table 4.5). This trend is not in keeping with maxBAIs themselves, which decrease with increasing temperatures at all sites in the summer model, increase with increasing temperature at all sites in the winter model, and increase or decrease depending on the site temperature level in the growth-year model, over the same time periods. Universal growth-trend response functions depicting maxBAIs of populations from cold and warm provenances relative to normal 1990s site climates illustrate the impacts of site precipitation on maximum radial growth (Figure 4.3). Separate models were generated using growth-year, summer and winter site climate variables. In the growth-year model, maxBAIs increase with increasing precipitation at all temperatures, except for a slight decline at the highest precipitation levels at mid-temperature sites. In the summer model, maxBAIs also increase with increasing precipitation, but the biggest radial-growth maxima occur at middle temperatures across all precipitation levels. In contrast, maxBAIs are largest at moderate precipitation levels in the winter model, increasing with temperature across the full winter temperature spectrum. Due to the negligible effect of provenance precipitation on population performance, only moist provenance precipitation levels were used to generate these URFs (Figure 4.3).  4.4 Discussion We demonstrate that the influences of provenance and growing-site climate on annual radial growth can be effectively modeled using universal growth-trend response functions. Our technique builds on the foundation of the universal response function created by Wang et al. (2010), by incorporating another critical dimension: tree age. Using our model, we can predict annual radial growth for trees from any provenance at any age, growing in any present or future growth-year or seasonal climate scenario within reason. Our approach provides new depth to the quantitative genetics response-function technique by illustrating potential changes in population 54  growth and competitiveness over time, as well as the age and size at which annual growth begins declining among populations and growing sites. These temporal-growth indices lend different perspectives on productivity than those inferred from cumulative growth alone.  4.4.1 Model evaluation Universal response functions (URFs) provide smoothed estimates of how populations respond to climate by generating performance estimates for missing populations and sites based on among-population relationships within the available dataset. By the same token, URFs mask the genetic signals of individual populations that do not follow dominant climatic and geographic trends (geographic variables are not included in our models, but see Wang et al. 2010). If these populations are of specific conservation or wood production interest, or as O‟Neill et al. (2008) predict, do not have the genetic capacity to respond to climate change, they are likely to stand out in population response functions but to go unnoticed in URFs. However, URFs avoid the biologically-improbable growth predictions generated by PRFs when sampled populations are grown in an insufficient number of sites for trend characterization. These types of errors can lead to inaccurate portrayals of population growth increases or failures under climate change. URFs are equally dependent on the size and scope of data sets, but because the population growth responses are smoothed, the predictions are buffered against extremes. Random-coefficient models are elegant tools for generating universal growth-trend response functions. These models enable examinations of temporal growth dimensions relative to site and provenance climate by allowing the growth-trend input parameters to vary, thereby extending the smoothing advantages of URFs to trends over time. Our models capture nearly two-thirds of the variance in the full data set, which we consider a high level of accuracy given that radial growth is under weaker genetic control (i.e. lower heritability) than height in lodgepole pine (Wang et al. 1999), and that the strength of the relationship between tree rings and climate is dependent on how strongly growth is limited by site climate (Fritts 1976). The Illingworth trial sites are mostly located in moderate to high-productivity environments for lodgepole pine, so tree-ring growth is expected to be only moderately sensitive to climatic signals. Given the effectiveness of our models, we recommend that universal growth-trend response functions be used to analyze other provenance-trial data sets. We excluded nonclimatic data (e.g., soil type, photoperiod, geographic coordinates) to highlight the dominant influence of climate on growth and to avoid multicollinearity issues, but these variables, as well 55  as other phenotypic or genotypic traits (e.g., phenology, candidate gene polymorphism, and gene expression) could be incorporated in future studies. The summer and winter URFs in Figure 4.3 illustrate that irregular data trends may remain visible in URFs. The irregularities in the surfaces of both graphics are primarily caused by low growth for all populations at one site, M451. The response surfaces were generated by smoothing maxBAI predictions among locations with similar temperature and precipitation levels, thereby simultaneously extending and moderating the influences of anomalous sites such as M451. These kinds of aberrations should be considered cautiously, bearing in mind that such graphics may over-emphasize the influence of precipitation, which has small data ranges in both the summer (168 mm) and winter (322 mm) models. In Tables 4.3 and 4.4 where precipitation is fixed at moderate levels across site temperature levels, no such anomalies are predicted.  4.4.2 Applications of the model A lack of test sites warm enough to estimate temperatures beyond which growth begins to decline has frequently limited past studies predicting growth responses to climate change using provenance trial data (e.g., Wang et al. 2006a). Using the universal growth-trend response function developed in this study, we successfully forecast temperatures at which radial growth begins to decline for all populations and sites, rather than only for populations grown across the widest range of site climates. Using growth-year climate data, we predict that radial growth will reach overall maximal levels for all populations by the 2020s at warm sites and by the 2050s at cool sites, while continuing to increase beyond the 2080s at cold sites. This finding reinforces previous predictions that radial growth will generally increase in northern portions of species ranges, while declining in central and southern locations as global temperatures rise throughout the 21st century (Lapointe-Garant et al. 2009). The corresponding growth-year temperatures above which growth is predicted to decline, 3.9 °C at cool sites and 5.1 °C at warm sites, also roughly parallel Wang et al.‟s (2010) finding that height growth is maximized in lodgepole pine trees growing at 4.5 °C normal mean annual temperature. Our temperature estimates are approximate and the comparison is loose, since radial (secondary) and height (primary) growth are distinct processes occurring during different, although overlapping, seasonal time periods. Primary and secondary growth do not necessarily respond to temperature in unison, but nonetheless, the similarity of the results bolsters our confidence in both studies‟ estimates,  56  particularly because total wood volume, biomass, and carbon sequestration are functions of both types of growth. Radial growth in temperate and boreal conifers is often sensitive to summer drought and winter snow (Brooks et al. 1998; Yeh et al. 2000), and indeed, our radial-growth trends modeled using summer and winter climate variables yielded different growth-potential patterns than were generated using growth-year climate data. Our finding that radial growth is expected to decrease relative to hotter summers and increase relative to warmer winters across all site-temperature levels and time periods was not unto itself surprising. Of greater interest was that warmer winters appear to have a stronger positive influence on maximum radial growth than hotter summers have a negative influence. Should this prediction prove accurate, then growth increases due to warmer winters could outpace growth declines caused by hotter summers, rendering the growth declines forecasted by our growth-year climate model overestimates. However, we suggest interpreting the exact numerical predictions of the summer model with caution, because the range of summer temperatures and precipitation levels is small (4.5 ºC, 168 mm) compared to the ranges encompassed by both the growth-year (6.9 ºC, 1050 mm) and winter (14 ºC, 322 mm) models. Because of this compressed climate spectrum, the gradient of growth responses is less well defined in the summer model. These sorts of intraannual climate – growth interactions merit further investigation, as they will impact the role of Canada‟s forests as globally-significant carbon sinks (Kurz et al. 2008). Among populations, we found that the warmer the provenance, the greater the maximum radial growth level, except for populations growing in cold locations in present climates using growth-year climate variables. This matches previous findings that populations from the geographic center of the species range have the greatest growth potential across a wide spectrum of mean annual temperatures (Rehfeldt et al. 1999; Wang et al. 2006a, 2010). Unfortunately, we were not able to sample populations from the southern portion of the species range because relatively few were planted and survived in the Illingworth trials. Due to this data deficiency, we were not able to deduce provenance-climate levels at which radial growth is maximized among populations. We observed one population (Zigzag, 119) with provenance temperature 8.9 °C and precipitation 2518 mm, to have only mid-sized maxBAIs at most sampling sites, indicating that population growth rates may maximize between provenance temperatures 5.9 (provenance 1) and 8.9 °C. However, the large (3 °C) temperature difference between provenances 119 and 1, as  57  well as 119‟s anomalously high precipitation level, prompted us to exclude this population from the analysis. The impact of precipitation on radial growth was deemphasized in our study for a few reasons: provenance precipitation was found not to influence predicted BAIs; climate-modelpredicted precipitation levels are less accurate than temperature estimates; and changes in precipitation under climate change scenarios are relatively small and have higher variances among global circulation models and carbon dioxide scenarios than temperature predictions. However, we found that greater precipitation is associated with larger maxBAIs at all but the coldest temperatures in the growth-year model and at all temperatures in the summer models, while moderate precipitation levels are best for growth in the winter model. These trends make sense under the pretext that both extreme summer drought and excess winter snow limit radial growth. Age of BAI maximization decreased with increasing site temperature across temperature levels and time periods and in all seasons. This indicates that trees may reach their maximum radial-growth rates earlier as global temperatures rise, regardless of how much radial biomass they accumulate at their maxBAI levels. On the positive side, this could mean shorter harvest rotations for wood grown for bioenergy or carbon sequestration purposes. On the negative side, it could portend earlier growth failures for natural stands, which could increase the periodicity of wildfires and negatively impact wildlife that are obligate to later-seral stands.  4.5 Conclusions Predicting how trees will respond to temperature increases and other climatic shifts associated with climate change is an increasingly pressing priority. Forests cover ~30% of Earth‟s land surface and store ~45% of terrestrial carbon (Bonan 2008). Canada alone harbours ~7% of the world‟s total forest cover, and yet models indicate that Canada‟s forests are more likely to act as net carbon sources than sinks over the coming century due to heat-induced pest and pathogen outbreaks and associated wildfires (Kurz et al. 2008), as well as drought stress (Allen et al. 2010). Despite continuing advances in carbon-cycle modeling, feedback loops involving forested landscapes remain one of the largest uncertainties in predicting climateinduced environmental changes. In this study, we show that universal growth-trend response functions, created using random-coefficient models, successfully capture the genetic and agebased underpinnings of radial-growth responses to current climates, and facilitate predictions of 58  how trees will respond to future climates. We recommend that similar studies be performed in other provenance trials, and that the data be made publically available. An open-access growthtrend data set encompassing numerous biomes, species and provenances would contribute substantially to predicting forest productivity under future-climate scenarios. In future URF studies, particular emphasis should be placed on forecasting growth responses to intra-annual climate, because tree growth is not equally sensitive to climatic conditions over the course of a year, nor is climate change predicted to affect all seasons in the same manner. Expanding new provenance trials to include warmer test sites will also help improve the certainty of growthresponse models across the range of future-predicted climates.  59  Table 4.1: Geographic and climatic descriptions of sites and provenances sampled from the Illingworth lodgepole pine (Pinus contorta) experiment. Latitude Longitude Elevation Temperature (°C) (°N) (°W) (m) Ann Sum Win  Precipitation (mm) Ann Sum Win  Name Code Planting site Watson Lake WATS 60.08 128.83 700 -1.9 13.8 -19.0 396 155 77 Blue River BLUE 59.78 129.13 730 -0.9 13.2 -15.7 468 163 109 Mile 451 M451 58.83 125.72 1100 -0.6 11.2 -11.3 592 276 91 Whitehorse WHRS 60.80 135.18 663 -0.3 13.0 -13.5 303 123 59 Lussier River LUSS 49.80 115.50 1650 1.3 11.6 -8.9 1353 291 381 Salmon Lake SAMN 54.85 123.92 950 1.9 12.4 -8.6 671 172 172 Mons Lake MONS 51.67 123.00 1280 2.3 12.2 -7.8 362 150 66 McLatchie Creek MCLA 49.35 114.68 1550 2.3 12.6 -7.7 1015 239 278 Ootsa Lake OTSA 53.77 126.83 1040 2.7 12.0 -6.8 721 165 221 Bosk Lake BOSK 52.18 120.80 1100 2.9 12.6 -6.8 814 232 201 Chuwhels Lake CHUW 50.58 120.62 1430 3.4 13.1 -5.7 471 155 104 Suskwa River SUSK 55.32 127.27 640 3.5 13.2 -6.2 736 208 157 70 Mile MI70 51.28 121.33 1070 3.7 13.6 -6.4 361 144 64 Lassie Lake LASI 49.62 118.92 1370 3.7 13.4 -5.5 651 171 169 Cuisson Lake CUIS 52.50 122.38 850 4.7 14.6 -5.7 472 178 82 Wigwam WIGW 50.82 117.98 790 5.3 15.7 -5.0 1236 270 374 Provenance Lower Post 30 59.98 128.55 640 -2.3 432 Petitot River 163 59.90 122.08 396 -1.3 455 Atlin 35 59.80 133.78 789 0.1 361 Tower Lake 26 56.02 120.62 792 1.4 477 Nina Creek 100 55.80 124.82 762 1.6 521 Chilco 18 51.98 123.75 1059 2.5 328 Swan Hills 142 54.30 116.58 823 2.6 594 Fly Hills 71 50.72 119.45 1524 2.7 869 Marl Creek 44 51.52 117.18 945 3.4 622 Albreda 63 52.58 119.17 975 3.7 915 Trapping Creek 1 49.58 119.02 1006 4.7 568 Larch Hills 72 50.70 119.18 777 5.9 754 Notes: Sites and provenances are ordered from coldest to warmest normal growth-year (sites) or annual (provenances) (both abbreviated as Ann) temperature. Normal growth-year precipitation is also given, as well as normal summer (Sum) and winter (Win) site temperature and precipitation. We did not calculate summer and winter temperatures or precipitation levels for the provenances.  