{"http:\/\/dx.doi.org\/10.14288\/1.0435560":{"http:\/\/vivoweb.org\/ontology\/core#departmentOrSchool":[{"value":"Forestry, Faculty of","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/dataProvider":[{"value":"DSpace","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#degreeCampus":[{"value":"UBCV","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/creator":[{"value":"Degner, Jonathan","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/issued":[{"value":"2023-08-23T21:55:06Z","type":"literal","lang":"en"},{"value":"2023","type":"literal","lang":"en"}],"http:\/\/vivoweb.org\/ontology\/core#relatedDegree":[{"value":"Doctor of Philosophy - PhD","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#degreeGrantor":[{"value":"University of British Columbia","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/description":[{"value":"The interior spruce of western Canada comprises complex hybrids of white spruce (Picea glauca) and Engelmann spruce (Picea engelmannii). These trees are a keystone of their ecosystems and are economically important to forestry in western Canada. As rapid climate change is a particular threat to long-lived forest trees, understanding climate adaptation in this species complex is critical to anticipate and mitigate impacts to our forests.\r\nUsing an existing dataset of genotyped seedlings representing 254 provenances throughout British Columbia and Alberta, I assessed the role of hybridization in climate adaptation within this species complex. Using a novel climatic similarity index, I determined that the hybrid zone is a climatic mosaic of its parent species\u2019 ranges, with a precipitation regime more similar to white spruce and a winter temperature regime more similar to Engelmann spruce.\r\nI identified SNPs with an excess of alleles from one parent species across a broad range of hybrids. SNPs with a bias towards Engelmann spruce alleles tended to be strongly correlated with temperature variables, while SNPs with higher white spruce allele frequencies were more correlated with precipitation variables, suggesting that favorable parental alleles efficiently introgress into hybrids and allow them to adapt to novel climates.\r\nUsing the strong relationship between hybrid index and climate, I combined genomic and climate data with existing ecological plot data to model hybrid index throughout western Canada. These models were applied to paleoclimatic simulations which, combined with mitochondrial haplotypes, allowed me to infer the colonization history of spruce in western Canada following deglaciation. I also used these hybrid index models to estimate the impacts of contemporary and future climate change on the hybrid zone, predicting large shifts in species ranges and large genomic shifts in areas where spruce persists.\r\nTogether, these results demonstrate that hybridization has greatly contributed to local climate adaptation in interior spruce, and that climate change poses an existential threat to forests in the finely tuned adaptive landscape of this hybrid zone. This knowledge of local adaptation can inform our reforestation practices, as human intervention will be necessary to align genotypes to climates over the coming decades.","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/aggregatedCHO":[{"value":"https:\/\/circle.library.ubc.ca\/rest\/handle\/2429\/85601?expand=metadata","type":"literal","lang":"en"}],"http:\/\/www.w3.org\/2009\/08\/skos-reference\/skos.html#note":[{"value":"GENOMICS OF ADAPTATION IN INTERIOR SPRUCE TO PAST, PRESENT, AND FUTURE CLIMATES OF WESTERN CANADA by  Jonathan Degner  B.Sc., The University of British Columbia, 2014  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2023  \u00a9 Jonathan Degner, 2023  ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled:  Genomics of adaptation in interior spruce to past, present, and future climates of western Canada  submitted by Jonathan Degner in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Forestry  Examining Committee: Sally Aitken, Professor, Forest & Conservation Sciences, UBC Co-supervisor Loren Rieseberg, University Killam Professor, Botany, UBC Co-supervisor  Sam Yeaman, Associate Professor, Biological Sciences, University of Calgary Supervisory Committee Member Tongli Wang, Associate Professor, Forest & Conservation Sciences, UBC University Examiner Darren Irwin, Professor, Zoology, UBC University Examiner   Additional Supervisory Committee Members: Robert Guy, Professor, Forest & Conservation Sciences, UBC Supervisory Committee Member  iii  Abstract  The interior spruce of western Canada comprises complex hybrids of white spruce (Picea glauca) and Engelmann spruce (Picea engelmannii). These trees are a keystone of their ecosystems and are economically important to forestry in western Canada. As rapid climate change is a particular threat to long-lived forest trees, understanding climate adaptation in this species complex is critical to anticipate and mitigate impacts to our forests. Using an existing dataset of genotyped seedlings representing 254 provenances throughout British Columbia and Alberta, I assessed the role of hybridization in climate adaptation within this species complex. Using a novel climatic similarity index, I determined that the hybrid zone is a climatic mosaic of its parent species\u2019 ranges, with a precipitation regime more similar to white spruce and a winter temperature regime more similar to Engelmann spruce. I identified SNPs with an excess of alleles from one parent species across a broad range of hybrids. SNPs with a bias towards Engelmann spruce alleles tended to be strongly correlated with temperature variables, while SNPs with higher white spruce allele frequencies were more correlated with precipitation variables, suggesting that favorable parental alleles efficiently introgress into hybrids and allow them to adapt to novel climates. Using the strong relationship between hybrid index and climate, I combined genomic and climate data with existing ecological plot data to model hybrid index throughout western Canada. These models were applied to paleoclimatic simulations which, combined with mitochondrial haplotypes, allowed me to infer the colonization history of spruce in western Canada following deglaciation. I also used these hybrid index models to estimate the impacts of contemporary and iv  future climate change on the hybrid zone, predicting large shifts in species ranges and large genomic shifts in areas where spruce persists. Together, these results demonstrate that hybridization has greatly contributed to local climate adaptation in interior spruce, and that climate change poses an existential threat to forests in the finely tuned adaptive landscape of this hybrid zone. This knowledge of local adaptation can inform our reforestation practices, as human intervention will be necessary to align genotypes to climates over the coming decades.    v  Lay Summary  We used genomic data and climate models to predict how interior hybrid spruce, an important forest tree, adapted to the climates of western Canada. We found that the hybrids, the offspring between white and Engelmann spruce, are genetically more like white spruce in climates preferred by white spruce and more like Engelmann spruce in climates preferred by Engelmann spruce. This pattern suggests that hybridization may aid evolutionary adaptation to climate. We also modeled how different kinds of hybrids exist in different climates within western Canada, and how that may have influenced the ancient migration of these trees after glaciers receded. Applying these same methods to future climates, we predict large shifts in where trees are now and where they will grow best over the next hundred years, necessitating human assisted migration for the long-term prosperity of these trees. vi  Preface  The genomic data used throughout this dissertation was generated as part of the AdapTree large-scale applied genomics project, led by Drs. Sally Aitken and Andreas Hamann, with additional data for allopatric Sitka spruce provided by Joane Elleouet. The sequence capture protocol to generate the data used in Chapter 3 is described by Suren et al. (2016). The SNP array used to produce the genotypic data for Chapter 4 was designed by Drs. Sam Yeaman, Kay Hodgins, and Katie Lotterhos, with assistance and input from others. Seedlings were grown by Dr. Pia Smets and Christine Chourmouzis. Tissue collection was performed by me, with assistance from others. DNA extraction and sequence capture was primarily performed by Kristin Nurkowski, with assistance from me. I was responsible for DNA quality control and normalization. The research ideas explored in Chapters 3 and 4 were developed  by me. I performed all bioinformatics, with substantial assistance from Drs. Sam Yeaman and Kay Hodgins. I performed all analyses, generated all figures, and wrote initial drafts of all chapters in this dissertation. My supervisors and committee, Sally Aitken, Loren Rieseberg, Sam Yeaman, and Rob Guy, provided feedback, guidance and editing on the text for production of the final drafts. A modified version of Chapter 2 has been published as a chapter within \u201cThe Spruce Genome\u201d, which is part of the Compendium of Plant Genomes series.   Degner, J.C. 2020. Local Adaptation in the Interior Spruce Hybrid Complex. In The Spruce Genome. Edited by I. Porth and A. De La Torre. Springer, Dordrecht. pp. 155-176.    The work reviewed in Chapter 2 arose in part from other studies I contributed to over the course of my program. These include: vii    Conte, G.L., Hodgins, K.A., Yeaman, S., Degner, J.C., Aitken, S.N., Rieseberg, L.H., and Whitlock, M.C. 2017. Bioinformatically predicted deleterious mutations reveal complementation in the interior spruce hybrid complex. BMC Genomics 18(1): 1\u201312.   In this paper, I identified a subset of samples suitable for deleterious allele analysis, estimated hybrid index for sampled individuals, and contextualized the results within the ecology of the species system. I also developed figures, provided laboratory, greenhouse, and bioinformatic support, and provided methodological input. Gina Conte conducted analyses and led writing the manuscript. Kay Hodgins conducted analyses. Sam Yeaman and Kay Hodgins provided bioinformatic support. Gina Conte, Mike Whitlock, Loren Rieseberg and Sally Aitken conceived the study. All authors contributed to writing and editing the manuscript. This paper is discussed in Section 2.3.6.1.   Lotterhos, K.E., Yeaman, S., Degner, J., Aitken, S., and Hodgins, K.A. 2018. Modularity of genes involved in local adaptation to climate despite physical linkage. Genome Biol. 19(1): 1\u201324.   My contributions to this paper were to develop geographic and visual analyses of allele frequencies for specific loci of interest and to provide discussion of loci that had been previously implicated in convergent selection with interior spruce. I also developed figures and provided laboratory support. Katie Lotterhos conceived the study, conducted analyses, and led writing the manuscript. Kay Hodgins and Sam Yeaman provided bioinformatic support and conducted analyses. Sally Aitken led the project to generate data. All authors contributed to writing and editing the manuscript.    viii  Lu, M., Hodgins, K.A., Degner, J.C., and Yeaman, S. 2019. Purifying selection does not drive signatures of convergent local adaptation of lodgepole pine and interior spruce. BMC Evol. Biol. 19(1): 1\u201310.   In this project, I performed population structure and admixture analyses in interior spruce samples to correct for the effects of hybridization in downstream analyses, and to identify and remove putative lodgepole and jack pine hybrids from the dataset. I also provided background on the interior spruce species complex, and discussion on the putative causes of environmental clines observed in the data. I also developed figures, provided laboratory and bioinformatic support, and provided methodological input. Mengmeng Lu analyzed the data and led writing the manuscript. Sam Yeaman and Kay Hodgins conceived and designed the study and coordinated the research. All authors contributed to writing and editing the manuscript.    Yeaman, S., Hodgins, K.A., Lotterhos, K.E., Suren, H., Nadeau, S., Degner, J.C., Nurkowski, K.A., Smets, P., Wang, T., Gray, L.K., Liepe, K.J., Hamann, A., Holliday, J.A., Whitlock, M.C., Rieseberg, L.H., and Aitken, S.N. 2016. Convergent local adaptation to climate in distantly related conifers. Science 353(6306): 23\u201326.   In this paper, I performed population structure analyses to adjust for the effects of hybridization on the strength of associations identified in interior spruce samples, and identified a major sample labelling error that had a large impact on the results. I also developed figures, provided laboratory support, and provided methodological input. Sam Yeaman and Kay Hodgins conceived the study, analyzed data and led writing the manuscript. Katie Lotterhos, Simon Nadeau, Tongli Wang, Laura Gray, and Katharina Liepe conducted analyses. Kristin Nurkowski coordinated laboratory work. Pia Smets coordinated biological work and data collection. Haktan Suren provided bioinformatic support. Andreas Hamann, Jason Holliday, Mike Whitlock, Loren Rieseberg, and Sally Aitken led the project to generate data. All authors contributed to writing and editing the manuscript. This paper is discussed throughout Section 2.3.  ix  Table of Contents  Abstract ......................................................................................................................................... iii Lay Summary .................................................................................................................................v Preface ........................................................................................................................................... vi Table of Contents ......................................................................................................................... ix List of Tables .............................................................................................................................. xiv List of Figures ...............................................................................................................................xv Acknowledgements .................................................................................................................. xviii Dedication .....................................................................................................................................xx Chapter 1: Introduction ................................................................................................................1 1.1 Local adaptation to climate ............................................................................................. 1 1.2 Interspecific hybridization .............................................................................................. 4 1.3 The interior spruce species complex ............................................................................... 7 Chapter 2: Review - Local adaptation in the interior spruce hybrid complex. .....................10 2.1 Introduction ................................................................................................................... 10 2.2 Phenotypic evidence for local adaptation ..................................................................... 12 2.2.1 Methods for detection ............................................................................................... 12 2.2.2 Phenotypic evidence for local adaptation in allopatric white and Engelmann spruce ....................................................................................................................................14 2.2.3 Phenotypic evidence for local adaptation interior spruce ......................................... 15 2.3 Genomic evidence for local adaptation ......................................................................... 17 2.3.1 Methods for detection ............................................................................................... 17 x  2.3.2 Genomic divergence approaches .............................................................................. 19 2.3.3 Genotype-environment clines ................................................................................... 20 2.3.4 Genotype-phenotype clines ....................................................................................... 24 2.3.5 Genomic architecture of local adaptation ................................................................. 26 2.3.6 Hybrid fitness ............................................................................................................ 28 2.3.6.1 Complementation .............................................................................................. 28 2.3.7 Plasticity .................................................................................................................... 29 2.4 Conclusions ................................................................................................................... 30 Chapter 3: Research - Genome-environment associations and genomic clines in divergent SNPs suggest parental species\u2019 adaptations combine to generate fine-scale local adaptation to climate within a spruce hybrid zone. .....................................................................................33 3.1 Introduction ................................................................................................................... 33 3.1.1 Genomic clines.......................................................................................................... 34 3.1.2 Genome-environment analysis.................................................................................. 35 3.1.3 Sequence capture ...................................................................................................... 36 3.1.4 Objectives ................................................................................................................. 37 3.2 Methods......................................................................................................................... 38 3.2.1 Sampling ................................................................................................................... 38 3.2.2 Genotypic data .......................................................................................................... 39 3.2.3 Bioinformatics........................................................................................................... 40 3.2.4 Genomic hybrid index............................................................................................... 42 3.2.5 Genomic cline analysis ............................................................................................. 43 3.2.6 Genotype-environment analysis................................................................................ 45 xi  3.2.6.1 Climatic variables analyzed .............................................................................. 45 3.2.6.2 Climatic similarity index................................................................................... 47 3.2.6.3 Bayenv2 analysis ............................................................................................... 50 3.2.7 Signatures of local adaptation ................................................................................... 51 3.3 Results ........................................................................................................................... 53 3.3.1 Genomic hybrid index............................................................................................... 53 3.3.2 Climate similarity index ............................................................................................ 54 3.3.3 Genomic cline analysis ............................................................................................. 57 3.3.4 Genome-environment association ............................................................................. 58 3.4 Discussion ..................................................................................................................... 64 3.5 Conclusions ................................................................................................................... 67 Chapter 4: Research - Climatic modeling of genomic ancestry in a spruce hybrid zone predicts past gene flow dynamics and future adaptive lag. .....................................................68 4.1 Introduction ................................................................................................................... 68 4.1.1 Data-driven climate prediction ................................................................................. 68 4.1.2 Model-driven climate prediction............................................................................... 69 4.1.3 Paleoclimatic modeling ............................................................................................. 70 4.1.4 Species climatic niche models .................................................................................. 71 4.1.5 Mitotypes and species migration .............................................................................. 73 4.1.6 Adaptation lag ........................................................................................................... 74 4.1.7 Objectives ................................................................................................................. 75 4.2 Methods......................................................................................................................... 76 4.2.1 Genotypic data .......................................................................................................... 76 xii  4.2.2 Genomic hybrid index............................................................................................... 77 4.2.3 Climate data .............................................................................................................. 78 4.2.3.1 Contemporary and future climatic data............................................................. 78 4.2.3.2 Paleoclimatic data ............................................................................................. 79 4.2.4 Climatic prediction of hybrid index .......................................................................... 80 4.2.5 Species climatic niche modelling.............................................................................. 82 4.2.6 Mitotype assignment ................................................................................................. 85 4.3 Results ........................................................................................................................... 85 4.3.1 Genomic hybrid index............................................................................................... 85 4.3.2 Species climatic niche modelling.............................................................................. 85 4.3.3 Climatic prediction of hybrid index .......................................................................... 91 4.3.4 Mitotype assignment ................................................................................................. 95 4.4 Discussion ..................................................................................................................... 98 4.4.1 Characterization of the current hybrid zone.............................................................. 98 4.4.2 Paleoclimatic predictions of the hybrid zone and post-glacial colonization of western Canada..................................................................................................................................102 4.4.2.1 Paleoclimatic models of species niches and hybrid index .............................. 102 4.4.2.2 Post-glacial colonization routes ...................................................................... 103 4.4.2.3 Comparison with paleobotanical records ........................................................ 105 4.4.3 Present and future adaptive lag ............................................................................... 108 4.4.3.1 Present adaptive lag (1961-1990 reference period vs. 1991-2020 period) ..... 108 4.4.3.2 Future climates and the fate of interior spruce in western Canada ................. 110 4.5 Conclusions ................................................................................................................. 111 xiii  Chapter 5: Conclusions .............................................................................................................113 5.1 Summary ..................................................................................................................... 113 5.2 Limitations .................................................................................................................. 115 5.3 Future research directions ........................................................................................... 117 References ...................................................................................................................................121 Appendix .....................................................................................................................................144  xiv  List of Tables  Table 3.1: Climatic variables used in subsequent analyses, as predicted using ClimateWNA and ClimateNA. Means, standard deviations (SD), and range are shown for the 254 spruce provenance locations used in this dissertation. ............................................................................. 47 Table A.1: SNP outlier enrichment in genotype-environment associations for SNPs identified as Engelmann- or white-skewed based on genomic cline center. ................................................... 146  xv  List of Figures Figure 2.1: Geographic distribution of 254 interior spruce provenances, showing provenance mean species ancestry ................................................................................................................... 11 Figure 3.1: Distribution of 254 interior spruce provenances in relation to the number of degree days below 0\u00b0 C. ........................................................................................................................... 49 Figure 3.2: Ordination of 10,000 random geographic locations within the climatic niches of white spruce (blue), Engelmann spruce (red), and hybrid spruce (black and green) along principal component axes separately calculated for temperature and precipitation variables ...... 56 Figure 3.3: Distribution of genomic cline centers for 14,009 SNPs analyzed using bgc. ............ 57 Figure 3.4: Enrichment of hybrid index-neutral (gray) and parent-species-skewed (black) SNPs in the 99th percentile of Bayenv2 climatic associations as measured by odds ratio against all other SNPs in the dataset. ............................................................................................................. 59 Figure 3.5: Bayenv2 outlier enrichment for SNPs identified as Engelmann-skewed (red) or white-skewed (blue) cline center outliers in bgc, compared to SNPs that covary linearly with hybrid index. ............................................................................................................................................. 60 Figure 3.6: Proportion of SNPs identified as genomic cline outliers biased towards Engelmann spruce in the top percentiles of Bayenv2 genotype-environment associations for 22 provenance climate variables. .......................................................................................................................... 61 Figure 3.7: Mean genotype-environment correlations of SNPs identified as genomic cline outliers for Engelmann or white spruce, and those with no bias towards one or the other across 19 climate variables. ..................................................................................................................... 63 Figure 3.8: Difference in correlation strength between SNPs identified as genomic cline outliers for Engelmann and white spruce across 19 climate variables in Bayenv2. .................................. 63 xvi  Figure 4.1: Geographic distribution of ecological sample plots used to generate species niche models. .......................................................................................................................................... 83 Figure 4.2: Prediction accuracy for species climatic niche models for white, hybrid, and Engelmann spruce. ........................................................................................................................ 87 Figure 4.3: Relative variable importance for climatic prediction of hybrid index and climatic niche models for white, Engelmann, and hybrid spruce. .............................................................. 87 Figure 4.4: Predicted climatic niches for interior spruce from 15kyp to present. ........................ 88 Figure 4.5: Species climatic niches calculated for the 1961-1990 reference period compared to the 1991-2020 reference period. ................................................................................................... 89 Figure 4.6: Predicted species climatic niches for white, Engelmann, and hybrid spruce under three different emissions scenarios and future time periods. ........................................................ 90 Figure 4.7: Relationship between P. engelmannii ancestry as estimated through genomic data and predicted by provenance climate in RandomForests. ................................................................... 91 Figure 4.8: Predicted hybrid index of suitable habitat for interior spruce from 15kyp to present.92 Figure 4.9: Change in predicted genomic ancestry between the 1961-1990 normal period and the 1991-2020 normal period for areas that are within the species climatic niches for both climatic periods. .......................................................................................................................................... 93 Figure 4.10: Predicted hybrid index of interior spruce under three different emissions scenarios and future time periods. ................................................................................................................ 94 Figure 4.11: Network maps depicting the relationship between mitotypes identified in the sequence capture and SNP array datasets. .................................................................................... 96 xvii  Figure 4.12: Geographic distribution of spruce provenances showing the proportion of mitotypes identified in each provenance within the sequence capture dataset, overlaid on the species climatic niche models for 12kybp. ................................................................................................ 97 Figure 4.13: Predicted abundance of spruce species in North America from 15kybp to present, based on palynological data. ....................................................................................................... 106 Figure A.1: Predicted hybrid index for all environments in the 1961-1990 climatic reference period, without constraints based on species climatic niche models. ......................................... 144 Figure A.2: Species climatic niche models for 15kyp-present, including Sitka spruce. ............ 145 Figure A.3: Distribution of climatic differences between the 1991-2020 and 1961-1990 climatic reference periods across the area of their shared climatic niches. .............................................. 146   xviii  Acknowledgements My sincere and heartfelt thanks go out first and foremost to those who gave me the time and space I needed to complete this dissertation. My co-supervisor Sally Aitken has been nothing but supportive throughout the entirety of my program and has helped immensely at every step. First, she was my inspiration to pursue forest genetics as a field of study when I took her course as an undergraduate. She gave me the opportunity to pursue that interest, first as an undergraduate, then as a lab assistant, and finally as a graduate student. She introduced me to the forest genetics programs in the BC Ministry of Forests, opening my eyes to a field I hadn\u2019t previously considered, but where I am now happily established in a career that I love. I genuinely cannot thank her enough for her support and mentorship throughout the \u2013 at this point quite a few \u2013 years I have known her. I would also like to thank those who I have worked closely under in the Ministry: my supervisor, Keith Thomas, and mentors, Alvin Yanchuk and Michael Stoehr. They all gave me the time and space to complete this dissertation while working full time taking over multiple research programs. Though my progress was slow, I got there in the end in no small part thanks to their support and guidance, constantly adding wax to my candle burning at both ends. Additionally, I\u2019d like to thank Lise van der Merwe, who took on a huge amount of administrative work in the research programs that I was unable to shoulder as I completed this dissertation. The final person who I\u2019d like to thank for giving me time and space is my wife, Paige. She never once complained about how long I\u2019d been making tuition payments, or balked at the many, many times I pushed back my completion date. She has only ever been loving and compassionate as I faced mental hurdles and I have been incredibly lucky to have her emotional support for over fourteen years as I worked my way through my studies. xix  There have been many other people who have helped me immensely through my doctoral study. My co-supervisor Loren Rieseberg has been an absolute wellspring of ideas and brilliant comments, and I thank him for welcoming me into his lab full of likewise brilliant students who never failed to give me great advice from beyond the world of forestry. I\u2019d like to thank Sam Yeaman for his advice as a committee member and for his bioinformatic support in getting me up and running with this massive genomic dataset. I\u2019d like to give special thanks to some lab mates who helped me over the years in analysis and writing, and for letting me participate in their spectacular fieldwork: Ian MacLachlan, Joane Elleouet, and Vincent Hanlon. I will forever treasure our trials, tribulations, and spectacular views from the forest. Additional thanks to Joane for allowing me the use of her Sitka spruce data for estimating Sitka ancestry. I\u2019d like to give special thanks to all of the team members on the AdapTree project for helping to develop this incredible dataset that I\u2019ve had the privilege to study. I\u2019d like to thank Pia Smets, Christine Chourmouzis, and Joanne Tuytel for their incredible work in growing so many thousands of seedlings from so far afield. Finally, I\u2019d like to thank Kristin Nurkowski for training me in laboratory methods and being patient with me as I muddled my way through wetlab work. The world is less without her. My work would not have been possible without the generous financial support of the sponsors of the AdapTree project (https:\/\/adaptree.forestry.ubc.ca\/sponsors\/), and the NSERC scholarship programs from which I have benefited over the years (USRA, CGS-M, and PGS-D), as well as UBC for providing me with additional scholarship funding through the 4YF program. It has been an intense honor to complete many years of study with little concern for my financial health. xx  Dedication  I dedicate this dissertation to my grandfather, Jim G. Roberts (25 Dec. 1935 \u2013 3 Jun. 2023). Our uncountable trips to Little North Fork and Cultus Lake sparked a love of the woods and the natural world that has been my beacon through every storm. I will forever remember you in the sweet smell of pine duff on the still August air. 1  Chapter 1: Introduction  1.1 Local adaptation to climate In Darwin\u2019s On the Origin of Species, he noted that pine seeds acquired from different elevations in the Himalayas differed in their resistances to cold when grown in England, concluding that populations within species can sometimes diverge over the course of generations to become acclimated to different temperatures (Darwin 1859). Though other prominent botanists also noted this phenomenon, it cast doubt on the prevailing idea that all individuals of a species were more-or-less capable of the same growth under all conditions in which that species is found. This observation was initially largely ignored. Less than two decades later, several botanists reciprocally transplanted seedlings of herbaceous species that were found at drastically differing elevations or geographies, noting that local seed tended to produce healthier plants (Naudin and Radlkofer 1876; Kerner von Marilaun 1895) These early observational experiments all hinted at an unexplored aspect of natural selection, and it would be several decades before their results were first effectively synthesized (Turesson 1922; Clausen et al. 1940). Through these early experiments involving reciprocally transplanting plants among multiple environments, we began to understand that many plant species are composed of natural populations that are not only genetically adapted to their habitats, but that are uniquely adapted to that habitat. This concept has been codified in the present use of the term \u2018local adaptation\u2019 (Kawecki and Ebert 2004). Current evolutionary theory suggests that local adaptation is, in most cases, produced by divergent selection acting across one or more selective gradients throughout a species\u2019 range (Linhart and Grant 1996; but see Blanquart et al. 2012). If selection pressures do not change 2  substantially over short temporal scales but differ spatially, populations should become increasingly well-adapted to the selective environment in which they exist, and that adaptation may come at the cost of reduced fitness in other environments (Felsenstein 1976). This is generally thought to occur through the selection of alleles that are favorable in one environment but neutral or detrimental in another (Colautti et al. 2012). As allele frequencies gradually shift in response to variable strengths of selection across these environmental gradients, clinal relationships in adaptive phenotypes and alleles are expected to form (Huxley 1939; Barton 1999), and are commonly cited as evidence of local adaptation (Rellstab et al. 2015).  As many species experience differing selective pressures across their distribution (Mckay and Latta 2002), local adaptation is often viewed as a default path for natural selection, one which is bound to occur if it is not constrained by other pressures. Theoretical and empirical studies have identified many factors that can prevent species from becoming locally adapted, including a lack of standing genetic variation, high gene flow, small population size, or high temporal heterogeneity in selection pressures (Brown and Pavlovic 1992; Kassen 2002; Willi et al. 2006; Leimu and Fischer 2008). Additionally, some theoretical models predict that even if a species is capable of becoming locally adapted, phenotypic plasticity or generalism may be favorable under some evolutionary scenarios (Sultan and Spencer 2002; Spichtig and Kawecki 2004). Despite these caveats, local adaptation is common in nature in general (Hereford 2009) and in plants (Leimu and Fischer 2008) and specifically in forest trees (Leites and Benito 2023). Local adaptation was first identified in plants, where it may be viewed as an obvious solution to some of the challenges of being a plant: a sessile growth form that must tolerate its environment, a general correlation in habitat between parents and offspring due to limited dispersal capabilities, flexibility in mating systems, and a propensity for intra- and interspecific 3  hybridization (Bradshaw 1972; Linhart and Grant 1996).  