60  Table 4.2: Number of lodgepole pine trees per population sampled at each of the 16 sites. Population 30 163 35 26 100 18 142 71 44 63 1 72 WATS 12 11 11 12 BLUE 12 10 12 12 9 13 6 * M451 10 6 10 12 12 11 7 7 11 WHRS 12 12 11 12 LUSS 12 12 * 12 12 12 12 12 SAMN 12 10 12 12 * 12 11 11 * MONS 12 12 12 * 11 11 12 12 * MCLA 11 11 12 12 12 12 11 12 * OTSA 12 12 12 12 12 12 12 BOSK 10 12 * 12 12 12 12 * CHUW 12 11 12 12 12 12 12 * SUSK 10 6 12 12 12 11 12 11 10 MI70 10 * 12 6 12 11 12 12 LASI 12 11 11 * 12 12 12 12 CUIS 12 * 12 8 12 12 9 11 * WIGW * * 12 9 12 12 11 9 12 Notes: Asterisks (*) indicate locations where the study populations were planted but sample sizes were too low for analysis due to poor survival. Blank cells indicate populations were not planted at that site. Sites and populations are ordered from coldest to warmest normal growth-year (sites) or annual (provenances) temperature. Site  61  Table 4.3: Current-observed and future-predicted normal site, and current-observed normal provenance, temperature and precipitation represented by 10th, 50th and 90th percentile levels. Climate Time period variable  Sites  Provenances 10th % (cold, dry) 50th % (cool, moist) 90th % (warm, wet) 10th % 50th % 90th % 1990s 2020s 2050s 2080s 1990s 2020s 2050s 2080s 1990s 2020s 2050s 2080s 1990s  GrowthTemp year -0.7 0.2 1.1 2.2 2.5 3.0 3.9 4.8 4.2 5.1 6.1 6.9 -1.2 2.6 4.6 (°C) Summer 11.8 12.8 13.7 14.5 13.0 14.0 14.9 15.7 14.2 15.2 16.1 16.9 N/A Winter -14.6 -13.8 -12.6 -10.9 -8.0 -7.3 -6.7 -5.6 -5.6 -4.9 -3.5 -2.5 GrowthPrecip year 361 358 362 365 621 678 686 692 1125 1173 1187 1196 365 545 848 (mm) Summer 147 135 136 137 172 184 186 188 273 262 264 266 N/A Winter 65 80 82 82 133 192 195 197 326 398 404 407 Notes: Percentile levels (cold, cool and warm for temperature; dry, moist, wet for precipitation) were derived by combining the growth-year, summer and winter climate variables, respectively, for the 16 test sites and 12 provenances, respectively.  62  Table 4.4: Predicted maximum basal area increments (maxBAIs) for lodgepole pine populations grown at cold, cool and warm sites (i.e. 10th, 50th and 90th percentile site temperature levels) under present and future climate scenarios using (a) growth-year, (b) summer and (c) winter climate variables. Precipitation is held at the moist (i.e. 50th percentile) level across the three sitetemperature levels for consistency, but varies by provenance and climate scenario. The maxBAIs are shaded from light (small) to dark (large) to illustrate how climate differentially impacts BAIs among populations, sites, time periods, and climate models. Populations are ordered from coldest to warmest normal provenance temperature.  63  64  Table 4.5: Predicted ages at which basal area increments are maximized for 12 lodgepole pine populations grown at cold, cool and warm sites (i.e. 10th, 50th and 90th percentile site-temperature levels) under present and future climate scenarios using (a) growth-year, (b) summer and (c) winter climate variables. Precipitation is held at the moist (i.e. 50th percentile) level across the three site-temperature levels for consistency, but varies by provenance and climate scenario. The ages are shaded from light (small) to dark (large) to illustrate how climate differentially impacts BAIs among populations, sites, time periods, and climate models. Populations are ordered from coldest to warmest normal provenance temperature.  65  66  Figure 4.1: Locations of sites and provenances relative to the current lodgepole pine (Pinus contorta) species range.  67  Figure 4.2: Basal area increments (BAIs) relative to tree age for populations from cold (-1.2 °C, i.e. 10th percentile) and warm (4.6 °C, i.e. 90th percentile) provenances growing at cold (-0.7 °C, i.e. 10th percentile) and warm (4.2 °C, i.e. 90th percentile) sites with dry (361 mm, i.e. 10th percentile), moist (621 mm, i.e. 50th percentile) and wet (1125 mm, i.e. 90th percentile) precipitation levels. Growth-year climate measures were used. Cold sites were depicted using the moist rather than the wet precipitation level due to a lack of cold-wet sampling sites. Provenance precipitation was held at the moist (545 mm, i.e. 50th percentile) level because it had little impact on predicted BAIs.  68  Figure 4.3: Universal growth-trend response functions depicting maximum basal area increments (maxBAIs, in cm2) relative to normal 1990s temperature and precipitation for populations from cold (-1.2 °C, i.e. 10th percentile) and warm (4.6 °C, i.e. 90th percentile) provenances. Separate models were generated using growth-year, summer and winter site climate variables. Provenance precipitation was held at the moist (545 mm, i.e. 50th percentile) level because it had little impact on predicted BAIs.  69  Chapter 5: Climate impacts on lodgepole pine (Pinus contorta) radial growth in a provenance experiment3 5.1 Introduction The responsiveness of tree rings to monthly, seasonal and annual climate is frequently capitalized upon by dendrochronologists to estimate climatic trends prior to recorded history (Fritts 1976). More recently, similar principles have been used to predict how climatic warming, induced by 20th and 21st century greenhouse-gas emissions, may impact tree growth and survival (Hughes et al. 2002). Tree-ring series for such studies are typically collected at the latitudinal and altitudinal margins of species ranges where variation in ring-width is primarily climatedriven. In dendrochronology, trees with high variance in annual radial growth are considered sensitive, while those displaying low variance are described as complacent (Fritts 1976). While radial-growth patterns are known to vary among species, less is known about differences in radial-growth trends among populations relative to the climatic conditions of their growing locations (hereafter sites) and geographic places of origin (provenances). Populations from provenances located near the middle of the species range are generally referred to as central, while those from the latitudinal and altitudinal range peripheries are referred to as marginal. From a dendrochronology perspective, populations are expected to demonstrate a sensitive growth response to climate in marginal habitats where climate is the primary limiting factor (e.g. Daniels and Veblen 2004; Danby and Hick 2007; but see Briffa et al. 1998). From a genetics perspective, marginal populations may either be adapted to respond sensitively to interannual climate fluctuations, taking advantage of favourable conditions when available, or they may be adapted toward complacency, growing only a small amount regardless of climatic conditions in a particular year before ceasing growth in order to avoid cold or drought-induced mortality (Savolainen et al. 2007, Kawecki 2008). Likewise, central populations may grow more overall because they are adapted to take advantage of years with good conditions, or they may be complacent due to generally-favourable climate conditions in the centre of the species range nullifying the adaptive need to respond sensitively to climate.  3  McLane, S. C., L. D. Daniels, and S. N. Aitken. Under review. Climate impacts on lodgepole pine (Pinus contorta) radial growth in a provenance experiment. Forest Ecology and Management. 70  The relative influences of environment and genetics on the sensitivity of growth to interannual climate has not previously been examined, to our knowledge. An ideal venue for testing this question exists in the form of long-term provenance experiments, where numerous populations are grown in common gardens spanning large geographic areas to determine optimal seed sources for reforestation. Differences in cumulative (e.g. Rehfeldt et al. 1999; Andalo et al. 2005; Chuine et al. 2006; O‟Neill et al. 2008; Reich and Oleksyn 2008; Wang et al. 2010) and incremental radial (McLane et al. 2011) growth among populations relative to normal climate variables have already been assessed from these types of trials. Radial-growth responses of populations to site and provenance climate have also been examined on a single-site basis (Oleksyn et al. 1998; Savva et al. 2002, 2007, 2008), but a comparison of multiple populations growing at multiple sites has not been reported. In this study, we test whether lodgepole pine (Pinus contorta var. latifolia Engelm.) trees from climatically-diverse provenances vary in sensitivity and in the strength of their responses to seasonal and annual climate variables when grown across a climatically-diverse spectrum of sites. We use tree-ring data from 12 populations of 34-year-old lodgepole pine trees planted at 16 sites in the Illingworth provenance trial in British Columbia (BC) and Yukon, Canada. Quantitative genetics models called universal response functions (Wang et al. 2010; McLane et al. 2011), plus correlations of climate variables with ring-width chronologies (Fritts 1976), are used to evaluate the combined impacts of site and provenance climate on radial growth. Our objectives are to: 1) examine absolute radial-growth differences among populations relative to provenance and site climate; 2) assess whether populations differ in sensitivity relative to provenance and site climate; and 3) evaluate the correlations between seasonal and annual climate variables and annual radial growth for provenances grouped regionally. We use the results of our analyses to discuss how radial growth may be affected by climate change.  5.2 Methods  5.2.1 Plant materials We collected tree cores from the large (153 populations; 60 sites) Illingworth lodgepole pine provenance trial. The Illingworth trial is a BC Ministry of Forests, Mines and Lands experiment with sites located throughout the interior of BC and two sites in Yukon (Illingworth 1978). The trial was established in 1974 using three-year-old seedlings from bulked seed 71  collections in which numerous parent trees represented each provenance. Two randomized complete blocks were established at each site. Within each block, nine seedlings per population were planted in three-by-three tree plots. All trees were planted 2.5 m apart. Competing vegetation was controlled in the early years of the study, but no thinning of experimental trees was performed. For the purposes of our study, this uniform tree spacing rendered intra-specific competition for light and other resources relatively consistent for all populations and sites through time, which minimized confounding growth trends associated with uneven stand density. Sixteen sites and 12 populations were chosen to represent a broad spectrum of temperature and precipitation levels and geographic locations (Figure 5.1, Table 5.1). Survey data from 2005 were used in consultation with the BC Ministry of Forests, Mines and Lands to verify that mountain pine beetle and other insect and pathogen damage was minimal, and that there was adequate survivorship (12+ trees whenever possible) of the selected populations, at the selected sites. In 2006, increment cores were sampled from up to six of the nine trees per population in each block at each site, depending on survival, with a preference for sampling trees with six live neighbours to maintain uniform competitive effects. Thus, up to 12 trees per population were cored at each of the 16 selected sites. Trees were cored cross-slope between 30 and 40 cm above ground using a 5 mm increment borer.  5.2.2 Chronology development and statistics We processed the cores using the protocol of McLane et al. (2011). Cores were dried, mounted and sanded following Stokes and Smiley (1968) prior to being scanned using a Hewlett Packard Scanjet 2300c scanner and measured using a WinDENDRO digital tree-ring image processor (Régent Instruments, Quebec). Ring widths were recorded at a resolution of 0.001 mm. The tree-ring series for every tree was visually cross-dated using the graphical output provided by WinDENDRO, and then cross-dated and quality checked using the program COFECHA (Holmes 1986). COFECHA was programmed to correlate each tree core against the other cores from the same population and site based on 20-year segments with a five-year lag. Although some cores had rings dating to the mid-1970s, all rings prior to 1982 were excluded from the final data set to remove establishment-related growth effects. Site-population combinations (SPs) with fewer than six live trees were also eliminated from the analyses. The final data set consisted of data from 1,246 trees representing 110 SPs.  72  We calculated average absolute ring widths and standardized ring-width chronologies for each SP. Average absolute ring widths were deduced by calculating the average ring width for the years 1982 to 2005 for each of the 1,246 tree-ring series, then grouping the series by SP and calculating the grand average. This procedure resulted in one measure of tree growth for each of the 110 SPs, analogous to total tree height or diameter at a given age used in previous provenance-trial assessments. Residual ring-width chronologies were developed by using the program ARSTAN (Cook and Holmes 1986) to run cubic smoothing splines through each of the 1,246 individual tree-ring series and then averaging by year, site and provenance to derive an average residual chronology for each SP. The splines were programmed to retain 50% of the variance in the data over a 20year moving window. Running flexible spline curves through ring-width data accounts for growth trends associated with age, and reduces low-frequency, non-climatic signals such as stand-density aberrations caused by non-uniform growth and mortality in the area surrounding the sampled trees, while retaining high-frequency, interannual growth variation (Savva et al. 2002, 2008). Cubic smoothing splines also eliminate growth responses relative to lowfrequency, long-term climate trends, a desired outcome given our intention of examining annualgrowth responses to annual and seasonal climate fluctuations. Long-term climate signals were minimal in our study due to the short time period analyzed (24 years). Chronology statistics, including sensitivity, first-order autocorrelation, and expressed population signal (EPS), were calculated for each SP using ARSTAN. Sensitivity measures the average amount of variability between consecutive pairs of ring widths. Values approaching 1.0 indicate a high degree of variability, while lower values (< 0.5) are common for young trees and species that tend to be complacent such as lodgepole pine (Fritts 1976). Since only highfrequency signals are retained in residual chronologies, the assumption is that the remaining variability is largely climate-driven, thereby providing an indication of how sensitive each chronology is to climate. First-order autocorrelation provides a broader-scale indication of how closely growth in the current year is related to growth in the previous year over the entire chronology period. EPS shows how well the residual chronology compares with a theoretically noise-free chronology. An EPS of 1.0 indicates that the sample chronology expresses pure signal, while an EPS of 0.0 indicates pure noise. Chronologies with EPSs below 0.85 are often eliminated. However, because our intention is to examine growth sensitivity and the correlation between growth and 73  climate among populations, including evaluating the responses of populations for which individual trees do not respond synchronously, we retained all chronologies.  5.2.3 Climate variables Climate data were estimated for both the 16 planting sites and 12 provenances using ClimateWNA (an extended version of ClimateBC (Wang et al. 2006b) for western North America, www.genetics.forestry.ubc.ca/cfcg/ClimateWNA/ClimateWNA.html). ClimateWNA interpolates weather station data using high-resolution digital elevation models that accurately capture climatic variance in western North America‟s mountainous terrain. Normal (1971-2000) provenance temperature and precipitation variables were estimated for each provenance. Monthly temperature and precipitation variables were also estimated for each of the sites for the years 1982 to 2005. Corresponding heat:moisture indices (aridity) were calculated as: ((temperature+30)/precipitation)*1000 (after Wang et al. 2006b). Aridity is a biologicallymeaningful representation of the interaction between temperature and precipitation. All subsequent analyses were performed using SAS software version 9.2 (SAS Institute Inc. 2008). Correlations were examined between all SP residual chronologies and temperature, precipitation and aridity of individual months from April of the previous year through August of the growth year. These preliminary analyses revealed that climate in June and July of the year of ring formation significantly impacted radial growth at many of the sites. A corresponding „summer‟ climate variable was created based on these trends by averaging monthly temperatures and summing monthly precipitation values. Similarly, „biological-year‟ variables, representing the twelve-month period from August previous through July of the year of ring formation, were calculated. Normal site-climate variables were derived by averaging the summer and biologicalyear temperature and precipitation data by site for the study period (1982-2005). This period represents the climatic baseline in this study.  5.2.4 Absolute and detrended radial-growth responses to climate We created universal response functions (URFs) to evaluate the impacts of site and provenance climate on radial growth. URFs are a quantitative genetics technique used to simultaneously incorporate both site and provenance environmental variables in multiple regression models (Wang et al. 2010, McLane et al. 2011). First, a URF was developed to assess  74  the effects of normal biological-year site and normal provenance climate on average absolute radial growth: Yij = 0 + 1 to 6 x site climvarsi + 7 to 12 x prov climvarsj + 13 x site tempi * prov tempj + 14 x site precipi * prov precipj + 15 x site aridi * prov aridj + ij  Eq. 5.1  where Yij is the average absolute ring width for population j at site i; s are the regression coefficients; site climvarsi are normal biological-year temperature, precipitation, aridity and their squares for site i; prov climvarsj are normal temperature, precipitation, aridity and their squares for provenance j; site tempi * prov tempj, site precipi * prov precipj and site aridi * prov aridj are interactions between analogous site-provenance variables for site i and provenance j; and ij is the residual. A stepwise selection procedure allowed variables with significance levels of p < 0.05 to stay in the model. We subsequently created a simplified URF to visually illustrate the impacts of normal biological-year site and normal provenance temperature on average absolute radial growth: Yij = 0 + 1 x site tempi + 2 x site temp2i + 3 x prov tempj + 4 x prov temp2j + 5 x site tempi * prov tempj + ij  Eq. 5.2  where Yij is average absolute ring width for population j at site i; s are the regression coefficients; site tempi and site temp2i are normal biological-year temperature and temperature squared for site i; prov tempj and prov temp2j are normal temperature and temperature squared for provenance j; site tempi * prov tempj is the interaction between temperature at site i and temperature for provenance j; and ij is the residual. Temperature variables were used in this illustration model because they were shown in other models within this study to explain the majority of the impact of climate on radial growth. Lastly, we developed a URF to evaluate whether chronology sensitivity varies relative to normal biological-year and summer site and normal provenance climate. The model was built using Eq. 5.1, with Yij representing sensitivity for population j at site i, and „site climvars‟ representing biological-year and summer site climate variables in separate models. A simplified URF illustrating the impacts of normal site and normal provenance temperature, plus their squares and interactions, on sensitivity was also created using Eq. 5.2. 5.2.5 Growth – climate correlations We evaluated the directionality and strength of the relationships between radial growth and site and provenance climate by calculating correlation coefficients between annual 75  biological-year and summer climate variables and the annual residual ring widths for all SPs. Based on trends from this analysis, we grouped the sites and provenances into three „geoclimatic regions‟ by biological-year temperature for ease of interpretation: „cold‟ (< 1.0 °C), „cool‟ (1 to 3 °C) and „warm‟ (> 3 °C) (Table 5.1). The four cold sites and three cold provenances are located in northern BC and the southern Yukon, the six cool sites and five cool provenances are in northcentral BC or at high elevations in southern BC, and the six warm sites and four warm provenances are at lower elevations in central BC or in southern BC. Correlation coefficients were calculated for each pair of site-provenance geoclimatic regions, yielding nine correlations per climate variable. We used temperature to group the sites and provenances because we are primarily interested in population responses to temperature fluctuations, and correspondingly, how populations may respond to climatic warming. However geoclimatic-group precipitation indices were also calculated and analyzed (Table 5.1).  5.3 Results  5.3.1 Absolute and detrended radial-growth responses to climate Absolute ring widths ranged from 0.48 to 4.84 mm among all years and SPs. Averaged by SP, absolute average ring widths ranged from 1.08 to 3.36 mm, with an overall average value of 2.39 mm (Appendix 5). The variance in ring widths between 1982 and 2005 ranged from 0.71 to 2.98 mm among SPs, averaging 1.59 mm. Both sensitivity (0.01 to 0.35) and EPS (0.42 to 0.94) values varied greatly among SPs. Average absolute ring widths varied significantly (p < 0.0001) relative to normal site and provenance climate variables (Table 5.2a). The predictor variables collectively explained 58% of the total variance in average ring width, with site and provenance temperature and their squares and interactions accounted for two-thirds of the explained variance. The simplified URF depicting the impacts of normal site and provenance temperature on average absolute ring widths illustrated that the largest radial increments occur at moderately warm sites and in moderately warm years, and among populations from the warmest provenances (p < 0.0001, R2 = 0.43) (Figure 5.2). Sensitivity varied significantly (p < 0.0001) relative to normal biological-year siteclimate and normal provenance-climate variables, with the normal biological-year sensitivity URF explaining 28% of the total variance in average ring widths (Table 5.2b). In this model, 76  normal site temperature and its square accounted for three-quarters of the explained variance. The simplified URF depicting the impacts of normal site and provenance temperature on sensitivity illustrated that sensitivity was highest among populations from warm provenances growing at low site temperatures (p < 0.0001, R2 = 0.26) (Figure 5.3). Sensitivity did not vary significantly relative to normal summer site-climate variables. 5.3.2 Growth – climate correlations Correlation coefficients between residual ring-width chronologies and climate data revealed significant relationships between biological-year and summer-climate variables and radial growth for some regional site-provenance combinations (Table 5.3). Residual ring widths were positively correlated with biological-year temperature for all provenances in all regions, but the correlations were not significant for populations from warm provenances growing in cold or warm sites. Biological-year precipitation and aridity did not correlate significantly with growth. Using summer-climate variables, ring widths were correlated with precipitation and aridity but not temperature, except in the case of populations from warm provenances growing in warm sites, for which temperature and growth were negatively correlated. Precipitation was positively correlated with annual radial growth for populations growing at cold and warm sites, but negatively correlated with growth for populations growing at cool sites. Aridity was negatively correlated with annual radial growth for populations growing at cold and warm sites. All of these correlations were significant, except in the cases of populations from warm provenances growing at cold sites and populations from cold provenances growing at warm sites.  5.4 Discussion  5.4.1 Absolute and detrended radial-growth responses to climate Average ring widths were well described by site normal biological-year and provenance normal climate variables. This was expected, as average ring-width trends should be similar to total volume measurement trends, which have been characterized previously (Wang et al. 2006a). Site and provenance normal climate explained less of the total variance in average radial growth than the proportion explained for height growth (Wang et al. 2010), which was not surprising given that radial growth has a considerably lower heritability than height in lodgepole pine (Wang et al. 1999). Additionally, the strength of the relationship between tree rings and 77  climate is dependent on how strongly growth is limited by climate at the study site (Fritts 1976). Since the Illingworth trial sites are mostly located in moderate to high-productivity environments for lodgepole pine, annual radial growth is expected to be only moderately sensitive to climatic signals. From a site perspective and using biological-year site climate variables, trees were more sensitive to interannual climate fluctuations at cold sites than at warm sites, while from a genetics perspective, populations from warmer provenances were more sensitive than those from colder provenances. The interaction between these two trends resulted in the greatest sensitivity occurring among populations from warm provenances growing at cold sites, followed by cold provenances growing at warm sites. Thus, at first glance, it appears that populations from warmer, more central provenances are more sensitive to climate than those from colder, more marginal provenances and that, genetics aside, trees are more sensitive to climate at colder, more marginal sites than at warmer, more central sites. However, the regional breakdown of growth responses to climate provided by the growth – climate correlations provides a more nuanced story (see below). The lack of growth sensitivity to summer climate likely results from there being less variance among sites in the summer climate variables than the corresponding biological-year variables (Table 5.1). This indicates that climatic processes in seasons other than summer, such as average temperature, winter drought stress or duration of snowpack, cause the majority of the overall climatic differences among the sites. We may have found growth to be more sensitive to summer climate variables had sites experiencing more heat stress been established (Illingworth 1978). The trial was largely restricted to locations within BC, meaning that moderate and cold locations have good coverage, while hot locations do not. 5.4.2 Growth – climate correlations Our growth – climate correlations indicate that annual growth is positively correlated with biological-year temperature for all populations at all sites. However, the significance of the positive correlations varies regionally for both provenances and sites. We found that trees from warm provenances growing in cold sites show substantial growth fluctuations from year to year and respond positively, but not significantly so, to warmer biological-year temperatures. In contrast, trees from cold provenances are most sensitive to climate when growing in warm locations and demonstrate significant positive responses to warmer biological-year temperatures. 78  This may indicate that populations from cold provenances are more adapted to take advantage of warm conditions than populations from warm provenances. From a dendrochronology perspective, this could mean that some of the sensitivity attributed to climate acting as a limiting factor at marginal sites is due to local adaptation. This finding could partially explain the mixed results of studies such as Wilmking et al. (2004), who found diverse radial-growth responses to 20th-century climate change among natural populations of white spruce (Picea glauca) growing throughout Alaska. North-central Alaska was a glacial refuge for white spruce, with the species inhabiting a heterogeneous landscape throughout and subsequent to the glacial period (Anderson et al. 2006). Local adaptation may manifest in the form of some populations demonstrating more phenotypic plasticity in growth responses to climate than others. From a genetics perspective, our results provide interesting evidence of local adaptation to not only climate, but climate fluctuations. While phenotypic plasticity is often considered an alternative strategy to adaptation, the range of variation in growth that a population can display in response to climate within a site is, in fact, adaptive and subject to natural selection (Visser 2008). Our results provide preliminary evidence that as temperatures warm due to climate change, populations growing in cold locations will not only react positively to warmer overall temperatures, but also possess the ability to capitalize on optimal growing seasons when available. The degree of phenotypic plasticity could be under differential selection among populations, just as birds in the United Kingdom (Crick et al. 1999) and North American red squirrels (Tamiasciurus hudsonicus) in Yukon, Canada (Réale et al. 2003) that can breed earlier are being favoured under climate change. Radial growth responded to summer-climate variables, but did not fluctuate substantially from year to year. The seeming discrepancy between the summer-climate correlations for the cold and warm vs. cool site groups may be explained by the cool group‟s normal summer temperature being lower and precipitation higher than those of the cold and warm groups, despite the cool group‟s biological-year temperature being intermediate between that of the cold and warm groups (Table 5.1). For the cool sites, which are located in north-central BC and at high elevations in southern BC, summer aridity levels may simply not get high enough to negatively impact radial growth. Regional climate change models forecast that winter temperatures in general, and winter minimums in particular, will increase more in northern BC and the Yukon than in southern BC (Ministry of Water, Land and Air Protection 2002). Likewise, winter precipitation is predicted 79  to increase, but summer precipitation to decrease within the study region (Christensen et al. 2007). Based on the growth – climate relationships that we deduced, these seasonal and regionspecific climatic changes may result in increased radial growth in cool and warm regions as annual temperatures rise, but decreased radial growth in cold and warm regions as summer aridity increases. Specifically, our results suggest that radial growth is negatively affected once normal June-July temperatures rise to ~11.9 to 13.2 ºC (the cold and warm group average summer temperatures, respectively), given precipitation levels comparable to those found in this study.  5.4.3 Comparison with complimentary study using same data set Using the same Illingworth experiment ring-width data set used in this study, we previously used a random-coefficient modeling technique to create universal growth-trend response functions illustrating the impacts of climate on radial-growth trends over time (McLane et al. 2011). That technique models the age and basal area increment at which radial growth begins declining for trees from a wide diversity of provenances growing across a spectrum of sites and under a range of climate scenarios. It also illustrates changes in population raking over time. In contrast, the goals of the current study are to assess the responsiveness of ring widths to annual and seasonal climatic variables across a range of sites, and to quantify the influences of site and provenance climate on annual-growth sensitivity. Despite having markedly different objectives, methodologies and results, both studies indicate that radial growth will likely decline in southern and central BC due to rising summer aridity levels, but may increase in northern areas as temperatures continue rising over the 21st century.  