It is unsurprising, then, that local adaptation has been repeatedly identified in plants (Savolainen et al. 2007; Leimu and Fischer 2008), and less-commonly in animals (Hereford 2009; Sanford and Kelly 2011).  Many plants are locally adapted to climate, particularly widespread trees (Savolainen et al. 2007). As a consequence of their long lifespans, trees must endure their local climate for decades or centuries, including all of the extreme climatic events that may occur over that timescale. As injury and mortality from climatic events can often be mitigated through phenological timing (Richardson et al. 2013), local adaptation in phenology and seasonal phenotypes is common (Howe et al. 2003; Alberto et al. 2013). Given climate\u2019s importance to local adaptation and recent tools allowing many climatic variables to be quickly estimated for precise geographic locations (e.g., Wang et al. 2016; Fick and Hijmans 2017), local adaptation to climate has been studied in many temperate tree species (Alberto et al. 2013), and is the primary driver of local adaptation that will be discussed here.  Local adaptation plays a central role in our understanding of ecological speciation, and more broadly in our understanding of the role natural selection plays in a species\u2019 evolutionary history (Nosil 2010). It is also of critical concern to conservation biologists, as species that are locally adapted to climate may suffer substantially due to climate change without assistance (Aitken et al. 2008), and translocations of individuals for restoration and genetic rescue should be designed with knowledge of local adaptation. In particular, assisted migration of forest trees to mitigate climate change has received considerable attention in recent years, given the large impact of forests on global climate (Aitken and Whitlock 2013; Aitken and Bemmels 2016; S\u00e1enz-Romero et al. 2016). Having a firm understanding of local adaptation is critical to effectively achieving these aims.  4  Detecting selection at the genetic level generally involves inferring a single population\u2019s genotypic response to an experimentally-applied stressor (e.g., Guttman and Dykhuizen 1994) or measuring differences in allele frequencies across multiple populations that co-vary along an environmental gradient (e.g., Coop et al. 2010). While the first approach is ideal for direct study, it is generally intractable in non-model organisms. The second approach includes several confounding issues, including incidental collinearity between selective and non-selective gradients (Rellstab et al. 2015) and the relative difficulty of measuring non-directional selection (Excoffier et al. 2009b).  Local adaptation will generally act to genetically distinguish populations (Kawecki and Ebert 2004). In a locally-adapted population, gene flow will likely reduce the mean fitness of individuals, generating a \u201cmigration load\u201d (Garc\u00eda-Ramos and Kirkpatrick 1997), that selection may act upon (Lenormand 2002; Bolnick and Nosil 2007) This selective pressure against gene flow can lead to the accumulation of reproductive incompatibilities among populations, resulting in speciation (Nosil 2010). However, such speciation may occur even in the absence of reproductive incompatibilities if selection is strong enough (Feder et al. 2012).   1.2 Interspecific hybridization Hybridization, the mating between individuals of genetically-distinct populations (Barton and Hewitt 1985), may occur during many speciation processes, excepting instantaneous or completely allopatric speciation (Abbott et al. 2013). Hybridization (hereafter referring only to interspecific hybridization, i.e., reproduction between taxonomically defined species) is particularly common among plants. A survey of 3,202 genera across 282 families of vascular 5  plants revealed that 16% of genera across 40% of families contained currently hybridizing species (Whitney et al. 2010).  Hybridization is a process that has long fascinated taxonomists, as it stands in the face of the biological species concept. If species can successfully hybridize, and hybrids can backcross with parent species, then under the strict definitions of biological species they are populations that have not yet completed the process of speciation and should not be considered separate species. However, the frequent occurrence of hybridization and the observation that most hybridizing taxa maintain phenotypic divergence despite this potential gene flow led the developers of the biological species concept to assume that hybridization was typically an evolutionary nuisance with little impact (Dobzhansky 1940; Mayr 1942). Depending on the amount of divergence between hybridizing species, and the amount of reproductive isolation that has accumulated between them, hybridization may have positive or negative outcomes on the fitness of offspring.  Much of the early discussion surrounding hybridization considered it to be a generally unfavorable or inconsequential evolutionary process (Mayr 1942; Bigelow 1965). This early thought was based upon experimental crosses of various species and races, often in Drosophila, that produced viable but unfit offspring (e.g., Stern 1936; Dobzhansky 1937), and several documented animal hybrid zones e.g., Corvus (Mayr 1942) and Mus (Ursin 1952). As a result, stable hybrid zones were largely assumed to exist due to a balance between interspecific gene flow forming hybrids and selection acting to remove them i.e., tension zones (Key 1968). It is worth mentioning that some biologists of the time viewed hybridization as a process of great evolutionary potential and importance, despite the prevailing sentiments against (Anderson 1948, 1949; Anderson and Stebbins 1954). 6  Many stable hybrid zones have since been discovered and well-studied, shedding new insights on the potential dynamics of persistent hybrid zones. In the threespine stickleback (Gasterosteus aculeatus), a model natural system for studying ecological speciation (reviewed in McKinnon and Rundle 2002), multiple hybrid zones have been studied to understand the physiological and genomic bases for reproductive isolation between morphs (Hagen 1967; Jones et al. 2006). Sunflowers (Helianthus sp.) have become a valuable system for understanding hybrid speciation (Gross and Rieseberg 2005; Mallet 2007), with three ecologically-divergent species (Helianthus anomalus, H. deserticola, and H. paradoxus) arising from ancient hybridization between two divergent clades of annual sunflowers  (Rieseberg 1991; Owens et al. 2023). Model hybrid zones are often small tension zones, but other models of hybrid zone formation and maintenance have been proposed and supported by natural systems. While the tension zone model assumes that hybrids are less fit than parental genotypes in all environments (Barton and Hewitt 1985), the bounded hybrid superiority model allows for hybrids that are more fit than parental genotypes in an intermediate or novel environment (Moore 1977). Many plant hybrid zones appear to follow this model e.g., Artemesia (Wang et al. 1997); Carex (Choler et al. 2004); and Picea (De La Torre et al. 2014c). The mosaic or bimodal hybrid zone model supposes that, for species with strong ecological divergence that come into contact in heterogeneous habitats with sharp boundaries e.g., soil type (Rand and Harrison 1989) or light availability (Cruzan and Arnold 1993), hybrids are rare and will be strongly selected against, with most populations consisting of parental genotypes or advanced backcrosses in their respective habitats (Harrison 1990; Jiggins and Mallet 2000). This model of hybrid zone was first described in field crickets (Gryllus sp.) (Harrison 1986), which has since become a common system for studying 7  reproductive isolation (Larson et al. 2013) and divergent selection (Ross and Harrison 2002) in these zones. Many mosaic hybrid zones show evidence of prezygotic (e.g., Chorthippus; Bailey et al. 2004) or postzygotic (e.g., Populus; Lindtke et al. 2014) reproductive isolation.  Interspecific gene flow does not affect all regions of the genome equally (Hunt and Selander 1973; Payseur 2010). While all first generation (F1) hybrids contain complete chromosomes from their parent taxa, successive hybridization and back-crossing allows recombination, and selection for and against loci from each parental taxon. Genomic regions containing broadly adaptive alleles are likely to quickly sweep through hybrid populations, while those containing loci responsible for reproductive isolation will be filtered through selection and will not progress far through hybrid zones, reaching a balance between gene flow and selection (Barton and Hewitt 1985). As a result of this process, hybrid genomes are likely to be mosaics of parental ancestry, with measurable differences in the ability of parental alleles to introgress into backcross hybrids.  1.3 The interior spruce species complex From the vast expanse of Canada and Alaska\u2019s boreal forest to the ragged treeline of the US and Canadian Rocky Mountains, spruces thrive where few other trees can survive. White spruce (Picea glauca (Moench) Voss) and Engelmann spruce (Picea engelmannii Parry ex Engelm.) are North American trees that occupy distinct ecological niches, where they are typically dominant or co-dominant canopy species. White spruce is a boreal species, distributed widely across North America, covering 26\u00b0 of latitude and 111\u00b0 of longitude, generally at lower elevations (Little 1971). Engelmann spruce is a subalpine species, growing from intermediate elevations to treeline throughout many of the mountainous regions of Western North America, a 8  broad and highly fragmented area spanning 26\u00b0 latitude and 23\u00b0 longitude (Little 1971; Hope et al. 1991). The climate of these regions is generally cool and humid, with short growing seasons and high snowfall, and populations are often geographically isolated on mountaintops (Alexander and Shepperd 1984). Relative to white spruce, the habitat of Engelmann spruce tends to be warmer and wetter (Liepe et al. 2016). Molecular phylogenetic analyses have consistently recognized white and Engelmann spruce as sister taxa (Sigurgeirsson and Szmidt 1993; Ran et al. 2006, 2015; Bouill\u00e9 et al. 2011; Lockwood et al. 2013; Feng et al. 2019; Shao et al. 2019), with the most recent estimate of divergence around 7.3 Mya based on a transcriptomic comparison of 1141 orthogroups (Shao et al. 2019). Much of the present range of these species was covered by ice during the last glaciation (Shafer et al. 2010), and both species were forced south of their present distribution, where they may have come into secondary contact during the last glacial maximum (Ledig et al. 2006; De La Torre et al. 2014b). Since the last glacial maximum, both species recolonized large areas made available by glacial retreat and have formed several stable hybrid zones, covering an extensive area in the central plateaus of British Columbia, as well as the Rocky Mountains, extending eastward to Alberta and southward to Wyoming (Roche 1969; Daubenmire 1974; Haselhorst and Buerkle 2013). Hybrids between white and Engelmann spruce, referred to as \u201cinterior spruce\u201d in British Columbia, have been identified through morphological markers such as cone and branch morphology (e.g., Garman 1957; Daubenmire 1974) and confirmed through genetic markers (Rajora and Dancik 2000; Haselhorst and Buerkle 2013; De La Torre et al. 2014c; Conte et al. 2017; MacLachlan et al. 2018). The spatial and climatic extent of the zone is extensive, spanning at least 13\u00b0 of latitude and 20\u00b0 of longitude (Haselhorst and Buerkle 2013; Conte et al. 2017), 9  with mean annual temperatures ranging from -3 to +6\u00b0C, and mean annual precipitation regimes from 350 to 1500 mm in Western Canada (Yeaman et al. 2016). The hybrid zone spans all degrees of admixture, with most of the central zone containing hybrids with more white spruce ancestry, and hybrids in southern montane regions with more Engelmann ancestry. Both white spruce and interior spruce further hybridize with Sitka spruce (Picea sitchensis (Bong.) Carr) in coastal and submaritime river drainages in British Columbia and Alaska (Little 1953; Daubenmire 1968), producing three-way hybrids in some western parts of central British Columbia (Sutton et al. 1994; Hamilton et al. 2015). Local adaptation has been less studied in this system and will not be covered here (O\u2019Neill et al. 2002; Hamilton et al. 2013; Hamilton and Aitken 2013).     10  Chapter 2: Review - Local adaptation in the interior spruce hybrid complex.  2.1 Introduction Given the relatively recent range expansions of white and Engelmann spruce into western Canada, and subsequent secondary contact, it is possible that populations still bear strong signatures of these events and may not have inhabited their present geographic ranges long enough to locally adapt to those environments. However, analyses of interspecific heterozygosity throughout the hybrid zone suggest that these zones are at least 5000 years old, with populations mostly composed of multigenerational hybrids and backcrosses (De La Torre et al. 2015). Paleoclimatic modelling suggests that these species may have been in secondary contact even during the last glacial maximum (De La Torre et al. 2014b). Recent research, discussed later in this chapter and elsewhere in this dissertation, has shown that the geographic extent of the hybrid zone is immense, and that the genomic ancestry of hybrids is complex and varies over short geographic distances (Figure 2.1), indicating that gene flow and selection have been actively shaping the genomic landscape of the species complex since these areas were recolonized. Although some traces of demographic processes and recent introgression may still exist and obscure genomic patterns of local adaptation, there is still potential for local adaptation in this system.  11   Local adaptation has been of interest to foresters for centuries to ensure maximum tree health and timber value (Langlet 1971), and has received more attention over recent decades due to a reliance on replanting seed that is sourced from relatively broad geographic areas and the need to develop geographic seed zones or seed transfer guidelines (e.g., Ying and Yanchuk Figure 2.1: Geographic distribution of 254 interior spruce provenances, showing provenance mean species ancestry among Picea engelmannii (Engelmann spruce; dark red), P. glauca (white spruce; light blue), and P. sitchensis (Sitka spruce; green). Ancestry was estimated with ADMIXTURE v1.3 using 291,094 SNPs generated from exome capture data.  12  2006). As interior spruce is an important timber tree in western Canada, and white spruce is economically important across much of eastern Canada, considerable study has been given to local adaptation in this species complex, producing valuable insight into climatic adaptation. Given pressing concerns about local adaptation in these species in the face of climate change (O\u2019Neill et al. 2014), genomic research has been pursued in this system to rapidly gain insights into the underlying adaptive drivers present in this species complex, and to inform land managers and conservationists about the potential need for assistance to maintain healthy and productive forests.  2.2 Phenotypic evidence for local adaptation 2.2.1 Methods for detection In a strict sense (sensu Kawecki and Ebert 2004), local adaptation is a form of genotype-by-environment interaction that can only be directly inferred from the reciprocal transplantation of genotypes across two or more environments, comparing fitness or some of its components. As Darwinian fitness is notoriously difficult to estimate for any but the simplest systems (Benton and Grant 2000), phenotypic proxies for fitness are often used, especially for long-lived species like trees. In plants these include parameters related to survival, growth, reproductive output, or stress from biotic and abiotic factors (e.g., Clausen et al. 1940; Bradshaw 1960; Waser and Price 1985; Gr\u00f8ndahl and Ehlers 2008; Kim and Donohue 2013). Several statistical methods for characterizing local adaptation in quantitative traits have been developed and compared across species in various biological systems (Hoeksema and Forde 2008; Leimu and Fischer 2008; Fraser et al. 2011). These fall into two general categories (Kawecki and Ebert 2004; Blanquart et al. 2013). \u2018Home vs. away\u2019 analyses measure local 13  adaptation by comparing the fitness of genotypes grown in their local environment to those same genotypes grown in one or more foreign environments. \u2018Local vs. foreign\u2019 analyses instead compare the fitness of genotypes grown in their local environment to foreign genotypes grown in the same environment. Depending on the fitness measure used and the species system under study, the choice of analysis may result in differing interpretations of identical data (Blanquart et al. 2013), though either is typically acceptable to confirm the presence of local adaptation. In the absence of a reciprocal transplant experiment, there are other methods to infer local adaptation, though without the strong evidence provided by reciprocal transplant. As populations of locally-adapted species are expected to show divergence in fitness or its components under any environment, a single common garden, where individuals from multiple sampled locations (often referred to in the literature as provenances) are raised and evaluated in a common environment, can still yield evidence of local adaptation (M\u00e1ty\u00e1s 1996; Wadgymar et al. 2022). Locally-adapted species should show strong clines in fitness corresponding to selective gradients across the tested populations (Huxley 1939), and there is expected to be a relationship between an individual\u2019s fitness and the distance of transplanting i.e., populations closer to the common garden (either geographically or in relationship to selective gradients) should perform better than more distant populations (Williams 1966). The presence and strength of these clinal relationships has been used to infer local adaptation across many taxa, particularly plants given the ease of coaxing them into a common garden relative to animals (many plant taxa reviewed in Linhart and Grant 1996; Alberto et al. 2013).   14  2.2.2 Phenotypic evidence for local adaptation in allopatric white and Engelmann spruce As white spruce is an ecologically- and economically-important tree species in eastern Canada, local adaptation has been assessed extensively in its allopatric range. The results of these studies are mixed, generally suggesting weak but extant local adaptation. In eastern Canada, height and diameter of mature trees is best for provenances with climates similar to the test sites, although climatically-distant provenances still performed relatively well (Andalo et al. 2005). Rweyongeza et al. (2011) found similar patterns for a series of provenance trials in western Canada, although with generally steeper clines in performance. Clinal relationships have also been observed for traits related to photosynthetic capacity, water use efficiency, and cold hardiness in relation to provenance climates (Benomar et al. 2015; Sebastian-Azcona et al. 2018).  The most direct evidence of local adaptation in white spruce comes from a large series of provenance trials across eastern Canada, analyzed explicitly for local adaptation by Lu et al. (2014, 2016). They found in general that local provenances did not outperform non-local ones i.e., in a local-foreign comparison. However, when provenances were combined into southern, central, and northern groups, divergent clines emerged, showing that provenances tended to grow best nearer their provenance mean annual temperature, indicative of local adaptation in the home-away contrast. The results of Lu et al. echo those of Li et al. (1997) and Lesser and Parker (2004), finding that northern provenances tend to perform poorly across all test environments while southern provenances excel, which the authors attributed to reduced genetic variation in northern populations. There is substantially less data available for local adaptation in allopatric Engelmann spruce, but it generally suggests weak or no local adaptation. In a trial established in Colorado, 15  near the center of the species range, provenances from the northern and westernmost portions of the range outperformed local provenances for growth after ten years (Shepperd et al. 1981). However, a series of provenance trials farther north found that southwestern provenances grew exceptionally well at four years, and that clines in growth were present for elevational and climatic gradients relating to frost-free period, suggesting adaptation along these axes (Rehfeldt 1994, 2004). This early age of measurement leaves open the possibility that well-performing provenances may be susceptible to infrequent episodes of cold injury that would cause mortality or hinder long-term growth. A recent study also found no local adaptation for recruitment success in a reciprocal transplant across an elevational gradient, which could be limiting the development of local adaptation in these highly heterogeneous environments (Kueppers et al. 2017).  2.2.3 Phenotypic evidence for local adaptation interior spruce There is strong evidence for local adaptation in interior spruce via multi-environment experiments. Liepe et al. (Liepe et al. 2016) partitioned the variance of growth, phenology, and cold injury traits in seedlings from across Western Canada grown in growth chambers under various climatic regimes. They identified 11 adaptive syndromes \u2013groups of provenances with similar patterns across the measured traits\u2013 that correspond to provenance climates and ecosystems. They found the strongest clines for cold injury, suggesting strong local adaptation for this trait in particular. MacLachlan et al. (MacLachlan et al. 2018) likewise identified strong clinal variation in cold hardiness in a single outdoor seedling common garden trial, also noting strong clines in this and other phenotypic variables associated with provenance mean annual temperature. 16  Ukrainetz et al. (2011) and O\u2019Neill et al. (2014) assessed local adaptation via seed transfer distance, a metric reflecting the climatic distance at which provenances become maladapted, across 4 and 15 provenance trials, respectively. They found that transfer distances depended heavily on the climate of the test site, where sites with moderate climates showed little provenance differentiation, and sites with more extreme climates showed strong local adaptation for height and survival (O\u2019Neill et al. 2014) and for a composite fitness measure (Ukrainetz et al. 2011). This suggests that local adaptation in this species complex may be strong in the allopatric ranges of white and Engelmann spruce, but weaker in the hybrid zones. Interestingly, Xie (2003), found little evidence of local growth superiority for ten-year-old trees across a large series of provenance trials in interior spruce, despite significant genotype-by-environment interactions.  In summary, these experiments suggest relatively weak local adaptation in allopatric white and Engelmann spruce, but stronger local adaptation among introgressed interior spruce populations. Several explanations for this trend have been put forth, including selective pressures being higher in the highly heterogenous montane environments faced by interior spruce, differential selection acting upon interspecific hybrids, or an increase in standing genetic variation generated by the multi-species complex (Ledig et al. 2006; De La Torre et al. 2015; Liepe et al. 2016). In general, local adaptation appears stronger for growth, autumn phenology-related traits, and cold hardiness than for survival or spring phenology, a pattern observed across many other tree species(Alberto et al. 2013).  17  2.3 Genomic evidence for local adaptation 2.3.1 Methods for detection While reciprocal transplant experiments constitute the \u2018gold standard\u2019 for measuring local adaptation, this may be difficult or impossible to implement for many species due to their life histories or ecological constraints (Savolainen et al. 2013). Also, the age at which local adaptation is measured can have a profound impact on results, with studies measuring the same individuals frequently finding weak or inverse relationships over time (e.g., Conkle 1973; Hannerz et al. 1999; Weng et al. 2007; Stackpole et al. 2010; De La Torre et al. 2014c). Fortunately, the processes that generate local adaptation leave signatures throughout the genome, providing an alternate means to detection that should be comparable regardless of age (Fournier-Level et al. 2011).  Although genomic data alone cannot conclusively confirm local adaptation, an understanding of the genetic underpinnings of local adaptation can broaden our understanding adaptation beyond what can be derived from reciprocal transplant experiments (Sork 2018). For example, genetic information may be required to fully describe local adaptation when the selective drivers are uncertain, when identifying parallel evolution of local adaptation between taxa, or when determining the genomic architecture of local adaptation (Savolainen et al. 2013; Rausher and Delph 2015; Hoban et al. 2016). Additionally, many selective pressures are expected to act over long temporal windows, especially for long-lived species including trees (M\u00e1ty\u00e1s 1996), and rare, extreme events may be important. Local adaptation to these pressures may be undetectable in the relatively short time span over which transplant experiments are conducted but should still be detectable in genomic data. Additionally, plants in common gardens can only be phenotyped for a small number of traits over a short timescale (relative to 18  the lifespan of many trees), whereas genome-wide analyses can detect signals of local adaptation for unknown traits that may be selected for at any point in a tree\u2019s life.  As divergent selection occurs across a species\u2019 range, markers in association with that selection should shift in allele frequency, while neutral markers should follow prevailing patterns of drift or gene flow and show little divergence between populations (Wright 1931). This divergence in allele frequency between populations is often measured by FST (Wright 1949) or one of its analogues (Meirmans and Hedrick 2011). Markers showing unusually high FST may be local adaptation candidates for causal or linked variation, and the proportion and divergence of outliers may be evidence of the strength of selection occurring there (Lotterhos and Whitlock 2014). The detection of local adaptation through these analyses may be tenuous, as they are easily confounded by demography (Hoban et al. 2016), and may not be generalizable or reproducible due to sampling and linkage effects (\u0106ali\u0107 et al. 2016). In much the same way that polygenic traits tend to show strong clines along environmental gradients under local adaptation, the alleles responsible for those phenotypes should show similar patterns (Huxley 1939). Associations between allele frequencies and environmental variables have been used to infer the presence of local adaptation across a variety of taxa (e.g., Evans et al. 2014; Schweizer et al. 2016; Lind et al. 2017). Genotype-environment clines suffer similar drawbacks to outlier analysis in terms of confounding by demography and reproducibility but have the added benefit of highlighting environmental gradients that may be responsible for local adaptation. As such, the presence of strong clinal variation should suggest candidate local adaptation or support other lines of evidence. It should be noted that, while genomic divergence and clinal associations are not in and of themselves strong evidence for local 19  adaptation, a lack of these patterns is generally seen as strong evidence against local adaptation (Barton 1999).  2.3.2 Genomic divergence approaches Among the first successful attempts at identifying genomic markers associated with local adaptation in this study system came from candidate SNP panels. Namroud et al. (2008) successfully genotyped 534 SNPs distributed across 345 genes previously identified as candidates for regulatory functions, growth, phenology and wood formation (Pavy et al. 2005). Across six populations in as many distinct ecosystems in southern Quebec, there was extremely low divergence across all SNPs (FST=0.006), but the authors identified 20 SNPs across 19 genes with exceptionally high FST (range = 0.04 \u2013 0.13). Forty-nine SNPs were considered putative candidates for local adaptation based on their abnormally high or low allele frequency in a particular population, as measured by the interaction between locus and population effects on FST for a given locus (Beaumont and Balding 2004). None of these interactions were significant at 95 or 99% confidence levels, which the authors attributed to the limited size of the study area relative to the species range of white spruce and the highly polygenic nature of the traits associated with the candidate SNP panel used.  De La Torre et al. (2015) developed a SNP panel based on candidate genes for cold hardiness and herbivory resistance in Sitka spruce (Holliday et al. 2008). They were able to successfully genotype a panel of 86 candidate SNPs across interior spruce and allopatric populations. Approximately 6.5% of these candidate SNPs showed signals of selection based on FST outlier analysis or genotype-environment associations, although there was little overlap between these analyses (De La Torre et al. 2014c). Most SNPs showed signals of divergent 20  selection, although a small number showed signatures of balancing selection. This detection rate is similar in scale to allopatric white spruce and many other plant species (Namroud et al. 2008; Strasburg et al. 2012). Unsurprisingly, the FST values for outliers within the hybrid zone (range = 0.08 \u2013 0.18) and across the hybrid zone and both parental species (range = 0.04 \u2013 0.29) are substantially higher than those found for white spruce alone. While this is certainly in part due to the multi-species nature of interior spruce, it may also suggest stronger divergent selection among interior spruce populations than for white spruce. A separate approach of genomic divergence has been employed in white spruce, comparing population divergence in phenotypes (QST) to overall genomic divergence, as measured by GST, a multi-allelic FST analogue (Jaramillo-Correa et al. 2001). They found high phenotypic divergence, particularly in bud set, compared to low genetic divergence measured by 11 expressed sequence tags. Although this pattern is suggestive of local adaptation (Mckay and Latta 2002; S\u00e6ther et al. 2007; but see Whitlock 2008), the low genetic resolution of the molecular markers did not allow for direct inference of local adaptation.  2.3.3 Genotype-environment clines  Early studies of molecular variation identified a number of clinal relationships for interior spruce and its parent species. Latitudinal clines in allele frequency were identified for 12 of 24 isozyme markers in Engelmann spruce, as well as a general latitudinal cline in allelic diversity across all loci (Ledig et al. 2006). This supports the earlier conclusions of Shepperd et al. (1981), that genetic diversity in this species increases northward, with both authors suggesting hybridization as a putative cause of this relationship. These patterns were generally attributed to 21  population structure, as isozymes are expected to behave neutrally under most circumstances (Weeden and Wendel 1989; but see Falkenhagen 1985). The first genome-wide studies of clinal variation in this species complex came from Hornoy et al. (Hornoy et al. 2015), in allopatric white spruce. Using a custom SNP array, they genotyped 11,085 markers across 7,819 genes, likely representing between one quarter and one half of the transcriptome (Yeaman et al. 2014; Warren et al. 2015; Suren et al. 2016). Using a linear and quadratic regression approach and population structure as a covariate, the authors identified 31 genes showing strong clinal relationships with mean annual temperature, two of which were also associated with precipitation. Although the proportion of genes showing a clinal relationship with temperature is very low (0.004%) compared to the FST-based approaches discussed above, this study was the first to use a genome-wide panel, rather than a candidate gene approach which would presumably enrich the panel for targets under selection. Additionally, the weak local adaptation detected in phenotypic studies of white spruce may be underlain by smaller shifts in allele frequency across many loci than were detectable within the study (Le Corre and Kremer 2012).  The most comprehensive study of genome-wide adaptation to the environment in interior spruce was performed by Yeaman et al. (2016), a study which I co-authored (see preface). With 888,000 SNPs across nearly 23,000 genes and 254 provenances, we truly characterized genomic signatures of climatic adaptation in this species complex. Using a simple univariate marker-environment raw correlation method, Yeaman et al. identified 854 candidate genes for local adaptation (based on enrichment for SNP-environment association outliers compared to a null distribution). As our approach did not account for population structure, it is expected to have a high false positive rate, so this is an overestimate of the number of genes with a true effect on 22  local adaptation to climate. However, we expected adaptive variation to follow the same environmental gradients as population structure due to the hybrid nature of interior spruce, and statistically removing that population structure was expected to reduce our power to detect local adaptation (Lotterhos and Whitlock 2015). Like De La Torre et al. (2014c), Yeaman et al. found strong clinal relationships with precipitation as snow and mean annual precipitation, as well as strong relationships with winter temperatures, supporting phenotypic results showing strong local adaptation for cold hardiness in this species complex (Liepe et al. 2016; Yeaman et al. 2016; MacLachlan et al. 2018) and in other temperate tree species (Savolainen et al. 2007). After these correlations were corrected for population structure, the number of detected candidate genes dropped by an order of magnitude, as expected. Associations with winter temperatures remained strong, while associations with precipitation variables did not. This may suggest that local adaptation is stronger for temperature-associated genes, or that temperature is simply more orthogonal to population structure in this species complex. Yeaman et al. also directly compared results to those of Hornoy et al. 2015, identifying shared genes between the two datasets. We were able to compare 38 of 43 genes identified by Hornoy et al., though none of these appeared as top candidates in our analyses. However, 17 genes had weak-to-moderate associations with climatic variables. This may indicate that allopatric white spruce faces differing selective pressures than interior spruce, or that the different analyses employed by the two studies are not sensitive to the same adaptive signatures. The correlation between individual markers and environmental variables may be ineffective without sampling a large proportion of the genome and can be prone to false-positives, especially in systems with strong population structure or that recently underwent range expansion (de Villemereuil et al. 2014; Lotterhos and Whitlock 2015; Hoban et al. 2016). To 23  circumvent these concerns, a popular approach in interior spruce has been to correlate hybrid index, a measure of the overall proportion of ancestry an individual shares with one parental species or the other, with provenance environments. Hybrid index is, in this sense, a composite of population structure produced by migration and gene flow, as well as adaptation through natural selection (Barton and Gale 1993). While this index has the benefit of averaging the effects of many markers, it also confounds the effects of local adaptation and population structure.  De La Torre et al. (2014c) identified a cline in hybrid index with elevation in their candidate SNP data (86 markers), in agreement with relationships based on morphological features (Garman 1957) and allozymes (Rajora and Dancik 2000). This cline was not detected by De La Torre et al. (2015) with a separate estimate of hybrid index based on 10 microsatellite loci, emphasizing the importance of marker density in accurately describing genomic ancestry and therefore environmental clines in this species complex. This SNP-based estimate of hybrid index was also regressed against a suite of twenty measured and derived climatic variables (Wang et al. 2006), finding that variation in hybrid index was best described by a model containing two climatic variables: the amount of precipitation falling as snow, and a measure of summer aridity (De La Torre et al. 2014c). In this model, ancestry from Engelmann spruce increases as both snowfall and summer aridity increase, although snowfall was found to be the most important predictor in this system. A second study, using a 6,500 SNP panel and a broader range of provenances, again found a strong relationship between snowfall and hybrid index, but did not identify a relationship with summer aridity (MacLachlan et al. 2018). Instead, the authors found strong correlations between hybrid index and winter temperatures, likely as a result of more extensive sampling in the range of allopatric white spruce than De La Torre et al. The 24  results of MacLachlan et al. closely mirror those of Yeaman et al. (2016) in finding strong axes of adaptation associated with winter temperatures and weaker relationships with precipitation.  2.3.4 Genotype-phenotype clines Due to generally small effect sizes of loci underlying polygenic traits, studies linking genomic variants and these traits typically require large sample sizes, in terms of both loci and sampled individuals (Hong and Park 2012). Due to these limitations, only one study has comprehensively assessed these associations in interior spruce (Yeaman et al. 2016). In line with the environmental associations discussed above, the strongest relationships were found for traits related to cold adaptation. Only a handful of genes were identified as candidates for growth and phenology, but hundreds were identified in strong association with cold hardiness (Yeaman et al. 2016). Correcting for population structure identified candidate genes in association with more traits but had almost no overlap with candidates identified without correction for population structure. As discussed below, cold hardiness is tightly correlated with hybrid index, and as such many of the loci found to be associated with these traits are likely indicative of broader adaptive divergence between white and Engelmann spruce. This again highlights the coincident clines of adaptive variation and population structure in this species complex.  