5.5 Conclusions Provenance trials have been used by forest managers to assess genotype-by-environment interactions for decades, yet few researchers have examined what the annual radial-growth signatures of provenance-trial trees can tell us about climate and genetics. This study examines ring widths for multiple lodgepole pine populations grown across a climatically-diverse array of sites relative to site and provenance climate. Our models indicate that both central and peripheral populations demonstrate sensitive growth patterns when growing in an environment contrasting to their native conditions, but that peripheral populations respond to climatic signals more strongly than central populations in most environments. This could be evidence that 80  sensitivity to climate is a local adaptation possessed more strongly by peripheral populations than central populations, which is notable from a genetics standpoint and interesting from a dendrochronology standpoint, as well, given that growth sensitivity is typically attributed to climate acting as a limiting factor, rather than to genetic adaptation. Unfortunately, we were not able to corroborate our cold-limited range-margin trends with data from heat-limited rangemargin trees or sites because the Illingworth experiment included very few populations from, and no sites located, south of the Canadian border (Illingworth 1978). Non-uniform changes in climate trends among seasons may lead to regionally-divergent radial-growth trends over the coming century, as demonstrated by the correlations between radial growth and site climate varying among regions and seasons. Radial growth may increase in cool and warm regions as annual temperatures rise, but decrease in cold and warm regions as summer aridity increases. The implications for forest productivity under climate change are therefore more positive for trees growing in cool locations, where warmer annual temperatures will lead to increased growth overall, than in warm locations, where the negative effects of more arid summers may counteract the overall positive warming effects. This finding is in keeping with other studies that forecast growth increases in the far north but declines at middle latitudes as rising temperatures lead to more arid summers and corresponding heat and drought stress (Reich and Oleksyn 2008; Allen et al. 2010; McLane et al. 2011).  81  Table 5.1: Geographic and climatic descriptions of sites and provenances sampled from the Illingworth lodgepole pine experiment in and adjacent to British Columbia, Canada.  Provenances  Sites  Geoc. Latitude Longitude Elevation Temp. (°C) Precip. (mm) Aridity region (°N) (°W) (m) Ann. Sum. Ann. Sum. Ann. Sum. Cold 60.08 128.83 700 -1.9 14.3 396 111 71.0 399.1 Cold 59.78 129.13 730 -0.9 13.6 468 113 62.2 385.0 Cold 58.83 125.72 1100 -0.6 11.4 592 187 49.7 221.1 Cold 60.80 135.18 663 -0.3 13.3 303 81 97.9 536.2 Cool 49.80 115.50 1650 1.3 11.3 1353 217 23.2 189.8 Cool 54.85 123.92 950 1.9 12.3 671 126 47.6 336.6 Cool 51.67 123.00 1280 2.3 11.9 362 108 89.2 389.7 Cool 49.35 114.68 1550 2.3 12.3 1015 184 31.8 229.9 Cool 53.77 126.83 1040 2.7 11.6 721 111 45.3 376.6 Cool 52.18 120.80 1100 2.9 12.3 814 171 40.4 246.9 Warm 50.58 120.62 1430 3.4 12.7 471 115 70.8 372.0 Warm 55.32 127.27 640 3.5 13.0 736 148 45.5 290.9 Warm 51.28 121.33 1070 3.7 13.4 361 103 93.3 419.5 Warm 49.62 118.92 1370 3.7 12.9 651 126 51.8 340.7 Warm 52.50 122.38 850 4.7 14.4 472 128 73.4 347.3 Warm 50.82 117.98 790 5.3 15.4 1236 193 28.6 235.5 11.45 20.50 1010 7.2 4.1 1049 137 35.4 249.5 Cold 59.87 129.72 798 -0.9 13.2 440 123 66.1 350.6 Site geoclimatic regions Cool 51.94 120.79 1262 2.2 11.9 823 153 39.2 274.6 Warm 51.69 121.42 1025 4.0 13.6 655 135 52.0 322.2 Lower Post 30 Cold 59.98 128.55 640 -2.3 432 64.1 Petitot River 163 Cold 59.90 122.08 396 -1.3 455 63.1 Atlin 35 Cold 59.80 133.78 789 0.1 361 83.4 Tower Lake 26 Cool 56.02 120.62 792 1.4 477 65.8 Nina Creek 100 Cool 55.80 124.82 762 1.6 521 60.7 Chilco 18 Cool 51.98 123.75 1059 2.5 328 99.1 Swan Hills, AB 142 Cool 54.30 116.58 823 2.6 594 54.9 Fly Hills 71 Cool 50.72 119.45 1524 2.7 869 37.6 Marl Creek 44 Warm 51.52 117.18 945 3.4 622 53.7 Albreda 63 Warm 52.58 119.17 975 3.7 915 36.8 Trapping Creek 1 Warm 49.58 119.02 1006 4.7 568 61.1 Larch Hills 72 Warm 50.70 119.18 777 5.9 754 47.6 Range among provenances 10.40 17.20 1128 8.2 587 65.1 Cold 59.89 128.14 608 -1.2 416 69.3 Provenance geoclimatic Cool 53.76 121.04 992 2.2 558 57.7 regions Warm 51.10 118.64 926 4.4 715 48.2 Notes: Sites are ordered from coldest to warmest average biological-year temperature (Temp.) while provenances are ordered from coldest to warmest normal annual temperature (both notated as Ann.). The sites and provenances are averaged into geoclimatic regions (Geoc. region) by these temperatures: ‘cold’ (< 1.0 °C), ‘cool’ (1 to 3 °C) and ‘warm’ (> 3 °C). Normal biological-year (Ann.) precipitation (Precip.) and aridity and normal summer (Sum.) site temperature, precipitation and aridity (sites only) are also listed. We did not calculate summer temperature or precipitation variables for the provenances. Name Code Watson Lake, YK WATS Blue River, YK BLUE Mile 451 M451 Whitehorse WHRS Lussier River LUSS Salmon Lake SAMN Mons Lake MONS McLatchie Creek MCLA Ootsa Lake OTSA Bosk Lake BOSK Chuwhels Lake CHUW Suskwa River SUSK 70 Mile MI70 Lassie Lake LASI Cuisson Lake CUIS Wigwam WIGW Range among sites  82  Table 5.2: Universal response function analysis predicting a) average raw ring width and b) chronology sensitivity from normal site and provenance climate variables. Models a) and b) are both significant overall (p < 0.0001). Independent variable Parameter estimate Partial R² Model R² Intercept -131.169 Site temperature X provenance temperature 0.030 0.298 0.298 Site temperature² 0.079 0.095 0.394 Site normal precipitation 0.020 0.084 0.478 Site aridity² -5.055 0.057 0.534 Site aridity 50.935 0.023 0.558 Site precipitation² 0.000 0.021 0.578 Intercept 0.132 Site temperature -0.020 0.181 0.181 Site temperature² 0.003 0.036 0.217 Site temperature X provenance temperature -0.004 0.035 0.252 Provenance temperature² 0.003 0.030 0.283 Notes : The model column indicates dependent and independent variables used.  (b) Sensitivity vs. normal biologicalyear site and normal provenance climate  (a) Average ring widths vs. normal biologicalyear site and normal provenance climate  Model  83  C(p)  F  P < 0.0001  81.402 45.92 < 0.0001 38.087 19.31 < 0.0001 13.391 21.13 < 0.0001 68.209 9.44 0.0027 33.836 4.93 0.0285 8.844 5.37 0.0224 < 0.0001 34.644 23.94 < 0.0001 23.273 5.25 0.0240 27.435 31.431  4.91 4.12  0.0288 0.0449  Table 5.3: Pearson‟s correlations between residual ring-widths and annual climate fluctuations for all site and provenance geoclimatic regions relative to biological-year and summer climate variables. Temp. = temperature; Precip. = precipitation. Negative correlations are highlighted in  Summer  Biological-year  gray and significant correlations (p < 0.05) are bolded. Climate Site variable region Cold Temp. Cool (°C) Warm Cold Precip. Cool (mm) Warm Cold Aridity Cool Warm Cold Temp. Cool (°C) Warm Cold Precip. Cool (mm) Warm Cold Aridity Cool Warm  Provenance region Cold Cool Warm 0.16 0.15 0.10 0.14 0.15 0.12 0.30 0.11 0.09 0.11 0.11 0.10 -0.01 -0.06 -0.03 -0.01 0.05 0.03 -0.10 -0.09 -0.09 0.03 0.08 0.04 0.03 -0.07 -0.04 -0.10 -0.09 -0.05 0.08 0.10 -0.02 0.06 -0.07 -0.14 0.15 0.14 0.16 -0.15 -0.11 -0.02 0.06 0.15 0.21 -0.16 -0.16 -0.12 0.12 0.10 0.00 -0.10 -0.19 -0.26  84  Figure 5.1: Locations of sampled Illingworth lodgepole pine provenance trial sites and provenances in and adjacent to British Columbia, Canada.  85  Figure 5.2: Universal response function modeling absolute lodgepole pine ring widths relative to normal site and provenance temperature (p < 0.0001, R2 = 0.43).  86  Figure 5.3: Universal response function modeling lodgepole pine radial-growth sensitivity relative to normal site and provenance temperature (p < 0.0001, R2 = 0.26).  87  Chapter 6: Conclusions 6.1 Introduction Climate change poses an unprecedented threat to forested ecosystems. The ability of tree populations to survive major climatic changes depends on their capacity to persist in situ via phenotypic plasticity, to adapt via natural selection, or to migrate to new, more suitable habitats (Aitken et al. 2008; Visser 2008). Some tree species with narrow ecological niches are predicted to lose most of their current climatically-suitable habitats by the end of the century (Thomas et al. 2004; Hamann and Wang 2006; Wang et al. in prep.), while even broad-ranging species are at risk of productivity losses due to populations decoupling from their climatic optima as temperatures rise (O‟Neill et al. 2008; Reich and Oleksyn 2008; McLane et al. 2011). Species distribution models, forest productivity models, and genetics tools must be integrated to predict future ranges and productivity levels for populations and species. In this dissertation, I integrated genetics, ecology, species distribution modeling and dendrochronology techniques to assess the impacts of climate on establishment and growth for populations of two North American pines under a range of climate scenarios. I demonstrated the use of these techniques for a keystone species of conservation concern, whitebark pine (Pinus albicaulis Engelm.), and a species that is widely harvested and valuable for carbon storage, lodgepole pine (Pinus contorta var. latifolia Engelm.). Through my research, I sought to inform forest professionals and conservationists in an era of climate change.  6.2 Whitebark pine may require assisted migration For whitebark pine, which grows in narrow bands at treeline throughout the North American west, persistence within the current species range will entail maintaining competitiveness relative to lower-elevation trees that may encroach as temperatures warm, including mountain hemlock (Tsuga mertensiana), subalpine fir (Abies lasiocarpa), lodgepole pine, and Engelmann spruce (Picea engelmannii) (Weaver 2001). Evolving to withstand warmer temperatures will require the existence of, and selection for, locally-adapted genes that confer growth or phenological advantages under new climatic conditions (Visser 2008). Migrating will necessitate relatively long-distance seed dispersal to available (unoccupied) niches in newly climatically-suitable locations (Hamann and Wang 2006), since areas uphill from whitebark pine habitat tend to have poorly developed soils, and because migrating through valleys is not an 88  option given the pine‟s poor competitive ability relative to its downslope neighbours (Weaver 2001, Chapter 3 of this dissertation). Using growth chambers, I demonstrated that whitebark pine has the genetic capacity to take advantage of warmer temperatures than typically experienced within its current range by observing seedlings increase growth through 18 °C average and 42 °C high growing-season (mid-May through early September) temperatures. Others have reported healthy whitebark pine stands growing in warmer-than-average conditions (Jacobs and Weaver 1990, Weaver 1994), but my experiment may to be the first to demonstrate increased growth capacity for multiple populations across a range of, and up to such high, temperatures. Initially, this appears to be good news for the species under climate change. However, I concurrently found that lodgepole pine seedlings exceeded whitebark pine in height growth above growing-season temperatures of 11 °C. Given that growing-season temperatures are expected to increase from 9.7 °C to ~14 °C on average in the current whitebark species range by the end of the 21st century (Wang et al. in prep.), it is therefore possible that whitebark pine will be outcompeted by faster-growing species within its current range under most 21st-century climate change scenarios. Adaptation is not considered a tenable option for whitebark pine given the predicted rapid rate of climate change and because of the species‟ 50 to 100-year generation times (COSEWIC 2010). One round of selection (corresponding to one generation) prior to the 22nd century is simply not sufficient to produce a major adaptive shift. Furthermore, my growth chamber data demonstrated that whitebark pine has little genetic variation among populations, meaning that populations naturally adapted to withstand warmer temperatures that could be utilized in restoration plantings may not exist. This is likely a result of stabilizing selective pressure resulting in relative genetic uniformity among populations throughout the species‟ geographically-broad but elevationally and functionally-narrow range. Other researchers using data from more populations have found slightly more genetic diversity among populations (Mahalovitch et al. 2006; Bower and Aitken 2008), particularly for cold-adaptation traits, but differentiation for growth traits is generally low. A breeding program could hypothetically be created to produce more competitive lines of whitebark pine, but this would be temporally and economically unfeasible given how long it would take to cultivate hybrids given the species‟ long generation intervals. Migration may not occur naturally for whitebark pine within the timeframe required given the rate and magnitude of climate change, but humans may choose to assist the migration 89  of the species northward. The practice of assisting species migration is legally and ethically complex, and, so far, has only been executed for conservation purposes in well-intentioned but scientifically unsanctioned manners, most notably by local citizens concerned with the survival of the Torreya pine (Torreya taxifolia) in the southeastern United States (Schwartz et al. 2009). Far more frequently, plant migration is facilitated unintentionally. This occurs when nurseries sell garden plants (van der Veken et al. 2008), municipalities and forestry companies plant trees (Woodall et al. 2010), and “stowaway” seeds are deposited unknowingly, in areas outside of species‟ native ranges. My intention was to execute a sanctioned assisted migration trial that would provide scientific information regarding the potential for whitebark pine to be successfully relocated north of its current range, as well as demonstrate techniques for safely testing speciesrelocation potential. This information will help forest and biology professionals make informed decisions regarding how to conserve the species in the likely event that its numbers continue declining. I demonstrated that whitebark pine can germinate, survive and grow in model-predicted climatically-suitable locations hundreds of kilometres north of the current species range. So far, I have only observed survival and growth over three growing seasons, but my results provide preliminary evidence of northward-migration potential for whitebark pine. Establishment was greatest in locations with moderate snow durations and among populations with well-developed seeds and from warmer geographic locations. This was largely consistent with the findings of previous researchers, including Weaver (1994) and Mellman-Brown (2005), who found the duration of snow cover vitally affects whitebark pine establishment and survival, and Berdeen et al. (2007), who found embryo maturity to affect germination potential. Due to the species‟ status as a candidate for endangered listing in Canada (COSEWIC 2010), I recommend that assisted migration be considered for whitebark pine. However, a program of assisted migration should not be implemented until the ethics of assisted migration have been further discussed by the biological community and the public, the species‟ natural migration potential has been assessed, and the species‟ capacity to establish long-term in the recipient locations has been determined. Continued monitoring of my assisted migration trials will contribute to these assessments.  90  6.3 Lodgepole pine productivity may decline except in the far north Lodgepole pine grows in near-monocultures and mixed-species stands across large portions of northwestern North American. Persistence within central and northern portions of the current species range is expected due to the species‟ extensive current and future-predicted geographic and climatic distributions (Hamann and Wang 2006). This is despite range-wide diebacks caused by mountain pine beetle (Safranyik and Wilson 2006; Axelson et al. 2009). However, expected changes in growth rates will affect timber supplies and carbon sequestration rates, and some populations at the lagging edge of migration are likely to be extirpated. Natural adaptation to warmer climates will likely occur slowly for the species, facilitated by the extensive adaptive genetic variation that exists both among and within populations (Xie and Ying 1995; Rehfeldt et al. 1999). Through this research, I found that lodgepole pine growth will decline throughout much of southern and central British Columbia (BC), which comprises the most productive growing area for the species within Canada. Specifically, I showed that radial growth increases for the species up to annual growth-year temperatures of 4 to 5° C, but that declines begin above this temperature threshold. This is true for populations spanning a large portion of the species range, with populations from warmer provenances performing best in all future-climate scenarios, but nonetheless experiencing declines in warmer growing locations. These results are in keeping with those of Wang et al. (2010) who predicted that lodgepole pine growing in the Illingworth provenance trials will experience height-growth declines above 4.5 °C. My results also parallel those of Reich and Oleksyn (2008), who predicted that Scots pine (Pinus sylvestris) growing in provenance trials in Europe and North America will experience growth declines everywhere but the far north as the climate warms. Using the same data set, I also demonstrated that lodgepole pine annual radial growth is generally positively correlated with annual temperatures and negatively correlated with summer aridity. However, this trend varied regionally, indicating that climate change will have different impacts on growth depending on latitude and elevation. I also found that populations from warm provenances growing at cold sites showed large inter-annual variations in ring widths (i.e. high growth sensitivity), but that these fluctuations were not significantly correlated with the biological-year climate variables tested. On the other hand, populations from cold provenances had overall lower interannual growth fluctuations, but the fluctuations were significantly correlated with climate at all locations. This result has interesting implications for 91  dendrochronologists, who have long recognized that tree growth is most climate-sensitive in marginal environments (Fritts 1976). My results demonstrate that sensitivity at species margins is not only climatically-induced, but may also be adaptive, resulting from genetic selection within populations growing in extreme environments. My results are valuable from a forest-productivity standpoint, but also contribute new applications of analytical methods addressing species responses to climate. To my knowledge, random coefficient models had not been previously used to analyze quantitative genetics data sets. My model quantified the temperature at which tree radial growth maximizes or begins to decline, the age of radial growth maximization, the size of growth rings over time, and the impacts of genetics and climate on these parameters. Random coefficient models can be used to assess any serially-recorded phenotypic trait relative to numerous input variables, making it a versatile tool for assessing forest productivity over time relative to genetics, climate, geographic location, soil type, and other variables.  6.4 Limitations of present work  6.4.1 Whitebark pine My whitebark pine work was subject to the challenges faced by most whitebark pine researchers: seeds are logistically difficult and expensive to collect, growth is slow in the best of circumstances, and accessing whitebark pine habitat is physically challenging and timeconsuming. Ideally, I would have installed dozens of common gardens, each populated by seeds from dozens of provenances, and the trees would have grown more than a few centimeters in three growing seasons. Additionally, I would have monitored the common gardens through full growing seasons at each site, observing differences in bud flush and bud set relative to the timing of snowmelt, and better identifying the timing and causation of seedling mortality. I collected copious data despite these limitations, but undoubtedly my projects were subject to the inherent challenges of growing whitebark pine from seed in field conditions. It also would have been beneficial if the newer-generation whitebark pine species distribution model currently being created by Wang et al. (in prep.) had been available at the time that the common gardens were installed. My gardens only broadly coincide with locations predicted by the first-generation whitebark pine species distribution model (Wang pers. comm., created using the methods of Hamann and Wang 2006) to be habitable under both present and 92  2055 climate scenarios. Had the newer model been available, I could have conducted more sophisticated assessments of the model‟s accuracy by comparing climate variables associated with the current range of whitebark pine to variables representing areas not predicted to be habitable for the species.  6.4.2 Lodgepole pine In analyzing the lodgepole pine radial-growth data from the Illingworth provenance trials, I originally hoped that regressing annual growth indices against annual climate data would provide a more refined illustration of among-population growth differences than is generated using total growth and normal climate indices (Wang et al. 2006a). Using annual climate data would also hypothetically allow among-population growth-climate relationships to be assessed over a wider spectrum of temperature and precipitation levels than is possible using normal climate data. Given the dearth of provenance-trial sites representing temperatures as warm as those predicted under climate change scenarios, I hoped annual radial-growth indices might provide a more comprehensive picture of how populations may respond to predicted climatic changes by the end of the 21st century than is possible using growth-yield data. Unfortunately, the refined annual-growth annual-climate URF that I sought to create did not prove feasible. The variance in normal temperature (7.2 ºC) and precipitation (1049 mm) among sites exceeded the average variance in annual temperature (2.9 ºC) and precipitation (378 mm) within sites, rendering within-site annual climate fluctuations of relatively low consequence compared to the fundamental climatic differences among sites (Table 5.1). Because of this, between-site climate differences continued to drive the quadratic population responses, rendering the population response curves created using annual growth data almost indistinguishable from those created using yield data. Thus, in pursuing an improved model, I affirmed that models generated using normal climate and total growth measures are the most efficient and effective method of assessing differences in population cumulative growth responses to climate. However, the universal growth-trend response function that I created using a random coefficient model is a very useful and comprehensive way of assessing changes in growth trends over a variety of climatic conditions and relative to genetics and other variables.  93  6.5 Future research This thesis represents significant advancements in our knowledge of how populations of whitebark and lodgepole pine respond to climate, and contributes a new long-term provenance trial as well as new methods for assessing growth trends over time. Arising from these contributions are several questions that have and will provide the basis for future research projects.  Can whitebark pine persist as an understory species as the climate warms? The ability of whitebark pine to persist as an understory species, i.e. its shade tolerance, could prove critical to the species‟ ability to persist in situ as the climate changes. As of September 2010, this question is being addressed in a long-term experiment in Whistler, British Columbia (S.C. McLane, B. Brett, J. Krakowski and S.N. Aitken unpublished data). Whitebark pine seedlings representing the same six provenances used in this thesis were planted in six cleared sites spanning 1,000 m elevation and 3+ °C temperature change under the new Whistler Blackcomb Ski Resort Peak-to-Peak gondola. The seedlings will be monitored for survival and growth as vegetative competition increases over the coming decades.  What restricts the northern range of whitebark pine? Little is known about dispersal and recruitment dynamics at the northern boundary of the whitebark pine species range, largely due to the region being physically difficult to access. Data indicating the coordinates of individual whitebark pine trees growing along the species boundaries, and a corresponding assessment of the climatic parameters associated with locations within vs. directly north of the northernmost range boundary, would be immensely valuable for determining what ultimately limits the northern edge of the species range. Knowing the fecundity of the northernmost mature individuals, and the extent of nutcracker caching activity and seedling recruitment north of those locations, would also be critical to understanding the causes of the current northern range limit. I hypothesize that recruitment, and ultimately migration, is limited north of the current species range by a snow-duration variable that is not captured in current species distribution models. It has also been suggested that whitebark may be limited due to a lack of additional seed resources for nutcrackers north of whitebark pine‟s current range, e.g. due to Douglas-fir‟s (Pseudotsuga menziesii) northern limit occurring at approximately the same latitude (J. Vinnedge, Ministry of Environment, pers. comm.). Another 94  theory is that whitebark pine may be limited by mycorrhizal associates (Mohatt 2006) that exist within but not north the current species range. These hypotheses cannot be tested without a robust assessment of the factors limiting the species‟ northernmost populations.  Can duration of snow cover be added to species distribution models? Duration of snow cover is a major limiting factor for whitebark pine within its current range (Weaver 1994; Mellman-Brown 2005) and may ultimately limit the northward migration of the species. However, this variable is not well captured by current-generation species distribution models, most of which are created using climate data that is interpolated from distant weather stations. The best available snow variable is often “precipitation as snow”, but precipitation varies widely by region, and whether the precipitation falls as snow or rain varies even more so based on microclimate, topography and weather-system dynamics. However, a snowmelt timing variable could potentially be added to SDMs by incorporating slope and aspect, satellite-derived snowpack data (see Canadian Meteorological Centre National Snow and Ice Data Centre website: http://nsidc.org/data/nsidc-0447.html), and regional snowmelt-timing predictions (e.g. Beniston et al. 2003). Further research into the impacts of climate change on snow-accumulation patterns and subsequent consequences for ecological communities is also needed.  How much of the phenotypic variation observed among whitebark pine populations growing in common environments is genetic vs. maternal in origin? Whitebark pine is very slow to mature reproductively (~50 - 100 years) and is not valuable from a commercial standpoint, and as such there has been no incentive to cultivate an offspring generation from parent populations growing in a common environment in order to eliminate the phenotypic variation caused by maternal (Falconer and Mackay 1996) effects. This limits our ability to accurately estimate heritability and QST for the species. Maternal effects impact phenotypic traits for all species, but they are particularly pronounced for whitebark pine due to its unusually large seeds and short seed-development season (Moles and Westoby 2003).  What are the genomics of local adaptation for whitebark and lodgepole pine? Newly accessible and affordable genomics procedures are revolutionizing the ability of geneticists to evaluate the genetic underpinnings of local adaptation. A technique called 95  association mapping links individual genes or single nucleotide polymorphisms (SNPs) within genes with phenological or phenotypic traits such as timing of budset or cold hardiness (Neale and Savolainen 2007). When this procedure is performed using multiple populations representing broad geographic areas, it becomes possible to associate clinal variation in geneotypes with clinal variation in environmental characteristics such as temperature (Holliday et al. 2010). For slow-growing species such as whitebark pine, this technique could allow much faster evaluations of local adaptation for traits associated with tolerance of future-predicted climates than is possible using common garden studies. For timber species such as lodgepole pine, productivity could be maximized by identifying populations not currently associated with the breeding process that combine high growth and yield with adaptation to warmer environments. Overall, the new genomics techniques may allow faster, more comprehensive quantification of standing genetic variation for purposes of conserving genetic resources and quantifying adaptive potential in a changing environment (Barrett and Schluter 2008), and the cost of accessing these technologies has dropped precipitously in recent years. Studies of economically important conifers may also yield lists of candidate genes for local adaptation that are worth pursuing in species like whitebark pine.  How steeply does growth decline for lodgepole pine growing in warmer-than-optimal locations? The Illingworth lodgepole pine provenance trial is one of the largest and most extensively phenotyped trials of its kind in the world. However, its Achilles heel from a climate-change growth-prediction perspective is that it does not include trial sites where many populations are growing above the climatic optimum for the species (Rehfeldt 1999; Wang et al. 2006a). This is because the trial was originally established with the intention of deducing which populations and locations are optimal for growth within British Columbia in order to maximize the growth and timber value of the species. Because of this limitation, growth trends for populations growing in sites above ~5 °C mean annual temperature are estimated using quadratic curves that mirror the growth curves of the same populations growing at low temperatures. However, there is no biological reason that these curves should, in fact, be mirror images. This data gap renders our predictions of productivity losses under climate change imprecise at best. To address the warmer test-site deficit, Greg O‟Neill and others in the Ministry of Forests, Mines and Lands in BC recently initiated an ambitious new provenance trial experiment, planting populations from 16 96  species across 48 sites encompassing a range of environments representative of climate change projections (www.for.gov.bc.ca/HRE/forgen/interior/AMAT.htm). I hope to see similar trials initiated elsewhere for species of both economic and conservation concern. How will disturbance-induced mortality affect our ability to accurately predict population persistence and growth? My work has focused on predicting germination and growth capacities of tree populations in a changing climate. However, it is equally crucial to understand the genetic and environmental capacity of populations to survive disturbances that are predicted to be exacerbated by climate change, most notably fire, climate-related insect and pathogen outbreaks, and ice and wind storms (Dale et al. 2001; Kurz et al. 2008). Incorporating predictions of these disturbances, as well as the genetic capacity of populations to withstand such events, into species distribution and growth models will further improve our ability to predict the impacts of climate change on forests.  