As with the environmental association analyses discussed above, some studies have used hybrid index as a composite genomic measure to identify clinal relationships in phenotypes. De La Torre et al. (2014c) split provenances from eleven field common gardens into parental and hybrid classes based on hybrid index, finding higher survival for provenances matching the environment of the common garden e.g., pure white spruce provenances performed best in common gardens planted in the allopatric range of white spruce. This provides good evidence of 25  an adaptive advantage of hybridization in this species complex, as provenances with hybrid ancestry generally outperformed pure parental species within the present hybrid zone. Additionally, De La Torre et al. (2014c) identified clinal relationships between hybrid index and growth, with a higher proportion of ancestry from Engelmann spruce generally translating to less growth across most of the common gardens.  MacLachlan et al. (2018) used a similar approach, finding a very strong cline in cold hardiness with hybrid index, with a higher proportion of Engelmann ancestry corresponding to lower cold tolerance. However, there was only a weak trade-off with growth (r2 = 0.25), with the cline giving opposing results to those of De La Torre et al. (2014c). This is interesting, as Engelmann ancestry has also been associated with depressed growth as well as lower cold tolerance (Liepe et al. 2016). These discrepancies may be, in part, due to differences in age of testing and growing conditions between the studies. MacLachlan et al. and Liepe et al. tested younger individuals than De La Torre et al., and all studies differed with respect to their growing environments. The common garden used by MacLachlan et al. was in Vancouver, BC, well outside the range of white or Engelmann spruce with warm, mild temperatures, Bud phenology has been implicated in local adaptation for many temperate tree species (Alberto et al. 2013), including within the genus Picea (Mimura and Aitken 2007; Chen et al. 2012), and bud break and bud set are critical components of cold hardiness (Aitken and Hannerz 2001). Interestingly, despite strong evidence for local adaptation to cold, little genomic evidence of phenological adaptation has been observed in interior spruce. Although clinal relationships have been observed between autumn bud phenology and climate in interior spruce (De La Torre et al. 2014c; Liepe et al. 2016; MacLachlan et al. 2018) and allopatric white spruce (Lesser and Parker 2004), no genomic studies have identified relationships between bud phenology and 26  hybrid index, or strong associations with genomic markers (De La Torre et al. 2014c; Yeaman et al. 2016; MacLachlan et al. 2018).  2.3.5 Genomic architecture of local adaptation The potentially locally adaptive phenotypes that have received attention above include survival, growth, phenology, and cold hardiness. Survival and growth are both so polygenic that most genes in the genome may have some near-immeasurably small effect, causing these traits to potentially be \u201comnigenic\u201d (Boyle et al. 2017). As such, it is unsurprising that Yeaman et al. (2016) identified few genomic loci in relation to height. Despite the adaptive significance of faster growth within the limitations of a given environment, the ability to detect associated loci declines with increasing polygenic control of traits (Yeaman 2015). In allopatric white spruce, quantitative trait loci for height are numerous, with small effect sizes and little consistency between crosses (Pelgas et al. 2011), also an expected outcome of highly polygenic traits (Mckay and Latta 2002).  Across most of the genome, linkage disequilibrium (LD) decays rapidly in allopatric white spruce and other conifers (Pavy et al. 2012). However, interior spruce does have detectable amounts of LD across many loci due to hybridization (Yeaman et al. 2016). As such, it is difficult to disentangle linkage caused by introgression from that caused by selective sweeps or epistasis. However, approximately one sixth of the genes associated with cold hardiness show LD above the 95th percentile of background (Yeaman et al. 2016), suggesting that while the trait is still highly polygenic and composed of largely independent genes, there may be more linked or strongly-selected gene complexes present for this trait than for growth. As cold hardiness is produced by the confluence of many phenological traits coordinating growth initiation and 27  cessation (Bigras et al. 2001), and is highly polygenic (Howe et al. 2003), some of the component phenotypes have simpler genomic architectures (Cooke et al. 2012; Singh et al. 2017) that could be responsible for the pattern seen here. However, these architectures have not been directly identified for interior spruce.  A paradox of widespread, outcrossing tree species is that they frequently exhibit local adaptation in the face of high gene flow through wind-dispersed pollen (Savolainen et al. 2007). It has long been predicted that high levels of gene flow should homogenize genetic variation, limiting local adaptation (Lenormand 2002). However, local adaptation appears nearly ubiquitous in these species (Alberto et al. 2013; Lind et al. 2018) and there is strong evidence for its presence in interior spruce. Recent theoretical work has suggested models that allow for the maintenance of local adaptation in the presence of high gene flow (Yeaman and Otto 2011; Yeaman 2015; Tigano and Friesen 2016). Factors that can act to produce or maintain local adaptation in the face of high gene flow include adaptive introgression, selection of linked loci, low recombination, and highly polygenic architectures (Yeaman 2015; Tigano and Friesen 2016). Adaptive introgression seems likely in this study system. Out of 86 SNP markers tested, 7 showed narrow geographic clines in allele frequency across the hybrid zone (De La Torre et al. 2014c), a pattern expected under local adaptation (Bridle et al. 2010). Of these seven SNPs, four also exhibited signs of adaptive introgression, generally in the form of directional selection for one parental allele over the other. While these only constitute a few observations, detection of any adaptive introgression, combined with other evidence for hybrid superiority and complementation (De La Torre et al. 2014c; Conte et al. 2017), suggests that novel combinations of white and Engelmann spruce alleles or parental alleles in novel environments, are contributing 28  to increased hybrid fitness and providing adaptive fodder for local adaptation. The large number of loci associated with cold hardiness and high background LD in interior spruce suggest that, while selection may be acting on some groups of linked loci, adaptation is primarily occurring through selection on independent genes (Yeaman et al. 2016). Together, these lines of evidence suggest that, despite high levels of gene flow, interior spruce has the requisite genomic capacity to adapt to local conditions.  2.3.6 Hybrid fitness  As many hybrid zones are geographically narrow and mobile (Hewitt 1988; Buggs 2007), and many interspecific hybrids are expected to show reduced fitness compared to their parents (Arnold and Hodges 1995), the apparent persistence and scale of the interior spruce zone begs an explanation. As discussed above, De La Torre et al. (2014c) found some phenotypic evidence for bounded hybrid superiority, i.e., hybrids having higher fitness than either parent within the hybrid zone but lower fitness outside of it, a model under which stable hybrid zones can form (Moore 1977). 2.3.6.1 Complementation Another potential explanation for hybrid superiority is complementation, wherein the effects of recessive deleterious alleles present in a parental species are mitigated in hybrids by the addition of non-deleterious alleles from the other parent species (Crow 1948; Birchler et al. 2006). As the effects of any single deleterious allele are expected to be small (Agrawal and Whitlock 2012), their detection and impact on fitness has been challenging to measure. However, the effects of nonsynonymous amino acid changes on protein function are reasonably predictable across taxa (Choi et al. 2012). Using the dataset of Yeaman et al. (2016) discussed above, Conte 29  et al. (2017) identified 165,576 potential non-synonymous SNP variants, of which about 13% contained a putatively deleterious allele. We found a relationship between the proportion of loci predicted to be homozygous for deleterious alleles and hybrid index, with hybrids having generally lower values than the parental species, as predicted under complementation. Two-year seedling biomass was used as a fitness proxy, showing a weak but significant parabolic relationship with hybrid index (R2 = 0.04, p < 0.001), lending some evidence to a fitness effect of hybridization in this system.  2.3.7 Plasticity  Phenotypic plasticity provides an alternative strategy to local adaptation, allowing the expression of traits within a single individual to vary depending on its environment. The degree of plasticity differs between traits (Sultan 2000), species (Funk 2008), and often between populations (M\u00e4gi et al. 2011; Molina-Montenegro and Naya 2012), and is difficult to predict based on species characteristics or ecology (Nicotra et al. 2010). Despite the importance of plasticity in understanding species\u2019 historical adaptation and potential responses to climate change (Aitken et al. 2008; Valladares et al. 2014), it has received limited direct study in interior spruce. However, the effects of phenotypic plasticity may be inferred from other experiments measuring phenotypes across environmental gradients. For example, while Liepe et al. (Liepe et al. 2016) partitioned phenotypic variance in several seedling traits across multiple environments in the interest of determining genetic effects, plasticity can also be inferred here from the environmental effects on these phenotypes. The results suggest wide variation in plasticity among traits. Bud break, for example, was found to be almost entirely plastic with environmental variation explaining 93.3% of the observed phenotypic variation, and only 0.5% attributable to 30  population variation. Fall cold hardiness showed almost no plasticity, with 36.6% of variation attributable to genetic variation but only 0.3% to environmental variation.   Another approach to studying plasticity has been through plasticity in gene expression. Yeaman et al. (2014) analyzed whole transcriptomic data from 39 interior spruce seedlings from a single population grown under seven climatic regimes, identifying 6,413 genes (27%) differing in expression across treatments. These genes clustered into 15 co-expression networks, suggesting a wide variety of traits may have experienced plastic responses across treatments. The largest differences in gene expression appeared in response to combined heat plus drought stress, suggesting a large plastic response to drought, with more moderate levels of plasticity for temperature treatments alone. To determine the relationship between plasticity and local adaptation in this species complex, Yeaman et al. (Yeaman et al. 2016) compared the expression patterns of top candidate genes for environmental and phenotypic associations (discussed above) with non-candidate genes, finding no difference between these groups (p = 0.34). This suggests that genes involved in local adaptation may also play roles in adaptive plasticity and vice versa, and that the genomic architectures underlying these evolutionary processes are not necessarily at odds in interior spruce.  2.4 Conclusions  To date, the strongest evidence for local adaptation in this species complex comes from separate analyses identifying similar environmental drivers associated with adaptive genomic variation and phenotypic traits. The research reviewed here suggests that, for interior spruce, these tend to largely be adaptations to cold temperature. Analyses of cold hardiness phenotypes in common garden experiments have revealed very strong clines in association with provenance 31  winter temperatures (Ukrainetz et al. 2011; Liepe et al. 2016; Yeaman et al. 2016; MacLachlan et al. 2018; Sebastian-Azcona et al. 2018), while genomic signatures of adaptation are also strong for these variables (Yeaman et al. 2016; MacLachlan et al. 2018). Additionally, cold hardiness appears to be under strong genetic control, with almost no plasticity (Liepe et al. 2016).   Adaptation to precipitation and aridity is less clear. Hybrid index appears to be correlated with precipitation regimes and snowfall (De La Torre et al. 2014c; MacLachlan et al. 2018), and aridity was found to be correlated with cold tolerance in one study (Ukrainetz et al. 2011), but only weakly in others (Yeaman et al. 2016; MacLachlan et al. 2018; Sebastian-Azcona et al. 2018). Precipitation may be a stronger driver of adaptation in allopatric white spruce, showing clines in bud set, cold hardiness and survival (Jaramillo-Correa et al. 2001; Sebastian-Azcona et al. 2018).  Trade-offs between growth and cold acclimation have been identified in many conifer species, as trees that grow for longer generally grow larger, but have delayed phenology and cold acclimation (Howe et al. 2003). This provides a strong fitness driver for local adaptation, as individuals must sacrifice growth for the ability to withstand colder temperatures when necessary, or neglect these adaptations to increase growth potential and competitive ability. However, this trade-off is not well characterized in interior spruce. In seedlings, most studies have found weak trade-offs (Ukrainetz et al. 2011; De La Torre et al. 2014c; Yeaman et al. 2016), while MacLachlan et al. (MacLachlan et al. 2018) found strong trade-offs. However, there is a clear trade-off in mature allopatric white spruce (Sebastian-Azcona et al. 2018). Unfortunately, no comparable data exists for mature interior spruce, so the true extent of trade-offs between these traits in natural systems remains unknown. If trends identified in seedlings 32  persist to adulthood, a lack of trade-offs between these ecologically relevant traits may explain some of the adaptive advantages of hybridization in this system. Additionally, spring and autumn cold hardiness appear to have differing environmental drivers and levels of genetic control (Bigras et al. 2001; Alberto et al. 2013), and may show opposing clinal variation (Simpson 1994), complicating the relationship of these trade-offs. Genomic data has provided unprecedented insight into the patterns and processes underpinning local adaptation in interior spruce. Through an increased understanding of the dynamics of hybridization and a better grasp on the climatic drivers that have shaped populations for millennia, genomic data for interior spruce has proved to be an invaluable companion to phenotypic experiments measuring local adaptation in this complex. Specifically, the many adaptations to cold temperature that are only at times apparent in phenotypic data become abundantly clear in the genomic signal. While there is much yet to learn about the nature and extent of local adaptation in the interior spruce species complex, the countless hours of effort contributed to understanding this species complex at both the phenotypic and genomic levels have made it a strong study system for local adaptation to climate by interspecific hybrids. The knowledge gained from the studies discussed above contribute not just to forest management and conservation practices, but also to our understanding of climate adaptation in trees and hybrid evolution more broadly. 33  Chapter 3: Research - Genome-environment associations and genomic clines in divergent SNPs suggest parental species\u2019 adaptations combine to generate fine-scale local adaptation to climate within a spruce hybrid zone.  3.1 Introduction As discussed in Chapter 2, relative species ancestry (i.e., hybrid index) in interior spruce follows several putatively adaptive climatic and phenotypic clines (De La Torre et al. 2014c; Liepe et al. 2016; Yeaman et al. 2016; MacLachlan et al. 2018), suggestive that species ancestry is, itself, locally adaptive in this species complex. This is likely due to selection in highly polygenic traits that draw large amounts of adaptive variation from both parental species to adapt to the varying climates faced by hybrids across the species range (Yeaman et al. 2016). However, there are still important questions to answer regarding the genomic nature of this adaptation. Under an infinitesimal model of selection, we would expect most allele frequencies to align to this primary axis of shared ancestry and selection, i.e., for alleles fixed between species, we would expect allele frequencies to mirror hybrid index at a population level. However, as even traits under polygenic selection tend to include some genes of larger effect (e.g., Zhang et al. 2018), it may be that hybrid index is driven by a few genes under strong selection pulling along large, physically-linked stretches of variation from one parent species or the other, shifting the overall hybrid index towards that parent. One method for distinguishing these patterns is to analyze genomic clines among differentiated SNPs.    34  3.1.1 Genomic clines The change in trait values or allele frequencies measured across a hybrid zone transect can be modeled as a geographic cline (Endler 1977), where biological relevance can be obtained from the width and steepness of the cline. This concept has been translated to genetic analysis of hybrid zones (Barton 1983; Gompert and Buerkle 2011), where a change in allele frequency for one locus can be plotted as a function of mean genomic ancestry. In this framework, an individual\u2019s genotype at a locus (expressed as probability of carrying an allele derived from a focal parent species) is modeled as a monotonically-increasing sigmoidal function of background genomic ancestry i.e., hybrid index, after Szymura and Barton (1986). The cline is dictated by two shape parameters carrying biological significance: (1) an overall shift of the cline center (\u03b1) based on the degree of introgression into hybrid populations from the focal parent or the alternate parent, presumably as a result of directional selection; and (2) the cline\u2019s rate or steepness of the gradient (\u03b2), which reflects how rapidly genotypic ratios change across the spectrum of hybrid indices, and may also carry adaptive significance. Simulated data suggest that \u03b1 outliers tend to be more pronounced and less prone to false positives than \u03b2 outliers (Gompert and Buerkle 2011), so they are the focus of this research.  The presence of genomic cline outliers suggests that differing degrees of introgression are occurring at the genomic level. In general, extreme values of genomic cline center can be generated by alleles that are widely adaptive across many landscapes i.e., one allele is more abundant than expected across all hybrid indices. However, in the special case where hybrid index follows an adaptive gradient, outliers should also be generated by alleles that are strongly selected at one end of that gradient e.g., if all hybrids with a hybrid index > 0.5 face a particular selective pressure that is not present where hybrid index < 0.5, one allele should be favored 35  above intermediacy in many hybrids, skewing the cline center parameter. As alleles that are common across a broad environmental gradient would suggest either neutrality or global adaptiveness of that allele, rather than local adaptiveness, the correlation between allele frequencies and selective environmental gradients should be elevated in locally-adaptive SNPs.  3.1.2 Genome-environment analysis  As discussed in Chapter 1, selection will act to closely align favorable genotypes to environmental pressures under a model of local adaptation (Williams 1966; Savolainen et al. 2013). This can be observed at the landscape level, as population allele frequencies at selected loci are expected to correlate with environmental gradients that cause or are associated with selective pressures (Haldane 1948). This process, often referred to as genotype-environment association (GEA), is a fundamental concept in landscape genetics and has been widely-used to attempt to identify the bases of local adaptation for many species (Rellstab et al. 2015). These associations should be exceptionally strong in plants as compared to animals, as plants are sessile and must endure the environmental stressors present at their site of germination. All resistance to hostile environments must ultimately stem from genetic adaptations.  One important caveat in GEA analysis is that demographic processes can produce clines similar to those expected by local adaptation, e.g., allele surfing during range expansion or migration barriers coincident with environmental clines (Hoban et al. 2016). GEA analyses generally incorporates some degree of correction for these confounding factors, either by comparing patterns against a neutral model, e.g., partial Mantel tests (Smouse et al. 1986), modelling spatial autocorrelation e.g., spatial generalized linear mixed models (Guillot et al. 36  2014), or generating a covariance matrix of population structure or individual relatedness; e.g., Bayenv (Coop et al. 2010).   Accounting for demographic processes is especially challenging when analyzing trends of adaptation within hybrid zones. As adaptive gradients tend to mirror demographic gradients, structure correction may \u201cwash out\u201d a great deal of adaptive variation (Yeaman et al. 2016). However, this correction is useful in the context of describing genomic cline outliers. After correcting for population structure, the SNPs that still show strong adaptive signals are those that run most orthogonal to hybrid index, i.e., they represent SNPs that show local adaptation beyond what is expected by hybrid index. These are the SNPs of interest in this research, as they can indicate whether alleles of a particular species are over-represented among SNPs showing exceptional adaptive patterns.   3.1.3 Sequence capture Many conifers have extremely large genomes (De La Torre et al. 2014a). Interior spruce is no exception, with an approximately 20gb genome size (Birol et al. 2013). This produces many challenges to genomic analysis. Whole genome shotgun sequencing (WGS) is extremely cost-ineffective, as the genome is composed largely of repetitive elements (De La Torre et al. 2014a) that are of little evolutionary significance (but see Wegrzyn et al. 2014; Schrader and Schmitz 2019). As selection is expected to work most strongly in and near coding regions (Pulkkinen and Metzler 2013; Enard et al. 2014; but see Clyde 2020), it is desirable to focus sequencing efforts on these regions. While many methods have been developed to enrich genomic libraries in coding regions (Puritz and Lotterhos 2018), some of these approaches suffer from low consistency between samples. Exome sequence capture (seqcap) is a reasonably cost-37  effective method for consistently sampling the coding region of large genomes (Meek and Larson 2019) that can achieve high consistency in conifers (Suren et al. 2016). The fundamental idea behind seqcap is to use \u201cbaits\u201d i.e., short, unique sequences of single-stranded DNA that will bind specifically to a targeted region, to capture strands of genomic DNA around the targeted region (Jones and Good 2016). Although this method is attractive, it requires challenging and finicky laboratory methods, and requires pre-existing knowledge about the genome; i.e., the targeted sequences must be known.   3.1.4 Objectives Previous research projects have attempted to identify the environmental drivers of adaptation in the spruce hybrid zones of BC using genomic data (De La Torre et al. 2014c; Hamilton et al. 2015; MacLachlan et al. 2018). However, these efforts were limited in sampling, either geographically or genomically. By sampling many white and Engelmann spruce populations in western Canada, and assessing variation across a large portion of the exome, we will be able to more fully and accurately identify genomic patterns of adaptation to climate in this hybrid zone.  In this chapter, we will assess the geographic extent of hybridization between white, Engelmann, and Sitka spruce. This has important implications for understanding the extent and degree of gene flow between naturally hybridizing species in this region, and to identify regions of allopatry. As previous research has suggested that hybrid index shows clinal relationships with climate, we expect that hybridization contributes to local adaptation in this species complex, and sampling the broadest possible range of climates will help to refine those relationships and 38  potentially reveal novel relationships that were not present in the areas sampled in prior research projects. Using genotype-environment associations, we will identify putative climatic drivers of genomic variation throughout western Canada. This is important for understanding the selective pressures that have allowed the interior spruce hybrid zone to form and persist, but also to determine the climatic drivers that may shape the genomic makeup of the hybrid zone in the future. By analyzing genomic clines in SNPs with divergent allele frequencies between white and Engelmann spruce, we can determine the extent of introgression into the hybrids, and identify putatively adaptive alleles that allow hybrids to thrive in the specific environments of the hybrid zone. By specifically analyzing the genotype-environment associations of these loci, we can identify climatic drivers of hybridization and adaptive introgression of parental alleles into hybrids.  This research has obvious implications specifically for understanding adaptation in these species and their hybrids. However, the broad climatic and geographic scales sampled here will deepen our understanding of the genomic consequences of adaptive hybridization, and potentially identify climatic drivers of adaptation in other temperate species.   3.2 Methods  3.2.1 Sampling In order to strike a balance between sampling the broadest possible range of climates and ensuring that sampling locations are reasonably representative within the limitations of sequencing costs, a \u201cbroad and shallow\u201d sampling scheme was employed for the seqcap dataset. 39  One to four individuals per provenance were sampled across 254 provenances (totaling 579 genotyped individuals), chosen from wildstand seedlots available from the BC and National Tree Seed Centers. Seedlots were selected to achieve good spatial distribution across western Canada, and to fully sample precipitation and temperature gradients for interior spruce and its parent species in western Canada. Seeds were stratified at 4\u00b0C for four weeks, then germinated and grown in a Conviron growth chamber with a simulated climate with a mean annual temperature of 6\u00b0C. Juvenile needles were sampled after eleven weeks of growth and used for subsequent DNA extraction.  3.2.2 Genotypic data The genotypic data used in this chapter was generated using a sequence capture approach described by Suren et al. (2016). DNA was extracted using Machery-Nagel Nucleospring 96 Plant II kits. DNA quantity and quality was assessed using a Thermo Scientific NanoDrop ND-2000 spectrophotometer, with spot checks performed using an Invitrogen Qubit 2.0 fluorometer, a more accurate but time- and reagent-consuming machine. An important consideration in sequence capture is the target DNA fragment size. Smaller fragments will have a higher rate of success but are more challenging to align. A target size of 350bp was chosen as a compromise between these competing aims. To size select fragments, DNA was run on agar gels alongside a ladder, and DNA between 300-400bp was extracted from these gels and re-purified. These libraries were spot checked using an Agilent 2100 Bioanalyzer to ensure that size selection was generating normally-distributed fragments centered around 350bp. To generate a series of genomic baits that could be used to sample the spruce exome, a transcriptome of interior spruce generated as part of a previous study (Yeaman et al. 2014) was 40  sampled extensively to saturate the exome. Approximately 812,000 60-90bp baits were designed to capture 28,649 putative genes as well as other unmapped transcriptome contigs and non-coding sequences (Suren et al. 2016). Magnetic microbeads were coated with the baits to extract the desired genomic regions. After allowing de-annealed genomic DNA to hybridize with the baits, magnets were used to filter the baited sequences and wash out the remainder. For the sake of efficiency in time and reagents, 1ug DNA samples were individually barcoded and pooled with five other samples for this step. These pools were again barcoded and paired for sequencing (for ultimately 12-plex sequencing). Sequencing was performed on an Illumina HiSeq 2000, with 100bp paired-end reads (200bp total sequenced per read).   3.2.3 Bioinformatics Sequences were aligned to a modified version of the February 2013 draft interior spruce genome PG29 (Birol et al. 2013). To facilitate alignment against the large and highly fragmented genome assembly, the genome was reduced to contigs with baits designed to target them, and any other contigs with SNPs called in at least 50% of samples based on a preliminary alignment using 84 individuals selected to represent a broad geographic distribution. These contigs were linked together into 1000 \u201cpseudoscaffolds\u201d, with contigs separated by 20 \u201cA\u201d nucleotides to avoid alignment across contigs. Prior to alignment, sequence reads were trimmed and filtered using FASTX toolkit, with end-read bases dropped if their Phred scores were below 10 and entire reads dropped if 10% of the read had bases below 10.  Sequences were aligned to the genome using the Burrows-Wheeler alignment \u201cBWA-MEM\u201d algorithm using default parameters, with PCR duplicates flagged and removed using 41  Picard MarkDuplicates. GATK v3.3 was used to realign reads around indels, recalibrate SNP scores, and call SNPs (using the GATK Unified Genotyper), using default parameters.  After SNP calling, SNPs were initially filtered using available best practice guidelines. This generated a large dataset of ~6 million SNPs, which were used for preliminary genotype-environment and genotype-phenotype analyses. The results of these preliminary analyses were used to design an Affymetrix 55k SNP array (see Section 4.2.1 of this dissertation), which was in turn used to generate data-driven filtering parameters for the seqcap data, based on the performance of SNPs on the SNP array. Parameters were visually assessed in plots showing parameter scores against genotyping success rate on the SNP array. Cutoffs were chosen to maximize successful calls while minimizing variant exclusion. Based on this exercise, SNPs were retained if they met the following criteria: SNP quality > 20, mapping quality > 40, Fisher strand score (based on the ratio of reference\/alternate alleles on forward\/reverse strands) < 40, haplotype score (a measure of how well the SNP is predicted using a 10bp window around the SNP) < 13, mapping quality ranksum (a Wilcoxon rank test assessing if reads are consistently of poorer quality for the alternate allele than the reference allele) < -12.5, and read position ranksum (a Mann-Whitney rank sum test assessing if the SNP is more commonly found at the edges of reads compared to the middle) > -8, a heterozygote allele balance > 2.3, genotype missing rate < 0.9, and minor allele frequency > 0.05. Overall, this generated a dataset of 899,596 SNPs. A total of 573 samples were genotyped using the seqcap method and used for subsequent analyses.   42  3.2.4 Genomic hybrid index ADMIXTURE (Alexander et al. 2009; Alexander and Lange 2011) is a software tool designed to estimate the probability that individuals are derived from some one or more pre-determined number of (sampled or unsampled) ancestral populations, using marker data. When the pre-determined number of ancestral populations is equivalent to the number of hybridizing species in a dataset, and when relatively pure parental species genotypes are available, these ancestry probabilities tend to closely approximate hybrid index (Alexander and Lange 2011; but see Lawson et al. 2018). ADMIXTURE uses the same set of equations as the more well-known STRUCTURE algorithm (Pritchard et al. 2000), but instead of using Bayesian estimation methods, uses accelerated maximum likelihood methods. This allows ADMIXTURE to estimate ancestry proportions using hundreds of thousands of markers across hundreds to thousands of individuals in minutes or a few hours, compared to the days or weeks necessary to estimate these proportions using STRUCTURE on only a few thousand loci (Alexander et al. 2009). While this method performs relatively poorly on small datasets or complex population histories (Lawson et al. 2018), it can reliably estimate ancestry proportions when using many thousands of loci and when the ancestral populations are known (Meisner and Albrechtsen 2018).  An important confounding factor in the interior spruce hybrid zone is that Sitka spruce ancestry is present in more westerly populations (Sutton et al. 1991; Hamilton et al. 2015). Naively, one might assume that running ADMIXTURE with three ancestral populations would recover the Sitka ancestry in some individuals. This may be the case when large amounts of Sitka ancestry are present, but in the seqcap dataset, the predominant ancestries are from white and Engelmann spruce, with pure parental species present for both, and Sitka ancestry tends to be confounded with Engelmann ancestry (results not shown). To account for potential Sitka 43  ancestry, 28 additional parental Sitka spruce genotypes were obtained from an unrelated study using the same seqcap methods (Elleouet 2018), and the full marker dataset was pruned to only include markers present in at least 70% of Sitka and AdapTree seqcap samples. This produced a dataset of 291,094 SNPs, which were then used to run ADMIXTURE v1.3, assuming three ancestral populations (K=3).  Individuals were classified based on their estimated ancestry proportions. Any individual with a proportion >0.875 for any individual species was considered a pure parental genotype. This proportion would be equivalent to a second-generation backcross between an F1 hybrid and parental genotypes, and accounts for error in estimating the proportions due to incomplete parent species range sampling and genotyping error. Using this rationale, any individual with >=0.125 ancestry for a given species was considered a hybrid of that species. Genotypes with >= 0.125 for all three species were considered three-way hybrids.   3.2.5 Genomic cline analysis Bayesian analysis of genomic clines (Gompert and Buerkle 2011, 2012) was selected for this analysis as it was able to simultaneously calculate genomic clines on several thousand unlinked loci (Gompert and Buerkle 2012). The input data required for bgc are a classification of individuals into parent or unknown, and a list of loci. To determine a panel of trees belonging to the parent species, ADMIXTURE v1.3 was run on the full set of samples and SNPs, with ancestry proportions bootstrapped using 20 replicates. With ancestry proportions ranging from 0 (pure white spruce) to 1 (pure Engelmann spruce), white spruce parents were those with an upper bootstrapped confidence bound of 0.05. This ultimately yielded a parental panel of 102 samples. For Engelmann spruce, there were fewer samples with high ancestry proportions, so a cutoff was 44  imposed at a lower bootstrapped confidence bound of 0.9. This yielded a parental panel of 84 samples. As bgc attempts to classify alleles within loci to parent species, it works best on loci that have separate fixed alleles in each parent species, or at least show high divergence between species. The parental panels were used to filter the full SNP dataset to those only showing fixed differences or strong divergence between species. As very few loci showed fixed differences between white and Engelmann parental panels, an allele frequency differential of 0.6 was chosen to reduce the number of loci to a computationally manageable number while still selecting SNPs that had alleles that could accurately be classified as being derived from, or under selection in, a given parent species. This produced a set of 14,009 SNPs across 3,832 unique contigs that were used for further analysis. Once internal parameters had been tweaked to consistently produce log-likelihood convergence and a stochastic spread around the converged likelihood (Gompert and Buerkle 2012), two chains were generated and used to estimate the genomic cline center for each SNP. As the results between the two chains were extremely similar, no more chains were generated.   Outlier SNPs were determined by calculating a bootstrapped p-value of the genomic cline center parameter, with any 95% confidence intervals falling outside of 0 classed as outliers. Due to the directionality of hybrid index used (0 = white spruce, 1 = Engelmann spruce), negative values represent SNPs with a higher-than-expected proportion of white spruce alleles at a given hybrid index, and positive values represent SNPs with a higher-than-expected proportion of Engelmann spruce alleles at a given hybrid index. Following outlier detection, SNPs were classified into three groups: \u201cEngelmann-skewed\u201d SNPs are those with positive genomic cline centers i.e., the Engelmann allele is more abundant than expected across the spectrum of hybrid indices; \u201cwhite-skewed\u201d SNPs are those with negative genomic cline centers i.e., the white allele 45  is more abundant than expected across the spectrum of hybrid indices; \u201cindex-neutral\u201d SNPs are those with a cline center near 0 i.e., allele frequencies vary closely with hybrid index. The term \u201cparent-skewed\u201d refers to all SNPs with a cline center differing from 0.  3.2.6 Genotype-environment analysis 3.2.6.1 Climatic variables analyzed  The climatic variables analyzed here were derived from ClimateWNA (Wang et al. 2016). To limit the number of covariates, only annual climatic variables were analyzed. For each provenance, 22 annual climatic variables were estimated based on the latitude, longitude, and elevation of each provenance. Briefly, ClimateWNA predicts climates at specific points by downscaling PRISM (Daly et al. 1994) raster data (at 4km resolution), using bilinear interpolation to create a seamless surface and a local elevation lapse rate adjustment based on regressing climatic anomalies against elevational differences between nearby grid cells (Wang et al. 2012, 2016) The annual climatic variables generated by ClimateWNA include 8 variables calculated as simple averages or simple equations derived directly from data: mean annual temperature (MAT), mean temperature of the warmest month (MWMT), mean temperature of the coldest month (MCMT), temperature differential (TD; the difference between MWMT and MCMT), mean annual precipitation (MAP), mean summer precipitation (MSP; calculated from May \u2013 Sept), annual heat-moisture index (AHM; calculated as (MAT + 10)\/(MAP\/1000)), and summer heat-moisture index (SHM; calculated as MWMT\/(MSP\/1000)).  