97  Literature cited Adams H. D., T. E. Kolb. 2005. Tree growth response to drought and temperature in a mountain landscape in northern Arizona, USA. Journal of Biogeography 32:1629-1640. 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. Gonzalez, R. Fensham, Z. Zhang, J. Castro, N. Demidova, J. H. 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. Andalo C., J. Beaulieu, and J. Bousquet. 2005. The impact of climate change on growth of local white spruce populations in Quebec, Canada. Forest Ecology and Management 205:169182. Anderson L. L., F. S. Hu, D. M. Nelson, R. J. Petit, and K. N. Paige. 2006. Ice-age endurance: DNA evidence of a white spruce refugium in Alaska. Proceedings of the National Academy of Sciences of the United States of America 103:12447-12450. Araujo M. B., A. Guisan. 2006. Five (or so) challenges for species distribution modelling. Journal of Biogeography 33:1677-1688. Araujo M. B., M. Cabeza, W. Thuiller, L. Hannah, and P. Williams. 2004. Would climate change drive species out of reserves? An assessment of existing reserve-selection methods. Global Change Biology 10:1618-1626. Axelson J. N., R. I. Alfaro, and B. C. Hawkes. 2009. Influence of fire and mountain pine beetle on the dynamics of lodgepole pine stands in British Columbia, Canada. Forest Ecology and Management 257:1874-1882. Barber V. A., G. P. Juday, and B. P. Finney. 2000. Reduced growth of Alaskan white spruce in the twentieth century from temperature-induced drought stress. Nature 405:668-673. Barrett R. D. H., D. Schluter. 2008. Adaptation from standing genetic variation. Trends in Ecology & Evolution 23:38-44. Baskin C. C., J. M. Baskin. 1998. Seeds, Ecology, Biogeography and Evolution of Dormancy, and Germination. Academic Press, San Diego, CA, USA. Bell G., A. Gonzalez. 2009. Evolutionary rescue can prevent extinction following environmental change. Ecology Letters 12:942-948. Beniston M., F. Keller, and S. Goyette. 2003. Snow pack in the Swiss Alps under changing climatic conditions: an empirical approach for climate impacts studies. Theoretical and Applied Climatology 74:19-31. Berdeen J. C., L. E. Riley, R. S. Sniezko. 2007. Whitebark pine seed storage and germination: a follow-up look at seedlots from Oregon and Washington. Department of Agriculture, Forest Service. Report R6-NR-FHP-2007-01. 113 p. Berry P. M., T. P. Dawson, P. A. Harrison, and R. G. Pearson. 2002. Modelling potential impacts of climate change on the bioclimatic envelope of species in Britain and Ireland. Global Ecology and Biogeography 11:453-462. Boisvenue C., S. W. Running. 2006. Impacts of climate change on natural forest productivity evidence since the middle of the 20th century. Global Change Biology 12:862-882. Bonan G. B. 2008. Forests and climate change: Forcings, feedbacks, and the climate benefits of forests. Science 320:1444-1449. 98  Bossdorf O., C. L. Richards, and M. Pigliucci. 2008. Epigenetics for ecologists. Ecology Letters 11:106-115. Bower A. D., S. N. Aitken. 2008. Ecological genetics and seed transfer guidelines for Pinus albicaulis (Pinaceae). American Journal of Botany 95:66-76. Bower A. D., S. N. Aitken. 2006. Geographic and seasonal variation in cold hardiness of whitebark pine. Canadian Journal of Forest Research 36:1842-1850. Bower A. D., S. C. McLane, A. Eckert, S. Jorgensen, A. Schoettle, S. N. Aitken. In Press. Conservation genetics of high elevation five-needle pines. In High-Five Symposium: the future of high-elevation five-needle white pines in western North America. USDA Forest Service. Missoula, MT, USA. Bradshaw W. E., C. M. Holzapfel. 2006. Evolutionary response to rapid climate change. Science 312:1477-1478. Briffa K. R., F. H. Schweingruber, P. D. Jones, T. J. Osborn, S. G. Shiyatov, and E. A. Vaganov. 1998. Reduced sensitivity of recent tree-growth to temperature at high northern latitudes. Nature 391:678-682. Brooks J. R., L. B. Flanagan, and J. R. Ehleringer. 1998. Responses of boreal conifers to climate fluctuations: indications from tree-ring widths and carbon isotope analyses. Canadian Journal of Forest Research 28:524-533. Caccianiga M., S. Payette. 2006. Recent advance of white spruce (Picea glauca) in the coastal tundra of the eastern shore of Hudson Bay (Quebec, Canada). Journal of Biogeography 33:2120-2135. Callaway R. M. 1998. Competition and facilitation on elevation gradients in subalpine forests of the northern Rocky Mountains, USA. Oikos 82:561-573. Christensen JH, Hewitson B, Busuioc A, et al. Regional climate projections. 2007. Climate Change 2007: the Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom. 94 p. Chuine I., G. E. Rehfeldt, and S. N. Aitken. 2006. Height growth determinants in pines: a case study of Pinus contorta and Pinus monticola. Canadian Journal of Forest Research 36:10591066. Cook E. R., R. L. Holmes. 1986. User's manual for Program ARSTAN. Pages 50-65 In R. L. Holmes, R. K. Adams, and H. C. Fritts, editors. Tree-ring chronologies of western North America: California, Eastern Oregon and Northern Great Basin, Laboratory of Tree Ring Research, University of Arizona, Tucson, Arizona, USA. Cornelius J. 1994. Heritabilities and additive genetic coefficients of variation in forest trees. Canadian Journal of Forest Research 24:372-379. COSEWIC (Committee on the Status of Endangered Wildlife in Canada). 2010. COSEWIC assessment and status report on the Whitebark Pine Pinus albicaulis in Canada. Ottawa, Ontario, Canada: Committee on the Status of Endangered Wildlife in Canada. 44 p. Crick H. Q. P., T. H. Sparks. 1999. Climate change related to egg-laying trends. Nature 399:423424. Danby R. K., D. S. Hik. 2007. Variability, contingency and rapid change in recent subarctic alpine tree line dynamics. Journal of Ecology 95:352-363. Daniels L. D., T. T. Veblen. 2004. Spatiotemporal influences of climate on altitudinal treeline in northern Patagonia. Ecology 85:1284-1296. D'Arrigo R. D., R. K. Kaufmann, N. Davi, G. C. Jacoby, C. Laskowski, R. B. Myneni, and P. Cherubini. 2004. Thresholds for warming-induced growth decline at elevational tree line in the Yukon Territory, Canada. Global Biogeochemical Cycles 18:GB3021. 99  Davis A. J., L. S. Jenkinson, J. H. Lawton, B. Shorrocks, and S. Wood. 1998. Making mistakes when predicting shifts in species range in response to global warming. Nature 391:783-786. Davis M. B., R. G. Shaw. 2001. Range shifts and adaptive responses to Quaternary climate change. Science 292:673-679. Davis M. B., R. G. Shaw, and J. R. Etterson. 2005. Evolutionary responses to changing climate. Ecology 86:1704-1714. Dormann C. F. 2007. Promising the future? Global change projections of species distributions. Basic and Applied Ecology 8:387-397. Ettl G. J., D. L. Peterson. 1995. Growth-response of sub-alpine fir (Abies Lasiocarpa) to climate in the Olympic Mountains, Washington, USA. Global Change Biology 1:213-230. Falconer D. S., T. F. C. Mackay. 1996. Introduction to Quantitative Genetics. Fourth Edition. Addison Wesley Longman, Harlow, Essex, United Kingdom. Flato G. M., G. J. Boer, W. G. Lee, N. A. McFarlane, D. Ramsden, M. C. Reader, and A. J. Weaver. 2000. The Canadian Centre for Climate Modelling and Analysis global coupled model and its climate. Climate Dynamics 16:451-467. Fritts H. C. 1976. Tree Rings and Climate. Academic Press, San Francisco, California, USA. Gamache I., S. Payette. 2005. Latitudinal response of subarctic tree lines to recent climate change in eastern Canada. Journal of Biogeography 32:849-862. Gamache I., S. Payette. 2004. Height growth response of tree line black spruce to recent climate warming across the forest-tundra of eastern Canada. Journal of Ecology 92:835-845. Germino M. J., W. K. Smith, and A. C. Resor. 2002. Conifer seedling distribution and survival in an alpine-treeline ecotone. Plant Ecology 162:157-168. Guisan A., W. Thuiller. 2005. Predicting species distribution: offering more than simple habitat models. Ecology Letters 8:993-1009. Hamann A., T. Wang. 2006. Potential effects of climate change on ecosystem and tree species distribution in British Columbia. Ecology 87:2773-2786. Hamann A., P. Smets, A. D. Yanchuk, and S. N. Aitken. 2005. An ecogeographic framework for in situ conservation of forest trees in British Columbia. Canadian Journal of Forest Research 35:2553-2561. Hampe A., R. J. Petit. 2005. Conserving biodiversity under climate change: the rear edge matters. Ecology Letters 8:461-467. Hartl D. L., A. G. Clark. 1997. Principles of Population Genetics. Sinauer Associates, Inc., Sunderland, MA. Heineman J. L., D. L. Sachs, W. J. Mather, and S. W. Simard. 2010. Investigating the influence of climate, site, location, and treatment factors on damage to young lodgepole pine stands in southern British Columbia. Canadian Journal of Forest Research 40:1109:1127. Hobbie S. E., F. S. Chapin. 1998. An experimental test of limits to tree establishment in Arctic tundra. Journal of Ecology 86:449-461. Hoegh-Guldberg O., L. Hughes, S. McIntyre, D. B. Lindenmayer, C. Parmesan, H. P. Possingham, and C. D. Thomas. 2008. Assisted colonization and rapid climate change. Science 321:345-346. Holliday J. A., K. Ritland, and S. N. Aitken. 2010. Widespread, ecologically relevant genetic markers developed from association mapping of climate-related traits in Sitka spruce (Picea sitchensis). New Phytologist 188:501-514. Holmes R. L. 1986. Quality control of crossdating and measuring: a user's manual for program COFECHA. Pages 41-49 In R. L. Holmes, R. K. Adams, and H. C. Fritts, editors. Tree-ring chronologies of western North America: California, Eastern Oregon and Northern Great Basin, Laboratory of Tree-Ring Research, University of Arizona, Tucson, Arizona, USA. 100  Howe G. T., S. N. Aitken, D. B. Neale, K. D. Jermstad, N. C. Wheeler, and T. H. H. Chen. 2003. From genotype to phenotype: unraveling the complexities of cold adaptation in forest trees. Canadian Journal of Botany 81:1247-1266. Hughes M. K. 2002. Dendrochronology in climatology: the state of the art. Dendrochronologia 20:95-116. Husch B., C. I. Miller, and G. Bell. 1972. Growth of the Tree. John Wiley & Sons, Toronto. 402 p. Illingworth K. Study of lodgepole pine genotype-environment interaction in BC. Victoria, British Columbia, Canada: Ministry of Forests, Mines and Lands, Information Services Branch; 1978. Report nr Proceedings International Union of Forestry Research Organizations (IUFRO) joint meeting of working parties: Douglas-fir provenances, lodgepole pine provenances, Sitka spruce provenances, and Abies provenances. 151 p. IPCC. 2007. Climate change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC). Cambridge University Press, New York, NY, USA. Iverson L. R., M. W. Schwartz, and A. M. Prasad. 2004. How fast and far might tree species migrate in the eastern United States due to climate change? Global Ecology and Biogeography 13:209-219. Jacobs J., T. Weaver. 1990. Effects of temperature and temperature preconditioning on seedling performance of whitebark pine. Pages 134-139 In W. C. Schmidt and K. J. McDonald, editors. Whitebark pine ecosystems - Ecology and management of a high mountain resource, USDA Forest Service, Intermountain Research Station, General Technical Report INT-270, Ogden, Utah. Johnstone J. F., F. S. Chapin. 2003. Non-equilibrium succession dynamics indicate continued northern migration of lodgepole pine. Global Change Biology 9:1401-1409. Kawecki T. J. 2008. Adaptation to Marginal Habitats. Annual Review of Ecology Evolution and Systematics 39:321-342. Keane R. E., D. Tomback, C. Aubry, A. Bower, E. Campbell, M. Jenkins, M. Manning, S. McKinney, M. Murray, D. Perkins, D. Reinhart, C. Ryan, A. Schoettle, C. Smith. In Press. A range-wide restoration strategy for whitebark pine (Pinus albicaulis). Department of Agriculture, Forest Service, Rocky Mountain Research Station, Fort Collins, CO, USA. General Technical Report RMRS-GTR-xxx. Kendall K. C., R. E. Keane. 2001. Whitebark pine decline: infection, mortality and population trends. Pages 221-242 In D. F. Tomback, S. F. Arno, and R. E. Keane, editors. Whitebark pine communities: ecology and restoration, Island Press, Washington DC, USA. Kilpelainen A., H. Peltola, I. Rouvinen, and S. Kellomaki. 2006. Dynamics of daily height growth in Scots pine trees at elevated temperature and CO2. Trees-Structure and Function 20:16-27. Korner C. 2003. Alpine plant life: functional plant ecology of high mountain ecosystems. Springer, New York. Kozlowski T. T., P. J. Kramer, and S. G. Pallardy. 1991. The Physiological Ecology of Woody Plants. Academic Press, Inc., San Diego. Kurz W. A., G. Stinson, G. J. Rampley, C. C. Dymond, and E. T. Neilson. 2008. Risk of natural disturbances makes future contribution of Canada's forests to the global carbon cycle highly uncerain. Proceedings of the National Academy of Sciences of the United States of America 105:1551-1555.  101  Lahti M., P. J. Aphalo, L. Finer, A. Ryyppo, T. Lehto, and H. Mannerkoski. 2005. Effects of soil temperature on shoot and root growth and nutrient uptake of 5-year-old Norway spruce seedlings. Tree physiology 25:115-122. Langlet O. 1971. Two hundred years genecology. Taxon 20:653-721. Lapointe-Garant M., J. Huang, G. Gea-Izquierdo, F. Raulier, P. Bernier, and P. M. Berry. 2009. Use of tree rings to study the effect of climate change on trembling aspen in Quebec. Global Change Biology 16:2039-2051. Littell R. C., G. A. Milliken, W. W. Stroup, R. D. Wolfinger, and O. Schabenberger. 2006. Random Coefficients Models. SAS Institute, Inc, Cary, NC, USA. Loarie S. R., P. B. Duffy, H. Hamilton, G. P. Asner, C. B. Field, and D. D. Ackerly. 2009. The velocity of climate change. Nature 462:1052-U111. Logan J. A., W. W. Macfarlane, and L. Willcox. 2010. Whitebark pine vulnerability to climatedriven mountain pine beetle disturbance in the Greater Yellowstone Ecosystem. Ecological Applications 20:895-902. Lotan J. E., W. B. Critchfield. 1990. Pinus contorta Dougl. ex. Loud. Lodgepole pine. Pages 302-315 In R. M. Burns and B. H. Honkala, editors. Silvics of North America. Vol. 1, Conifers, USDA Forest Service Agricultural Handbook. 665 p. Mahalovich M. F., K. E. Burr, D. L. Foushee, and USDA. 2006. Whitebark pine germination, rust resistance, and cold hardiness among seed sources in the inland northwest: Planting strategies for restoration. National Proceedings: Forest and Conservation Nursery Associations 2005 43:91-101. Maher E. L., M. J. Germino. 2006. Microsite differentiation among conifer species during seedling establishment at alpine treeline. Ecoscience 13:334-341. Malcolm J. R., A. Markham, R. P. Neilson, and M. Garaci. 2002. Estimated migration rates under scenarios of global climate change. Journal of Biogeography 29:835-849. Masek J. G. 2001. Stability of Boreal Forest Stands during Recent Climate Change: Evidence from Landsat Satellite Imagery. Journal of Biogeography 28:967-976. Mattson D. J., B. M. Blanchard, and R. R. Knight. 1992. Yellowstone grizzly bear mortality, human habituation, and whitebark-pine seed crops. Journal of Wildlife Management 56:432442. McCaughey W. W., D. F. Tomback. 