The inputs for ClimateWNA are monthly precipitation and temperature variables. However, many biologically relevant climatic variables require daily data. To estimate these 46  data, ClimateWNA estimates empirical relationships between daily weather station data from 4,891 weather stations across North America and the monthly climatic variables generated from ClimateWNA. This results in 14 derived climatic variables: Degree-days below 0\u00b0C (DD<0 i.e., chilling degree-days), degree-days above 5\u00b0C (DD>5 i.e., growing degree-days), degree-days below 18\u00b0C (DD<18 i.e., cooling degree-days), degree-days above 18\u00b0C (DD>18 i.e., heating degree-days), number of frost-free days (NFFD), length of frost-free period (FFP), beginning date of frost-free period (bFFP), end date of frost-free period (eFFP), precipitation falling as snow (PAS), 30-year extreme minimum temperature (EMT), 30-year extreme maximum temperature (EXT), Hargreaves reference evaporation (Eref), Hargreaves climatic moisture deficit (CMD), and relative humidity (RH). Units for all climatic variables, as well as summary statistics for each across the dataset, can be found in Table 3.1.            47  Table 3.1: Climatic variables used in subsequent analyses, as predicted using ClimateWNA and ClimateNA. Means, standard deviations (SD), and range are shown for the 254 spruce provenance locations used in this dissertation. Climatic variable Abbreviation Mean (SD) Range Mean annual temperature (\u00b0C) MAT 1.6 (1.8) -3.3 - 6.4 Mean temperature of the warmest month (\u00b0C) MWMT 14.5 (1.7) 10 - 19 Mean temperature of the coldest month (\u00b0C) MCMT -12.8 (5.3) -25.2 - -3.2 Temperature differential (\u00b0C) TD 27.3 (6.2) 17.6 - 42.1 Mean annual precipitation (mm) MAP 668 (282) 194 - 2184 Mean summer precipitation (mm) MSP 305 (73) 81 - 530 Annual heat-moisture index AHM 19.4 (5.9) 6.2 - 52.5 Summer heat-moisture index SHM 50.8 (17.4) 21.8 - 227.6 Degree-days below 0\u00b0C (Celsius degree-days) DD_0 1436 (566) 436 - 2953 Degree-days above 5\u00b0C (Celsius degree-days) DD5 1085 (238) 497 - 1773 Degree-days below 18\u00b0C (Celsius degree-days) DD_18 5975 (664) 4295 - 7736 Degree-days above 18\u00b0C (Celsius degree-days) DD18 26 (21) 3 - 115 Number of frost-free days (days) NFFD 148 (16) 108 - 207 Beginning of frost-free period (Julian date) bFFP 156 (10) 128 - 181 End of frost-free period (Julian date) eFFP 252 (8) 233 - 278 Length of frost-free period (days) FFP 96 (16) 53 - 148 Precipitation as snow (mm) PAS 279 (189) 79 - 932 30-year extreme minimum temperature (\u00b0C) EMT -42.4 (4.1) -52.1 - -29 30-year extreme maximum temperature (\u00b0C) EXT 32.8 (1.9) 28 - 39 Hargreaves reference evaporation (mm) Eref 519 (59) 367 - 766 Hargreaves climatic moisture deficit (mm) CMD 170 (89) 0 - 541 Relative humidity (%) RH 60 (4) 46 - 73  3.2.6.2 Climatic similarity index  To climatically characterize the white-Engelmann hybrid zone, each climatic variable was individually assessed for its intermediacy between the parent species\u2019 ranges. Using the parental assignments described above to categorize individuals into white, Engelmann, or hybrid, a Fisher-Pitman permutation test was used to test the probability that the distribution of climates for hybrid provenances was drawn from the same distribution as each of the parental panels. An index was derived calculating the difference between log-transformed p-values of the hybrid-Engelmann test and the hybrid-white test. These indices were converted to a scale of -1 \u2013 1, 48  where -1 represents a climatic variable where the hybrid zone is very similar to the range of white spruce, and +1 represents a climatic variable where the hybrid zone is very similar to the range of Engelmann spruce. This range was anchored by the most extreme climatic variable (DD<0), where the hybrid zone is very similar to Engelmann spruce and very dissimilar to white spruce (Figure 3.1) Intermediate values represent climatic variables where either the hybrid zone is truly intermediate between the two parent species ranges, or where there is low differentiation between parent species.   Climatic variables were also analyzed using principal component analysis to determine major axes of climate variation and to characterize the hybrid zone in terms of these axes. For the areas covered by climatic niche models for white, Engelmann and hybrid spruce separately (as generated by Hamann and Wang (2006)), 10,000 random geographic points were selected to predict climatic variables using ClimateWNA. Exploratory analyses using all climatic variables together showed differentiation of the first two components into generally temperature and precipitation gradients, respectively. To separate these two major climatic gradients more clearly, two separate PCAs were completed using all precipitation or temperature variables to generate the first component of each independently. The values for the first component from each analysis was subsequently used for visualization.  49      Figure 3.1: Distribution of 254 interior spruce provenances in relation to the number of degree days below 0\u00b0 C. Species assignments were determined using genomic hybrid index.  50  3.2.6.3 Bayenv2 analysis  Bayenv2 (G\u00fcnther and Coop 2013) was selected for GEA analysis as comparisons at the time showed consistently high performance in systems with strong population structure relative to other GEA methods (Lotterhos and Whitlock 2015; but see Forester et al. 2018). First, to generate a covariance matrix between all samples that could be used to correct for population structure, 9,921 SNPs in non-coding regions were used to estimate population structure. Three replicates were generated using 100,000 iterations each and averaged to produce the final covariance matrix used by Bayenv2. Bayenv2 was then run 3 times with the full set of 899,596 SNPs against the full suite of annual climatic variables for 10,000 iterations each, with the Bayes factors and correlations averaged across the three runs. Rather than analyzing raw Bayes factors using an arbitrary threshold for what constitutes a \u201cstrong\u201d adaptive signature, all Bayes factors within each climatic variable were ranked and analyzed as percentiles. This approach was chosen to analyze where the bgc outlier SNPs of interest fall within the broader dataset, and to what extent were SNPs \u201cenriched\u201d for a particular climatic association, e.g., were Engelmann-skewed SNPs more commonly identified in the highest percentiles of associations than white-skewed or index-neutral SNPs. While this does not address the relative strength of association for each variable, this has been done elsewhere for this dataset (Yeaman et al. 2016). To test absolute strength of associations, the structure-corrected SNP-environment correlations were also analyzed and compared across the bgc SNP classes. While correcting for population structure does strongly bias the identification of outliers in this dataset (Yeaman et al. 2016), it is useful for assessing the strength of correlation in these pre-selected diverged SNPs that are, through their inclusion in this dataset, already strongly correlated with population structure.  51  3.2.7 Signatures of local adaptation  Local adaptation to climate was inferred through several lines of evidence. First, bgc outliers were quantified in terms of their \u201cenrichment\u201d in the top percentiles of SNPs identified for each climatic association. This was calculated by estimating the odds ratios of SNPs in a given group (i.e., Engelmann-skewed, white-skewed, parent-skewed, and index-neutral) for each climate variable and percentile, as compared to all other SNPs included at that percentile, and 95% confidence intervals were estimated using the fisher.exact implementation in R. For example, if 2,000 SNPs were identified as Engelmann-skewed, and 100 of those SNPs were identified at the 99th percentile of associations with a specific climatic variable (8996 total SNPs out of 889965 possible SNPs), the odds ratio would be (100\/1900) \/ (8896\/889069) = 5.26, indicating that Engelmann-skewed SNPs are 5.26 times as likely to be in that percentile of that analysis than other SNPs. This gives us an indication of how enriched a particular class of SNPs is compared against the entire dataset, but is of limited value. As the 14,009 SNPs selected for bgc analysis were selected a priori for divergence between provenances, it is very likely that they will generally be outliers in association analyses. Rather, a lack of enrichment is more informative here, as it indicates that, if any adaptation is occurring at all along that axis, hybridization is likely not playing a major role.   To determine the enrichment of parent-skewed SNPs, separate odds ratios were calculated for white- or Engelmann-skewed SNPs as compared to index-neutral SNPs. The index-neutral SNPs provide a null expectation of SNPs that, while highly diverged between parent species, covary closely with hybrid index. If parent-skewed SNPs are over- or under-represented in these analyses compared to index-neutral SNPs, it is suggestive of selective pressure for or against parental alleles with respect to that environmental gradient. In this 52  context, an odds ratio > 1 indicates that parent-skewed SNPs are overrepresented for a given climatic variable at a given percentile, suggesting that alleles from that parent are being favored at some point along that environmental gradient. Conversely, an odds ratio < 1 indicates that parent-skewed SNPs are underrepresented compared to index-neutral SNPs and may indicate selection against alleles from that parent along that climatic axis. An odds ratio \u2248 1 suggests that alleles from a given parent and not specifically favored along that axis, and any adaptation along it is likely to correlate closely with hybrid index. This enrichment analysis largely tests for congruence between the genomic cline analysis and genotype-environment association analysis in terms of identifying SNPs that fall outside the expected distributions in both analyses. Importantly, testing Engelmann- and white-skewed SNPs separately should allow us to determine which parent\u2019s alleles, if any, are particularly advantageous with regards to a specific climatic gradient. Bearing in mind that bgc outliers indicate that one or the other parent allele is more abundant than expected in hybrids, strong correlations between these SNPs and climatic variables suggest that selection has acted to enrich hybrids with alleles from a given parent in response to climatic stress, or at least from an axis that is strongly correlated to the environment. Although hybrid index for a given provenance may itself be indicative of selective pressures (Barton and Gale 1993; Kruuk et al. 1999; Feder et al. 2012; Taylor et al. 2015), bgc outliers indicate allelic abundance even beyond this adaptive or gene-flow-related gradient. However, strong associations with parent-skewed SNPs does not necessarily mean that the alleles of that parent are being selected for. Selection against those alleles along certain climatic axes would also show strong enrichment in the top percentiles of genotype-environment analyses. To determine whether the bgc outliers were suggestive of one parent allele being favored over the 53  other, the population-structure-corrected correlations between each SNP and each environmental variable as output by Bayenv2 were compared between Engelmann- and white-skewed SNPs. For a given selective gradient (as represented here by climatic variables), if Engelmann alleles are generally favored, then we would expect positive correlations between Engelmann alleles in Engelmann-skewed SNPs, and negative or negligible correlations for white alleles in white-skewed SNPs. If this is the case, we would expect that climatic variables acting as differential selective pressures i.e., favoring alleles from one species or the other, will exhibit a difference in mean correlation between Engelmann- and white-skewed SNPs. If selection is indeed acting in the way described above, climatic variables with extreme climate similarity indices i.e., where the climate of the hybrid zone is much more similar to one parental species than the other, should show more pronounced trends, as hybrids have selectively colonized areas that are non-intermediate between the parent species in this aspect.   3.3 Results  3.3.1 Genomic hybrid index  Of the 579 genotypes analyzed, 204 were identified as pure white spruce, 79 as pure Engelmann spruce, and 277 as white-Engelmann hybrids. These 560 individuals were used for subsequent genomic cline analyses, as the incorporation of individuals with Sitka ancestry would confound the analyses, which can only incorporate univariate clines; i.e., clines between two parent species. Additionally, 11 Engelmann-Sitka hybrids were identified, primarily from populations in the southern coastal submaritime region of BC, which have climates intermediate between coastal and interior regions (Green and Klinka 1994). A further eight individuals were 54  classified as white-Engelmann-Sitka hybrids, occurring primarily in the northwestern interior rainforest region of BC. The spruce in this region has long been suspected to be hybrid (Roche 1969), and three-way hybrids have been previously reported in this area (Sutton et al. 1994; Hamilton et al. 2015). In general, ancestry proportions for hybrid populations were consistent within provenances, with an average standard deviation of 0.032. Average hybrid indices for each provenance are shown in Figure 2.1. A wide range of hybrid indices are present at high levels (Figure 2.1). This is valuable for subsequent analyses, as genomic clines can only accurately be evaluated across the range of hybrid indices that exist within a dataset (Gompert and Buerkle 2011).   3.3.2 Climate similarity index  Climatically, the white-Engelmann hybrid zone varies in characterization depending upon which climatic variables are analyzed. Based on the concept of bounded hybrid superiority (Moore 1977), it was anticipated that many climatic variables for hybrid provenances would show intermediacy between the parental species ranges. The earliest reports of this hybrid zone remarked upon the elevational intermediacy of the hybrid zone (Garman 1957), and subsequent reports used this as justification that the zone is likely one of bounded hybrid superiority. However, the reality appears far more complex. The values of \u0394log(p) for individual climatic variables ranged from -8.5 (AHM) to +10.2 (DD<0). The largest absolute value (10.2) was used to anchor the scale of possible similarity indices, with \u0394log(p) = -10.2 corresponding to a similarity index of \u20131; i.e., the hybrid zone is very similar to the allopatric range of white spruce, \u0394log(p) = 0 corresponding to a similarity index of 0, and \u0394log(p) = +10.2 corresponding to a 55  hybrid index of +1; i.e., the hybrid zone is very similar to the allopatric range of Engelmann spruce. For climatic variables related to precipitation (MAP, MSP, PAS) and composite heat-moisture indices (AHM, SHM), the hybrid zone is much more similar to allopatric white spruce than Engelmann spruce (\u0394log(p) ranges from -3.5 to -8.5, relating to climatic similarity indices of -0.83 to -0.34, average index = -0.6). For temperature-related variables (MAT, MWMT, MCMT, TD, DD<0, DD>5, EMT, EXT), the hybrid zone is more similar to allopatric Engelmann spruce (\u0394log(p) ranges from +0.4 to +10.2, relating to climate similarity indices of 0.04 to 1, average hybrid index = 0.6). The hybrid zone is most intermediate for climatic variables relating to the length and characteristics of the growing season (MWMT, MSP, SHM, DD>5, NFFD, BFFP, EFFP, FFP, EREF, CMD). These variables showed \u0394log(p) ranging from -3.8 to +8.0, relating to climate similarity indices of -0.38 to 0.78, average similarity index = 0.1).  Principal component analysis (Figure 3.2) revealed that the parent species exist in largely distinct climate spaces with regard to precipitation and temperature. Along these axes, low temperature PC values correspond to generally warmer temperatures, and low precipitation PC values correspond to generally higher precipitation. White spruce spans a broad temperature gradient with relatively little variation in precipitation and vice versa for Engelmann spruce. Hybrid spruce tends to occupy the climatic intersection of the two, at the warmest and driest end of the respective parental species\u2019 niches. The geographic area covered only by the hybrid spruce niche includes much of the extreme end of this intersection.  56     Figure 3.2: Ordination of 10,000 random geographic locations within the climatic niches of white spruce (blue), Engelmann spruce (red), and hybrid spruce (black and green) along principal component axes separately calculated for temperature and precipitation variables. The percentages on the axes represent the percentage of variances explained by the first component of each analysis. Black points represent areas of the hybrid niche that overlap with at least one of the parent species, and green points (enlarged for emphasis) represent geographic areas only covered by the hybrid niche.  57  3.3.3 Genomic cline analysis   Of the 14,009 SNPs tested, 4,944 had cline centers outside the neutral range at the 95% confidence level (Figure 3.3), 3,303 at 99%, and 2,347 at 99.9%. A total of 1,703 of the 3,832 contigs contained at least one SNP with a significant cline center at the 95% level. This indicates that, despite filtering for SNPs with extreme allele frequency differences in the parent species, the majority of SNPs and contigs covary with hybrid index. In general, the number of SNPs with Figure 3.3: Distribution of genomic cline centers for 14,009 SNPs analyzed using bgc. SNPs highlighted in dark red have cline centers shifted towards P. engelmannii ancestry and SNPs highlighted in blue had cline centers shifted towards P. glauca ancestry.  58  allelic excesses favoring Engelmann spruce ancestry was approximately equal to the number with a white spruce excess at the 95% level (2,517 and 2,427, respectively). However, in general, SNPs with Engelmann allelic excess had slightly more extreme cline centers (mean alpha = 0.231 for Engelmann-skewed SNPs, mean alpha = -0.206 for white-skewed SNPs). Of the top 100 most skewed contigs, 61 contained predominantly or entirely Engelmann-skewed SNPs, though the average cline center value for the top 100 contigs was similar between Engelmann-skewed and white-skewed (mean alpha = 0.421 for Engelmann-skewed SNPs, mean alpha = -0.433 for white-skewed SNPs).  3.3.4 Genome-environment association   The SNPs selected for genomic cline analysis tended to be enriched in the top percentiles of Bayenv2 outliers, particularly in climatic variables with high climatic similarity indices (Figure 3.4). This is unsurprising, given that the subset of SNPs was selected based on high divergence of allele frequencies among provenances. Of the 14,009 SNPs were selected for bgc analysis out of the 899,596 SNPs analyzed with Bayenv2, we would naively expect ~140 SNPs to lie within the 99th percentile of associations for any given climatic variable. However, the number of bgc SNPs within the top percentile ranged from 170 (eFFP) to 1130 (PAS). For both index-neutral and parent-skewed SNPs, climatic variables with extreme climatic indices showed the highest enrichment (Figure 3.4). The trend was typically more pronounced for parent-skewed SNPs than index-neutral SNPs, particularly for SNPs with very high climate similarity indices i.e., for climatic variables where the hybrid zone is very similar to the allopatric range of Engelmann spruce. The lack of enrichment for climatic variables with moderate climatic similarity indices suggests that these variables either vary linearly with hybrid index (and therefore the genotype-59  environment associations are largely neutralized by population structure), or that any adaptation occurring along that axis does not differentiate along species lines.  When separating parent-skewed SNPs into white- or Engelmann-skewed, we see that enrichment in the top percentiles is strongly correlated with climate similarity index, with extreme climate similarity index values associated with increased enrichment of parent-skewed SNPs (Figure 3.4; Table A.1). What is of particular interest is the enrichment of white- and Engelmann-skewed SNPs relative to one another in the climatic variables with extreme climatic similarity indices. There is no reason to expect that, in the absence of selection, one group or the other should be more-strongly correlated with any particular environment. Both sets of SNPs have allele frequency distributions different than expected under neutral population structure and occur in very similar amounts as shown in the results above.         Figure 3.4: Enrichment of hybrid index-neutral (gray) and parent-species-skewed (black) SNPs in the 99th percentile of Bayenv2 climatic associations as measured by odds ratio against all other SNPs in the dataset. Each point represents a climatic variable, and climatic variables are arranged along the X axis according to their climatic similarity index, where -1 represents climatic variables where the hybrid zone strongly resembles the allopatric range of white spruce, and +1 represents climatic variables where the hybrid zone strongly resembles the allopatric range of Engelmann spruce. Error bars show 95% confidence intervals. 60   Figure 3.5: Bayenv2 outlier enrichment for SNPs identified as Engelmann-skewed (red) or white-skewed (blue) cline center outliers in bgc, compared to SNPs that covary linearly with hybrid index. Each plot corresponds to genotype-environment associations with a single climate variable, and each point within that plot denotes enrichment at a given association percentile (as determined by ranking the Bayes values of all 899,596 SNPs used in the Bayenv2 analysis). Values above 1 indicate that parent-skewed SNPs are overrepresented compared to index-neutral SNPs, while values below 1 indicate that parent-skewed SNPs are underrepresented. Error bars denote 95% confidence intervals. a-c: The climatic variables with the most positive climatic similarity indices; d-f: The climatic variables with climatic similarity indices closest to 0; g-i: The climatic variables with the most negative climatic similarity indices. 61  At any given SNP, an overabundance of one parental allele or the other should produce the same environmental correlation e.g., a white-skewed SNP with a cline center of -0.3 should not be expected to have a stronger association to a given climatic variable than an Engelmann-skewed SNP with a cline center of +0.3. However, the climatic variables differ heavily in the proportion of white- and Engelmann-skewed SNPs associated with each (Figure 3.5), and the difference is strongly correlated to climatic similarity index (overall R2=0.85, Figure 3.6). This difference tends to increase with increasing stringency in association level (Figure 3.5; Figure Figure 3.6: Proportion of SNPs identified as genomic cline outliers biased towards Engelmann spruce in the top percentiles of Bayenv2 genotype-environment associations for 22 provenance climate variables. Each point represents a combination of environmental variable and percentile threshold, and environmental variables are arranged along the X axis according to their climatic similarity index, where -1 represents climatic variables where the hybrid zone strongly resembles the allopatric range of white spruce, and +1 represents climatic variables where the hybrid zone strongly resembles the allopatric range of Engelmann spruce.   62  3.6), suggesting that environmental variables with extreme climatic similarity indices have much stronger associations with parental alleles aligning with that hybrid index. For example, in relation to mean temperature of the coldest month (MCMT), a climatic variable for which the hybrid zone is very similar to the allopatric range of Engelmann spruce (climatic similarity index = 0.94), Engelmann-skewed SNPs are much more strongly associated (Engelmann-skewed SNPs represent 93% of the bgc SNPs identified in the top 1% of Bayenv2 associations; odds ratio over index-neutral SNPs = 5.02). Additionally, the trend is much stronger for climatic variables with a high climatic similarity index than for those with a low climatic similarity index (Compare Figure 3.5a-c with Figure 3.5g-i). The same pattern is visible in the structure-corrected SNP-environment correlations themselves. Parent-skewed SNPs had overall stronger environmental correlations than index-neutral SNPs, and the trend is stronger for climatic variables with high climatic similarity indices (Figure 3.7). Again, this is unsurprising given that index-neutral SNPs should largely show correlations that follow hybrid index and would therefore be downscaled in the structure correction. Also, all correlations were typically stronger in climatic variables with extreme climate similarity indices, as these variables differ more strongly from the axis of population structure. We should not expect that SNPs with a skew toward one parent or the other should show stronger strengths of correlation, but this is what we see across most climatic variables, and again this pattern is well-explained by climate similarity index (R2=0.87, Figure 3.8). White-skewed SNPs are more strongly associated with low climate similarity index variables, and Engelmann-skewed SNPs are more strongly associated with high climate similarity index variables.   63    Figure 3.7: Mean genotype-environment correlations of SNPs identified as genomic cline outliers for Engelmann or white spruce, and those with no bias towards one or the other across 19 climate variables. Engelmann-biased outliers are shown in dark red, white-biased outliers shown in blue, and index-neutral SNPs shown in grey. Climate variables are ordered along the X axis by their climatic similarity index, where -1 represents climatic variables where the hybrid zone strongly resembles the allopatric range of white spruce, and +1 represents climatic variables where the hybrid zone strongly resembles the allopatric range of Engelmann spruce.  Figure 3.8: Difference in correlation strength between SNPs identified as genomic cline outliers for Engelmann and white spruce across 19 climate variables in Bayenv2. Climate variables are ordered along the X axis by their climatic similarity index, where -1 represents climatic variables where the hybrid zone strongly resembles the allopatric range of white spruce, and +1 represents climatic variables where the hybrid zone strongly resembles the allopatric range of Engelmann spruce.  64  3.4 Discussion As demonstrated by the estimates of genomic hybrid index, the interior spruce hybrid zone is complex and comprises hybrids of all form, from nearly pure parent species to truly intermediate hybrids and everything between. This agrees with previous genomic assessments of this hybrid zone (De La Torre et al. 2015; Hamilton et al. 2015; MacLachlan et al. 2018). However, contrary to De La Torre et al. (2014b; 2015), we did not identify an excess of Engelmann ancestry throughout the hybrid zone, but rather an excess of white spruce ancestry. This is likely due to our broader sampling that incorporates hybrids from areas more proximal to allopatric white spruce and covers more of BC\u2019s central interior as well as central and northern Alberta. Where our sampling overlaps with the sampling of De La Torre et al., there are indeed more Engelmann-like hybrids. Additionally, the identification of tri-species hybrids in the northwestern and southwestern portions of the study zone is in agreement with a prior effort in classifying the genomes of hybrids across BC (Hamilton et al. 2015), as well as previous cytoplasmic work (Sutton et al. 1994) which suggested that interior spruce in those regions contains Sitka spruce ancestry. Overall, these results suggest a hybrid zone with very weak reproductive barriers between species that has existed long enough for many generations of crossing and backcrossing to have occurred throughout.  Although there is clearly a visible relationship between hybrid index and proximity to allopatric parental provenances, many of the provenances distant from both parental species ranges tend to show a bias towards white spruce ancestry. As multiple studies have suggested that hybrid index itself may represent an adaptive gradient in this hybrid zone (De La Torre et al. 2014c; MacLachlan et al. 2018), this suggests that perhaps there is selection favoring white spruce ancestry across much of the central plateau of BC and southeastern Rockies. 65  The analysis of climatic variables presented here suggests that, where hybrids are present, many climatic variables that have been implicated in selection in interior spruce and other conifers (Ukrainetz et al. 2011; De La Torre et al. 2014c; O\u2019Neill et al. 2014; Liepe et al. 2016; Yeaman et al. 2016; MacLachlan et al. 2018) are not intermediate between the allopatric ranges of white and Engelmann spruce, but rather more similar to the allopatric range of one parent species or the other. Winter temperatures have been associated with adaptive clines in interior spruce and other northern conifers (Alberto et al. 2013; MacLachlan et al. 2018), and the hybrid zone is very similar to the range of Engelmann spruce in this regard. Precipitation, especially snowfall, has also been associated with adaptation in interior spruce (e.g., Ukrainetz et al. 2011; De La Torre et al. 2014c), and the hybrid zone is more similar to the range of white spruce in this regard. This pattern can be seen clearly in the PCA (Figure 3.2), showing that white and Engelmann spruce tend to exist in largely distinct climate spaces that intersect at the hybrid zone. Relative to one another, white spruce exists in a primarily dry niche spanning cold to warm temperatures while Engelmann spruce exists in a primarily warm niche spanning dry to wet precipitation regimes. The hybrid zone largely exists where these two niches intersect: in warm and dry regions of the climate space i.e., those that share characteristics of the extreme ends of both species\u2019 climatic niches.  The results of the genomic cline analysis suggest that thousands of genes have an excess of one parental allele across the hybrid zone, approximately evenly split between the parent species. This suggests that adaptation across the hybrid zone is not being driven by one parent species at the expense of the other, as may be expected under a tension zone model (Barton and Hewitt 1985), and that adaptation is highly polygenic. While the distribution of genomic cline 66  centers appears to be fairly normally distributed (Figure 3.3) with many index-neutral SNPs, it bears mentioning that many index-neutral SNPs are likely involved in climate adaptation. The 14,009 SNPs analyzed here are all highly diverged between white and Engelmann spruce, two closely-related species with evidently few to no reproductive barriers (Wright 1955). Given the presumably large effective population sizes in both of these species, it seems unlikely that many of these SNPs would arrive at their present frequencies through genetic drift. Both species have, however, experienced substantial range expansions over the last 20,000 years (Ritchie and Macdonald 1986; Hamrick et al. 1994; de Lafontaine et al. 2010) which may have caused some of these SNPs to arrive at their present diverged frequencies through \u201callele surfing\u201d as the parent species\u2019 ranges expanded following deglaciation and prior to secondary contact (Goodsman et al. 2014). Unfortunately, it is not possible, with the methods used here, to determine whether SNPs are index-neutral due to gene flow or selection that acts parallel to the gradient of population structure. It is for this reason that these SNPs are not discussed at length.  The strong correlation between allelic excess and climate similarity index (as indicated by both strength of correlation and enrichment of outliers in genotype-environment associations) demonstrates that parental alleles are associated with climates similar to their allopatric ranges in the hybrid zone. Given that any given point within the hybrid zone is a mosaic of white, Engelmann, and intermediate climates, it appears that selection is acting to differentially favor sets of parental alleles that are tailored to that particular environment, and that local adaptation is occurring at a fine scale within this hybrid zone. The existence of large numbers of genomic cline outliers suggests that, regardless of whether or not hybrid index is itself an adaptive gradient, that many genes are under directional selection from differing evolutionary origins, strongly suggesting the presence of local adaptation in this species complex. 67  3.5 Conclusions Overall, these results suggest several important lines of evidence for high levels of local adaptation to climate in the interior spruce hybrid zone. First, the climatic landscape across the vast geographic area covered by this hybrid zone is complex. Important climatic gradients are not strictly intermediate between the allopatric ranges of white and Engelmann spruce, allowing for directional selection to occur in orthogonal directions; i.e., white spruce alleles may be favored in the dry climate of the hybrid zone, while Engelmann spruce alleles may be favored in the relatively warm climate experienced in much of the hybrid zone. Given that adaptation to broad climatic gradients is typically highly polygenic (Barghi et al. 2020; Isabel et al. 2020), the underlying changes may be seen in genomic hybrid indices. For example, if a hybrid population is located in a geographic area that is broadly climatically similar to white spruce, then one would expect a majority of white spruce alleles to be selected across the genome, leading to an overall white spruce-biased hybrid index in that area. This phenomenon has been extensively observed and discussed in relation to this hybrid zone elsewhere (De La Torre et al. 2014c; MacLachlan et al. 2018). However, when important climatic variables in the geographic area described above are more similar to the range of allopatric Engelmann spruce, Engelmann spruce alleles appear to be favored, generating local adaptation beyond what is described by genomic hybrid index (i.e., SNPs with high genomic cline centers). A high degree of local adaptation in hybrid populations, as observed here, suggests that this hybrid zone is maintained by hybrid superiority (Moore 1977), wherein hybrids exhibit a high degree of fitness in environments intermediate to their parent species, corroborating the findings of De La Torre et al. (2014c). However, the high degree of genomic variability and the introgression of specific parental alleles in specific environments is a novel finding. 68  Chapter 4: Research - Climatic modeling of genomic ancestry in a spruce hybrid zone predicts past gene flow dynamics and future adaptive lag.  4.1 Introduction In chapter 3, I demonstrated that local adaptation of hybrid spruce populations to climate is complex, with hybrids favoring adaptive white or Engelmann spruce alleles depending on specific local temperature and precipitation conditions. This suggests that local adaptation to climate in the interior spruce hybrid complex is tightly linked to hybridization. Using this strong relationship between hybridization and climate, we can explore past evolutionary dynamics and investigate future implications of climate change for these populations. 4.1.1 Data-driven climate prediction The spruce provenances used in these studies cover a massive geographic area with extreme variation in topography. To assess climate adaptation within these provenances, one must be able to reliably model the historical climate at any arbitrary point within this area and over a long timescale. Given the sparse and uneven distribution of weather stations with varying record lengths, it is unreasonable to assume that the nearest weather station to each provenance would provide an accurate and comparable estimate of the climate within these locations, necessitating interpolation from a broader dataset. Although several interpolation methods have been developed to model climates in Canada, many perform poorly in mountainous terrain (Milewska et al. 2005). PRISM (Parameter-elevation Regressions on Independent Slopes Model) was designed specifically to interpolate climates within physiographically complex landscapes such as western North America (Daly et al. 1994). Using a combination of elevational 69  adjustment, coastal proximity, topographic position, boundary layer conditions, and manual expert knowledge adjustments, PRISM outperforms other interpolations for this region (Milewska et al. 2005). Although this interpolation provides estimates throughout the study area and at a reasonably-high resolution (800 m2), it is still insufficient to capture the variation in mountainous terrain. This resolution can be further downscaled by interpolating around a high-resolution digital elevation model, and accounting for variability in elevational changes across different environments (Wang et al. 2006). This process has been used to generate historical climate predictions for the entirety of North America from 1901 until the present (Wang et al. 2016).  4.1.2 Model-driven climate prediction While data-driven methods are ideal for predicting historical climates within the timespan of automated weather measurement, they are unable to predict the climates of the distant past or into the future, which instead require holistic modeling methods. The last two decades have seen a massive effort in developing global and regional climate models to predict the effects of climate change over the coming decades (Eyring et al. 2016; IPCC 2021). Although a variety of models exist to predict climates at differing spatial scales, all rely upon one or more foundational atmosphere-ocean general circulation models (GCMs). These models apply a set of inputs to interconnected physical, chemical, and biological process sub-models to predict climate within a three-dimensional grid comprising the land, atmosphere, and oceans across large geographic regions or the entirety of the planet. Due to the extreme extent and complexity of these models, they are necessarily computed at a coarse scale. Although advances in computing have allowed recent models to provide predictions at a comparatively high resolution (IPCC 2021), the 50km2 70  or lower resolution provided by these models is still insufficient to adequately explain climatic variation at ecologically-relevant scales (Kriticos and Leriche 2010; Farashi and Alizadeh-Noughani 2018). Fortunately, the same downscaling approaches described above apply equally well to the output data from GCMs and can be used to provide predictions at arbitrary locations (Wang et al. 2016). As a large number of GCMs are available for any given region, using a variety of differing inputs and sub-models (IPCC 2021), utilizing the outputs of many models is an approach often used to provide an ensemble mean expectation of future climates (e.g., Mahony et al. 2022). The number of models used and the weighting applied to each may vary depending on the region under study or the purpose of model predictions.   4.1.3 Paleoclimatic modeling While recent advances in climate modelling have been primarily developed for use in future climate prediction, the value has not been lost on those studying historical climates, as coupled GCMs should be able to predict paleoclimates equally as well as future climates, given proper constraints (Liu et al. 2009). These models offer the benefit of continuous process modeling, allowing analysis at any hypothetical timescale (Fordham et al. 2017). While these models have been used to accurately model paleoclimates at the global scale, they may underpredict changes at the regional level (Braconnot et al. 2012), and regional climates may change dramatically over short timescales (Steffensen et al. 2008), necessitating accurate input of known regional events over the timescales being studied.  Conversely, paleobotanical evidence can be used to infer past climates from present climatic tolerances of identified plant species (Wolfe 1978), particularly palynological data 71  (pollen assemblages) due to their presence in datable marine and lacustrine sediments. While these data can provide extremely fine-scale paleoclimatic estimates near sampling sites and may refine regional impacts of large-scale climatic changes (e.g., Goring 2012), they may not accurately reflect relative species abundance or richness (Goring et al. 2013) and rarely provide species-level taxonomic resolution (Birks and Birks 2000). Likewise, palynological data have limited predictive power for large wind-pollinated species, as these may deposit pollen far from the site of production (Ritchie and Lichti-Federovich 1967). Fortunately, widespread palynological data exists for British Columbia (Goring 2012) as well as a detailed understanding of post-glacial climate dynamics in the area (Clark et al. 2009; Lorenz et al. 2016). As such, many disparate data sources can be used to estimate and calibrate paleoclimatic models across the study area.  4.1.4 Species climatic niche models All species occupy a specific ecological niche within their respective ecosystems. For many plant species, that ecological niche is defined in part or wholly by climatic drivers (Thomas 2010). This is especially true of wind-pollinated species, where establishment success and long-term survival are defined largely by their ability to withstand or thrive in particular climates (e.g., McKenzie et al. 2003). In this case, the ecological niche of a species may be closely approximated by its climatic niche.  The climatic niche of a species can be further classified into a fundamental or realized niche (Colwell and Rangel 2009). The fundamental climatic niche of a species refers to the absolute climatic limitations of a species beyond which it cannot survive and reproduce (Hutchinson 1978). These are typically driven by physiological constraints e.g., heat, cold or 72  drought tolerances. The realized climatic niche refers to the climate space occupied by a species across its geographic range and is narrower or equal to its fundamental niche. There are myriad limitations that may prevent a species from occupying its fundamental climatic niche, including interspecific competition, pest and pathogen distributions, edaphic conditions, or topographic barriers (Colwell and Rangel 2009). While the fundamental climatic niche of a species must typically be experimentally determined, the realized niche can be inferred through ecological records i.e., presence and absence data or simply presence data (Jim\u00e9nez et al. 2019). Although it would be ideal to model a species\u2019 fundamental climatic niche for understanding past and future distributions, it is nearly impossible to determine the complete niche in practice due to the large and unknown number of climatic factors that may define such a niche. As the realized niche is always smaller than the fundamental niche, past and future projections based on this niche can generally be viewed as a conservative estimate of climatic suitability.  Species niche models are typically developed using a probabilistic model of species presence, which is then classified by a threshold criterion for inclusion or absence in the niche (Franklin 2010). This threshold can be determined by utilizing validation plots containing species presence and absence data beyond that which was used to generate the model. The two parameters most commonly used to assess the model at different thresholds are sensitivity (i.e., the proportion of plots where a species was correctly predicted to be present), and specificity (i.e., the proportion of plots where a species was correctly predicted to be absent). Several different metrics have been developed to balance sensitivity and specificity, and more than one may be used to select the most-appropriate model (Mouton et al. 2010).    73  4.1.5 Mitotypes and species migration Mitochondria and chloroplasts are specialized organelles that provide energy to plant cells through the production and consumption of various carbohydrates. These organelles contain their own haploid genomes, produced as a relic of endosymbiosis (Mar\u00e9chal 2018). Organellar genomes tend to show consistent inheritance patterns within taxa, with some variation among taxa. For most angiosperms, both organelles are inherited maternally (Reboud and Zeyl 1994). However, in most conifers, the mitochondrion is inherited maternally while the chloroplast is inherited paternally (Petit et al. 2005). This is the case in all studied members of the Pinaceae, including Picea (Neale and Sederoff 1989; Sutton et al. 1991). This biparental pattern of inheritance can be used to assess differential rates of gene flow from pollen and seed (e.g., Latta et al. 1998; Richardson et al. 2002; Jaramillo-Correa et al. 2006; Zou et al. 2012). Plant mitochondria tend to evolve at much slower rates than the nuclear genome in terms of single base-pair mutations (Wolfe et al. 1987; Laroche et al. 1995; Mower et al. 2007), which, combined with a reduced effective population size due to uniparental inheritance, leads to generally low genetic diversity within species, and countably few haplotypes within populations (Birky et al. 1983). As migration in established forests is overwhelmingly driven by pollen flow (Ellstrand 1992; Liepelt et al. 2002; Mitton and Williams 2006), maternal mitochondrial haplotypes tend to reflect the establishment conditions of a forest (Le Corre et al. 1997; Pyh\u00e4j\u00e4rvi et al. 2008; Hamilton and Aitken 2013; Semerikov et al. 2019). For species that have experienced substantial range shifts over time, as is common in areas covered by ice during the last glacial maximum, these mitochondrial haplotypes can be used to infer the most recent establishment histories in these species (e.g., Comes and Kadereit 1998; Jaramillo-Correa et al. 2004; Godbout et al. 2005; Semerikov et al. 2019).  74  4.1.6 Adaptation lag Due to the discrete nature of sexual reproduction, adaptation via changes in allele frequencies occur over timescales that may not align with changes in selective pressures (Kopp and Matuszewski 2014). This is particularly true of long-lived species (Yamamichi et al. 2019), which can result in a substantial mismatch between adaptation and selective pressure, referred to as adaptation lag (Matyas 1994). As the global climate changes, many tree species are experiencing substantial adaptive lag, particularly at species range margins (Aitken et al. 2008; Solarik et al. 2018; Fr\u00e9javille et al. 2020). Increasing temperatures may push out populations already at the limit of their thermal tolerances (Aitken and Bemmels 2016), while changes in precipitation regime may cause drought-related mortality across large areas of a species range (Pile et al. 2019) or favor populations or species better-adapted to wetter climates (Dunton et al. 2001). Changes in both temperature and precipitation can have multiplicative effects beyond what either would produce in isolation (Mccain and Colwell 2011). The tree populations in the present study represent mature forests that established under potentially different climatic conditions than they face at present, and the genotyped seedlings were grown from seed collected between 1959 and 2007, with a mean of 1987. For this reason, the 1961-1990 climatic normal period is considered the climate to which these provenances are putatively adapted. If these populations are indeed locally adapted to their respective climates, then we can assess the degree to which these areas have already experienced adaptation lag by comparing their historical climate to a more recent time period such as the 1991-2020 normal period.  75  4.1.7 Objectives Understanding the present state of adaptation in the interior spruce species complex requires an understanding of past conditions that led to its establishment. Previous research suggests that climatic conditions were suitable for white and Engelmann spruce to have shared refugia during the last glacial maximum, and may have hybridized during that time (De La Torre et al. 2014b). This could greatly extend the length of time that these species have been in secondary contact, and therefore the length of time that selective forces have had to shape the hybrid genome. If white and Engelmann spruce only came into secondary contact after the last glacial maximum, then stronger selective pressures may be required to explain observed patterns of local adaptation in relatively few generations. Additionally, if multiple glacial refugia contributed to the current genetic landscape in the species complex, they may have harbored different adaptations that could contribute to standing genetic variation in the species complex today. In this chapter, we will use population variation in mitochondrial haplotypes, in conjunction with paleoclimatic species climatic niche models, to infer the likely pathway(s) used by white and Engelmann spruce to colonize western Canada during deglaciation. Many studies have used species climatic niche models to predict the future distributions of species under climate change scenarios (Peterson et al. 2018). However, the binary nature of niche envelopes limits their ability to describe changes expected to occur within species\u2019 ranges or to explain the magnitude of impacts, compared to more granular approaches like Gradient Forests (Ellis et al. 2012; Fitzpatrick and Keller 2015; Fitzpatrick et al. 2021). As hybrid index is strongly correlated with climate in western Canada (Chapter 3 of this dissertation), we will use climatic variables to estimate hybrid index throughout western Canada, and project these estimates into future climate scenarios to predict the genomic composition of hybrids that will be ecologically suitable moving forward. This will allow us to not only infer areas of predicted loss or gain of the species\u2019 niches over time, but also to estimate the degree of maladaptation that may occur within portions of the species\u2019 niches that are expected to persist into the future. 76  4.2 Methods  4.2.1 Genotypic data While sequence capture is an attractive method for generating high quality sequence data for species with large genomes, it is still cost-prohibitive for large sampling schemes. The \u201cmany provenances, few samples\u201d approach works well for assessing range-wide patterns but gives little confidence in population-level representation. To achieve higher within-population sampling, a 55k Affymetrix SNP array was designed based on initial analyses using the sequence capture data. The initial array was designed to capture a mixture of SNPs with adaptive signatures (based on the top results of preliminary genotype-environment associations (GEA; Bayenv2) and genome wide association (GWA; TASSEL) association analyses, totaling about 10k SNPs), candidate genes identified in other studies (about 2.5k SNPs), putatively neutral SNPs identified in non-coding regions targeted by the seqcap design (~5k SNPs), with the remainder filled by SNPs with high quality metrics from the seqcap dataset. However, several issues arose during the initial genotyping attempts with this array, and approximately half of the SNPs failed to genotype successfully. It was later discovered that many of the putatively adaptive SNPs had been misidentified due to a sorting error in a large SNP database, and many of the high-quality seqcap SNPs were of lower quality than expected. However, 20,071 SNPs were eventually genotyped successfully and used for subsequent analyses presented here. The success or failure of SNPs on this array was used to filter SNPs for the final seqcap dataset (described above). A total of 2,102 samples were genotyped using this array for subsequent analysis.    77  4.2.2 Genomic hybrid index Genomic hybrid index was calculated as described in Chapter 2, across the largest possible number of individuals, to ensure that provenance-level estimates of genomic ancestry were as accurate as possible. To this end, we used both the SNP array dataset and seqcap datasets, representing a total of 2,681 individuals. As in the previous chapter, genomic hybrid index is presented on a scale of 0 to 1, with 0 representing pure white spruce and 1 representing pure Engelmann spruce.  The seqcap dataset was trimmed down to the 20,071 SNPs successfully genotyped on the array to ensure comparability between datasets. To ensure that the limited number of SNPs would not have a large influence on the accuracy of hybrid index estimation, hybrid index was compared in seqcap individuals using the full and reduced set of SNPs. This dataset comprises an average of 10 individuals per provenance (range 1-19), allowing for much more accurate estimates of provenance mean hybrid index as well as variability within provenances than the sample sizes from in Chapter 3. These estimates of genomic hybrid index were used to train the RandomForests model used to predict hybrid index based on climate, described below. Although ancestry from Sitka spruce is present in some provenances (see Chapter 3 of this dissertation), only white and Engelmann spruce ancestry were considered in this study for simplicity. For the relatively few individuals with Sitka ancestry, hybrid index was calculated using the relative proportions of white and Engelmann ancestry in those individuals.   78  4.2.3 Climate data 4.2.3.1 Contemporary and future climatic data For each of the 254 provenances analyzed within this study, ClimateNA version 7.20 (Wang et al. 2016) was used to predict seasonal climatic variables for contemporary and future time periods, specifically mean total precipitation, mean daily minimum temperature, and mean daily maximum temperature. Across the four seasons, this generates twelve climatic variables, which were used in all subsequent analyses. These variables are the primary outputs of global circulation models, allowing comparisons across multiple climate datasets. Climate data were estimated for a historical reference period (1961-1990), as this is the climate to which we expect the parent trees in these provenances have adapted. These climate data were employed to train the models used to predict hybrid index and species climatic niche models.  A recent time period (1991-2020) was used to assess potential contemporary adaptation lag; i.e., what is the potential extent of maladaptation already present within the hybrid zone? To predict future species suitability, a 13 GCM ensemble of CMIP6 models (Mahony et al. 2022) was used to predict future climates under three emissions scenarios (shared socioeconomic pathways; SSPs) provided by the IPCC 6th report (IPCC 2021): SSP1 represents an optimistic future of reduced emissions and more sustainable economic development, reflected in a radiative forcing of 2.6 Watts\/m2; SSP2 represents a pathway in which countries commit to present levels of decarbonization and incorporates a radiative forcing of 4.5 Watts\/m2 (analogous to the RCP4.5 pathway from IPCC5); and SSP3 represents a pessimistic future in which countries withdraw from climate commitments to shore up their own energy security, incorporating a radiative forcing of 7 Watts\/m2. These climate data were used to output predictions across the entirety of western Canada at 1km2 resolution. 79  4.2.3.2 Paleoclimatic data To predict historical climate data, a multi-step process was used. First, paleoclimatic data were obtained from the dataset generated by Lorenz et al. (2016). This dataset comprises the same twelve seasonal climatic variables discussed above, averaged across 1000-year intervals using CCSM3 (Liu et al. 2009) from 21kybp to 1kybp. Lorenz et al. downscaled their predictions from 4\u00b0 to 0.5\u00b0 resolution (corresponding to between approximately 50km x 28km and 50km x 36km within the study area). These input data are subsequently described as the \u201cLorenz\u201d dataset. While these data are already downscaled from their original GCM outputs, the resolution is insufficient for assessing variability within the hybrid zone, as it does not capture local variation in elevation. To further downscale these climate data to the 1km2 resolution provided by ClimateNA, a local elevational lapse rate masking layer was generated. To generate this layer, contemporary data were obtained from the Lorenz dataset across the reference climatic period. These data were downsampled to 1km2 resolution using bilinear interpolation in order to align them to the ClimateNA raster. Then, for each of the 8 temperature variables analyzed here, the difference between the ClimateNA and Lorenz rasters was used to generate a raster of anomalies at each grid cell. Effectively, this applies ClimateNA\u2019s localized elevation lapse rates to the Lorenz dataset for these variables. Simple subtraction is appropriate here as temperatures tend to change at a relatively linear rate with elevation (e.g., Minder et al. 2010). For the four precipitation variables, the mask was generated by dividing the ClimateNA predictions by the Lorenz predictions. This produces relative changes that avoid negative precipitation values and accounts for the logarithmic distribution of precipitation across landscapes (R\u00e4ty et al. 2014). 80  These masks were added to or multiplied by (for temperature and precipitation, respectively) each of the paleoclimatic rasters to generate 1km resolution rasters for every time period.  4.2.4 Climatic prediction of hybrid index  To predict hybrid index across western Canada using only climatic data, a RandomForests (Breiman 2001) machine learning model comprising 1000 decision trees was trained using the available genomic hybrid indices and predicted climatic variables from each provenance, using the cforest implementation in R. This model will subsequently be referred to as the climatic hybrid index model.  Provenance mean values of hybrid index as calculated above were used to provide the most accurate estimate of hybridity given a particular climate, and to ensure that the out-of-bag sampling method used by RandomForests was assessing model performance on unrelated samples. All seasonal climatic variables described above were used to train and assess the model, with precipitation log-transformed to normalize data and to ensure that predictions to other climate surfaces operated on proportional changes in precipitation, rather than absolute changes. The out-of-bag or \u201cleave-one-out\u201d approach employed by RandomForests has received some criticism for overfitting datasets (Strobl et al. 2008); therefore, the algorithm was tested with 10,000 iterations of training and evaluating subsets of our data. These subsets were generated by subsampling 90% of each of five hybrid index quantiles, and then downsampling to the least-represented quantile (in practice, hybrid indices between 0.6 and 0.8). This generated a subsample of 90 provenances (18 of each quantile) in each iteration. The model was trained on this subset i.e., \u201ctraining provenances\u201d and evaluated on the remaining 164 provenances, i.e., \u201cevaluation provenances\u201d. Model performance was assessed by calculating R2 and RMSE using 81  two evaluation datasets. Preliminary results revealed that the models were generally better at predicting extreme rather than intermediate values. To address this, the final model was assessed against both the full set of evaluation provenances, as well as the subset of evaluation provenances with a hybrid index between 0.1 and 0.9. This ensured that the model was accurately predicting intermediate hybrid indices, rather than just accurately assigning parental species provenances. The predicted hybrid index for each of the evaluation provenances was retained and averaged across all runs for which it was not included in the test subset (n = 466 \u2013 8,431 per provenance) to provide the final predicted hybrid index for that provenance. This was compared to the predicted hybrid index from the default out-of-bag estimation. With each model iteration, hybrid index was predicted across the rasterized climatic surfaces of western Canada (past, present, and future), with the average prediction from the 50 runs used for each raster cell.  Variable importance as calculated by decrease in model accuracy following permutation was used to subjectively identify climatic drivers of hybrid index, relative to those identified in the previous chapter. As variable importance tends to be overestimated in correlated predictor variables (Strobl et al. 2008), and climatic variables are often highly intercorrelated, variable importance was calculated using a conditional approach that tends to deflate the importance of correlated variables with negligible impacts on uncorrelated variables (Strobl et al. 2008). Calculating conditional variable importance is very computationally intensive and, as such, variable importance was estimated using the mean importance from 50 model iterations. The importance ratio i.e., the ratio between the variable importance of the most important variable and importance of a given other variable, was used to quantify relative importance between variables, as this is often more interpretable than the empirical values (Archer and Kimes 2008). 82  This ratio produces values between 0 and 1, where larger values indicate variables nearer in importance to the most important. These variable importance ratios were averaged across each of the 50 iterations, providing an estimate of mean variable importance.  4.2.5 Species climatic niche modelling  As the RandomForests model described above will predict a hybrid index for any set of climatic variables, the results were bounded by species climatic niche models to only predict hybrid indices for areas where the species could be expected to occur. To generate these species climatic niche models, presence and absence data from two extensive ecological plot datasets were used, which were developed separately for BC and Alberta but built upon shared methodologies (Alberta 2016). The BC dataset comprises 7,601 permanent sample plots with recorded presence and abundance of all plant species (MacKenzie and Meidinger 2018), and the Alberta dataset comprises 14,447 plots with similar information (Alberta 2016), for a total of 22,047 plots (Figure 4.1). Both datasets were designed to extensively and thoroughly sample all ecological gradients within their respective provinces and provide a comprehensive catalogue of plant distributions, particularly for widely distributed and conspicuous species. In both datasets, tree species are separately categorized as white, Engelmann, or hybrid spruce. For each plot location, seasonal climatic variables were predicted using ClimateNA across the reference climatic period as described previously. 83  These data were used to train probability of presence models for white, Engelmann, and hybrid spruce, with climatic variable importance for each model calculated as described in Section 4.2.4. As species climatic niche models must ultimately make a binary decision on presence or absence for each location, the true skill statistic (TSS; (Allouche et al. 2006)) was used to determine classification thresholds for each model independently. TSS balances sensitivity and specificity to predict a species\u2019 realized climatic niche. As sensitivity and specificity oppose one another across the range of threshold values, TSS produces a curve across prediction thresholds with a maximum value somewhere in the middle of predicted probabilities, where trade-offs between sensitivity and specificity are minimized (Figure 4.2). The interpretation of TSS is subjective, but a reasonable interpretation is that values < 0.4 are poor and values > 0.8 are very good, with intermediate values being acceptable models (Allouche et al. 2006).  Figure 4.1: Geographic distribution of ecological sample plots used to generate species niche models. 84  As with the hybrid index prediction, the species climatic niches were iteratively trained and evaluated on subsamples of the dataset. In each iteration, 80% of plots were subsampled, and then downsampled to produce a balanced dataset with an equal number of presences and absences. In practice, this produced datasets of 5,232 plots for Engelmann spruce, 7,424 plots for white spruce, and 3,878 plots for hybrid spruce. Probability of presence was predicted for another balanced sample of excluded plots (in practice this included the 20% of presence plots unused by the model and a random equally sized sample of absence plots), and used to calculate sensitivity and specificity for the model. Due to the larger datasets involved compared to the hybrid index model, each forest contained 300 trees, and 50 iterations were performed per species. Predicted values for each plot were averaged across the runs where it was not included in the training dataset (n = 2 \u2013 50 for each plot), and used to estimate specificity, sensitivity, and TSS. Using the mean model predictions for each sample plot, sensitivity, specificity, and TSS were calculated using threshold presence values between 0.1 and 0.9. The threshold with the highest TSS value was used to classify presence in all climate rasters for that species. Probability of presence was predicted for all rasters in each iteration, with the mean value for each raster cell used to generate final species climatic niche models. To further bound estimates of hybrid index and species climatic niches across the paleoclimatic period under study, spatial extents of ice sheets estimated by Dyke (2004) were overlaid on distribution maps to further limit the realistic ranges of species at any given historical time point.  85  4.2.6 Mitotype assignment Over the course of bioinformatic analysis of the sequence capture dataset, 54 markers were identified with conspicuously low heterozygosity. Subsequent alignment to the spruce organellar genomes (Jackman et al. 2016) revealed them to be haploid mitochondrial markers. Of these markers, 15 were also present on the SNP array. Markers were manually classified into mitochondrial haplotypes (mitotypes) and assigned a parental origin based on their abundance in parental provenances and correlation with other markers.   4.3 Results  4.3.1 Genomic hybrid index Predicting genomic hybrid index with the reduced set of SNPs on the array (20,071) produced very similar results to those predicted with the full set of SNPs (899,596) for the individuals genotyped using sequence capture (R2=0.999, 0.998, and 0.964 for white, Engelmann, and Sitka spruce ancestry, respectively). Within-provenance variation in admixture was generally low, with standard deviation of white spruce ancestry in provenances with n \u2265 10 ranging from 0.4% to 16.2% (mean 4.1%).  4.3.2 Species climatic niche modelling Species distribution models were moderately accurate as determined by TSS. Maximum TSS values for white, hybrid, and Engelmann spruce models were 0.69, 0.73, and 0.77, respectively (Figure 4.2). In all cases, models were more sensitive than specific at the point where TSS was maximized, with the largest difference occurring for the white spruce model (Figure 4.2a). This indicates that the niches produced by these models will be wider than 86  observed in reality and will generally overestimate species ranges. However, geographic distributions match well with what others have calculated for present distributions and climatic niches (Hamann and Wang 2006; De La Torre et al. 2014b) and with genomic classifications of provenances. Given the morphological similarity between white and Engelmann spruce and their hybrids (Garman 1957; Daubenmire 1974), it is likely that many of the species assignments in the ecological plot data used to generate these models were made based on ecological assumptions, rather than morphological distinction. This could have biased the resulting niche models, particularly the niche for hybrid spruce. However, the general geographic agreement between the climatic niche models and genomic estimates of hybrid index (compare Figure 2.1 and Figure 4.5a) provides confidence that the ecological plot classifications are, overall, representative of the true distributions of the parental species and hybrids. Variable importance differed substantially between species models. For white and hybrid spruce, precipitation variables were consistently important (Figure 4.3b,d). For white spruce, the low VIR values (maximum 0.68 for winter precipitation) indicate that there was not a single variable that was most important across models, whereas spring precipitation came across as most important for hybrid spruce (mean VIR = 0.87). For Engelmann spruce, winter precipitation was consistently ranked as the most important (mean VIR = 0.91; Figure 4.3c), but maximum temperature variables were the next-most important, in contrast to all other models.  87    Paleoclimatic predictions suggest that, 15kybp, unglaciated regions of western Canada could largely have been inhabited by white spruce, if the species had been able to migrate that far, with Engelmann and hybrid spruce habitat appearing as early as 13kybp. By 12kybp, as ice sheets had largely retreated from western Canada, the available niches for spruce were similar to Figure 4.2: Prediction accuracy for species climatic niche models for white, hybrid, and Engelmann spruce (a-c, respectively).  Figure 4.3: Relative variable importance for climatic prediction of hybrid index (a) and climatic niche models for white, Engelmann, and hybrid spruce (b-d, respectively).   88  today, with the major exception that much of the central interior plateau was only suitable for white spruce. By 10kybp, climatic niches were largely similar to the present-day predictions, with only minor fluctuations in the interim. The niches calculated using the reference climatic period (1961-1990) and the more recent period (1981-2010) show substantial changes over the short time scale (Figure 4.5). The range of white spruce in Alberta is already shown retreating northward, and becoming unsuitable in most of central BC, while northwest BC becomes more suitable for Engelmann spruce and less suitable for hybrid spruce.                  Figure 4.4: Predicted climatic niches for interior spruce from 15kyp to present. Grey areas correspond to areas covered by ice sheets and light blue areas correspond to glacial water bodies.  89            Future projections paint a grim picture for interior spruce in BC, with overall habitat reduction ranging from 38-59% for the 2071-2100 climatic period, relative to the 1961-1990 normal period (Figure 4.6). The reduction in suitable habitat was not equally distributed between species, with Engelmann spruce being least affected and white spruce being the most affected. For example, in the 2071-2100 time period for the SSP2 scenario, Engelmann spruce habitat is reduced by 31%, hybrid spruce by 57%, and white spruce by 91% (Figure 4.6h). These figures only take the absolute area of niches into account. The reductions in habitat are even more pronounced when looking at the existing habitat (in terms of the 1961-1990 reference period) compared to future climatic projections. Again, using the 2071-2100 period for SSP2 as an example, the amount of existing habitat from the 1961-1990 reference period remaining in the climatic niches are 36, 19, and 7% for Engelmann, hybrid, and white spruce, respectively.   Figure 4.5: Species climatic niches calculated for the 1961-1990 reference period (a) compared to the 1991-2020 reference period (b).  90       :  Figure 4.6: Predicted species climatic niches for white, Engelmann, and hybrid spruce under three different emissions scenarios and future time periods. Emissions scenarios represent the shared socioeconomic pathways described in the IPCC6 report, increasing in radiative forcing from SSP1 to SSP3. 91  4.3.3 Climatic prediction of hybrid index The RandomForests model was effective in predicting hybrid index from the limited number of climatic variables. Across all provenances, the model fit very well (R2 = 0.89, RMSE = 0.10; Figure 4.7), and moderately well for provenances with intermediate hybrid indices (R2 = 0.75, RMSE = 0.12). Variable importance for the hybrid index model differed substantially from those predicting species climatic niches, implicating winter minimum temperature and precipitation as the primary drivers of hybrid index, with mean VIR = 0.97 and 0.82, respectively (Figure 4.3). Paleoclimatic predictions of hybrid index suggest that pockets of habitat have been available for hybrids since 15kybp (Figure 4.8), contrary to what is suggested by the distribution models (Figure 4.4). Likewise, the model predicts intermediate hybrid indices throughout much of Alberta during the glacial retreat from 13-11kybp (Figure 4.8c-e). The predicted hybrid indices from 11kybp to present are remarkably stable, with only minor expansions of Engelmann ancestry in southern BC but little else changing.  Figure 4.7: Relationship between P. engelmannii ancestry as estimated through genomic data and predicted by provenance climate in RandomForests. 92  As with the species distribution models, there is a surprising amount of change predicted between the climatic reference period and the \u201cmodern\u201d climate (Figure 4.9). In general, provenances would be expected to have more Engelmann ancestry than at present in several parts of BC. As hybrid index tracks adaptive gradients in this hybrid zone (examples discussed in Chapter 1 of this dissertation), the disconnect between what is observed and what is predicted to be on the landscape given present climates suggests that adaptive lag is already occurring.  While the distribution models suggest large shifts in species suitability over the next 100 years, the hybrid index modeling paints a finer-grained picture. In general, as temperatures warm, Figure 4.8: Predicted hybrid index of suitable habitat for interior spruce from 15kyp to present. Light grey areas correspond to areas covered by ice sheets and blue-green areas correspond to glacial water bodies.  93  increased Engelmann ancestry is expected to become more climatically adaptive throughout much of western Canada, though white spruce is still predicted to remain in what habitat is available in Alberta.       Figure 4.9: Change in predicted genomic ancestry between the 1961-1990 normal period and the 1991-2020 normal period for areas that are within the species climatic niches for both climatic periods. Areas in green show regions that fall within the 1991-2020 period but not the 1961-1990 period. Areas in black show regions that fall within the 1961-1990 period but not the 1991-2020 period.  94       Figure 4.10: Predicted hybrid index of interior spruce under three different emissions scenarios and future time periods. Emissions scenarios represent the shared socioeconomic pathways described in the IPCC6 report, increasing in radiative forcing from SSP1 to SSP3. 95  4.3.4 Mitotype assignment Of the 54 identified mitochondrial markers, 42 were fixed for different alleles in all Engelmann and white spruce parental individuals, and the remainder showed large allele frequency differences between species, allowing alleles to be broadly classified as white- or Engelmann-derived. Eleven markers were polymorphic in some Engelmann spruce and hybrid spruce provenances, generating 18 mitotypes (Figure 4.11a). One marker was polymorphic in a single white spruce provenance, generating a total of 2 white spruce mitotypes. Sitka spruce parental individuals shared 52 alleles with the predominant Engelmann spruce mitotype, with the other two producing a diagnostic mitotype that was present in all Sitka spruce samples.  White spruce showed remarkably little haplotype variation (Figure 4.11a), with 345 individuals sharing an identical haplotype. Only two individuals exhibited a single alternate white spruce mitotypes, both from an isolated glacial relict population (Newsome 1963). Engelmann spruce had much more variation, though there was still a predominant mitotype shared by 146 individuals. The second-most prominent mitotype, shared by 28 individuals, was isolated to the Rocky Mountains. The remainder were present in small numbers (1-8 individuals\/mitotype). In general, the Rocky Mountains show much higher haplotype diversity than the rest of the study area.  Mitotypes were generally geographically contiguous at the species level (Figure 4.12), with Engelmann mitotypes in the south, Sitka spruce mitotypes in the northwest, and white spruce in the remainder. The distribution of Engelmann mitotypes was more diverse, with rare mitotypes scattered throughout the area. For the most part, provenances contained mitotypes of a 96  single species. However, some provenances in the southern Rocky Mountains, Caribou Mountains, and northern plateau contained mitotypes from two species.  All 15 SNPs that were included on the SNP array were polymorphic and produced a large number of mitotypes in the SNP array dataset (Figure 4.11b). Due to the reduced number of mitochondrial SNPs present on the array, some of the rare Engelmann mitotypes could not be reliably distinguished from one another. However, four major mitotypes (n=62-1409 each) and 20 minor mitotypes (n=1-7 each) were identified. Unfortunately, the Sitka spruce mitotype could Figure 4.11: Haplotype network depicting the relationship between mitotypes identified in the sequence capture (a) and SNP array (b) datasets. Mitotypes belonging to white spruce are shown in blue, Engelmann mitotypes are shown in red, orange, and yellow, and Sitka spruce is shown in green. The size of the circles is relative to the number of individuals in the dataset sharing that mitotype. The points along the connection lines indicate the number of variants separating mitotypes.   97  not be distinguished from the second most-common Engelmann mitotype. However, given the large geographic distance between these two mitotypes as found in the sequence capture dataset (Figure 4.12), the shared mitotype was assigned to Sitka in the northwest portion of the study area, and assigned to Engelmann in the southern portions of the study area. The larger sample size of the SNP array genotypes allows higher resolution of within-provenance variation in mitotypes, showing greater diversity than suggested by the small sample sizes of the sequence capture dataset. In particular, several rare white spruce mitotypes were identified that are distributed throughout the study area, and more provenances were identified with mixed mitochondrial ancestry.  Figure 4.12: Geographic distribution of spruce provenances showing the proportion of mitotypes identified in each provenance within the sequence capture dataset, overlaid on the species climatic niche models for 12kybp. Light grey areas correspond to areas covered by ice sheets and blue-green areas correspond to glacial water bodies. Mitotype colors correspond to those found in Figure 4.11a. Broadly, blue represents white spruce mitotypes, green represents Sitka spruce mitotypes, and other colors correspond to Engelmann spruce mitotypes. 