2001. The natural regeneration process. Pages 105-120 In D. F. Tomback, S. F. Arno, and R. E. Keane, editors. Whitebark pine communities: ecology and restoration, Island Press, Washington DC, USA. McCaughey W., G. L. Scott, and K. L. Izlar. 2009. Whitebark pine planting guidelines. Western Journal of Applied Forestry 24:163-166. McKenney D. W., J. H. Pedlar, K. Lawrence, K. Campbell, and M. F. Hutchinson. 2007. Potential impacts of climate change on the distribution of North American trees. Bioscience 57:939-948. McLachlan J. S., J. J. Hellmann, and M. W. Schwartz. 2007. A framework for debate of assisted migration in an era of climate change. Conservation Biology 21:297-302. McLachlan J. S., J. S. Clark, and P. S. Manos. 2005. Molecular indicators of tree migration capacity under rapid climate change. Ecology 86:2088-2098. McLane S. C., S. N. Aitken. Under review. Whitebark pine (Pinus albicaulis) assisted migration potential: testing establishment north of the species range. Ecological Applications. McLane S. C., V. M. LeMay, and S. N. Aitken. 2011. Modeling lodgepole pine radial growth relative to climate and genetics using universal growth-trend response functions. Ecological Applications. 102  Meidinger D., J. Pojar. 1991. Ecosystems of British Columbia. Special Report Series 6, British Columbia Ministry of Forests, Mines and Lands, Victoria, BC, Canada. Mellmann-Brown S. 2005. Regeneration of whitebark pine in the timberline ecotone of the Beartooth Plateau, U.S.A.: spatial distribution and responsible agents. Pages 97-115 In G. Broll and B. Keplin, editors. Mountain Ecosystems: Studies in Treeline Ecology, Springer Berlin-Heidelberg, Germany. Mellmann-Brown S. 2002. The regeneration of whitebark pine in the timberline ecotone of the beartooth Plateau, Montana and Wyoming. Unpublished dissertation. 157 p. Midgley G. F., L. Hannah, D. Millar, M. C. Rutherford, and L. W. Powrie. 2002. Assessing the vulnerability of species richness to anthropogenic climate change in a biodiversity hotspot. Global Ecology and Biogeography 11:445-451. Millar C. I., R. D. Westfall, D. L. Delany, J. C. King, and L. J. Graumlich. 2004. Response of subalpine conifers in the Sierra Nevada, California, USA, to 20th-century warming and decadal climate variability. Arctic Antarctic and Alpine Research 36:181-200. Mimura M., S. N. Aitken. 2007a. Increased selfing and decreased effective pollen donor number in peripheral relative to central populations in Picea sitchensis (Pinaceae). American Journal of Botany 94:991-998. Mimura M., S. N. Aitken. 2007b. Adaptive gradients and isolation-by-distance with postglacial migration in Picea sitchensis. Heredity 99:224-232. Ministry of Water, Land and Air Protection. 2002. Indicators of Climate Change for British Columbia. National Library of Canada, Victoria, BC, Canada. 50 p. Mohatt K. R. 2006. Ectomycorrhizal fungi of whitebark pine (Pinus albicaulis) in the northern greater Yellowstone ecosystem. Unpublished thesis. Montana State University, Bozeman, MT, USA. Mokany K., R. J. Raison, and A. S. Prokushkin. 2006. Critical analysis of root: shoot ratios in terrestrial biomes. Global Change Biology 12:84-96. Moles A. T., M. Westoby. 2003. Latitude, seed predation and seed mass. Journal of Biogeography 30:105-128. Morgenstern E. K. 1996. Geographic variation in forest trees: Genetic basis and application of knowledge in silviculture. UBC Press, Vancouver, BC, Canada. Mueller J. M., J. J. Hellmann. 2008. An assessment of invasion risk from assisted migration. Conservation Biology 22:562-567. Neale D. B., O. Savolainen. 2004. Association genetics of complex traits in conifers. Trends in plant science 9:325-330. Nogués-Bravo D., M. B. Araújo, M. P. Errea, and J. P. Martínez-Rica. 2007. Exposure of global mountain systems to climate warming during the 21st Century. Global Environmental Change 17:420-428. Ogilvie R. T. 1990. Distribution and ecology of whitebark pine in western Canada. Pages 40-48 In W. C. Schmidt and K. J. McDonald, editors. Proceedings - Symposium on whitebark pine ecosystems: Ecology and management of a high-mountain resources. USDA Forest Service, Intermountain Research Station, General Technical Report INT-270, Ogden, Utah. Oleksyn J., M. G. Tjoelker, and P. B. Reich. 1998. Adaptation to changing environment in Scots pine populations across a latitudinal gradient. Silvae Genetica 32:129-140. Olszyk D., C. Wise, E. VanEss, and D. Tingey. 1998. Elevated temperature but not elevated CO2 affects long-term patterns of stem diameter and height of Douglas-fir seedlings. Canadian Journal of Forest Research 28:1046-1054.  103  O'Neill G. A., A. Hamann, and T. Wang. 2008. Accounting for population variation improves estimates of the impact of climate change on species' growth and distribution. Journal of Applied Ecology 45:1040-1049. Parmesan C. 2006. Ecological and evolutionary responses to recent climate change. Annual Review of Ecology Evolution and Systematics 37:637-669. Parmesan C., G. Yohe. 2003. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37-42. Pauli H., M. Gottfried, K. Reier, C. Klettner, and G. Grabherr. 2007. Signals of range expansions and contractions of vascular plants in the high Alps: observations (1994-2004) at the GLORIA*master site Schrankogel, Tyrol, Austria. Global Change Biology 13:147-156. Pautasso M. 2009. Geographical genetics and the conservation of forest trees. Perspectives in Plant Ecology Evolution and Systematics 11:157-189. Pearson R. G., T. P. Dawson. 2003. Predicting the impacts of climate change on the distribution of species: are bioclimate envelope models useful? Global Ecology and Biogeography 12:361-371. Pearson R. G., T. P. Dawson, P. M. Berry, and P. A. Harrison. 2002. SPECIES: A Spatial Evaluation of Climate Impact on the Envelope of Species. Ecological Modelling 154:289300. Peltola H., A. Kilpelainen, and S. Kellomaki. 2002. Diameter growth of Scots pine (Pinus sylvestris) trees grown at elevated temperature and carbon dioxide concentration under boreal conditions. Tree physiology 22:963-972. Peterson D. W., D. L. Peterson. 2001. Mountain hemlock growth responds to climatic variability at annual and decadal time scales. Ecology 82:3330-3345. Reale D., A. G. McAdam, S. Boutin, and D. Berteaux. 2003. Genetic and plastic responses of a northern mammal to climate change. Proceedings of the Royal Society of London Series BBiological Sciences 270:591-596. Rehfeldt G. E., W. R. Wykoff, and C. C. Ying. 2001. Physiologic plasticity, evolution, and impacts of a changing climate on Pinus contorta. Climatic Change 50:355-376. Rehfeldt G. E., C. C. Ying, D. L. Spittlehouse, and D. A. Hamilton. 1999. Genetic responses to climate in Pinus contorta: Niche breadth, climate change, and reforestation. Ecological Monographs 69:375-407. Rehfeldt G. E., N. M. Tchebakova, Y. I. Parfenova, W. R. Wykoff, N. A. Kuzmina, and L. I. Milyutin. 2002. Intraspecific responses to climate in Pinus sylvestris. Global Change Biology 8:912-929. Reich P. B., J. Oleksyn. 2008. Climate warming will reduce growth and survival of Scots pine except in the far north. Ecology Letters 11:588-597. Ricciardi A., D. Simberloff. 2009. Assisted colonization is not a viable conservation strategy. Trends in Ecology & Evolution 24:248-253. Richardson D. M., M. Rejmanek. 2004. Conifers as invasive aliens: a global survey and predictive framework. Diversity and Distributions 10:321-331. Richardson D. M., J. J. Hellmann, J. S. McLachlan, D. F. Sax, M. W. Schwartz, P. Gonzalez, E. J. Brennan, A. Camacho, T. L. Root, O. E. Sala, S. H. Schneider, D. M. Ashe, J. R. Clark, R. Early, J. R. Etterson, E. D. Fielder, J. L. Gill, B. A. Minteer, S. Polasky, H. D. Safford, A. R. Thompson, and M. Vellend. 2009. Multidimensional evaluation of managed relocation. Proceedings of the National Academy of Sciences 106:9721-9724. Riley L. E., C. M. Coumas, J. F. Danielson, J. C. Berdeen. 2007. Seedling nursery culture of whitebark pine at Dorena Genetic Resources Center: headaches, successes and growing pains. USDA Forest Service Report R6-NR-FHP-2007-01. p. 122-131. 104  Root T. L., J. T. Price, K. R. Hall, S. H. Schneider, C. Rosenzweig, and J. A. Pounds. 2003. Fingerprints of global warming on wild animals and plants. Nature 421:57-60. Safranyik L., B. Wilson (eds.). 2006. The mountain pine beetle: a synthesis of biology, management, and impacts on lodgepole pine. Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, Victoria, BC, Canada. Savolainen O., T. Pyhajarvi, and T. Knurr. 2007. Gene flow and local adaptation in tees. Annual Review of Ecology Evolution and Systematics 38:595-619. Savva Y., F. Schweingruber, L. Milyutin, and E. Vaganov. 2002. Genetic and environmental signals in tree rings from different provenances of Pinus sylvestris L. planted in the southern taiga, central Siberia. Trees-Structure and Function 16:313-324. Savva Y., Y. Bergeron, B. Denneler, A. Koubaa, and F. Tremblay. 2008. Effect of interannual climate variations on radial growth of jack pine provenances in Petawawa, Ontario. Canadian Journal of Forest Research 38:619-630. Savva Y., B. Denneler, A. Koubaa, F. Tremblay, Y. Bergeron, and M. G. Tjoelker. 2007. Seed transfer and climate change effects on radial growth of jack pine populations in a common garden in Petawawa, Ontario, Canada. Forest Ecology and Management 242:636-647. Saxe H., M. G. R. Cannell, B. Johnsen, M. G. Ryan, and G. Vourlitis. 2001. Tree and forest functioning in response to global warming. New Phytologist 149:369-399. Schwartz M. W., J. J. Hellmann, and J. S. McLachlan. 2009. The precautionary principle in managed relocation is misguided advice. Trends in Ecology & Evolution 24:474-474. Sexton J. P., P. J. McIntyre, A. L. Angert, and K. J. Rice. 2009. Evolution and Ecology of Species Range Limits. Annual Review of Ecology Evolution and Systematics 40:415-436. Skelly D. K., L. N. Joseph, H. P. Possingham, L. K. Freidenburg, T. J. Farrugia, M. T. Kinnison, and A. P. Hendry. 2007. Evolutionary responses to climate change. Conservation Biology 21:1353-1355. Smith C. M., B. Wilson, S. Rasheed, R. C. Walker, T. Carolin, and B. Shepherd. 2008. Whitebark pine and white pine blister rust in the Rocky Mountains of Canada and northern Montana. Canadian Journal of Forest Research 38:982-995. Smith W. K., M. J. Germino, T. E. Hancock, and D. M. Johnson. 2003. Another perspective on altitudinal limits of alpine timberlines. Tree physiology 23:1101-1112. Stokes M. A., T. L. Smiley. 1968. An introduction to tree-ring dating. The University of Chicago Press, Chicago, IL. Stromgren M., S. Linder. 2002. Effects of nutrition and soil warming on stemwood production in a boreal Norway spruce stand. Global Change Biology 8:1195-1204. Taiz L., E. Zeiger. 2002. Plant Physiology, 3rd Edition. Sinauer Associates, Inc., Sunderland, MA, USA. Takenaka A. 2005. Local coexistence of tree species and the dynamics of global distribution pattern along an environmental gradient: a simulation study. Ecological Research 20:297304. Thomas C. D., A. Cameron, R. E. Green, M. Bakkenes, L. J. Beaumont, Y. C. Collingham, B. F. N. Erasmus, M. F. de Siqueira, A. Grainger, L. Hannah, L. Hughes, B. Huntley, A. S. van Jaarsveld, G. F. Midgley, L. Miles, M. A. Ortega-Huerta, A. T. Peterson, O. L. Phillips, and S. E. Williams. 2004. Extinction risk from climate change. Nature 427:145-148. Tomback D. F. 2001. Clark's nutcracker: agents of regeneration. Pages 89-104 In D. F. Tomback, S. F. Arno, and R. E. Keane, editors. Whitebark pine communities: ecology and restoration, Island Press, Washington DC, USA. Tomback D. F., P. Achuff. 2010. Blister rust and western forest biodiversity: ecology, values and outlook for white pines. Forest Pathology 40:186-225. 105  Tomback D. F., A. J. Anderies, K. S. Carsey, M. L. Powell, and S. Mellmann-Brown. 2001. Delayed seed germination in whitebark pine and regeneration patterns following the yellowstone fires. Ecology 82:2587-2600. Tomback D. F. 1982. Dispersal of whitebark pine seeds by Clark's Nutcracker: a mutualism hypothesis. Journal of Animal Ecology 51:451-467. Turesson G. 1922. The genotypical response of the plant species to the habitat. Hereditas 3:211350. Van der Veken S., M. Hermy, M. Vellend, A. Knapen, and K. Verheyen. 2008. Garden plants get a head start on climate change. Frontiers in Ecology and the Environment 6:212-216. Visser M. E. 2008. Keeping up with a warming world: Assessing the rate of adaptation to climate change. Proceedings of the Royal Society B-Biological Sciences 275:649-659. Walther G. R., E. Post, P. Convey, A. Menzel, C. Parmesan, T. J. C. Beebee, J. M. Fromentin, O. Hoegh-Guldberg, and F. Bairlein. 2002. Ecological responses to recent climate change. Nature 416:389-395. Wang T., G. A. O'Neill, and S. N. Aitken. 2010. Integrating environmental and genetic effects to predict responses of tree populations to climate. Ecological Applications 20:153-163. Wang T., A. Hamann, A. Yanchuk, G. A. O'Neill, and S. N. Aitken. 2006a. Use of response functions in selecting lodgepole pine populations for future climate. Global Change Biology 12:1-13. Wang T., A. Hamann, D. L. Spittlehouse, and S. N. Aitken. 2006b. Development of scale-free climate data for western Canada for use in resource management. International Journal of Climatology 26:383-397. Wang T. L., S. C. McLane, E. M. Campbell, and T. A. Schroeder. In preparation. Whitebark pine (Pinus albicaulis) may not survive climate change. Wang T., S. N. Aitken, P. Rozenberg, and M. R. Carlson. 1999. Selection for height growth and Pilodyn pin penetration in lodgepole pine: Effects on growth traits, wood properties, and their relationships. Canadian Journal of Forest Research 29:434-445. Warwell M. V., G. E. Rehfeldt, N. L. Crookston. 2007. Modeling contemporary climate profiles of whitebark pine (Pinus albicaulis) and predicting responses to global warming. Department of Agriculture Forest Service Report R6-NR-FHP-2007-01. p. 139-142. Way D. A., R. Oren. 2010. Differential responses to changes in growth temperature between trees from different functional groups and biomes: a review and synthesis of data. Tree physiology 30:669-688. Weaver T. 2001. Whitebark pine and its environment. Pages 41-73 In D. F. Tomback, S. F. Arno, and R. E. Keane, editors. Whitebark Pine Communities: Ecology and Restoration, Island Press, Washington DC, USA. Weaver T. 1994. Vegetation distribution and production in Rocky Mountain climates - with emphasis on whitebark pine. Pages 142-152 In W. C. Schmidt and F. K. Holtmeier, editors. Proceedings - International workshop on subalpine stone pines and their environments: The status of our knowledge. USDA Forest Service, Intermountain Research Station, General Technical Report INT-GTR-309, Ogden, Utah. Whitlock C., P. J. Bartlein. 1997. Vegetation and climate change in northwest America during the past 125 kyr. Nature 388:57-61. Wilmking M., G. P. Juday, V. A. Barber, and H. S. J. Zald. 2004. Recent climate warming forces contrasting growth responses of white spruce at treeline in Alaska through temperature thresholds. Global Change Biology 10:1724-1736.  106  Woodall C. W., D. J. Nowak, G. C. Liknes, and J. A. Westfall. 2010. Assessing the potential for urban trees to facilitate forest tree migration in the eastern United States. Forest Ecology and Management 259:1447-1454. Xie C. Y., C. C. Ying. 1995. Genetic architecture and adaptive landscape of interior lodgepole pine (Pinus contorta ssp latifolia) in Canada. Canadian Journal of Forest Research 25:20102021. Yeh H. Y., L. C. Wensel. 2000. The relationship between tree diameter growth and climate for coniferous species in northern California. Canadian Journal of Forest Research 30:14631471. Zeglen S. 2002. Whitebark pine and white pine blister rust in British Columbia, Canada. Canadian Journal of Forest Research 32:1265-1274.  107  Appendices Appendix 1 Methods for creating current-observed, current-predicted and futurepredicted species distribution models for whitebark pine within British Columbia, Canada.  