98  4.4 Discussion  4.4.1 Characterization of the current hybrid zone The high predictive ability of the Random Forests model for hybrid index suggests that species ancestry is tightly associated with climate, as has been demonstrated elsewhere, albeit in univariate analyses (e.g., De La Torre et al. 2014c; MacLachlan et al. 2018). The importance of winter climates in predicting hybrid index supports the results of De La Torre et al. (2014) and MacLachlan et al. (2018), who both identified strong clines between precipitation as snow and hybrid index. MacLachlan et al. also identified a clinal relationship with mean temperature of the coldest month. Using the climate data for vegetation plots containing white, Engelmann, or hybrid spruce as an example, mean minimum daily winter temperature is tightly correlated with mean coldest month temperature (R2=0.95), and precipitation as snow is strongly correlated with mean winter precipitation (R2=0.91). Interestingly, spring precipitation was also identified as an important variable, albeit less important than the first two. Spring precipitation has obvious biological importance to plants, providing a critical resource at the start of the growing season, and has been implicated in selective gradients elsewhere (e.g., Vizca\u00edno-Palomar et al. 2016; Warwell and Shaw 2017), but not in this species complex (De La Torre et al. 2014c; Liepe et al. 2016; Yeaman et al. 2016; MacLachlan et al. 2018). However, it is correlated with mean annual precipitation (R2=0.82 in the plot dataset described above), which has been associated with adaptive signatures (Liepe et al. 2016; Yeaman et al. 2016; MacLachlan et al. 2018; previous chapter). Interestingly, the studies cited above all used a similar set of climatic variables estimated in ClimateNA, which includes several precipitation variables (described in previous chapter), but these are all associated with either summer, winter, or annual precipitation. The 99  identification of spring precipitation as a potential driver of species range limits in this complex merits further investigation, as well as the inclusion of seasonal variables in future exploratory analyses of climatic association. The differences in variable importance between the hybrid index prediction model (Figure 4.3a) and species climatic niche models (Figure 4.3b-d) is striking, particularly with regard to winter minimum daily temperature, with a relative variable importance of 0.97 for the hybrid index prediction model and ~0.3 for the species climatic niche models. Minimum winter temperatures have been identified as the strongest drivers of genomic variation and clinal variation throughout the interior spruce species complex (De La Torre et al. 2014c; Liepe et al. 2016; Yeaman et al. 2016; MacLachlan et al. 2018). The hybrid index prediction model should explain differences within the geographic range of the species complex, whereas the species climatic niche models will highlight conditions at the limits of the species\u2019 ranges. The divergence identified here suggests that winter temperatures are highly influential within the species range, but do not necessarily define the boundaries of those species\u2019 ranges.  The geographic distribution of hybrid index across the 1961-1990 reference climatic period (Figure 4.8p) paints a complex picture of hybridity in spruce throughout western Canada. Much of the mountainous region in eastern and southeastern BC, once thought to be primarily pure Engelmann spruce (Garman 1957), appear to be largely Engelmann-like hybrids, with pure Engelmann isolated to only the highest peaks. While white spruce was largely thought to inhabit much of northern BC, it appears that pure white spruce should not be expected anywhere outside of true boreal ecosystems. In contrast to patterns identified in other studies (De La Torre et al. 2015; Hamilton et al. 2015), hybrid populations throughout the majority of the hybrid zone appear to have primarily white rather than Engelmann spruce ancestry.  100  Several curious patterns emerge in areas not covered by the genomic datasets (i.e., outside the range of input and validation data). Pure Engelmann spruce is predicted at the western limits of the ecological niche, which is expected based on the species distribution models, as well as previous species distributions generated for Engelmann spruce (Hamann and Wang 2006). This is expected, but does provide a welcome sanity check to the model. More interesting is that much of northwestern BC, north central BC, and the Rocky Mountains of Alberta are predicted to have intermediate hybrid indices. In the species distribution models, these areas are expected to only contain pure parental species.  Given that the climatic hybrid index models were generated via machine learning, it is not possible to assess precisely why these regions differ from ecological expectations. It is possible that these are indicative of climatic spaces novel to the model and are therefore poorly predicted. In a map of predictions unbounded by species distribution models (Figure A.1) we see that many novel ecosystems \u2013 e.g., the cold grasslands of southeastern Alberta or the warm, wet islands of Haida Gwaii (shown in boxes in Figure A.1) \u2013 are predicted to be intermediate. When a random forest encounters novel data, it is likely to generate intermediate values, as the data will randomly \u201ctumble\u201d through the decision trees and result in a wide range of outputs that are averaged to intermediacy (Breiman 2001).  It is also possible that these areas do, in fact, contain genomically-intermediate but phenotypically parental trees. Given the difficulties in accurately assessing introgression using phenotypic characters (Hamilton 2012), these regions may be incorrectly assigned to pure parental species on the basis of ecological assumptions. This has already occurred in other parts of central and northern BC that were thought to contain pure white spruce prior to genomic 101  analysis (Daubenmire 1974), so they do at minimum merit further genomic investigation to determine the true spatial extent of the hybrid zone.  The species distribution models seem to largely represent the species ranges accurately at present. They correctly reflect important geographic barriers affecting the distribution of parental species (e.g., Engelmann spruce only occurring on the east side of the coast mountains in BC and missing in extreme alpine areas of southeastern BC and Alberta (Hope et al. 1991), white spruce absent from the southern grasslands and much of Alberta\u2019s Caribou mountains (Nienstaedt and Zasada 1990)), and some unique features such as hybrids in the Rocky Mountain trench (Hope et al. 1991) and white spruce in Cypress Hills (Newsome 1963). Likewise, populations in areas where the range of hybrid spruce overlaps with parental ranges tend to have a broad genomic bias towards that parent; e.g., Engelmann-like hybrids in the southern interior of BC and white-like hybrids in the east of the central plateau (Figure 4.4p vs. Figure 4.8p). The models\u2019 imperfect performance, particularly their relative lack of specificity at optimal TSS thresholds may be due to inaccurate species assignment in areas of sympatry. Both parental species have presences recorded in areas largely documented to contain hybrids, and the hybrid classification appears to be applied broadly in areas east of the Rockies that likely are white spruce (Figure 4.1).      102  4.4.2 Paleoclimatic predictions of the hybrid zone and post-glacial colonization of western Canada  4.4.2.1 Paleoclimatic models of species niches and hybrid index  The paleoclimatic models of species niches and hybrid index suggest that suitable habitat has existed for white spruce in western Canada since at least 15kybp (Figure 4.8), as previously suggested by De La Torre et al. (2014b), and that suitable habitat for Engelmann and hybrid spruce has been available since at least 13kybp. The predictions across time capture several notable climatic events of the Holocene, including an increase in suitability for Engelmann and hybrid spruce during the B\u00f8lling\u2013Aller\u00f8d warming period (Figure 4.8c,d), followed by a decrease in suitability for these warmer-adapted genotypes during the colder period of the Younger Dryas (Figure 4.8e). As the saddle between the Cordilleran and Laurentide ice sheets broke down, the species climatic niche and climatic hybrid index models both predict suitable white spruce habitat in the unglaciated space between them. This could provide a colonization window for refugia from the north (Anderson et al. 2006) or south (Ritchie and Macdonald 1986).  Interestingly, the climatic hybrid index and ecological niche models disagree somewhat on the suitability of habitat during deglaciation. While the climatic hybrid index model suggests suitability for intermediate hybrids throughout much of central and eastern Alberta from 13-11kybp, the species climatic niche models predict only white spruce in these regions. This could again be indicative of poor prediction ability from the climatic hybrid index model due to novel climates in these regions. Elsewhere, the models are largely in agreement throughout the paleoclimatic period. 103  By the time the Cordilleran ice sheet had largely melted, the species climatic niche models cover much of western Canada, and suitable habitat was available for white, Engelmann, and hybrid spruce, in a distribution not dissimilar from today\u2019s hybrid zone (Figure 4.6d). The subsequent millennia are marked by a largely constant climate for BC, and a retreat of suitable habitat throughout the southern Alberta grasslands into today\u2019s distribution. Pollen records indicate that spruce was present throughout much of central BC by 11 kybp (Bennett et al. 2001; Gavin et al. 2009), suggesting that colonization likely closely followed the retreat of the glaciers. Unfortunately, spruce pollen is not visually distinguishable to the species level, and as such it is unclear from these data whether both parental species were present in the earliest stages of re-colonization and therefore if secondary contact occurred during recolonization or in glacial refugia.  4.4.2.2 Post-glacial colonization routes  The distribution of mitotypes cluster strongly throughout the study area (Figure 4.12) and may help clarify the routes that the parent species took to recolonization. In Figure 4.12, there appear to be four major mitotype clusters: One dominated by white spruce mitotypes in central and northern BC, and throughout much of Alberta; one dominated by Sitka spruce throughout the northwestern provenances; and two clusters of Engelmann mitotypes to the west and east of the Rocky Mountains. A handful of provenances in the Caribou and Monashee mountains of BC, at the boundary between the west Engelmann and white clusters, contain individuals with both white and Engelmann mitotypes, and several provenances near the boundary between the white and Sitka clusters have mixed mitotypes as well. This pattern points strongly to a \u201cfirst-come first-served\u201d pattern of colonization, by which the first mitotype to colonize an area becomes the 104  dominant or only mitotype present in that area over time due to the maternal inheritance of mitochondria. This seems plausible given the rarity of long-distance seed dispersal in conifers (e.g., Williams et al. 2006).  This suggests that each cluster is likely representative of a separate colonization pathway, and therefore that spruce recolonized western Canada via four separate refugia. The two dominant pathways are white spruce arriving from the northeast and Engelmann spruce from the south, meeting in central BC. This is in agreement with prior phylogeographies of these species and this region (Williams et al. 2004; Anderson et al. 2006; Roberts and Hamann 2015). The eastern Engelmann cluster could represent a separate colonization event from a more genetically-diverse refugium, or perhaps is merely closer to a shared refugium, and the more homogeneous western cluster may be indicative of allele surfing during colonization (Excoffier et al. 2009a). The presence of the Sitka mitotype cluster is more anomalous, suggesting colonization far into the interior of BC from a coastal refugium. While Sitka mitotypes have, in the past, been observed in hybrid spruce from this region (Sutton et al. 1994), the widespread extent shown here has not been reported elsewhere. When species climatic niche models for Sitka spruce are included in the paleoclimatic predictions (Figure A.2), at no point is the central plateau of BC climatically suitable for Sitka spruce. This could suggest early pollen flow from either white or hybrid spruce into a refugium on the central coast of BC, bringing in adaptive alleles that would allow expansion into the continental climates of BC\u2019s central interior. The provenances containing Sitka mitotypes also tend to have moderate components of Sitka ancestry in their genomes (Figure 2.1), which could point to some residual ancestry from this past gene flow event. The Nass and Skeena river drainages on the central coast of BC are a well-documented white-Sitka hybrid zone (Hamilton et al. 2013), which may be relictual of this time period. 105  4.4.2.3 Comparison with paleobotanical records While these inferences on species migration are somewhat speculative, they are supported by the pollen record. A comparison of spruce pollen abundance as reconstructed by Williams et al. (2004) at 13-11kybp (Figure 4.13), shows several disparate regions of spruce pollen. The rapid migration of white spruce out of an eastern US refugium supports the lack of mitotype diversity in white spruce and aligns with the first appearances of spruce pollen in the interior of BC. Williams et al. predict spruce spreading eastward from coastal BC, near Haida Gwaii, around 12kybp, which could correspond to the expansion of the Sitka mitotype observed in that area. Spruce pollen appears in sediments on Haida Gwaii dated to 26-25kybp (Mathewes and Clague 2017), and climatic models suggest that parts of Haida Gwaii could have hosted boreal and subalpine species during the last glacial maximum (Roberts and Hamann 2015). It is notable that the pollen records co-occur with Abies sp. and Tsuga mertensiana pollen (Mathewes and Clague 2017), which suggests a more subalpine climate than is typically associated with Sitka spruce. It is possible that hybrid spruce (carrying Sitka spruce mitotypes) had a refugium there, as hybrid spruce is more adapted to these climates (Hamilton et al. 2013). Two semi-isolated clusters of spruce pollen can be seen south of the Cordilleran ice sheet that could roughly correspond to the eastern and western Engelmann mitotype clusters observed in Figure 4.12. The southern and northern pollen clusters seen in the NAPD reconstruction (corresponding to the Engelmann and white mitotype clusters observed in my data) do not meet until approximately 7kybp (Figure 4.13), roughly in the geographic area where the multi-mitotype provenances are observed. Two palynological and macrofossil datasets from near the 106  area of mixed mitotype provenances support this, with spruce pollen first appearing around Figure 4.13: Predicted abundance of spruce species in North America from 15kybp to present, based on palynological data. Color intensity corresponds to pollen abundance. Grey: 1-5% of pollen profile, light green: 5-20%, medium green: 20-40%, dark green: >40%. Figures reproduced with permission from Williams et al. (2004).      107  10kybp and becoming abundant between 8-6kybp (Gavin et al. 2011; Schw\u00f6rer et al. 2017). Elucidating past habitat availability and migrations from the paleoclimatic hybrid index and species climatic niche models has several important limitations which bear discussion. First, the method of paleoclimatic modelling relies on a single climatic model (CCSM3), which may produce more extreme predictions than an ensemble (e.g., Mahony et al. 2022), and may not accurately model the climate in this particular part of the world. Additionally, downscaling climatic data assumes that elevational lapse rates in climatic variables have remained constant throughout time, and inferences based on current species occurrences and genomic ancestries assume that correlations between climate and hybrid index have remained constant over time and that species have fully realized their ecological niches. Another potential confounding influence on the timing of post-glacial recolonization is the uncertainty in the deglaciation chronology of the Cordilleran ice sheet. Due to the complex topography of BC, modeling this ice sheet has proved a major challenge. The figures used in this chapter superimpose the deglaciation chronology of Dkye et al. (2003), but there have been many subsequent efforts to estimate the margins of this ice sheet over time (Seguinot et al. 2016; Gombiner 2019; Dulfer et al. 2022), with widely varying remnant ice fields and chronologies. The species distribution reconstructions of Williams et al. (2004) use the ice sheet extents estimated by Dyke and Prest (1987), which predicted large remnant ice sheets in the coastal ranges of BC as well as the Kootenay ranges of southeastern BC from 12-8kybp. These barriers would have had major implications on the colonization of what is now BC, particularly inland from a coastal refugium.     108  4.4.3 Present and future adaptive lag  4.4.3.1 Present adaptive lag (1961-1990 reference period vs. 1991-2020 period) When comparing the species climatic niches and predicted hybrid indices for the 1961-1990 reference period with the 1991-2020 period, there are clear differences in the area predicted to be suitable for spruce at present (areas shown in black in Figure 4.9). In particular, parts of central Alberta show a retreat of suitable habitat from the southern range margin. This area coincides with the dry mixedwood ecosystem, an interface between the sub-boreal grasslands and parklands of southern Alberta and the colder boreal forests of the north. This area has seen notable mortality and declines in stand productivity over the last several decades (Hogg et al. 2017; Birch et al. 2019). This is expected behavior of the lagging edge of a species\u2019 range under climatic warming (Solarik et al. 2018; Fr\u00e9javille et al. 2020), supporting the model predictions of a northward shift in suitability for white spruce in Alberta (Figure 4.6 and Figure 4.10). While spruce has not been completely eliminated from the areas of predicted niche loss, it is unlikely that these forests will persist under this current climate. Future monitoring of natural seedling recruitment and mortality will be necessary to determine the total demographic effects of the warming that has already been observed over the last thirty years. There is also notable loss of habitat in central BC (Figure 4.5), corresponding primarily to the dry, cool interior Douglas-fir ecosystems. These ecosystems hold some similarity to the dry mixedwood forests of Alberta, with sparse forest canopy, extensive parkland, and where spruce only ever existed as a minor species (Hope et al. 1991). These ecosystems are no longer considered suitable for planting interior spruce under the new reforestation guidelines for BC (MacKenzie and Mahony 2021). While much of the combined climatic niche overlaps between 109  the reference and present time periods, there are important although subtle differences revealed by the predicted change in hybrid index. Throughout much of British Columbia, the predicted hybrid index from the 1991-2020 time period is substantially higher than the 1961-1990 period, ranging from +0.1-0.4 in favor of Engelmann spruce ancestry (Figure 4.9). These areas presently contain intermediate hybrids, but are now predicted to be favorable to Engelmann-like hybrids or purely Engelmann spruce. The results of the previous chapter suggest that this shift should be indicative of warming winter temperatures or increasing precipitation. Throughout the study area, precipitation is predicted to stay constant or decline, whereas winter temperatures are uniformly higher (Figure A.3), suggesting this to be the likely driver of these predicted shifts. Natural stands of interior spruce in central BC have shown reductions in growth over the last several decades, which has been broadly ascribed to warming trends (Wiley et al. 2018; Ivanusic et al. 2020), suggesting that these predicted differences in hybrid index may indeed correspond to maladaptation of extant genotypes. Although predicted changes in precipitation regimes are not large drivers of the shifts observed in these models, precipitation variables are important in the climatic niche models (Figure 4.3) and may indeed be primary drivers of species range shifts in the future. Precipitation variables are poorly predicted relative to temperature variables in climatic models, and changes in precipitation can vary widely between models (Mahony et al. 2022). Disagreement between climatic models tends to produce intermediate results for model ensembles (Tebaldi and Knutti 2007), which limits their value for predictive purposes in this regard. 110  4.4.3.2 Future climates and the fate of interior spruce in western Canada Future predictions of both climatic niches and hybrid index suggest that the climates of the next 80 years will mark a massive departure from previous millennia (Figure 4.6 and Figure 4.10). Even under the most optimistic climate change scenario (SSP1), the predicted hybrid zone under near-term future climates (2011-2040) is nearly unrecognizable, with massive areas of central Alberta no longer suitable for white spruce, a recession of suitable habitat in southern BC, and a northward shift of the hybrid zone into the Skeena and Omineca mountains of northern BC. While all of the future climatic pathways are similar in the near term, they show some divergence as time progresses. While central BC is still largely suitable for spruce in the 2071-2100 time period under SSP1, SSP2 shows a large reduction of suitable habitat by this time, and under SSP3 there is little suitable habitat remaining (Figure 4.10).  In all cases, predicted warming correlates to a shift toward more Engelmann-like genotypes within the hybrid zone. In the long-term, this could potentially generate an adaptive gradient, pushing southern genotypes with more warmly adapted alleles northward (Kremer et al. 2012). In the short-term this is likely to result in widespread maladaptation of extant populations. However, given the wide climatic tolerances that interior spruce can endure (Ukrainetz et al. 2011), this maladaptation is more likely to manifest in reduced growth rather than a complete replacement of the species by other, better-adapted tree species in the short-term. In the portions of the species distribution that are predicted to become unsuitable, this maladaptation may be substantially more severe, as increased severity and frequency of drought drive the species towards extirpation along its southern range margin. Where spruce is commercially planted for forestry, these areas will likely be prone to high mortality, reduced growth, and increased insect and disease incidence (Kolb et al. 2016; Renwick et al. 2016; Nelson et al. 2021). 111  The difference between the most-optimistic climate change scenario and the \u201cbusiness-as-usual\u201d scenario paints a foreboding image of the future of interior spruce in western Canada. Interior spruce is a slow-growing and long-lived species with an operational rotation length of approximately 80 years for second-growth forests (Xie and Yanchuk 2002; Boateng et al. 2009), meaning that trees planted today will be facing the climates of 2100 and beyond. The climatic niche for SSP1 in the 2071-2100 time period covers 524,618 km2, compared to the 310,634 km2 of suitable habitat for the SSP2 scenario, a 41% reduction. However, even the SSP1 scenario is approximately half the area covered by the 1961-1990 normal period. While both models show a worrying change in the adaptive landscape for interior spruce, tangible gains in climate policy will be necessary to save the productive timberlands of western Canada. Given the importance of forestry to the economy of this region, these simulations serve as an important warning that climate change adaptation at a massive geographic scale will be necessary no matter our path forward, but also that truly dramatic change will be necessary if we do not massively reduce carbon emissions in the coming decades.   4.5 Conclusions Paleoclimatic models suggest that climatically suitable habitat for spruce has existed around glacial margins since at least 15kybp, supporting the hypothesis that spruce quickly colonized following deglaciation. The geographic segregation of mitotypes suggests several colonization routes into central British Columbia. Though previous research suggests that white and Engelmann spruce could have co-existed in glacial refugia (De La Torre et al. 2014b; Roberts and Hamann 2015), the spruces that colonized northern and southern BC appear to have done so exclusively on a mitochondrial background of white and Engelmann spruce, respectively. This suggests that the species may not have come into secondary contact in western 112  Canada since the last glacial maximum until approximately 7kybp, based on congruence between palynological data and provenances with both white and Engelmann mitotypes. If the species entered secondary contact during or before the last glacial maximum, strong selection against hybrids in the ensuing millennia would be required to explain the trace amounts of hybrid ancestry observed in southern BC and boreal regions at present (Figure 2.1). Both the climatic prediction of hybrid index and species climatic niche models point to a vastly different future for spruce in western Canada than it has experienced since re-colonizing the landscape. Many areas where spruce is a keystone species are expected to be climatically unsuitable for the species within the next hundred years, regardless of climate scenario. Many other regions where spruce is expected to persist will likewise be unsuitable for the forests presently growing there, with large shifts in hybrid index predicted across much of the current hybrid zone. This signals the need for intensive monitoring and management of these forests. Assisted migration will be necessary across much of BC to ensure that warmly-adapted alleles and genotypes are able to rapidly migrate northward, and assisted range expansion in Alberta where much of the historical range of white spruce is already expected to have shifted northward, and where it will continue to do so. 113  Chapter 5: Conclusions  5.1 Summary In Chapter 2, we demonstrated that substantial evidence exists for local adaptation within the interior spruce hybrid zone. Multiple lines of phenotypic and genomic evidence point to hybridization fueling adaptation to the intermediate climates of the hybrid zone compared to the allopatric ranges of its parent species. At a more granular level, the relative genomic proportions of the two primary parent species (i.e., hybrid index) appears to be a good proxy for adaptation along several climatic axes, as demonstrated by multiple genomic studies.  In Chapter 3, we further explored the idea of hybridization providing unique adaptations to the climates of the hybrid zone. Although white and Engelmann spruce are very closely related, we identified thousands of divergent alleles within many genes. Genomic cline analysis revealed that many of these alleles introgress through the hybrid zone beyond what one would expect by looking at only hybrid index. These loci are frequently tightly correlated with climatic variables implicated in selection within this species complex. Additionally, these climatic variables do not vary linearly across the hybrid zone, but rather are more often skewed towards one parental species\u2019 range or the other throughout the hybrid zone. We identified a strong congruence between genes with allelic skew to one parent species and climatic variables that are likewise skewed. This suggests that local adaptation to climate in the interior spruce hybrid complex is tightly linked to hybridization, and that beneficial parental alleles have been able to efficiently introgress into the hybrid genome in the geographic areas where they are beneficial. In Chapter 4, we modeled the putative colonization history of western Canada by spruce following deglaciation, using a combination of mitochondrial haplotypes and species climatic 114  niche models. The proposed colonization routes match well with palynological records, suggesting that several populations of spruce have migrated into BC through multiple refugia. Although we could not conclusively determine whether or not white and Engelmann spruce hybridized in shared refugia, the present distribution of genomic and mitochondrial ancestry suggests that the two species arrived in BC separately. This indicates that the strong local adaptation documented in the previous chapters might have developed over as few as 7000 years. This translates to relatively few generations for spruces, which can have generation lengths ranging from tens to hundreds of years. This suggests a strong adaptive capacity within the species complex, likely bolstered by adaptive hybridization. We used the strong association between hybrid index and present-day climate to model hybrid index throughout western Canada. This revealed putative areas of hybrid spruce beyond what we were able to sample genomically, broadening the potential boundaries of the interior spruce hybrid zone. We also used these data to predict the past and future extent and genomic composition of the hybrid zone, revealing a largely stable distribution of suitable environments over the past several millennia that is already rapidly changing, and is predicted to continue changing radically over the coming decades. Between large reductions in predicted suitable habitat, and large genomic shifts predicted for most persisting areas, these projections suggest that many of western Canada\u2019s spruce forests are unlikely to thrive without substantial human assistance.  115  5.2 Limitations There are several important limitations to the research described in this dissertation that bear mention. First, the genomic resources presently available for white and Engelmann spruce limit the depth of genomic study in this species complex. Conifer genomes tend to be very large with high repeat content (De La Torre et al. 2014a). This has limited our ability to generate high-quality, chromosome-level assemblies for these species (though see Niu et al. 2022). This limits our ability to assess the physical relationship of putatively adaptive SNPs, which could help us identify important gene complexes or the presence of adaptive sweeps. Also, this lack of resources limits our ability to identify structural variants such as inversions and copy-number variations, which may play important roles in adaptation and partial reproductive isolation (Prunier et al. 2017; Walsh et al. 2023). Additionally, the majority of conifer genes lack high-quality annotation, severely limiting our ability to describe the functional mechanisms by which these trees adapt to their environments. Even among extant annotations, the vast majority have been computationally predicted based on homology to distantly related taxa, with little to nothing known about their function or pathways (Wegrzyn et al. 2014; Warren et al. 2015). This has limited the practical identification of genes related to climate adaptation, with many studies instead characterizing genes by shared ontologies (e.g., Namroud et al. 2008; Yeaman et al. 2014). While this has the benefit of characterizing large numbers of genes and being more generalizable than annotations from distant relatives (Ashburner et al. 2000), they are difficult to interpret in an adaptive context, as these broad categories do little to explain a given gene\u2019s role in specific traits, physiological responses to climate, or developmental pathways.  116  The incomplete genomic resources for interior spruce limit the scope and applicability of the analyses presented in Chapter 3. These analyses (i.e., the Bayenv2 association analyses and bgc cline analyses) focus on individual loci, though adaptation to climate in trees is presumed to be highly polygenic (Lind et al. 2018; Sork 2018). The effect size of any given locus analyzed here is expected to be very small, but when summarized across many thousands of loci should at least demonstrate the highly polygenic nature of climate adaptation in this species complex. This high-level approach was used in lieu of attempting to classify the number of significant loci using a false discovery rate approach. We used a lax threshold for determining a skewed locus in the bgc analyses and used percentile-based thresholds for the Bayenv2 analyses as means to cast the widest-possible net. This was done under the assumption that the large number of variants included in the dataset should capture a reasonably large proportion of the adaptive variation and that in this way we are analyzing variation throughout the genome. We acknowledge that, on the level of an individual SNP, there are likely many false positives. Much of this dissertation is focused on climate adaptation, as represented by sets of long-term climatic variables predicted across the landscape. In reality, selection for climate adaptation is likely mostly driven by individual weather events (e.g., Montw\u00e9 et al. 2018; Vitasse et al. 2019). The climatic variables analyzed in this dissertation represent proxies for the probability of weather events that may be driving selection in this species complex. Those weather events are not known, and climatic variables may be proxies for multiple distinct selective agents. For example, mean temperature of the coldest month, one of the most consistently important variables in the analyses presented here, could represent adaptation to one or many distinct causative agents such as growing season frost events, winter drought stress, or tissue damage due to extreme cold (Howe et al. 2003). The actual selective agents at play in this species complex 117  have been investigated elsewhere (Liepe et al. 2016; MacLachlan et al. 2018), and are discussed in Chapter 2 of this dissertation, but have not been the focus of Chapters 3 and 4. Bearing this in mind, it is important to limit the interpretations of the climatic variables identified in this dissertation as being implicated in adaptation.  All of the analyses presented in this dissertation are based on genomic data. As discussed in Chapters 1 and 2, genomic data has several advantages over phenotypic data in quickly assessing long-term adaptation to climate. However, phenotypic data is still invaluable for the study of adaptation. Many adaptive traits are highly polygenic and may be driven by subtle shifts across many loci and by complex interactions between loci that are undetectable with univariate measures of divergence (Yeaman 2022). Additionally, genomic signals of adaptation may be confounded by signatures of recent migration or fluctuations in population size (Fang et al. 2013; Zhao et al. 2020). Genomic studies likewise do little to elucidate the mechanisms of adaptation, which may be readily apparent in even short-term phenotypic studies (e.g., Liepe et al. 2016; MacLachlan et al. 2018).   5.3 Future research directions While our understanding of local adaptation in interior spruce is much more complete than for many other non-model wild species, there are still large gaps that remain unfilled. While many studies have attempted to infer the presence of local adaptation through clinal variation, only Lu et al. (2014, 2016) have directly tested for the presence of local adaptation within this species system, and no studies in interior spruce have fully synthesized genomic data with reciprocal transplant data. Additional studies combining the inferential power of reciprocal transplants with genomic data could reveal genes directly associated with genotype-by-118  environment interactions and therefore likely candidates for local adaptation. Furthermore, few studies have assessed local adaptation in allopatric Engelmann spruce. While the evidence gathered so far suggests that interior spruce is more strongly locally adapted than its parent species, this may be due simply to a lack of comparable studies in the allopatric range of white and Engelmann spruce. Plasticity also remains under-explored in this species complex (but see Benomar et al. 2016). In general, plasticity is the assumed response to environmental variation unless evidence exists for genetic control, instead of being directly quantified (Meril\u00e4 and Hendry 2014). Studies coherently synthesizing analyses of adaptive plasticity and genetic variation will produce a more holistic understanding of adaptation to climate in interior spruce, and may have profound impacts on our ability to predict future population responses to climate (Valladares et al. 2014). Though we have learned that plasticity and adaptation are not mutually exclusive at the gene level (Yeaman et al. 2016), we still do not understand the interplay of these forces at the landscape level.  As adaptation in the interior spruce hybrid zone appears to have historically been in relation to cold temperatures, climate change has the potential to massively disrupt the adaptive strategies of natural populations. As discussed in Chapter 4, most areas of the interior spruce hybrid zone will experience large shifts in adaptive optima within the next century, potentially resulting in widespread maladaptation. The approach we took to assess putative maladaptation via shifts in optimal hybrid index is similar to the genomic offsets predicted by Gradient Forests (Ellis et al. 2012; Fitzpatrick and Keller 2015), which has recently gained popularity for predicting species\u2019 responses to climate change (e.g., Gugger et al. 2018; Sang et al. 2022; Yuan 119  et al. 2023). It may be worthwhile to compare the approach presented here to GradientForests to assess the consistency of these offset approaches. Regardless of the approach used to genomically predict maladaptation, long-term assisted migration trials will be critical to validate these results, tracking the effects of ongoing climate change and informing reforestation and conservation practices. Some of these trials have already been established for interior spruce in British Columbia (O\u2019Neill et al. 2014), but additional trials should be established frequently to assess changing climatic conditions on early life stages. The genomic resources available for interior spruce are developing rapidly, but there are still critical areas limiting our ability to understand the genomic basis of local adaptation. The lack of a complete reference genome is a major hurdle, as analyses of the genomic architecture of local adaptation to date have relied on statistical linkage, and generally lack information on physical linkage beyond a few kilobases (Yeaman et al. 2016). Emerging technologies, such as long-read sequencing, have enabled chromosome-level genome assemblies in other conifers (Scott et al. 2020; Niu et al. 2022; Fujino et al. 2023). This work is already underway for interior spruce, with a recent effort generating chromosome-level maps comprising multiple gigabases for white, Engelmann, Sitka, and interior spruce genomes (Gagalova et al. 2022). Understanding how candidate local adaptation genes cluster across the genome may help narrow our focus to particular genomic regions enriched for adaptive genes and better understand the genomic architectures of selection for traits and their interactions across the genome.  An unresolved question of this dissertation that might be answered with a complete reference genome is the number of generations that white and Engelmann spruce have been in secondary contact. By estimating the number and distribution of interspecific recombination events throughout the genomes of hybrid individuals, one can estimate the number of generations 120  that these species have been in secondary contact (e.g., Medina et al. 2018). This can allow us to determine whether white and Engelmann spruce were in secondary contact during the last glacial maximum, and if present allopatric populations still contain traces of hybrid ancestry.    Though the research presented in this dissertation improves upon our understanding of the interior spruce hybrid complex, there is still much to learn about the nature of this hybrid zone, and about climate adaptation in plants more broadly. Across the world, the coming decades will present unprecedented challenges to plants that have spent millennia adapting to the spaces they inhabit. Hybrid zones provide us with an incredible opportunity to understand how distinct populations can pool their genetic variation to adapt to a broader range of climates. Armed with this knowledge, we have the opportunity to accelerate climate adaptation in our tree species through scientifically informed reforestation efforts.    121  References Abbott, R., Albach, D., Ansell, S., Arntzen, J.W., Baird, S.J.E., Bierne, N., Boughman, J., Brelsford, A., Buerkle, C.A., Buggs, R., Butlin, R.K., Dieckmann, U., Eroukhmanoff, F., Grill, A., Cahan, S.H., Hermansen, J.S., Hewitt, G., Hudson, A.G., Jiggins, C., Jones, J., Keller, B., Marczewski, T., Mallet, J., Martinez-Rodriguez, P., M\u00f6st, M., Mullen, S., Nichols, R., Nolte, A.W., Parisod, C., Pfennig, K., Rice, A.M., Ritchie, M.G., Seifert, B., Smadja, C.M., Stelkens, R., Szymura, J.M., V\u00e4in\u00f6l\u00e4, R., Wolf, J.B.W., and Zinner, D. 2013, February. Hybridization and speciation. doi:10.1111\/j.1420-9101.2012.02599.x. Agrawal, A.F., and Whitlock, M.C. 2012. Mutation load: The fitness of individuals in populations where deleterious alleles are abundant. Annu. Rev. Ecol. Evol. Syst. 43: 115\u2013135. Aitken, S.N., and Bemmels, J.B. 2016. Time to get moving: Assisted gene flow of forest trees. Evol. Appl. 9(1): 271\u2013290. doi:10.1111\/eva.12293. Aitken, S.N., and Hannerz, M. 2001. Genecology and gene resource management strategies for conifer cold hardiness. In Conifer Cold Hardiness. Edited by F.J. Bigras and S.J. Colombo. Springer, Dordrecht. pp. 23\u201353. doi:10.1007\/978-94-015-9650-3_2. Aitken, S.N., and Whitlock, M.C. 2013. Assisted gene flow to facilitate local adaptation to climate change. Annu. Rev. Ecol. Evol. Syst. 44(1): 367\u2013388. doi:10.1146\/annurev-ecolsys-110512-135747. Aitken, S.N., Yeaman, S., Holliday, J.A., Wang, T., and Curtis-McLane, S. 2008. Adaptation, migration or extirpation: climate change outcomes for tree populations. Evol. Appl. 1(1): 95\u2013111. doi:10.1111\/j.1752-4571.2007.00013.x. Alberta. 2016. Introduction to Ecological Information System ECOSYS. Alberto, F.J., Aitken, S.N., Al\u00eda, R., Gonz\u00e1lez-Mart\u00ednez, S.C., H\u00e4nninen, H., Kremer, A., Lef\u00e8vre, F., Lenormand, T., Yeaman, S., Whetten, R., and Savolainen, O. 2013. Potential for evolutionary responses to climate change - evidence from tree populations. Glob. Chang. Biol. 19(6): 1645\u20131661. doi:10.1111\/gcb.12181. Alexander, D.H., and Lange, K. 2011. Enhancements to the ADMIXTURE algorithm for individual ancestry estimation. BMC Bioinformatics 12. doi:10.1186\/1471-2105-12-246. Alexander, D.H., Novembre, J., and Lange, K. 2009. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 19: 1655\u20131664. doi:10.1101\/gr.094052.109. Alexander, R.R., and Shepperd, W.D. 1984. Silvical characteristics of Engelmann spruce. USDA Forest Service, Fort Collins. Allouche, O., Tsoar, A., and Kadmon, R. 2006. Assessing the accuracy of species distribution models: Prevalence, kappa and the true skill statistic (TSS). J. Appl. Ecol. 43(6): 1223\u20131232. doi:10.1111\/j.1365-2664.2006.01214.x. Andalo, C., Beaulieu, J., and Bousquet, J. 2005. The impact of climate change on growth of local white spruce populations in Qu\u00e9bec, Canada. For. Ecol. Manage. 205(1\u20133): 169\u2013182. doi:10.1016\/j.foreco.2004.10.045. Anderson, E. 1948. Hybridization of the Habitat. Evolution (N. Y). 2(1): 1\u20139. Anderson, E. 1949. Introgressive Hybridization. John Wiley and Sons, New York. Anderson, E., and Stebbins, G.L. 1954. Hybridization as an evolutionary stimulus. Evolution (N. Y). 8(4): 378\u2013388. 122  Anderson, L.L., Hu, F.S., Nelson, D.M., Petit, R.J., and Paige, K.N. 2006. Ice-age endurance: DNA evidence of a white spruce refugium in Alaska. Proc. Natl. Acad. Sci. U. S. A. 103(33): 12447\u201312450. doi:10.1073\/pnas.0605310103. Archer, K.J., and Kimes, R. V. 2008. Empirical characterization of random forest variable importance measures. Comput. Stat. Data Anal. 52(4): 2249\u20132260. doi:10.1016\/j.csda.2007.08.015. Arnold, M.L., and Hodges, S. a. 1995. Are natural hybrids fit or unfit relative to their parents? Trends Ecol. Evol. 10(2): 67\u201371. Ashburner, M., Ball, C.A., Blake, J.A., Botstein, D., Butler, H., Cherry, M.J., Davis, A.P., Dolinski, K., Dwight, S.S., Eppig, J.T., Harris, M.A., Hill, D.P., Issel-Tarver, L., Kasarskis, A., Lewis, S., Matese, J.C., Richardson, J.E., Ringwald, M., Rubin, G.M., and Sherlock, G. 2000. Gene Ontology: tool for the unification of biology. Nat. Genet. 25(May): 25\u201329. doi:10.1038\/75556. Bailey, R.I., Thomas, C.D., and Butlin, R.K. 2004. Premating barriers to gene exchange and their implications for the structure of a mosaic hybrid zone between Chorthippus brunneus and C. jacobsi (Orthoptera: Acrididae). J. Evol. Biol. 17(1): 108\u2013119. doi:10.1046\/j.1420-9101.2003.00648.x. Barghi, N., Hermisson, J., and Schl\u00f6tterer, C. 2020. Polygenic adaptation: a unifying framework to understand positive selection. Nat. Rev. Genet. 21(12): 769\u2013781. Springer US. doi:10.1038\/s41576-020-0250-z. Barton, N.H. 1983. Multilocus clines. Evolution (N. Y). 37(3): 454\u2013471. Barton, N.H. 1999. Clines in polygenic traits. Genet. Res. (Camb). 74(3): 223\u2013236. doi:10.1017\/S001667239900422X. Barton, N.H., and Gale, K.S. 1993. Genetic analysis of hybrid zones. In Hybrid Zones and the Evolutionary Process. Edited by R. Harrison. Oxford University Press, Oxford, UK. Barton, N.H., and Hewitt, G.M. 1985. Analysis of hybrid zones. Annu. Rev. Ecol. Syst. 16: 113\u2013148. Beaumont, M.A., and Balding, D.J. 2004. Identifying adaptive genetic divergence among populations from genome scans. Mol. Ecol. 13: 969\u2013980. doi:10.1111\/j.1365-294X.2004.02125.x. Bennett, J.R., Cumming, B.F., Leavitt, P.R., Chiu, M., Smol, J.P., and Szeicz, J. 2001. Diatom, pollen, and chemical evidence of postglacial climatic change at Big Lake, south-central British Columbia, Canada. Quat. Res. 55(3): 332\u2013343. Benomar, L., Lamhamedi, M.S., Rainville, A., Beaulieu, J., Bousquet, J., and Margolis, H.A. 2016. Genetic adaptation vs. ecophysiological plasticity of photosynthetic-related traits in young Picea glauca trees along a regional climatic gradient. Front. Plant Sci. 7(February): 1\u201315. doi:10.3389\/fpls.2016.00048. Benomar, L., Lamhamedi, M.S., Villeneuve, I., Rainville, A., Beaulieu, J., Bousquet, J., and Margolis, H.A. 2015. Fine-scale geographic variation in photosynthetic-related traits of Picea glauca seedlings indicates local adaptation to climate. Tree Physiol. 35(8): 864\u2013878. doi:10.1093\/treephys\/tpv054. Benton, T.G., and Grant, A. 2000. Evolutionary fitness in ecology: Comparing measures of fitness in stochastic, density-dependent environments. Evol. Ecol. Res. 2(6): 769\u2013789. doi:10.1145\/1835428.1835432. Bigelow, R.S. 1965. Hybrid zones and reproductive isolation. Evolution (N. Y). 4: 449\u2013458. Bigras, F.J., Ryypp\u00f6, A., Lindstr\u00f6m, A., and Stattin, E. 2001. Cold acclimation and deacclimation of shoots and roots of conifer seedlings. In Conifer Cold Hardiness. Edited by F.J. Bigras and S.J. Colombo. Springer, 123  Dordrecht. pp. 57\u201388. doi:10.1007\/978-94-015-9650-3_3. Birch, J.D., Lutz, J.A., Hogg, E.H., Simard, S.W., Pelletier, R., LaRoi, G.H., and Karst, J. 2019. Decline of an ecotone forest: 50 years of demography in the southern boreal forest. Ecosphere 10(4). doi:10.1002\/ecs2.2698. Birchler, J., Yao, H., and Chudalayandi, S. 2006. Unraveling the genetic basis of hybrid vigor. Proc. Natl. Acad. Sci. 103(35): 12957\u201312958. doi:10.1073\/pnas.0605627103. Birks, H.H.J.B., and Birks, H.H.J.B. 2000. Future uses of pollen analysis must include plant macrofossils. J. Biogeogr. 27(1): 31\u201335. doi:10.1046\/j.1365-2699.2000.00375.x. Birky, C.W., Maruyama, T., and Fuerst, P. 1983. An approach to population and evolutionary genetic theory for genes in mitochondria and chloroplasts, and some results. Genetics 103(3): 513\u2013527. Birol, I., Raymond, A., Jackman, S.D., Pleasance, S., Coope, R., Taylor, G.A., Yuen, M.M. Saint, Keeling, C.I., Brand, D., Vandervalk, B.P., Kirk, H., Pandoh, P., Moore, R.A., Zhao, Y., Mungall, A.J., Jaquish, B., Yanchuk, A., Ritland, C., Boyle, B., Bousquet, J., Ritland, K., Mackay, J., Bohlmann, J., and Jones, S.J.M. 2013. Assembling the 20 Gb white spruce (Picea glauca) genome from whole-genome shotgun sequencing data. Bioinformatics 29(12): 1492\u20137. doi:10.1093\/bioinformatics\/btt178. Blanquart, F., Gandon, S., and Nuismer, S.L. 2012. The effects of migration and drift on local adaptation to a heterogeneous environment. J. Evol. Biol. 25(7): 1351\u20131363. doi:10.1111\/j.1420-9101.2012.02524.x. Blanquart, F., Kaltz, O., Nuismer, S.L., and Gandon, S. 2013. A practical guide to measuring local adaptation. Ecol. Lett. 16(9): 1195\u20131205. doi:10.1111\/ele.12150. Boateng, J.O., Heineman, J.L., Bedford, L., Harper, G.J., and Nemec, A.F.L. 2009. Long-term effects of site preparation and postplanting vegetation control on Picea glauca survival, growth and predicted yield in boreal British Columbia. Scand. J. For. Res. 24(2): 111\u2013129. doi:10.1080\/02827580902759685. Bolnick, D.I., and Nosil, P. 2007. Natural selection in populations subject to a migration load. Evolution (N. Y). 61(9): 2229\u20132243. doi:10.1111\/j.1558-5646.2007.00179.x. Bouill\u00e9, M., Senneville, S., and Bousquet, J. 2011. Discordant mtDNA and cpDNA phylogenies indicate geographic speciation and reticulation as driving factors for the diversification of the genus Picea. Tree Genet. Genomes 7(3): 469\u2013484. doi:10.1007\/s11295-010-0349-z. Boyle, E.A., Li, Y.I., and Pritchard, J.K. 2017. An Expanded View of Complex Traits: From Polygenic to Omnigenic. Cell 169(7): 1177\u20131186. Elsevier. doi:10.1016\/j.cell.2017.05.038. Braconnot, P., Harrison, S.P., Kageyama, M., Bartlein, P.J., Masson-Delmotte, V., Abe-Ouchi, A., Otto-Bliesner, B., and Zhao, Y. 2012. Evaluation of climate models using palaeoclimatic data. Nat. Clim. Chang. 2(6): 417\u2013424. Nature Publishing Group. doi:10.1038\/nclimate1456. Bradshaw, A.D. 1960. Population differentiation in Agrostis tenuis Sibth. III. Populations in varied environments. New Phytol. 59(1): 92\u2013103. Bradshaw, A.D. 1972. Some evolutionary consequences of being a plant. In Evolutionary Biology. Edited by T. Dobzhansky, M.K. Hecht, and W.C. Steere. Appleton-Century-Crofts, New York. pp. 25\u201347. Breiman, L. 2001. Random forests. Mach. Learn. 45: 5\u201332. doi:10.1109\/ICCECE51280.2021.9342376. Bridle, J.R., Polechov\u00e1, J., Kawata, M., and Butlin, R.R. 2010. Why is adaptation prevented at ecological margins\u202f? New insights from individual-based simulations. Ecol. Lett. 13: 485\u2013494. doi:10.1111\/j.1461-124  0248.2010.01442.x. Brown, J.S., and Pavlovic, N.B. 1992. Evolution in heterogeneous environments: effects of migration on habitat specialization. Evol. Ecol. 6: 360\u2013382. Buggs, R.J.A. 2007. Empirical study of hybrid zone movement. Heredity (Edinb). 99(3): 301\u2013312. doi:10.1038\/sj.hdy.6800997. \u0106ali\u0107, I., Bussotti, F., Mart\u00ednez-Garc\u00eda, P.J., and Neale, D.B. 2016. Recent landscape genomics studies in forest trees\u2014what can we believe? Tree Genet. Genomes 12(1): 1\u20137. doi:10.1007\/s11295-015-0960-0. Chen, J., K\u00e4llman, T., Ma, X., Gyllenstrang, N., Zaina, G., Morgante, M., Bousquet, J., Eckert, A., Wegrzyn, J., Neale, D., Lagercrantz, U., and Lascoux, M. 2012. Disentangling the roles of history and local selection in shaping clinal variation of allele frequencies and gene expression in Norway spruce (Picea abies). Genetics 191(3): 865\u2013881. doi:10.5061\/dryad.82201.1. Choi, Y., Sims, G.E., Murphy, S., Miller, J.R., and Chan, A.P. 2012. Predicting the functional effect of amino acid substitutions and indels. PLoS One 7(10). doi:10.1371\/journal.pone.0046688. Choler, P., Erschbamer, B., Tribsch,  a, Gielly, L., and Taberlet, P. 2004. Genetic introgression as a potential to widen a species\u2019 niche: insights from alpine Carex curvula. Proc. Natl. Acad. Sci. U. S. A. 101(1): 171\u2013176. doi:10.1073\/pnas.2237235100. Clark, P.U., Dyke, A.S., Shakun, J.D., Carlson, A.E., Clark, J., Wohlfarth, B., Mitrovica, J.X., Hostetler, S.W., and McCabe, A.M. 2009. The Last Glacial Maximum. Science (80-. ). 325(5941): 710\u2013714. doi:10.1126\/science.1172873. Clausen, J., Keck, D.D., and Hiesey, W.M. 1940. Experimental studies on the nature of species. I. Effect of varied environments on western North American plants. Carnegie Institution of Washington, Washington. doi:Genetic Considerations in Ecological Restoration. Clyde, D. 2020. From a distance \u2014 gene regulation in plants. Nat. Rev. Genet. 21(2): 68\u201369. doi:10.1038\/s41576-019-0201-8. Colautti, R.I., Lee, C.R., and Mitchell-Olds, T. 2012. Origin, fate, and architecture of ecologically relevant genetic variation. Curr. Opin. Plant Biol. 15(2): 199\u2013204. Elsevier Ltd. doi:10.1016\/j.pbi.2012.01.016. Colwell, R.K., and Rangel, T.F. 2009. Hutchinson\u2019s duality: The once and future niche. Proc. Natl. Acad. Sci. U. S. A. 106(SUPPL. 2): 19651\u201319658. doi:10.1073\/pnas.0901650106. Comes, H.P., and Kadereit, J.W. 1998. The effect of quaternary climatic changes on plant distribution and evolution. Trends Plant Sci. 3(11): 432\u2013438. doi:10.1016\/S1360-1385(98)01327-2. Conkle, M.T. 1973. Growth data for 29 years from the California elevational transect study of ponderosa pine. For. Sci. 19(1): 31\u201339. Conte, G.L., Hodgins, K.A., Yeaman, S., Degner, J.C., Aitken, S.N., Rieseberg, L.H., and Whitlock, M.C. 2017. Bioinformatically predicted deleterious mutations reveal complementation in the interior spruce hybrid complex. BMC Genomics 18(1): 1\u201312. BMC Genomics. doi:10.1186\/s12864-017-4344-8. Cooke, J.E.K., Eriksson, M.E., and Junttila, O. 2012. The dynamic nature of bud dormancy in trees: Environmental control and molecular mechanisms. Plant, Cell Environ. 35(10): 1707\u20131728. doi:10.1111\/j.1365-3040.2012.02552.x. 125  Coop, G., Witonsky, D., Di Rienzo, A., and Pritchard, J.K. 2010. Using environmental correlations to identify loci underlying local adaptation. Genetics 185(4): 1411\u20131423. doi:10.1534\/genetics.110.114819. Le Corre, V., and Kremer, A. 2012. The genetic differentiation at quantitative trait loci under local adaptation. Mol. Ecol. 21(7): 1548\u20131566. doi:10.1111\/j.1365-294X.2012.05479.x. Le Corre, V., Machon, N., Petit, R.J., and Kremer, A. 1997. Colonization with long-distance seed dispersal and genetic structure of maternally inherited genes in forest trees: a simulation study. Genet. Res. (Camb). 69: 117\u2013125. Crow, J.F. 1948. Alternative hypotheses of hybrid vigor. Genetics 33: 447\u2013487. Cruzan, M.B., and Arnold, M.L. 1993. Ecological and genetic associations in an iris hybrid zone. Evolution (N. Y). 47(5): 1432\u20131445. Daly, C., Neilson, R.P., and Phillips, D.L. 1994. A statistical-topographic model for mapping climatological precipitation over mountainous terrain. J. Appl. Meteorol. 33(2): 140\u2013158. doi:10.1175\/1520-0450(1994)033<0140:ASTMFM>2.0.CO;2. Darwin, C. 1859. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. Murray, London. Daubenmire, R.F. 1968. Some geographic variations in Picea sitchensis and their ecological interpretation. Can. J. Bot. 46: 787\u2013798. Daubenmire, R.F. 1974. Taxonomic and ecologic relationships between Picea glauca and Picea engelmannii. Can. J. Bot. 52(7): 1545\u20131560. doi:10.1139\/b74-203. Dobzhansky, T. 1937. Genetics and the Origin of Species. Columbia University Press, New York. Dobzhansky, T. 1940. Speciation as a stage of evolutionary divergence. Am. Nat. 74(753): 312\u2013321. Dulfer, H.E., Margold, M., Darvill, C.M., and Stroeven, A.P. 2022. Reconstructing the advance and retreat dynamics of the central sector of the last Cordilleran Ice Sheet. Quat. Sci. Rev. 284: 107465. The Authors. doi:10.1016\/j.quascirev.2022.107465. Dunton, K.H., Hardegree, B., and Whitledge, T.E. 2001. Response of estuarine marsh vegetation to interannual variations in precipitation. Estuaries 24(6): 851\u2013861. doi:10.2307\/1353176. Dyke, A., Moore, A., and Robertson, L. 2003. Deglaciation of North America. Dyke, A.S. 2004. An outline of North American deglaciation with emphasis on central and northern Canada. In Quaternary glaciations - extent and chronology, part II. North America. Edited by J. Ehlers and P.L. Gibbard. Elsevier. pp. 373\u2013424. Dyke, A.S., and Prest, V.K. 1987. Late Wisconsonian and Holocene history of the Laurentide ice sheet. G\u00e9ographie Phys. Quat. 41(2): 237\u2013263. Elleouet, J.S. 2018. Linking demographic history and evolution at the expanding range edge of Sitka spruce (Picea sitchensis). University of British Columbia. Available from https:\/\/open.library.ubc.ca\/cIRcle\/collections\/ubctheses\/24\/items\/1.0372322. Ellis, N., Smith, S.J., and Roland Pitcher, C. 2012. Gradient forests: Calculating importance gradients on physical predictors. Ecology 93(1): 156\u2013168. doi:10.1890\/11-0252.1. 126  Ellstrand, C. 1992. Gene flow among seed plant populations. New For. 6: 241\u2013256. Enard, D., Messer, P.W., and Petrov, D.A. 2014. Genome-wide signals of positive selection in human evolution. Genome Res. 24(6): 885\u2013895. doi:10.1101\/gr.164822.113. Endler, J.A. 1977. Geographic variation, Speciation, and Clines. Princeton University Press, Princeton, NJ. Evans, L.M., Slavov, G.T., Rodgers-Melnick, E., Martin, J., Ranjan, P., Muchero, W., Brunner, A.M., Schackwitz, W., Gunter, L., Chen, J.G., Tuskan, G.A., and Difazio, S.P. 2014. Population genomics of Populus trichocarpa identifies signatures of selection and adaptive trait associations. Nat. Genet. 46(10): 1089\u20131096. Nature Publishing Group. doi:10.1038\/ng.3075. Excoffier, L., Foll, M., and Petit, R.J. 2009a. Genetic consequences of range expansions. Annu. Rev. Ecol. Evol. Syst. 40(May 2014): 481\u2013501. doi:10.1146\/annurev.ecolsys.39.110707.173414. Excoffier, L., Hofer, T., and Foll, M. 2009b. Detecting loci under selection in a hierarchically structured population. Heredity (Edinb). 103(4): 285\u2013298. Nature Publishing Group. doi:10.1038\/hdy.2009.74. Eyring, V., Bony, S., Meehl, G.A., Senior, C.A., Stevens, B., Stouffer, R.J., and Taylor, K.E. 2016. Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev. 9(5): 1937\u20131958. doi:10.5194\/gmd-9-1937-2016. Falkenhagen, E.R. 1985. Isozyme studies in provenance research of forest trees. Theor. Appl. Genet. 69(4): 335\u2013347. doi:10.1007\/BF00570897. Fang, J.Y., Chung, J. Der, Chiang, Y.C., Chang, C. Te, Chen, C.Y., and Hwang, S.Y. 2013. Divergent Selection and Local Adaptation in Disjunct Populations of an Endangered Conifer, Keteleeria davidiana var. formosana (Pinaceae). PLoS One 8(7): 16\u201323. doi:10.1371\/journal.pone.0070162. Farashi, A., and Alizadeh-Noughani, M. 2018. Effects of models and spatial resolutions on the species distribution model performance. Model. Earth Syst. Environ. 4(1): 263\u2013268. Springer International Publishing. doi:10.1007\/s40808-018-0422-4. Feder, J.L., Egan, S.P., and Nosil, P. 2012. The genomics of speciation-with-gene-flow. Trends Genet. 28(7): 342\u201350. Elsevier Ltd. doi:10.1016\/j.tig.2012.03.009. Felsenstein, J. 1976. The theoretical population genetics of variable selection and migration. Annu. Rev. Genet. 10: 253\u2013280. Feng, S., Ru, D., Sun, Y., Mao, K., Milne, R., and Liu, J. 2019. Trans-lineage polymorphism and nonbifurcating diversification of the genus Picea. New Phytol. 222(1): 576\u2013587. doi:10.1111\/nph.15590. Fick, S.E., and Hijmans, R.J. 2017. WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Int. J. Climatol. 37(12): 4302\u20134315. doi:10.1002\/joc.5086. Fitzpatrick, M.C., Chhatre, V.E., Soolanayakanahally, R.Y., and Keller, S.R. 2021. Experimental support for genomic prediction of climate maladaptation using the machine learning approach Gradient Forests. Mol. Ecol. Resour. 21(8): 2749\u20132765. doi:10.1111\/1755-0998.13374. Fitzpatrick, M.C., and Keller, S.R. 2015. Ecological genomics meets community-level modelling of biodiversity: Mapping the genomic landscape of current and future environmental adaptation. Ecol. Lett. 18(1): 1\u201316. doi:10.1111\/ele.12376. Fordham, D.A., Saltr\u00e9, F., Haythorne, S., Wigley, T.M.L., Otto-Bliesner, B.L., Chan, K.C., and Brook, B.W. 2017. 127  PaleoView: a tool for generating continuous climate projections spanning the last 21 000 years at regional and global scales. Ecography (Cop.). 40(11): 1348\u20131358. doi:10.1111\/ecog.03031. Forester, B.R., Lasky, J.R., Wagner, H.H., and Urban, D.L. 2018. Comparing methods for detecting multilocus adaptation with multivariate genotype\u2013environment associations. Mol. Ecol. 27(9): 2215\u20132233. doi:10.1111\/mec.14584. Fournier-Level, A., Korte, A., Cooper, M.D., Nordborg, M., Schmitt, J., and Wilczek, A.M. 2011. A map of local adaptation in Arabidopsis thaliana. Science (80-. ). 334: 86\u201390. Franklin, J. 2010. Mapping Species Distributions: Spatial Inference and Prediction. Cambridge University Press, Cambridge. Fraser, D.J., Weir, L.K., Bernatchez, L., Hansen, M.M., and Taylor, E.B. 2011. Extent and scale of local adaptation in salmonid fishes\u202f: Review and meta-analysis. Heredity (Edinb). 106(3): 404\u2013420. Nature Publishing Group. doi:10.1038\/hdy.2010.167. Fr\u00e9javille, T., Vizca\u00edno-Palomar, N., Fady, B., Kremer, A., and Benito Garz\u00f3n, M. 2020. Range margin populations show high climate adaptation lags in European trees. Glob. Chang. Biol. 26(2): 484\u2013495. doi:10.1111\/gcb.14881. Fujino, T., Katsushi, Y., Yokoyama, T.T., Hamanaka, T., Harazono, Y., Kamada, H., Kobayashi, W., Ujino-Ihara, T., Uchiyama, K., Matsumoto, A., Izuno, A., Tsumura, Y., Toyoda, A., Shigenobu, S., Moriguchi, Y., Ueno, S., and Kasahara, M. 2023. A chromosome-level genome assembly of a model conifer plant, the Japanese cedar, Cryptomeria japonica D. Don. bioRxiv: 2023.02.24.529822. Available from https:\/\/www.biorxiv.org\/content\/10.1101\/2023.02.24.529822v1%0Ahttps:\/\/www.biorxiv.org\/content\/10.1101\/2023.02.24.529822v1.abstract. Funk, J.L. 2008. Differences in plasticity between invasive and native plants from a low resource environment. J. Ecol. 96(6): 1162\u20131173. doi:10.1111\/j.1365-2745.2008.01435.x. Gagalova, K.K., Warren, R.L., Coombe, L., Wong, J., Nip, K.M., Yuen, M.M. Saint, Whitehill, J.G.A., Celedon, J.M., Ritland, C., Taylor, G.A., Cheng, D., Plettner, P., Hammond, S.A., Mohamadi, H., Zhao, Y., Moore, R.A., Mungall, A.J., Boyle, B., Laroche, J., Cottrell, J., Mackay, J.J., Lamothe, M., G\u00e9rardi, S., Isabel, N., Pavy, N., Jones, S.J.M., Bohlmann, J., Bousquet, J., and Birol, I. 2022. Spruce giga-genomes: structurally similar yet distinctive with differentially expanding gene families and rapidly evolving genes. Plant J. 111(5): 1469\u20131485. doi:10.1111\/tpj.15889. Garc\u00eda-Ramos, G., and Kirkpatrick, M. 1997. Genetic models of adaptation and gene flow in peripheral populations. Evolution (N. Y). 51(1): 21\u201328. doi:10.1111\/j.1558-5646.1997.tb02384.x. Garman, E.H. 1957. The occurrence of spruce in the interior of British Columbia. Victoria, BC. Gavin, D.G., Henderson, A.C.G., Westover, K.S., Fritz, S.C., Walker, I.R., Leng, M.J., and Sheng, F. 2011. Abrupt Holocene climate change and potential response to solar forcing in western Canada. Quat. Sci. Rev. 30: 1243\u20131255. Elsevier Ltd. doi:10.1016\/j.quascirev.2011.03.003. Gavin, D.G., Hu, F.S., Walker, I.R., and Westover, K. 2009. The northern inland temperate rainforest of british columbia: Old forests with a young history? Northwest Sci. 83(1): 70\u201378. doi:10.3955\/046.083.0107. Godbout, J., Jaramillo-Correa, J.P., Beaulieu, J., and Bousquet, J. 2005. A mitochondrial DNA minisatellite reveals the postglacial history of jack pine (Pinus banksiana), a broad-range North American conifer. Mol. Ecol. 14: 3497\u20133512. doi:10.1111\/j.1365-294X.2005.02674.x. 128  Gombiner, J. 2019. Post-glacial radiocarbon ages for the southern cordilleran ice sheet. Open Quat. 5(1): 1\u20136. doi:10.5334\/oq.55. Gompert, Z., and Buerkle, C.A. 2011. Bayesian estimation of genomic clines. Mol. Ecol. 20: 2111\u20132127. doi:10.1111\/j.1365-294X.2011.05074.x. Gompert, Z., and Buerkle, C.A. 2012. bgc: Software for Bayesian estimation of genomic clines. Mol. Ecol. Resour. 12(6): 1168\u201376. doi:10.1111\/1755-0998.12009.x. Goodsman, D.W., Cooke, B., Coltman, D.W., and Lewis, M.A. 2014. The genetic signature of rapid range expansions: How dispersal, growth and invasion speed impact heterozygosity and allele surfing. Theor. Popul. Biol. 98: 1\u201310. Elsevier Inc. doi:10.1016\/j.tpb.2014.08.005. Goring, S. 2012. Holocene climate history of British Columbia using pollen-based climate reconstruction techniques. Simon Fraser University. Goring, S., Lacourse, T., Pellatt, M.G., and Mathewes, R.W. 2013. Pollen assemblage richness does not reflect regional plant species richness: A cautionary tale. J. Ecol. 101(5): 1137\u20131145. doi:10.1111\/1365-2745.12135. Green, R.N., and Klinka, K. 1994. A field guise to site identification and interpretation for the Vancouver Forest Region. Victoria BC. Gr\u00f8ndahl, E., and Ehlers, B.K. 2008. Local adaptation to biotic factors: reciprocal transplants of four species associated with aromatic Thymus pulegioides and T. serpyllum. J. Ecol. 96: 981\u2013992. doi:10.1111\/j.1365-2745.2008.01407.x. Gross, B.L., and Rieseberg, L.H. 2005. The ecological genetics of homoploid hybrid speciation. J. Hered. 96(3): 241\u2013252. doi:10.1093\/jhered\/esi026. Gugger, P.F., Liang, C.T., Sork, V.L., Hodgskiss, P., and Wright, J.W. 2018. Applying landscape genomic tools to forest management and restoration of Hawaiian koa (Acacia koa) in a changing environment. Evol. Appl. 11(2): 231\u2013242. doi:10.1111\/eva.12534. Guillot, G., Vitalis, R., le Rouzic, A., and Gautier, M. 2014. Detecting correlation between allele frequencies and environmental variables as a signature of selection. A fast computation approach for genome-wide studies. Spat. Stat. 8: 145\u2013155. Available from http:\/\/arxiv.org\/abs\/1206.0889. G\u00fcnther, T., and Coop, G. 2013. Robust identification of local adaptation from allele frequencies. Genetics 195(1): 205\u2013220. doi:10.1534\/genetics.113.152462. Guttman, D.S., and Dykhuizen, D.E. 1994. Detecting selective sweeps in naturally occurring Escherichia coli. Genetics 138(4): 993\u20131003. Hagen, D.W. 1967. Isolating mechanisms in threespine sticklebacks (Gasterosteus). J. Fish. Res. Board Canada 24(8): 1637\u20131692. Haldane, J.B.S. 1948. The theory of a cline. J. Genet. 48(3): 277\u2013284. doi:10.1007\/BF02986626. Hamann, A., and Wang, T. 2006. Potential Effects of climate chane on ecosystem. Ecol. Soc. Am. 87(11): 2773\u20132786. Hamilton, J.A. 2012. Genomic and phenotypic architecture of a spruce hybrid zone. University of British Columbia. Hamilton, J.A., and Aitken, S.N. 2013. Genetic and morphological structure of a spruce hybrid (Picea sitchensis x P. 129  glauca) zone along a climatic gradient. Am. J. Bot. 100(8): 1651\u201362. doi:10.3732\/ajb.1200654. Hamilton, J.A., De la Torre, A.R., and Aitken, S.N. 2015. Fine-scale environmental variation contributes to introgression in a three-species spruce hybrid complex. Tree Genet. Genomes 11(1): 817. doi:10.1007\/s11295-014-0817-y. Hamilton, J.A., Lexer, C., and Aitken, S.N. 2013. Genomic and phenotypic architecture of a spruce hybrid zone (Picea sitchensis \u00d7 P. glauca). Mol. Ecol. 22(3): 827\u2013841. doi:10.1111\/mec.12007. Hamrick, J.L., Schnabel, A.F., and Wells, P. V. 1994. Distribution of genetic diversity within and among populations of Great Basin conifers. In Natural history of the Colorado Plateau and Great Basin. Edited by K.T. Harper, L.L. St.clair, K.H. Thorne, and W.M. Hess. University Press of Colorado, Niwot, CO. pp. 147\u2013162. doi:10.2307\/3985651. Hannerz, M., Sonesson, J., and Ekberg, I. 1999. Genetic correlations between growth and growth rhythm observed in a short-term test and performance in long-term field trials of Norway spruce. Can. J. For. Res. 29(6): 768\u2013778. doi:10.1139\/x99-056. Harrison, R.G. 1986. Pattern and process in a narrow hybrid zone. Heredity (Edinb). 56(3): 337\u2013349. doi:10.1038\/hdy.1986.55. Harrison, R.G. 1990. Hybrid zones: Windows on evolutionary process. Oxford Surv. Evol. Biol. 7: 69\u2013128. Haselhorst, M.S.H., and Buerkle, C.A. 2013. Population genetic structure of Picea engelmannii, P. glauca and their previously unrecognized hybrids in the central Rocky Mountains. Tree Genet. Genomes 9(3): 669\u2013681. doi:10.1007\/s11295-012-0583-7. Hereford, J. 2009. A quantitative survey of local adaptation and fitness trade\u2010offs. Am. Nat. 173(5): 579\u2013588. doi:10.1086\/597611. Hewitt, G.M. 1988. Hybrid zones - natural laboratories for evolutionary studies. Trends Ecol. Evol. 3(7): 158\u2013167. Hoban, S., Kelley, J.L., Lotterhos, K.E., Antolin, M.F., Bradburd, G., Lowry, D.B., Poss, M.L., Reed, L.K., Storfer, A., and Whitlock, M.C. 2016. Finding the genomic basis of local adaptation: Pitfalls, practical solutions, and future directions. Am. Nat. 188(4): 379\u2013397. doi:10.1086\/688018. Hoeksema, J.D., and Forde, S.E. 2008. A meta-analysis of factors affecting local adaptation between interacting species. Am. Nat. 171(3): 275\u2013290. doi:10.1086\/527496. Hogg, E.H., Michaelian, M., Hook, T.I., and Undershultz, M.E. 2017. Recent climatic drying leads to age-independent growth reductions of white spruce stands in western Canada. Glob. Chang. Biol. 23(12): 5297\u20135308. doi:10.1111\/gcb.13795. Holliday, J.A., Ralph, S.G., White, R., Bohlmann, J., and Aitken, S.N. 2008. Global monitoring of autumn gene expression within and among phenotypically divergent populations of Sitka spruce (Picea sitchensis). New Phytol. 178(1): 103\u2013122. doi:10.1111\/j.1469-8137.2007.02346.x. Hong, E.P., and Park, J.W. 2012. Sample size and statistical power calculation in genetic association studies. Genomics Inform. 10(2): 117. doi:10.5808\/GI.2012.10.2.117. Hope, G.D., Lloyd, D.A., Mitchell, W.R., Erickson, W.R., Harper, W.L., and Wikeem, B.M. 1991. Interior Douglas-fir zone. In Ecosystems of British Columbia. Edited by D. Meidinger and J. Pojar. BC Ministry of Forests, Victoria BC. Available from http:\/\/www.jasso.go.jp\/ryugaku\/related\/kouryu\/2017\/__icsFiles\/afieldfile\/2017\/05\/08\/201705shihominorikawa130  .pdf. Hornoy, B., Pavy, N., G\u00e9rardi, S., Beaulieu, J., and Bousquet, J. 2015. Genetic adaptation to climate in white spruce involves small to moderate allele frequency shifts in functionally diverse genes. Genome Biol. Evol. 7(12): 3269\u20133285. doi:10.1093\/gbe\/evv218. Howe, G.T., Aitken, S.N., Neale, D.B., Jermstad, K.D., Wheeler, N.C., and Chen, T.H. 2003. From genotype to phenotype: unraveling the complexities of cold adaptation in forest trees. Can. J. Bot. 81(12): 1247\u20131266. doi:10.1139\/b03-141. Hunt, W.G., and Selander, R.K. 1973. Biochemical genetics of hybridisation in European house mice. Heredity (Edinb). 31(1): 11\u201333. doi:10.1038\/hdy.1973.56. Hutchinson, G.E. 1978. An Introduction to Population Biology. Yale University Press, New Haven, CT. Huxley, J.S. 1939. Clines\u202f: An auxiliary method in taxonomy. Bijdr. tot Dierkd. 27(5): 491\u2013520. IPCC. 2021. Climate Change 2021: The Physical Science Basis. Contributions of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Edited ByV. [Masson-Delmotte, P. Zhai, A. Pirani, S.L. Connors, C. P\u00e9an, S. Berger, N. Caud, Y. Chen, L. Goldfarb, M.I. Gomis, M. Huang, K. Leitzell, E. Lonnoy, J.B.R. Matthews, T.K. Maycock, T. Waterfield, O. Yelek\u00e7i, R. Yu, and B. Zhou. Cambridge University Press, Cambridge. Isabel, N., Holliday, J.A., and Aitken, S.N. 2020. Forest genomics: Advancing climate adaptation, forest health, productivity, and conservation. Evol. Appl. 13(1): 3\u201310. doi:10.1111\/eva.12902. Ivanusic, A., Wood, L.J., and Lewis, K. 2020. Structural carbon allocation and wood growth reflect climate variation in stands of hybrid white spruce in central interior British Columbia, Canada. Forests 11(8). doi:10.3390\/f11080879. Jackman, S.D., Warren, R.L., Gibb, E.A., Vandervalk, B.P., Mohamadi, H., Chu, J., Raymond, A., Pleasance, S., Coope, R., Wildung, M.R., Ritland, C.E., Bousquet, J., Jones, S.J.M.M., Bohlmann, J., and Birol, I. 2016. Organellar genomes of white spruce (picea glauca): Assembly and annotation. Genome Biol. Evol. 8(1): 29\u201341. doi:10.1093\/gbe\/evv244. Jaramillo-Correa, J.P., Beaulieu, J., and Bousquet, J. 2001. Contrasting evolutionary forces driving population structure at expressed sequence tag polymorphisms, allozymes and quantitative traits in white spruce. Mol. Ecol. 10(11): 2729\u20132740. doi:10.1046\/j.0962-1083.2001.01386.x. Jaramillo-Correa, J.P., Beaulieu, J., and Bousquet, J. 2004. Variation in mitochondrial DNA reveals multiple distant glacial refugia in black spruce (Picea mariana), a transcontinental North American conifer. Mol. Ecol. 13: 2735\u20132747. doi:10.1111\/j.1365-294X.2004.02258.x. Jaramillo-Correa, J.P., Beaulieu, J., Ledig, F.T., and Bousquet, J. 2006. Decoupled mitochondrial and chloroplast DNA population structure reveals Holocene collapse and population isolation in a threatened Mexican-endemic conifer. Mol. Ecol. 15(10): 2787\u20132800. doi:10.1111\/j.1365-294X.2006.02974.x. Jiggins, C.D., and Mallet, J. 2000. Bimodal hybrid zones and speciation. Trends Ecol. Evol. 15(6): 250\u2013255. doi:10.1016\/S0169-5347(00)01873-5. Jim\u00e9nez, L., Sober\u00f3n, J., Christen, J.A., and Soto, D. 2019. On the problem of modeling a fundamental niche from occurrence data. Ecol. Modell. 397(May 2018): 74\u201383. Elsevier. doi:10.1016\/j.ecolmodel.2019.01.020. Jones, F.C., Brown, C., Pemberton, J.M., and Braithwaite, V.A. 2006. Reproductive isolation in a threespine 131  stickleback hybrid zone. J. Evol. Biol. 19(5): 1531\u20131544. doi:10.1111\/j.1420-9101.2006.01122.x. Jones, M.R., and Good, J.M. 2016. Targeted capture in evolutionary and ecological genomics. Mol. Ecol. 25(1): 185\u2013202. doi:10.1111\/mec.13304. Kassen, R. 2002. The experimental evolution of specialists, generalists, and the maintenance of diversity. J. Evol. Biol. 15(2): 173\u2013190. doi:10.1046\/j.1420-9101.2002.00377.x. Kawecki, T.J., and Ebert, D. 2004. Conceptual issues in local adaptation. Ecol. Lett. 7(12): 1225\u20131241. doi:10.1111\/j.1461-0248.2004.00684.x. Kerner von Marilaun, A. 1895. Dependence of plant form on soil and climate. In The Natural History of Plants. Edited by F.W. Oliver. Blackie, London. pp. 495\u2013514. Key, K.H.L. 1968. The concept of stasipatric speciation. Syst. Zool. 17(1): 14\u201322. Kim, E., and Donohue, K. 2013. Local adaptation and plasticity of Erysimum capitatum to altitude: Its implications for responses to climate change. J. Ecol. 101(3): 796\u2013805. doi:10.1111\/1365-2745.12077. Kolb, T.E., Grady, K.C., McEttrick, M.P., and Herrero, A. 2016. Local-scale drought adaptation of ponderosa pine seedlings at habitat ecotones. For. Sci. 62(6): 641\u2013651. doi:10.5849\/forsci.16-049. Kopp, M., and Matuszewski, S. 2014. Rapid evolution of quantitative traits: Theoretical perspectives. Evol. Appl. 7(1): 169\u2013191. doi:10.1111\/eva.12127. Kremer, A., Ronce, O., Robledo-Arnuncio, J.J., Guillaume, F., Bohrer, G., Nathan, R., Bridle, J.R., Gomulkiewicz, R., Klein, E.K., Ritland, K., Kuparinen, A., Gerber, S., and Schueler, S. 2012. Long-distance gene flow and adaptation of forest trees to rapid climate change. Ecol. Lett. 15(4): 378\u2013392. doi:10.1111\/j.1461-0248.2012.01746.x. Kriticos, D.J., and Leriche, A. 2010. The effects of climate data precision on fitting and projecting species niche models. Ecography (Cop.). 33(1): 115\u2013127. doi:10.1111\/j.1600-0587.2009.06042.x. Kruuk, L.E.B., Baird, S.J.E., Gale, K.S., and Barton, N.H. 1999. A comparison of multilocus clines maintained by environmental adaptation or by selection against hybrids. Genetics 153(4): 1959\u20131971. Kueppers, L.M., Conlisk, E., Castanha, C., Moyes, A.B., Germino, M.J., de Valpine, P., Torn, M.S., and Mitton, J.B. 2017. Warming and provenance limit tree recruitment across and beyond the elevation range of subalpine forest. Glob. Chang. Biol. 23(6): 2383\u20132395. doi:10.1111\/gcb.13561. De La Torre, A., Ingvarsson, P.K., and Aitken, S.N. 2015. Genetic architecture and genomic patterns of gene flow between hybridizing species of Picea. Heredity (Edinb). 115(2): 153\u2013164. Nature Publishing Group. doi:10.1038\/hdy.2015.19. De La Torre, A.R., Birol, I., Bousquet, J., Ingvarsson, P.K., Jansson, S., Jones, S.J.M., Keeling, C.I., MacKay, J., Nilsson, O., Ritland, K., Street, N., Yanchuk, A., Zerbe, P., and Bohlmann, J. 2014a. Insights into conifer giga-genomes. Plant Physiol. 166(4): 1724\u20131732. doi:10.1104\/pp.114.248708. De La Torre, A.R., Roberts, D.R., and Aitken, S.N. 2014b. Genome-wide admixture and ecological niche modelling reveal the maintenance of species boundaries despite long history of interspecific gene flow. Mol. Ecol. 23(8): 2046\u20132059. Blackwell Publishing Ltd. doi:10.1111\/mec.12710. De La Torre, A.R., Wang, T., Jaquish, B., and Aitken, S.N. 2014c. Adaptation and exogenous selection in a Picea glauca x Picea engelmannii hybrid zone: implications for forest management under climate change. New 132  Phytol. 201(2): 687\u2013699. de Lafontaine, G., Turgeon, J., and Payette, S. 2010. Phylogeography of white spruce (Picea glauca) in eastern North America reveals contrasting ecological trajectories. J. Biogeogr. 37(4): 741\u2013751. doi:10.1111\/j.1365-2699.2009.02241.x. Langlet, O. 1971. Two hundred years genecology. Taxon 20(5): 653\u2013721. Laroche, J., Li, P., and Bousquet, J. 1995. Mitochondrial DNA and Monocot-Dicot Divergence Time. Larson, E.L., Andr\u00e9s, J.A., Bogdanowicz, S.M., and Harrison, R.G. 2013. Differential introgression in a mosaic hybrid zone reveals candidate barrier genes. Evolution (N. Y). 67(12): 3653\u20133661. doi:10.1111\/evo.12205. Latta, R.G., Linhart, Y.B., Fleck, D., and Elliot, M. 1998. Direct and indirect estimates of seed versus pollen movement within a population of Ponderosa pine. Evolution (N. Y). 52(1): 61\u201367. Lawson, D.J., van Dorp, L., and Falush, D. 2018. A tutorial on how not to over-interpret STRUCTURE and ADMIXTURE bar plots. Nat. Commun. 