In 2006, species distribution models for whitebark pine within British Columbia were created by Tongli Wang (University of British Columbia, unpublished) using the climateenvelope modeling technique of Hamann and Wang (2006). The current-observed range for whitebark pine was determined using 479 presence observations from the botanical inventory used to create the BC Ministry of Forests and Range‟s Biogeoclimatic Ecological Classification (BEC) system (Figure 1a). The BEC system is a hierarchical classification system that divides BC‟s landbase into 14 zones, 97 subzones and 152 variants based on vegetation, soil, climate and topography (Meidinger and Pojar 1991). The variant level describes land units comprising relatively homogeneous ecological and geoclimatic features. To create current-observed species ranges, Hamann and Wang (2006) extrapolated the occurrence data from one-dimensional observation points to two-dimensional BEC variant polygons, under the assumption that a species should be able to grow anywhere within a variant in which it is observed. Variant-level divisions had not been delineated for the alpine tundra BEC zone at the time that Hamann and Wang created the SDMs, so instead they created alpine tundra pseudo-variants based on geographic divisions between mountain ranges. High-elevation areas with permanent icefields were excluded, and polygons with low (<1/100 of average) predicted species frequencies eliminated. ClimateBC v3.1 (Wang et al. 2006b) was then used to generate biologically-relevant normal (1961-1990) climate variables associated with whitebark pine‟s current-observed range. ClimateBC interpolates weather station data using high-resolution digital elevation models that accurately capture climatic variance in BC‟s mountainous terrain. Whitebark pine‟s currentpredicted range (Figure 1b) was extrapolated by selecting all areas in the province with normal climate conditions in the range of those currently experienced by the species, accounting for climatic interactions. Future-predicted ranges for 2025, 2055 and 2085 were then projected using “middle of the road” (IS92a) Coupled Global Circulation Model ensemble mean (CGCM1 GAX) carbon scenarios developed by the Canadian Centre for Climate Modeling and Analysis (Flato et al. 2000, accessed using ClimateBC). Recently, whitebark pine‟s present and future 108  distributions were remodelled using a classification and regression-tree procedure called Random Forests, yielding broadly similar predictions (Wang et al. in prep.).  109  Appendix 2 Whitebark pine data summary  Table A.1: Whitebark pine germination, survival and growth summary by population, site and seed treatment. Survivors Total Germinants Germinants Survivors Height Seed (% of seeds (#) (total #) (total %) (total #) (cm) Population or site treatment germinants) Population Tatla Lake 1244 127 10% 94 74% 2.8 Fort St. James 1246 101 8% 79 78% 2.8 Penticton 1240 49 4% 37 76% 2.7 Untreated Whistler 1238 40 3% 24 60% 2.6 John Day, OR 1232 226 18% 181 80% 2.9 Entiat, WA 1250 218 17% 188 86% 2.8 Tatla Lake 1125 533 47% 291 55% 3.2 Fort St. James 1137 413 36% 230 56% 3.1 Penticton 1127 102 9% 42 41% 2.9 Treated Whistler 1121 171 15% 66 39% 2.8 John Day, OR 524 258 49% 160 62% 3.5 Entiat, WA 1104 362 33% 202 56% 3.3 Location and site Atlin 2 480 10 2% 7 70% 2.7 Atlin 1 478 7 1% 5 71% 2.1 Bell II 2 & 3 480 17 4% 16 94% 2.3 Bell II 1 480 31 6% 28 90% 2.6 Smithers 2 480 40 8% 35 88% 2.9 Smithers 1 250 42 17% 40 95% 2.5 Whistler 1 478 25 5% 11 44% 1.9 Untreated Whistler 2 482 15 3% 6 40% 2.2 Haines Junction 2 480 95 20% 86 91% 2.7 Terrace 2 480 64 13% 31 48% 3.6 Haines Junction 1 482 52 11% 39 75% 2.5 Stewart 1 480 13 3% 10 77% 2.6 Stewart 2 480 15 3% 15 100% 2.8 New Hazelton 1 480 135 28% 121 90% 3.1 New Hazelton 2 480 123 26% 100 81% 3.0 Terrace 1 480 77 16% 53 69% 2.7 Atlin 2 415 129 31% 10 8% 2.5 Atlin 1 408 136 33% 39 29% 2.6 Bell II 2 & 3 691 228 33% 73 32% 2.7 Smithers 2 440 117 27% 54 46% 3.1 Smithers 1 440 141 32% 81 57% 3.0 Whistler 1 424 89 21% 33 37% 2.8 Haines Junction 2 424 122 29% 100 82% 3.3 Treated Terrace 2 436 115 26% 79 69% 3.8 Haines Junction 1 424 135 32% 100 74% 3.1 Stewart 1 372 90 24% 55 61% 3.6 Stewart 2 371 70 19% 56 80% 3.0 New Hazelton 1 428 170 40% 104 61% 3.5 New Hazelton 2 429 139 32% 91 65% 3.4 Terrace 1 436 158 36% 116 73% 3.3 Notes : Populations and sites are ordered from coldest to warmest normal annual temperature.  110  Fascicles (#) 3.7 3.3 3.8 3.4 4.6 3.9 5.9 5.2 5.8 4.2 7.2 6.3 1.8 1.8 1.7 3.2 4.1 3.5 0.7 2.6 5.3 4.3 4.1 4.4 5.1 4.3 3.8 3.4 4.7 3.9 4.1 5.6 5.2 3.6 7.4 6.1 6.0 6.0 7.8 5.9 6.0 6.9  Appendix 3 Whitebark and lodgepole pine provenance data  Table A.2: Geographic and climatic data for the six whitebark pine and six lodgepole pine provenances.  Lodgepole pine  Whitebark pine  Latitude Longitude Elevation MAT PAS Seed weight (g Species Closest Town (°N) (°W) (m) (°C) (cm) per 100 seeds) Tatla Lake, BC 52.54 125.81 1541 0.1 524 16.6 Fort St. James, BC 54.88 125.37 1490 0.2 396 16.9 Penticton, BC 49.37 119.92 2148 0.7 336 10.2 Whistler, BC 50.10 122.90 1882 0.8 1290 9.7 John Day, OR 44.28 118.70 2438 3.7 572 12.2 Entiat, WA 47.99 120.41 1998 6.5 463 12.7 Gang Ranch, BC 51.47 122.60 1530 1.5 183 Kelowna, BC 49.83 118.33 1654 1.9 504 Burns Lake, BC 53.93 126.08 895 2.4 203 Princeton, BC 49.78 120.05 1605 2.5 320 Quesnel, BC 53.08 122.13 892 3.4 197 Bestwick, BC 50.55 120.12 1080 4.4 140 Notes : MAT (mean annual temperature) and PAS (precipitation as snow) were generated using ClimateWNA and represent 1971-2000 normals (Wang et al. 2006). Provenances are ordered from coldest to warmest MAT. BC = British Columbia, Canada; OR = Oregon, USA; WA = Washington, USA.  111  Appendix 4 Whitebark pine germination and survival data  Table A.3: Germination and survival of each whitebark pine population by the end of the second growing season. Total Survivors (# per chamber) Germinants Germinants Survivors seeds Survivors Provenance (# out of (% of sown 7.9 10.8 13.0 15.4 17.9 (% of sown (total #) sown seeds) seeds) (°C) (°C) (°C) (°C) (°C) germinants) (#) Tatla Lake, BC 563 236 42% 40 38 41 46 43 208 88% Fort St. James, BC 490 178 36% 41 25 25 40 20 151 85% Penticton, BC 284 30 11% 7 3 2 6 4 22 73% Whistler, BC 435 78 18% 12 14 15 13 16 70 90% John Day, OR 390 128 33% 18 20 22 27 29 116 91% Entiat, WA 273 154 56% 26 31 23 31 26 137 89% Overall 2,803 804 29% 144 131 128 163 138 704 88% Lodgepole pine N/A 57 54 56 42 22 N/A Notes : Provenances are ordered from coldest to warmest normal mean annual temperature. BC = British Columbia, Canada; OR = Oregon, USA; WA = Washington, USA.  112  Appendix 5 Lodgepole pine chronology statistics  Table A.4: Ring-width and chronology statistics for 12 lodgepole pine populations planted at 16 test sites. Prov = provenance; ave = average; incr = increment; min = minimum; max = maximum. Data are ordered from coldest to warmest site normal temperature, with populations within sites ordered from coldest to warmest provenance.  Site  Sample  Ave  Min  Max  Annual  First-order  size  incr  incr  incr  incr  Sensitivity  autocorrelation EPS of of residual  (no. of  width  width width  width  of residual  residual  Prov trees)  (mm)  (mm)  (mm)  range  chronology chronology  chronology  WATS  163  12  1.73  0.87  2.30  1.43  0.07  0.38  0.47  WATS  35  11  1.39  1.01  2.26  1.25  0.11  0.54  0.72  WATS  26  11  1.74  0.48  2.66  2.18  0.09  0.44  0.82  WATS  142  12  1.70  1.21  2.29  1.08  0.11  0.28  0.85  BLUE  30  12  2.63  2.10  3.63  1.53  0.17  0.30  0.83  BLUE  163  10  2.57  1.76  3.23  1.47  0.15  0.11  0.79  BLUE  35  12  2.58  1.91  3.59  1.68  0.18  0.30  0.90  BLUE  26  12  2.47  1.57  3.41  1.84  0.21  0.23  0.77  BLUE  100  9  2.11  0.88  2.99  2.11  0.25  0.43  0.77  BLUE  142  13  2.25  1.33  3.20  1.87  0.16  0.15  0.87  BLUE  44  6  1.61  0.89  2.72  1.83  0.29  0.15  0.57  M451  30  10  2.03  1.23  2.76  1.53  0.11  0.18  0.78  M451  163  6  1.33  0.97  1.94  0.97  0.34  0.43  0.85  M451  35  10  1.27  0.76  1.72  0.96  0.18  0.32  0.82  M451  26  12  1.78  0.94  3.09  2.16  0.16  0.38  0.64  M451  100  12  1.66  1.15  2.08  0.93  0.14  0.38  0.84  M451  142  11  1.81  1.25  2.67  1.42  0.14  0.18  0.64  M451  71  7  1.39  0.95  1.93  0.98  0.30  0.29  0.71  M451  44  7  1.22  0.84  1.77  0.93  0.30  -0.18  0.71  M451  1  11  1.08  0.58  1.83  1.25  0.35  0.37  0.89  WHRS  163  12  2.30  1.11  3.43  2.32  0.09  -0.25  0.88  WHRS  35  12  2.13  1.06  3.32  2.26  0.10  0.16  0.87  WHRS  26  11  2.34  1.86  3.31  1.45  0.14  -0.18  0.90  113  Table A.4 (cont.): Ring-width and chronology statistics for 12 lodgepole pine populations planted at 16 test sites. Prov = provenance; ave = average; incr = increment; min = minimum; max = maximum. Data are ordered from coldest to warmest site normal temperature, with populations within sites ordered from coldest to warmest provenance. Sample  Ave  Min  Max  Annual  First-order  size  incr  incr  incr  incr  Sensitivity  autocorrelation EPS of of residual  (no. of  width  width width  width  of residual  Site  Prov trees)  (mm)  (mm)  range  chronology chronology  WHRS  142  12  2.50  1.47  3.88  2.41  0.14  -0.35  0.91  LUSS  30  12  2.02  1.26  2.77  1.51  0.10  0.08  0.85  LUSS  100  12  2.01  1.22  2.65  1.43  0.11  0.19  0.86  LUSS  71  12  2.35  1.82  3.34  1.52  0.14  0.22  0.93  LUSS  44  12  2.38  1.42  3.65  2.23  0.18  0.00  0.89  LUSS  63  12  2.49  1.59  2.99  1.40  0.15  0.19  0.91  LUSS  1  12  2.47  2.09  3.83  1.75  0.11  0.21  0.81  LUSS  72  12  2.23  1.76  3.56  1.80  0.24  0.07  0.88  SAMN  30  12  2.34  1.76  2.84  1.08  0.06  0.36  0.72  SAMN  163  10  2.08  1.67  2.73  1.06  0.07  -0.07  0.49  SAMN  35  12  2.72  2.15  3.35  1.21  0.09  -0.27  0.77  SAMN  100  12  2.97  2.53  3.44  0.91  0.07  -0.45  0.81  SAMN  71  12  2.62  1.92  3.51  1.59  0.07  0.02  0.77  SAMN  44  11  2.87  2.15  3.63  1.48  0.09  -0.05  0.77  SAMN  63  11  2.74  1.65  3.46  1.80  0.07  -0.38  0.71  MCLA  30  10  2.54  2.11  3.02  0.91  0.10  0.36  0.85  MONS  30  12  1.80  1.38  2.17  0.79  0.09  -0.03  0.93  MCLA  35  11  2.78  1.79  3.46  1.67  0.11  0.38  0.80  MONS  35  12  1.51  1.20  2.02  0.83  0.11  0.07  0.94  MCLA  100  12  2.77  2.06  3.90  1.84  0.11  0.28  0.85  MONS  100  12  2.02  1.76  2.60  0.84  0.11  0.14  0.90  MCLA  18  12  2.49  2.03  3.02  0.98  0.13  0.27  0.83  MCLA  71  12  3.36  2.86  3.81  0.95  0.13  0.27  0.70  MONS  71  11  2.68  1.74  4.48  2.74  0.11  0.02  0.91  (mm)  114  residual chronology  Table A.4 (cont.): Ring-width and chronology statistics for 12 lodgepole pine populations planted at 16 test sites. Prov = provenance; ave = average; incr = increment; min = minimum; max = maximum. Data are ordered from coldest to warmest site normal temperature, with populations within sites ordered from coldest to warmest provenance.  Site  Sample  Ave  Min  Max  Annual  First-order  size  incr  incr  incr  incr  Sensitivity  autocorrelation EPS of of residual  (no. of  width  width width  width  of residual  Prov trees)  (mm)  (mm)  range  chronology chronology  (mm)  residual chronology  MCLA  44  12  3.25  2.43  3.93  1.50  0.11  0.31  0.76  MONS  44  11  2.54  1.51  3.72  2.21  0.08  -0.07  0.87  MCLA  63  11  3.17  2.33  4.01  1.68  0.11  0.23  0.84  MONS  63  12  2.19  1.29  2.70  1.41  0.10  0.08  0.92  MCLA  1  12  2.99  2.14  3.57  1.43  0.12  0.21  0.77  MONS  1  12  2.55  1.47  3.60  2.13  0.09  -0.16  0.93  OTSA  30  10  2.27  1.38  3.22  1.83  0.06  -0.06  0.87  OTSA  100  12  2.74  1.60  3.36  1.76  0.06  0.16  0.87  OTSA  18  12  2.42  1.96  2.81  0.85  0.06  0.43  0.87  OTSA  71  12  2.69  2.24  3.28  1.03  0.05  0.31  0.90  OTSA  44  12  2.54  1.92  3.25  1.33  0.05  0.23  0.91  OTSA  63  12  2.87  1.82  3.66  1.84  0.05  0.13  0.87  OTSA  1  12  3.18  2.37  3.79  1.42  0.05  -0.06  0.89  BOSK  30  10  1.94  1.33  2.52  1.20  0.10  0.34  0.73  BOSK  100  12  2.54  1.77  3.14  1.36  0.04  0.60  0.57  BOSK  71  12  2.21  1.39  2.95  1.57  0.13  0.55  0.60  BOSK  44  12  3.21  2.33  4.72  2.39  0.11  0.10  0.74  BOSK  63  12  2.59  1.69  3.29  1.59  0.13  0.21  0.87  BOSK  1  12  2.75  1.70  3.71  2.01  0.10  0.25  0.70  CHUW  30  12  2.18  1.31  3.12  1.80  0.12  0.18  0.85  CHUW  35  11  2.21  1.70  3.17  1.47  0.13  0.07  0.80  CHUW 100  12  2.71  2.14  3.45  1.32  0.13  0.14  0.85  CHUW  71  12  2.58  1.72  3.63  1.91  0.01  0.06  0.80  CHUW  44  12  2.43  1.25  4.23  2.98  0.13  0.23  0.66  115  Table A.4 (cont.): Ring-width and chronology statistics for 12 lodgepole pine populations planted at 16 test sites. Prov = provenance; ave = average; incr = increment; min = minimum; max = maximum. Data are ordered from coldest to warmest site normal temperature, with populations within sites ordered from coldest to warmest provenance.  Site  Sample  Ave  Min  Max  Annual  First-order  size  incr  incr  incr  incr  Sensitivity  autocorrelation EPS of of residual  (no. of  width  width width  width  of residual  Prov trees)  (mm)  (mm)  range  chronology chronology  (mm)  residual chronology  CHUW  63  12  2.63  1.07  3.89  2.82  0.21  0.30  0.62  CHUW  1  12  2.91  2.22  3.61  1.39  0.13  0.10  0.85  SUSK  30  10  2.54  2.04  3.18  1.15  0.08  -0.68  0.75  SUSK  163  6  2.47  1.98  3.16  1.18  0.11  0.49  -0.03  SUSK  26  12  2.67  2.16  3.09  0.94  0.06  0.53  0.82  SUSK  100  12  2.63  1.89  3.46  1.57  0.08  0.16  0.83  SUSK  142  12  2.84  2.14  3.83  1.69  0.07  0.02  0.64  SUSK  71  12  2.87  1.91  3.97  2.06  0.08  0.16  0.83  SUSK  44  12  3.13  2.27  3.89  1.62  0.06  0.28  0.67  SUSK  63  11  2.72  1.98  3.86  1.88  0.07  -0.03  0.53  SUSK  1  10  2.67  1.80  3.88  2.08  0.07  0.04  0.64  LASI  30  12  2.28  1.92  2.64  0.71  0.12  0.32  0.83  MI70  30  10  1.93  1.22  2.54  1.32  0.06  0.28  0.75  LASI  35  11  2.02  1.54  2.59  1.04  0.18  0.62  0.91  LASI  100  11  2.70  1.87  3.52  1.65  0.16  0.13  0.91  MI70  100  11  1.97  1.20  2.69  1.50  0.08  -0.05  0.87  MI70  18  6  2.25  1.70  2.77  1.07  0.12  0.40  0.67  LASI  71  12  2.63  2.11  3.33  1.22  0.17  0.39  0.92  MI70  71  12  1.76  0.99  2.54  1.55  0.09  0.27  0.83  LASI  44  12  2.97  1.77  4.63  2.86  0.13  -0.05  0.88  MI70  44  11  2.21  1.59  2.95  1.35  0.09  -0.14  0.89  LASI  1  11  3.32  2.65  4.37  1.72  0.14  -0.11  0.85  MI70  1  12  2.40  1.28  3.50  2.22  0.11  0.46  0.86  LASI  72  12  3.13  2.19  3.92  1.72  0.14  -0.08  0.79  116  Table A.4 (cont.): Ring-width and chronology statistics for 12 lodgepole pine populations planted at 16 test sites. Prov = provenance; ave = average; incr = increment; min = minimum; max = maximum. Data are ordered from coldest to warmest site normal temperature, with populations within sites ordered from coldest to warmest provenance.  Site  Sample  Ave  Min  Max  Annual  First-order  size  incr  incr  incr  incr  Sensitivity  autocorrelation EPS of of residual  (no. of  width  width width  width  of residual  Prov trees)  (mm)  (mm)  range  chronology chronology  (mm)  residual chronology  MI70  72  13  2.46  1.18  3.60  2.43  0.08  0.31  0.75  CUIS  30  11  1.77  1.42  2.46  1.04  0.12  0.50  0.42  CUIS  100  12  2.01  1.61  2.45  0.84  0.15  0.41  0.90  CUIS  18  8  1.89  1.58  2.38  0.80  0.18  0.38  0.71  CUIS  71  12  2.08  1.16  2.55  1.39  0.18  0.17  0.84  CUIS  44  12  2.68  1.86  4.35  2.50  0.15  0.01  0.85  CUIS  63  9  2.67  2.04  3.79  1.75  0.15  0.02  0.83  CUIS  1  11  2.52  1.83  3.20  1.37  0.16  -0.18  0.69  WIGW 100  12  2.56  1.86  3.12  1.25  0.07  0.46  0.83  WIGW  18  9  2.31  1.37  3.05  1.68  0.05  0.56  0.63  WIGW  71  12  2.03  1.22  3.30  2.08  0.08  0.68  0.90  WIGW  44  12  2.80  1.85  3.57  1.72  0.05  0.45  0.80  WIGW  63  11  3.21  2.18  4.84  2.66  0.05  0.27  0.80  WIGW  1  9  2.83  1.65  4.22  2.58  0.05  0.39  0.58  WIGW  72  12  3.13  1.99  4.16  2.17  0.06  0.20  0.82  Average among SPs  2.39  1.65  3.24  1.59  0.12  0.18  0.79  Range among SPs  2.28  2.37  3.13  2.27  0.34  1.36  0.97  117  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0071779/manifest

Comment

Related Items