9(1): 1\u201311. Springer US. doi:10.1038\/s41467-018-05257-7. Ledig, F.T., Hodgskiss, P.D., and Johnson, D.R. 2006. The structure of genetic diversity in Engelmann spruce and a comparison with blue spruce. Can. J. Bot. 84(12): 1806\u20131828. doi:10.1139\/b06-106. Leimu, R., and Fischer, M. 2008. A meta-analysis of local adaptation in plants. PLoS One 3(12): 1\u20138. doi:10.1371\/journal.pone.0004010. Leites, L., and Benito, M. 2023. Forest tree species adaptation to climate across biomes\u202f: Building on the legacy of ecological genetics to anticipate responses to climate change. Glob. Chang. Biol.: 1\u201320. doi:10.1111\/gcb.16711. Lenormand, T. 2002. Gene flow and the limits to natural selection. Trends Ecol. Evol. 17(4): 183\u2013189. doi:10.1016\/S0169-5347(02)02497-7. Lesser, M.R., and Parker, W.H. 2004. Genetic variation in Picea glauca for growth and phenological traits from provenance tests in Ontario. Silvae Genet. 53(4): 141\u2013148. doi:none. Li, P., Beaulieu, J., and Bousquet, J. 1997. Genetic structure and patterns of genetic variation among populations in eastern white spruce (Picea glauca). 198: 189\u2013198. Liepe, K.J., Hamann, A., Smets, P., Fitzpatrick, C.R., and Aitken, S.N. 2016. Adaptation of lodgepole pine and interior spruce to climate: Implications for reforestation in a warming world. Evol. Appl. 9(2): 409\u2013419. doi:10.1111\/eva.12345. Liepelt, S., Bialozyt, R., and Ziegenhagen, B. 2002. Wind-dispersed pollen mediates postglacial gene flow among refugia. Proc. Natl. Acad. Sci. 99(22): 14590\u201314594. Lind, B.M., Friedline, C.J., Wegrzyn, J.L., Maloney, P.E., Vogler, D.R., Neale, D.B., and Eckert, A.J. 2017. Water availability drives signatures of local adaptation in whitebark pine (Pinus albicaulis Engelm.) across fine spatial scales of the Lake Tahoe Basin, USA. Mol. Ecol. 26(12): 3168\u20133185. doi:10.1111\/mec.14106. Lind, B.M., Menon, M., Bolte, C.E., Faske, T.M., and Eckert, A.J. 2018. The genomics of local adaptation in trees: are we out of the woods yet? Tree Genet. Genomes 14(2). Tree Genetics & Genomes. doi:10.1007\/s11295-017-1224-y. Lindtke, D., Gompert, Z., Lexer, C., and Buerkle, C.A. 2014. Unexpected ancestry of Populus seedlings from a 133  hybrid zone implies a large role for postzygotic selection in the maintenance of species. Mol. Ecol. 23(17): 4316\u20134330. doi:10.1111\/mec.12759. Linhart, Y.B., and Grant, M.C. 1996. Evolutionary significance of local genetic differentation in plants. Annu. Rev. Ecol. Syst. 27(1996): 237\u2013277. Little, E. 1971. Atlas of United States Trees Volume 1: Conifers and Important Hardwoods. USDA Forest Service, Washington. Little, E.L. 1953. A natural hybrid species in Alaska. J. For. 41: 745\u2013746. Liu, Z., Otto-Bliesner, B.L., He, F., Brady, E.C., Tomas, R., Clark, P.U., Carlson, A.E., Lynch-Stieglitz, J., Curry, W., Brook, E., Erickson, D., Jacob, R., Kutzbach, J., and Cheng, J. 2009. Transient simulation of last deglaciation with a new mechanism for bolling-allerod warming. Science (80-. ). 325(5938): 310\u2013314. doi:10.1126\/science.1171041. Lockwood, J.D., Aleksi\u0107, J.M., Zou, J., Wang, J., Liu, J., and Renner, S.S. 2013. A new phylogeny for the genus Picea from plastid, mitochondrial, and nuclear sequences. Mol. Phylogenet. Evol. 69(3): 717\u2013727. doi:10.1016\/j.ympev.2013.07.004. Lorenz, D.J., Nieto-Lugilde, D., Blois, J.L., Fitzpatrick, M.C., and Williams, J.W. 2016. Downscaled and debiased climate simulations for North America from 21,000 years ago to 2100AD. Sci. Data 3: 1\u201319. doi:10.1038\/sdata.2016.48. Lotterhos, K.E., and Whitlock, M.C. 2014. Evaluation of demographic history and neutral parameterization on the performance of FST outlier tests. Mol. Ecol. 23(9): 2178\u20132192. doi:10.1111\/mec.12725. Lotterhos, K.E., and Whitlock, M.C. 2015. The relative power of genome scans to detect local adaptation depends on sampling design and statistical method. Mol. Ecol. 24: 1031\u20131046. doi:10.1111\/mec.13100. Lu, P., Parker, W.C., Colombo, S.J., and Man, R. 2016. Restructuring tree provenance test data to conform to reciprocal transplant experiments for detecting local adaptation. J. Appl. Ecol. 53(4): 1088\u20131097. doi:10.1111\/1365-2664.12647. Lu, P., Parker, W.H., Cherry, M., Colombo, S., Parker, W.C., Man, R., and Roubal, N. 2014. Survival and growth patterns of white spruce (Picea glauca [Moench] Voss) rangewide provenances and their implications for climate change adaptation. Ecol. Evol. 4(12): 2360\u20132374. doi:10.1002\/ece3.1100. MacKenzie, W.H., and Mahony, C.R. 2021. An ecological approach to climate change-informed tree species selection for reforestation. For. Ecol. Manage. 481(November 2020): 118705. Elsevier B.V. doi:10.1016\/j.foreco.2020.118705. MacKenzie, W.H., and Meidinger, D. V. 2018. The biogeoclimatic ecosystem classification approach: An ecological framework for vegetation classification. Phytocoenologia 48(2): 203\u2013213. doi:10.1127\/phyto\/2017\/0160. MacLachlan, I.R., Yeaman, S., and Aitken, S.N. 2018. Growth gains from selective breeding in a spruce hybrid zone do not compromise local adaptation to climate. Evol. Appl. 11(2): 166\u2013181. doi:10.1111\/eva.12525. M\u00e4gi, M., Semchenko, M., Kalamees, R., and Zobel, K. 2011. Limited phenotypic plasticity in range-edge populations: A comparison of co-occurring populations of two Agrimonia species with different geographical distributions. Plant Biol. 13(1): 177\u2013184. doi:10.1111\/j.1438-8677.2010.00342.x. Mahony, C.R., Wang, T., Hamann, A., and Cannon, A.J. 2022. A global climate model ensemble for downscaled monthly climate normals over North America. Int. J. Climatol. 42(11): 5871\u20135891. doi:10.1002\/joc.7566. 134  Mallet, J. 2007. Hybrid speciation. Nature 446(7133): 279\u2013283. doi:10.1038\/nature05706. Mar\u00e9chal, E. 2018. Primary endosymboisis: Emergence of the primary chloroplast and the chromatophore, two independent events. In Plastids: Methods and Protocols. Edited by E. Mar\u00e9chal. Springer, Dordrecht. pp. 3\u201316. Mathewes, R.W., and Clague, J.J. 2017. Paleoecology and ice limits of the early Fraser glaciation (Marine Isotope Stage 2) on Haida Gwaii, British Columbia, Canada. Quat. Res. 88: 277\u2013292. doi:10.1017\/qua.2017.36. Matyas, C. 1994. Modeling climate change effects with provenance test data. Tree Physiol. 14(7\u20139): 797\u2013804. doi:10.1093\/treephys\/14.7-8-9.797. M\u00e1ty\u00e1s, C. 1996. Climatic adaptation of trees: rediscovering provenance tests. Euphytica 92(1): 45\u201354. doi:10.1007\/BF00022827. Mayr, E. 1942. Systematics and the Origin of Species. Columbia University Press, New York. Mccain, C.M., and Colwell, R.K. 2011. Assessing the threat to montane biodiversity from discordant shifts in temperature and precipitation in a changing climate. Ecol. Lett. 14(12): 1236\u20131245. doi:10.1111\/j.1461-0248.2011.01695.x. Mckay, J.K., and Latta, R.G. 2002. Adaptive population divergence: Markers, QTL and traits. Trends Ecol. Evol. 17(6): 285\u2013291. McKenzie, D., Peterson, D.W., and Peterson, D.L. 2003. Modelling conifer species distributions in mountain forests of Washington State, USA. For. Chron. 79(2): 253\u2013258. doi:10.5558\/tfc79253-2. McKinnon, J.S., and Rundle, H.D. 2002. Speciation in nature: The threespine stickleback model systems. Trends Ecol. Evol. 17(10): 480\u2013488. doi:10.1016\/S0169-5347(02)02579-X. Medina, P., Thornlow, B., Nielsen, R., and Corbett-Detig, R. 2018. Estimating the timing of multiple admixture pulses during local ancestry inference. Genetics 210(3): 1089\u20131107. doi:10.1534\/genetics.118.301411. Meek, M.H., and Larson, W.A. 2019. The future is now: Amplicon sequencing and sequence capture usher in the conservation genomics era. Mol. Ecol. Resour. (January): 795\u2013803. doi:10.1111\/1755-0998.12998. Meirmans, P.G., and Hedrick, P.W. 2011. Assessing population structure: FST and related measures. Mol. Ecol. Resour. 11(1): 5\u201318. doi:10.1111\/j.1755-0998.2010.02927.x. Meisner, J., and Albrechtsen, A. 2018. Inferring population structure and admixture proportions in low-depth NGS data. Genetics 210(2): 719\u2013731. doi:10.1534\/genetics.118.301336. Meril\u00e4, J., and Hendry, A.P. 2014. Climate change, adaptation, and phenotypic plasticity: The problem and the evidence. Evol. Appl. 7(1): 1\u201314. doi:10.1111\/eva.12137. Milewska, E.J., Hopkinson, R.F., and Niitsoo, A. 2005. Evaluation of geo-referenced grids of 1961\u20131990 canadian temperature and precipitation normals. Atmos. - Ocean 43(1): 49\u201375. doi:10.3137\/ao.430104. Mimura, M., and Aitken, S.N. 2007. Adaptive gradients and isolation-by-distance with postglacial migration in Picea sitchensis. Heredity (Edinb). 99(2): 224\u2013232. doi:10.1038\/sj.hdy.6800987. Minder, J.R., Mote, P.W., and Lundquist, J.D. 2010. Surface temperature lapse rates over complex terrain: Lessons from the Cascade Mountains. J. Geophys. Res. Atmos. 115(14): 1\u201313. doi:10.1029\/2009JD013493. 135  Mitton, J.B., and Williams, C.G. 2006. Gene Flow in Conifers. In Landscapes, Genomics, and Transgenic Conifers. Edited by C.G. Williams. Springer, Dordrecht. pp. 147\u2013168. doi:10.1007\/1-4020-3869-0. Molina-Montenegro, M.A., and Naya, D.E. 2012. Latitudinal patterns in phenotypic plasticity and fitness-related traits: assessing the climatic variability hypothesis (CVH) with an invasive plant species. PLoS One 7(10): 23\u201328. doi:10.1371\/journal.pone.0047620. Montw\u00e9, D., Isaac-Renton, M., Hamann, A., and Spiecker, H. 2018. Cold adaptation recorded in tree rings highlights risks associated with climate change and assisted migration. Nat. Commun. 9(1): 1\u20137. Springer US. doi:10.1038\/s41467-018-04039-5. Moore, W.S. 1977. An evaluation of narrow hybrid zones in vertebrates. Q. Rev. Biol. 52(3): 263\u2013277. Mouton, A.M., Baets, B. De, and Goethals, P.L.M. 2010. Ecological relevance of performance criteria for species distribution models. Ecol. Modell. 221: 1995\u20132002. doi:10.1016\/j.ecolmodel.2010.04.017. Mower, J.P., Touzet, P., Gummow, J.S., Delph, L.F., and Palmer, J.D. 2007. Extensive variation in synonymous substitution rates in mitochondrial genes of seed plants. BMC Evol. Biol. 7(1): 135. doi:10.1186\/1471-2148-7-135. Namroud, M.C., Beaulieu, J., Juge, N., Laroche, J., and Bousquet, J. 2008. Scanning the genome for gene single nucleotide polymorphisms involved in adaptive population differentiation in white spruce. Mol. Ecol. 17(16): 3599\u20133613. doi:10.1111\/j.1365-294X.2008.03840.x. Naudin, C., and Radlkofer, L. 1876. Recherches au sujet des influences que les changement de climats exercent sur les plantes. In Annales Des Sciences Naturelles Sixi\u00e8me S\u00e9rie; Botanique. Edited ByG. Masson. Libraire de L\u2019Acad\u00e9mie de M\u00e9decine, Paris. Neale, D.B., and Sederoff, R.R. 1989. Paternal inheritance of chloroplast DNA and maternal inheritance of mitochondrial DNA in loblolly pine. Theor. Appl. Genet. 77: 212\u2013216. Nelson, K.N., O\u2019Dean, E., Knapp, E.E., Parker, A.J., and Bisbing, S.M. 2021. Persistent yet vulnerable: resurvey of an Abies ecotone reveals few differences but vulnerability to climate change. Ecology 102(12): 1\u201316. doi:10.1002\/ecy.3525. Newsome, R.D. 1963. A Phytosociological study of the woody, forest vegetation of the Cypress Hills. University of Saskatchewan. Nicotra, A.B., Atkin, O.K., Bonser, S.P., Davidson, A.M., Finnegan, E.J., Mathesius, U., Poo, P., Purugganan, M.D., Richards, C.L., Valladares, F., and van Kleunen, M. 2010. Plant phenotypic plasticity in a changing climate. Trends Plant Sci. 12: 684\u2013692. doi:10.2478\/v10183-010-0006-0. Nienstaedt, H., and Zasada, J.C. 1990. Picea glauca (Moench) Voss. white spruce. In Silvics of North America Volume 1: Conifers. pp. 204\u2013226. Niu, S., Li, J., Bo, W., Yang, W., Zuccolo, A., Giacomello, S., Chen, X., Han, F., Yang, J., Song, Y., Nie, Y., Zhou, B., Wang, P., Zuo, Q., Zhang, H., Ma, J., Wang, J., Wang, L., Zhu, Q., Zhao, H., Liu, Z., Zhang, X., Liu, T., Pei, S., Li, Z., Hu, Y., Yang, Y., Li, W., Zan, Y., Zhou, L., Lin, J., Yuan, T., Li, W., Li, Y., Wei, H., and Wu, H.X. 2022. The Chinese pine genome and methylome unveil key features of conifer evolution. Cell 185(1): 204\u2013217.e14. The Authors. doi:10.1016\/j.cell.2021.12.006. Nosil, P. 2010. Ecological speciation. Oxford University Press, Oxford. O\u2019Neill, G.A., Aitken, S.N., King, J.N., and Alfaro, R.I. 2002. Geographic variation in resin canal defenses in 136  seedlings from the Sitka spruce \u00d7 white spruce introgression zone. Can. J. For. Res. 32(3): 390\u2013400. doi:10.1139\/x01-206. O\u2019Neill, G.A., Stoehr, M., and Jaquish, B. 2014. Forest ecology and management quantifying safe seed transfer distance and impacts of tree breeding on adaptation. For. Ecol. Manage. 328: 122\u2013130. Elsevier B.V. doi:10.1016\/j.foreco.2014.05.039. Owens, G.L., Huang, K., Todesco, M., and Rieseberg, L.H. 2023. Re-evaluating Homoploid Reticulate Evolution in Helianthus Sunflowers. Mol. Biol. Evol. 40(2): 1\u201316. Oxford University Press. doi:10.1093\/molbev\/msad013. Pavy, N., Namroud, M.C., Gagnon, F., Isabel, N., and Bousquet, J. 2012. The heterogeneous levels of linkage disequilibrium in white spruce genes and comparative analysis with other conifers. Heredity (Edinb). 108(3): 273\u2013284. Nature Publishing Group. doi:10.1038\/hdy.2011.72. Pavy, N., Paule, C., Parsons, L., Crow, J.A., Morency, M.J., Cooke, J., Johnson, J.E., Noumen, E., Guillet-Claude, C., Butterfield, Y., Barber, S., Yang, G., Liu, J., Stott, J., Kirkpatrick, R., Siddiqui, A., Holt, R., Marra, M., Seguin, A., Retzel, E., Bousquet, J., and MacKay, J. 2005. Generation, annotation, analysis and database integration of 16,500 white spruce EST clusters. BMC Genomics 6: 1\u201319. doi:10.1186\/1471-2164-6-144. Payseur, B.A. 2010. Using differential introgression in hybrid zones to identify genomic regions involved in speciation. Mol. Ecol. Resour. 10(5): 806\u2013820. doi:10.1111\/j.1755-0998.2010.02883.x. Pelgas, B., Bousquet, J., Meirmans, P.G., Ritland, K., and Isabel, N. 2011. QTL mapping in white spruce: Gene maps and genomic regions underlying adaptive traits across pedigrees, years and environments. BMC Genomics 12. doi:10.1186\/1471-2164-12-145. Peterson, A.T., Cobos, M.E., and Jim\u00e9nez-Garc\u00eda, D. 2018. Major challenges for correlational ecological niche model projections to future climate conditions. Ann. N. Y. Acad. Sci. 1429(1): 66\u201377. doi:10.1111\/nyas.13873. Petit, R.J., Duminil, J., Fineschi, S., Hampe, A., Salvini, D., and Vendramin, G.G. 2005. Comparative organization of chloroplast, mitochondrial and nuclear diversity in plant populations. Mol. Ecol. 14(3): 689\u2013701. doi:10.1111\/j.1365-294X.2004.02410.x. Pile, L.S., Meyer, M.D., Rojas, R., Roe, O., and Smith, M.T. 2019. Drought impacts and compounding mortality on forest trees in the southern sierra nevada. Forests 10(3). doi:10.3390\/f10030237. Pritchard, J.K., Stephens, M., and Donnelly, P. 2000. Inference of population structure using multilocus genotype data. Genetics 155: 945\u2013959. doi:10.1111\/j.1471-8286.2007.01758.x. Prunier, J., Caron, S., and MacKay, J. 2017. CNVs into the wild: screening the genomes of conifer trees (Picea spp.) reveals fewer gene copy number variations in hybrids and links to adaptation. BMC Genomics 18(1): 97. BMC Genomics. doi:10.1186\/s12864-016-3458-8. Pulkkinen, O., and Metzler, R. 2013. Distance matters: The impact of gene proximity in bacterial gene regulation. Phys. Rev. Lett. 110(19): 1\u20135. doi:10.1103\/PhysRevLett.110.198101. Puritz, J.B., and Lotterhos, K.E. 2018. Expressed exome capture sequencing: A method for cost-effective exome sequencing for all organisms. Mol. Ecol. Resour. 18(6): 1209\u20131222. doi:10.1111\/1755-0998.12905. Pyh\u00e4j\u00e4rvi, T., Salmela, M.J., and Savolainen, O. 2008. Colonization routes of Pinus sylvestris inferred from distribution of mitochondrial DNA variation. Tree Genet. Genomes 4: 247\u2013254. doi:10.1007\/s11295-007-0105-1. 137  Rajora, O.P., and Dancik, B.P. 2000. Population genetic variation, structure, and evolution in Engelmann spruce, white spruce, and their natural hybrid complex in Alberta. Can. J. Bot. 78(6): 768\u2013780. Ran, J.H., Shen, T.T., Liu, W.J., Wang, P.P., and Wang, X.Q. 2015. Mitochondrial introgression and complex biogeographic history of the genus Picea. Mol. Phylogenet. Evol. 93: 63\u201376. Elsevier Inc. doi:10.1016\/j.ympev.2015.07.020. Ran, J.H., Wei, X.X., and Wang, X.Q. 2006. Molecular phylogeny and biogeography of Picea (Pinaceae): Implications for phylogeographical studies using cytoplasmic haplotypes. Mol. Phylogenet. Evol. 41(2): 405\u2013419. doi:10.1016\/j.ympev.2006.05.039. Rand, D.M., and Harrison, R.G. 1989. Ecological genetics of a mosaic hybrid zone: Mitochondrial, nuclear, and reproductive differentiation of crickets by soil type. Evolution (N. Y). 43(2): 432\u2013449. R\u00e4ty, O., R\u00e4is\u00e4nen, J., and Ylh\u00e4isi, J.S. 2014. Evaluation of delta change and bias correction methods for future daily precipitation: Intermodel cross-validation using ENSEMBLES simulations. Clim. Dyn. 42(9\u201310): 2287\u20132303. doi:10.1007\/s00382-014-2130-8. Rausher, M.D., and Delph, L.F. 2015. Commentary\u202f: When does understanding phenotypic evolution require identification of the underlying genes? Evolution (N. Y). 69(7): 1655\u20131664. doi:10.1111\/evo.12687. Reboud, X., and Zeyl, C. 1994. Organelle inheritance in plants. Heredity (Edinb). 72(2): 132\u2013140. doi:10.1038\/hdy.1994.19. Rehfeldt, G.E. 1994. Adaptation of Picea engelmannii populations to the heterogeneous environments of the intermountain west. Can. J. Bot. 72(8): 1197\u20131208. doi:10.1139\/b94-146. Rehfeldt, G.E. 2004. Interspecific and intraspecific variation in Picea engelmannii and its congeneric cohorts: biosystematics, genecology, and climate change. USDA Forest Service, Fort Collins. doi:Genetic Considerations in Ecological Restoration. Rellstab, C., Gugerli, F., Eckert, A.J., Hancock, A.M., and Holderegger, R. 2015. A practical guide to environmental association analysis in landscape genomics. Mol. Ecol. 24(17): 4348\u20134370. doi:10.1111\/mec.13322. Renwick, K.M., Rocca, M.E., and Stohlgren, T.J. 2016. Biotic disturbance facilitates range shift at the trailing but not the leading edge of lodgepole pine\u2019s altitudinal distribution. J. Veg. Sci. 27(4): 780\u2013788. doi:10.1111\/jvs.12410. Richardson, A.D., Keenan, T.F., Migliavacca, M., Ryu, Y., Sonnentag, O., and Toomey, M. 2013. Climate change, phenology, and phenological control of vegetation feedbacks to the climate system. Agric. For. Meteorol. 169: 156\u2013173. Elsevier B.V. doi:10.1016\/j.agrformet.2012.09.012. Richardson, B.A., Brunsfeld, S.J., and Klopfenstein, N.B. 2002. DNA from bird-dispersed seed and wind-disseminated pollen provides insights into postglacial colonization and population genetic structure of whitebark pine (Pinus albicaulis). Mol. Ecol. 11(2): 215\u2013227. doi:10.1046\/j.1365-294X.2002.01435.x. Rieseberg, L.H. 1991. Homoploid reticulate evolution in Helianthus (Asteraceae): Evidence from ribosomal genes. Am. J. Bot. 78(9): 1218\u20131237. Ritchie, J.C., and Macdonald, G.M. 1986. The Patterns of Post-Glacial Spread of White Spruce. J. Biogeogr. 13(6): 527\u2013540. Ritchie, J.C.J., and Lichti-Federovich, S. 1967. Pollen dispersal phenomena in arctic-subarctic Canada. Rev. Palaeobot. Palynol. 3(1\u20134): 255\u2013266. doi:10.1016\/0034-6667(67)90058-9. 138  Roberts, D.R., and Hamann, A. 2015. Glacial refugia and modern genetic diversity of 22 western North American tree species. Proc. R. Soc. B Biol. Sci. 282(1804). doi:10.1098\/rspb.2014.2903. Roche, L. 1969. A genecological study of the genus Picea in British Columbia. New Phytol. 68(2): 505\u2013554. Ross, C.L., and Harrison, R.G. 2002. A Fine-Scale Spatial Analysis of the Mosaic Hybrid Zone between Gryllus firmus and Gryllus pennsylvanicus. Evolution (N. Y). 56(11): 2296\u20132312. Rweyongeza, D.M., Barnhardt, L.K., and Hansen, C.R. 2011. Patterns of optimal growth for white spruce provenances in Alberta. Alberta Tree Improvement and Seed Centre, Alberta Sustainable Resource Development, Smoky Lake. S\u00e1enz-Romero, C., Lindig-Cisneros, R.A., Joyce, D.G., Beaulieu, J., Bradley, J.S.C., and Jaquish, B.C. 2016. Assisted migration of forest populations for adapting trees to climate change. Rev. Chapingo, Ser. Ciencias For. y del Ambient. 22(3): 303\u2013323. doi:10.5154\/r.rchscfa.2014.10.052. S\u00e6ther, S.A., Fiske, P., K\u00e5l\u00e5s, J.A., Kuresoo, A., Luiguj\u00f5e, L., Piertney, S.B., Sahlman, T., and H\u00f6glund, J. 2007. Inferring local adaptation from QST-FST comparisons: Neutral genetic and quantitative trait variation in European populations of great snipe. J. Evol. Biol. 20(4): 1563\u20131576. doi:10.1111\/j.1420-9101.2007.01328.x. Sanford, E., and Kelly, M.W. 2011. Local adaptation in marine invertebrates. Ann. Rev. Mar. Sci. 3(1): 509\u2013535. doi:10.1146\/annurev-marine-120709-142756. Sang, Y., Long, Z., Dan, X., Feng, J., Shi, T., Jia, C., Zhang, X., Lai, Q., Yang, G., Zhang, H., Xu, X., Liu, H., Jiang, Y., Ingvarsson, P.K., Liu, J., Mao, K., and Wang, J. 2022. Genomic insights into local adaptation and future climate-induced vulnerability of a keystone forest tree in East Asia. Nat. Commun. 13(1). Springer US. doi:10.1038\/s41467-022-34206-8. Savolainen, O., Lascoux, M., and Meril\u00e4, J. 2013. Ecological genomics of local adaptation. Nat. Rev. Genet. 14(11): 807\u201320. Nature Publishing Group. doi:10.1038\/nrg3522. Savolainen, O., Pyh\u00e4j\u00e4rvi, T., and Kn\u00fcrr, T. 2007. Gene flow and local adaptation in trees. Annu. Rev. Ecol. Evol. Syst. 38(1): 595\u2013619. doi:10.1146\/annurev.ecolsys.38.091206.095646. Schrader, L., and Schmitz, J. 2019. The impact of transposable elements in adaptive evolution. Mol. Ecol. 28(6): 1537\u20131549. doi:10.1111\/mec.14794. Schweizer, R.M., Vonholdt, B.M., Harrigan, R., Knowles, J.C., Musiana, M., Coltman, D., Novembre, J., and Wayne, R.K. 2016. Genetic subdivision and candidate genes under selection in North American grey wolves. Mol. Ecol. 25: 380\u2013402. doi:10.1111\/mec.13364. Schw\u00f6rer, C., Gavin, D.G., Walker, I.R., and Hu, F.S. 2017. Holocene tree line changes in the Canadian Cordillera are controlled by climate and topography. J. Biogeogr. 44: 1148\u20131159. doi:10.1111\/jbi.12904. Scott, A.D., Zimin, A. V., Puiu, D., Workman, R., Britton, M., Zaman, S., Caballero, M., Read, A.C., Bogdanove, A.J., Burns, E., Wegrzyn, J., Timp, W., Salzberg, S.L., and Neale, D.B. 2020. A reference genome sequence for giant sequoia. G3 Genes, Genomes, Genet. 10(11): 3907\u20133919. doi:10.1534\/g3.120.401612. Sebastian-Azcona, J., Hacke, U.G., and Hamann, A. 2018. Adaptations of white spruce to climate: strong intraspecific differences in cold hardiness linked to survival. Ecol. Evol. 8(3): 1758\u20131768. doi:10.1002\/ece3.3796. Seguinot, J., Rogozhina, I., Stroeven, A.P., Margold, M., and Kleman, J. 2016. Numerical simulations of the Cordilleran ice sheet through the last glacial cycle. Cryosphere 10(2): 639\u2013664. doi:10.5194\/tc-10-639-2016. 139  Semerikov, V.L., Semerikova, S.A., Putintseva, Y.A., Oreshkova, N. V, and Krutovsky, K. V. 2019. Mitochondrial DNA in Siberian conifers indicates multiple postglacial colonization centers. Can. J. For. Res. 49: 875\u2013883. Shafer, A.B.A., Cullingham, C.I., C\u00f4t\u00e9, S.D., and Coltman, D.W. 2010. Of glaciers and refugia: A decade of study sheds new light on the phylogeography of northwestern North America. Mol. Ecol. 19(21): 4589\u20134621. doi:10.1111\/j.1365-294X.2010.04828.x. Shao, C.C., Shen, T.T., Jin, W.T., Mao, H.J., Ran, J.H., and Wang, X.Q. 2019. Phylotranscriptomics resolves interspecific relationships and indicates multiple historical out-of-North America dispersals through the Bering Land Bridge for the genus Picea (Pinaceae). Mol. Phylogenet. Evol. 141(August). doi:10.1016\/j.ympev.2019.106610. Shepperd, W.D., Jeffers, R.M., and Ronco, F. 1981. An Engelmann spruce seed source study in the Central Rockies. USDA Forest Service, Fort Collins. Sigurgeirsson, A., and Szmidt, A.E. 1993. Phylogenetic and biogeographic implications of chloroplast DNA variation in Picea. Nord. J. Bot. 13(3): 233\u2013246. doi:10.1111\/j.1756-1051.1993.tb00043.x. Simpson, D.G. 1994. Seasonal and geographic origin effects on cold hardiness of white spruce buds, foliage, and stems. Can. J. For. Res. 24: 1066\u20131070. Singh, R.K., Svystun, T., AlDahmash, B., J\u00f6nsson, A.M., and Bhalerao, R.P. 2017. Photoperiod- and temperature-mediated control of phenology in trees \u2013 a molecular perspective. New Phytol. 213(2): 511\u2013524. doi:10.1111\/nph.14346. Smouse, P.E., Long, J.C., and Sokal, R.R. 1986. Multiple regression and correlation extensions of the Mantel test of matrix correspondence. Syst. Zool. 35(4): 627\u2013632. Solarik, K.A., Messier, C., Ouimet, R., Bergeron, Y., and Gravel, D. 2018. Local adaptation of trees at the range margins impacts range shifts in the face of climate change. Glob. Ecol. Biogeogr. 27(12): 1507\u20131519. doi:10.1111\/geb.12829. Sork, V.L. 2018. Genomic studies of local adaptation in natural plant populations. J. Hered. 109(1): 3\u201315. doi:10.1093\/jhered\/esx091. Spichtig, M., and Kawecki, T.J. 2004. The maintenance (or not) of polygenic variation by soft selection in heterogeneous environments. Am. Nat. 164(1): 70\u201384. doi:10.1086\/421335. Stackpole, D.J., Vaillancourt, R.E., de Aguigar, M., and Potts, B.M. 2010. Age trends in genetic parameters for growth and wood density in Eucalyptus globulus. Tree Genet. Genomes 6(2): 179\u2013193. doi:10.1007\/s11295-009-0239-4. Steffensen, J.P., Andersen, K.K., Bigler, M., Clausen, H.B., Dahl-jensen, D., Fischer, H., Goto-azuma, K., Hansson, M., Johnsen, S.J., Jouzel, J., Masson-delmotte, V., Popp, T., Rasmussen, S.O., R\u00f6thlisberger, R., Ruth, U., Stauffer, B., Sveinbj\u00f6rnsd\u00f3ttir, \u00c1.E., Svensson, A., White, J.W.C.C., Siggaard-Andersen, M.L., Sveinbj\u00f6rnsd\u00f3ttir, \u00c1.E., Svensson, A., and White, J.W.C.C. 2008. High-resolution greenland ice core data show abrupt climate change happens in few years. Science (80-. ). 321(5889): 680\u2013684. doi:10.1126\/science.1157707. Stern, C. 1936. Interspecific sterility. Am. Nat. 70(727): 123\u2013142. Strasburg, J.L., Sherman, N.A., Wright, K.M., Moyle, L.C., Willis, J.H., and Rieseberg, L.H. 2012. What can patterns of differentiation across plant genomes tell us about adaptation and speciation? Philos. Trans. R. Soc. Lond. B. Biol. Sci. 367(1587): 364\u201373. doi:10.1098\/rstb.2011.0199. 140  Strobl, C., Boulesteix, A.L., Kneib, T., Augustin, T., and Zeileis, A. 2008. Conditional variable importance for random forests. BMC Bioinformatics 9: 1\u201311. doi:10.1186\/1471-2105-9-307. Sultan, S.E. 2000. Phenotypic plasticity for plant development, function and life history. Trends Plant Sci. 5(12): 537\u201342. doi:10.1016\/S1360-1385(00)01797-0. Sultan, S.E., and Spencer, H.G. 2002. Metapopulation structure favors plasticity over local adaptation. Am. Nat. 160(2): 271\u2013283. Suren, H., Hodgins, K.A., Yeaman, S., Nurkowski, K.A., Smets, P., Rieseberg, L.H., Aitken, S.N., and Holliday, J.A. 2016. Exome capture from the spruce and pine giga-genomes. Mol. Ecol. Resour. 16(5): 1136\u20131146. doi:10.1111\/1755-0998.12570. Sutton, B.C.S., Flanagan, D.J., Gawley, J.R., Newton, C.H., Lester, D.T., and El-Kassaby, Y.A. 1991. Inheritance of chloroplast and mitochondrial DNA in Picea and composition of hybrids from introgression zones. Theor. Appl. Genet. 82(2): 242\u2013248. doi:10.1007\/BF00226220. Sutton, B.C.S., Pritchard, S.C., Gawley, J.R., and Newton, C.H. 1994. Analysis of Sitka spruce - interior spruce introgression in British Columbia using cytoplasmic and nuclear DNA probes. Can. J. For. Res. 24(2): 278\u2013285. Szymura, J.M., and Barton, N.H. 1986. Genetic analysis of a hybrid zone between the fire-bellied Toads, Bombina bombina and B. variegata, near Cracow in southern Poland. Evolution (N. Y). 40(6): 1141\u20131159. Taylor, S.A., Larson, E.L., and Harrison, R.G. 2015. Hybrid zones: Windows on climate change. Trends Ecol. Evol. 30(7): 398\u2013406. doi:10.1016\/j.tree.2015.04.010. Tebaldi, C., and Knutti, R. 2007. The use of the multi-model ensemble in probabilistic climate projections. Philos. Trans. R. Soc. A Math. Phys. Eng. Sci. 365(1857): 2053\u20132075. doi:10.1098\/rsta.2007.2076. Thomas, C.D. 2010. Climate, climate change and range boundaries. Divers. Distrib. 16(3): 488\u2013495. doi:10.1111\/j.1472-4642.2010.00642.x. Tigano, A., and Friesen, V.L. 2016. Genomics of local adaptation with gene flow. Mol. Ecol. 25(10): 2144\u20132164. doi:10.1111\/mec.13606. Turesson, G. 1922. The genotypical response of the plant species to the habitat. Hereditas 3: 211\u2013350. doi:none. Ukrainetz, N.K., O\u2019Neill, G.A., and Jaquish, B. 2011. Comparison of fixed and focal point seed transfer systems for reforestation and assisted migration: a case study for interior spruce in British Columbia. Can. J. For. Res. 41(7): 1452\u20131464. doi:10.1139\/x11-060. Ursin, E. 1952. Occurrence of voles, mice, and rats (Muridae) in Denmark, with a special note on a zone of intergradation between two subspecies of the house mouse (Mus musculus L.). Vidensk. Meddelelser Dansk Naturhistorisk Foren. 114: 217\u2013244. Valladares, F., Matesanz, S., Guilhaumon, F., Ara\u00fajo, M.B., Balaguer, L., Benito-Garz\u00f3n, M., Cornwell, W., Gianoli, E., van Kleunen, M., Naya, D.E., Nicotra, A.B., Poorter, H., and Zavala, M.A. 2014. The effects of phenotypic plasticity and local adaptation on forecasts of species range shifts under climate change. Ecol. Lett. 17(11): 1351\u20131364. doi:10.1111\/ele.12348. de Villemereuil, P., Frichot, \u00c9., Bazin, \u00c9., Fran\u00e7ois, O., and Gaggiotti, O.E. 2014. Genome scan methods against more complex models: When and how much should we trust them? Mol. Ecol. 23(8): 2006\u20132019. doi:10.1111\/mec.12705. 141  Vitasse, Y., Bottero, A., Cailleret, M., Bigler, C., Fonti, P., Gessler, A., L\u00e9vesque, M., Rohner, B., Weber, P., Rigling, A., and Wohlgemuth, T. 2019. Contrasting resistance and resilience to extreme drought and late spring frost in five major European tree species. Glob. Chang. Biol. 25(11): 3781\u20133792. doi:10.1111\/gcb.14803. Vizca\u00edno-Palomar, N., Ib\u00e1\u00f1ez, I., Gonz\u00e1lez-Mart\u00ednez, S.C., Zavala, M.A., and Al\u00eda, R. 2016. Adaptation and plasticity in aboveground allometry variation of four pine species along environmental gradients. Ecol. Evol. 6(21): 7561\u20137573. doi:10.1002\/ece3.2153. Wadgymar, S.M., Demarche, M.L., Josephs, E.B., Sheth, S.N., and Anderson, J.T. 2022. Local Adaptation: Causal Agents of Selection and Adaptive Trait Divergence. Annu. Rev. Ecol. Evol. Syst. 53: 87\u2013111. doi:10.1146\/annurev-ecolsys-012722-035231. Walsh, J., Billerman, S.M., Butcher, B.G., Rohwer, V.G., Toews, D.P.L., Vila-Coury, V., and Lovette, I.J. 2023. A complex genomic architecture underlies reproductive isolation in a North American oriole hybrid zone. Commun. Biol. 6(1): 1\u201311. Springer US. doi:10.1038\/s42003-023-04532-8. Wang, H., Mcarthur, E.D., Sanderson, S.C., Graham, J.H., and Freeman, D.C. 1997. Narrow hybrid zone between two subspecies of big sagebrush (Artemisia tridentata: Asteraceae). IV. Reciprocal transplant experiments. Evolution (N. Y). 51(1): 95\u2013102. Wang, T., Hamann, A., Spittlehouse, D., and Carroll, C. 2016. Locally downscaled and spatially customizable climate data for historical and future periods in North America. PLoS One 11(6): e0156720. doi:10.1371\/journal.pone.0156720. Wang, T., Hamann, A., Spittlehouse, D.L., and Aitken, S.N. 2006. Development of scale-free climate data for western Canada for use in resource management. Int. J. Climatol. 26(3): 383\u2013397. doi:10.1002\/joc.1247. Wang, T., Hamann, A., Spittlehouse, D.L., and Murdock, T.Q. 2012. ClimateWNA-high-resolution spatial climate data for western North America. J. Appl. Meteorol. Climatol. 51(1): 16\u201329. doi:10.1175\/JAMC-D-11-043.1. Warren, R.L., Keeling, C.I., Yuen, M.M. Saint, Raymond, A., Taylor, G.A., Vandervalk, B.P., Mohamadi, H., Paulino, D., Chiu, R., Jackman, S.D., Robertson, G., Yang, C., Boyle, B., Hoffmann, M., Weigel, D., Nelson, D.R., Ritland, C., Isabel, N., Jaquish, B., Yanchuk, A., Bousquet, J., Jones, S.J.M., Mackay, J., Birol, I., and Bohlmann, J. 2015. Improved white spruce (Picea glauca) genome assemblies and annotation of large gene families of conifer terpenoid and phenolic defense metabolism. Plant J. 83(2): 189\u2013212. doi:10.1111\/tpj.12886. Warwell, M. V., and Shaw, R.G. 2017. Climate-related genetic variation in a threatened tree species, pinus albicaulis. Am. J. Bot. 104(8): 1205\u20131218. doi:10.3732\/ajb.1700139. Waser, N.M., and Price, M. V. 1985. Reciprocal transplant experiments with Delphinium nelsonii (Ranunculaceae): Evidence for local adaptation. Am. J. Bot. 72(11): 1726\u20131732. Weeden, N.F., and Wendel, J.F. 1989. Genetics of plant isozymes. In Isozymes in Plant Biology. Edited by D. Soltis and P. Soltis. Chapman and Hall, London. pp. 46\u201372. doi:10.1007\/978-94-009-1840-5. Wegrzyn, J.L., Liechty, J.D., Stevens, K.A., Wu, L.S., Loopstra, C.A., Vasquez-Gross, H.A., Dougherty, W.M., Lin, B.Y., Zieve, J.J., Mart\u00ednez-Garc\u00eda, P.J., Holt, C., Yandell, M., Zimin, A. V., Yorke, J.A., Crepeau, M.W., Puiu, D., Salzberg, S.L., de Jong, P.J., Mockaitis, K., Main, D., Langley, C.H., and Neale, D.B. 2014. Unique features of the loblolly pine (Pinus taeda L.) megagenome revealed through sequence annotation. Genetics 196(3): 891\u2013909. doi:10.1534\/genetics.113.159996. Weng, Y.H., Tosh, K.J., Park, Y.S., and Fullarton, M.S. 2007. Age-related trends in genetic parameters for jack pine 142  and their implications for early selection. Silvae Genet. 56(5): 242\u2013252. doi:10.1515\/sg-2007-0035. Whitlock, M.C. 2008. Evolutionary inference from QST. Mol. Ecol. 17(8): 1885\u20131896. doi:10.1111\/j.1365-294X.2008.03712.x. Whitney, K.D., Ahern, J.R., Campbell, L.G., Albert, L.P., and King, M.S. 2010. Patterns of hybridization in plants. Perspect. Plant Ecol. Evol. Syst. 12(3): 175\u2013182. Elsevier. doi:10.1016\/j.ppees.2010.02.002. Wiley, E., Rogers, B.J., Griesbauer, H.P., and Landh\u00e4usser, S.M. 2018. Spruce shows greater sensitivity to recent warming than Douglas-fir in central British Columbia. Ecosphere 9(5). doi:10.1002\/ecs2.2221. Willi, Y., Van Buskirk, J., and Hoffmann, A.A. 2006. Limits to the adaptive potential of small populations. Annu. Rev. Ecol. Evol. Syst. 37(1): 433\u2013458. doi:10.1146\/annurev.ecolsys.37.091305.110145. Williams, C.G., Ladeau, S.L., Oren, R., and Katul, G.G. 2006. Modeling seed dispersal distances: Implications for transgenic Pinus taeda. Ecol. Appl. 16(1): 117\u2013124. doi:10.1890\/04-1901. Williams, G.C. 1966. Adaptation and Natural Selection. Princeton University Press, Princeton, NJ. Williams, J.W., Shuman, B.N., Webb, T.I., Bartlein, P.J., and Leduc, P.J. 2004. Late-quaternary vgetation dynamics in North America: Scaling from taxa to biomes. Ecol. Monogr. 74(2): 309\u2013334. Wolfe, J.A. 1978. A paleobotanical interpretation of Tertiary climates in the Northern Hemisphere. Am. Sci. 66(6): 694\u2013703. doi:10.2307\/27848958. Wolfe, K.H., Li, W.-H., and Sharp, P.M. 1987. Rates of nucleotide substitution vary greatly among plant mitochondrial, chloroplast, and nuclear DNAs. Proc. Natl. Acad. Sci. U. S. A. 84(24): 9054\u20139058. doi:10.1073\/pnas.84.24.9054. Wright, J.W. 1955. Species crossability in spruce in relation to distribution and taxonomy. For. Sci. 1(4): 319\u2013349. Wright, S. 1931. Evolution in Mendelian populations. Genetics 16: 97\u2013158. doi:10.1016\/S0092-8240(05)80011-4. Wright, S. 1949. The genetical structure of populations. Ann. Eugen. 15(1): 323\u2013354. doi:10.1360\/zd-2013-43-6-1064. Xie, C.-Y. 2003. Genotype by environment interaction and its implications for genetic improvement of interior spruce in British Columbia. Can. J. For. Res. 33(9): 1635\u20131643. doi:10.1139\/x03-082. Xie, C.Y., and Yanchuk, A.D. 2002. Genetic parameters of height and diameter of interior spruce in British Columbia. For. Genet. 9(1): 1\u201310. Yamamichi, M., Hairston, N.G., Rees, M., and Ellner, S.P. 2019. Rapid evolution with generation overlap: the double-edged effect of dormancy. Theor. Ecol. 12(2): 179\u2013195. Theoretical Ecology. doi:10.1007\/s12080-019-0414-7. Yeaman, S. 2015. Local adaptation by alleles of small effect. Am. Nat. 186(S1): S74\u2013S89. doi:10.1086\/682405. Yeaman, S. 2022. Evolution of polygenic traits under global vs local adaptation. Genetics 220(1). doi:10.1093\/genetics\/iyab134. Yeaman, S., Hodgins, K.A., Lotterhos, K.E., Suren, H., Nadeau, S., Degner, J.C., Nurkowski, K.A., Smets, P., Wang, T., Gray, L.K., Liepe, K.J., Hamann, A., Holliday, J.A., Whitlock, M.C., Rieseberg, L.H., and Aitken, 143  S.N. 2016. Convergent local adaptation to climate in distantly related conifers. Science (80-. ). 353(6306): 23\u201326. Yeaman, S., Hodgins, K.A., Suren, H., Nurkowski, K.A., Rieseberg, L.H., Holliday, J.A., and Aitken, S.N. 2014. Conservation and divergence of gene expression plasticity following c. 140 million years of evolution in lodgepole pine (Pinus contorta) and interior spruce (Picea glauca x Picea engelmannii). New Phytol. 203: 578\u2013591. Available from http:\/\/www.scielo.org.ar\/scielo.php?script=sci_arttext&pid=S1852-73372014000300005&lng=en&nrm=iso&tlng=es. Yeaman, S., and Otto, S.P. 2011. Establishment and maintenance of adaptive genetic divergence under migration, selection, and drift. Evolution (N. Y). 65(7): 2123\u20132129. doi:10.1111\/j.1558-5646.2011.01277.x. Ying, C.C., and Yanchuk, A.D. 2006. The development of British Columbia\u2019s tree seed transfer guidelines: Purpose, concept, methodology, and implementation. For. Ecol. Manage. 227(1\u20132): 1\u201313. doi:10.1016\/j.foreco.2006.02.028. Yuan, S., Shi, Y., Zhou, B.F., Liang, Y.Y., Chen, X.Y., An, Q.Q., Fan, Y.R., Shen, Z., Ingvarsson, P.K., and Wang, B. 2023. Genomic vulnerability to climate change in Quercus acutissima, a dominant tree species in East Asian deciduous forests. Mol. Ecol. (July 2022): 1639\u20131655. doi:10.1111\/mec.16843. Zhang, Y., Qi, G., Park, J.H., and Chatterjee, N. 2018. Estimation of complex effect-size distributions using summary-level statistics from genome-wide association studies across 32 complex traits. Nat. Genet. 50(9): 1318\u20131326. Springer US. doi:10.1038\/s41588-018-0193-x. Zhao, W., Sun, Y.Q., Pan, J., Sullivan, A.R., Arnold, M.L., Mao, J.F., and Wang, X.R. 2020. Effects of landscapes and range expansion on population structure and local adaptation. New Phytol. 228(1): 330\u2013343. doi:10.1111\/nph.16619. Zou, J. Bin, Peng, X.L., Li, L., Liu, J.Q., Miehe, G., and Opgenoorth, L. 2012. Molecular phylogeography and evolutionary history of Picea likiangensis in the Qinghai-Tibetan Plateau inferred from mitochondrial and chloroplast DNA sequence variation. J. Syst. Evol. 50(4): 341\u2013350. doi:10.1111\/j.1759-6831.2012.00207.x.   144  Appendix    Figure A.1: Predicted hybrid index for all environments in the 1961-1990 climatic reference period, without constraints based on species climatic niche models. 145        Figure A.2: Species climatic niche models for 15kyp-present, including Sitka spruce. 146   Table A.1: SNP outlier enrichment in genotype-environment associations for SNPs identified as Engelmann- or white-skewed based on genomic cline center. Enrichment is expressed as the odds ratio for skewed SNPs compared to index neutral SNPs within a given percentile of genotype-environment associations (90th \u2013 99.9th percentiles of associations). Numbers in bold indicate ratios where the 95% confidence interval does not include 1. Figure A2: Distribution of climatic differences between the 1991-2020 and 1961-1990 climatic reference periods across the area of their shared climatic niches. .3: Distribution of cli atic di ferences between the 1991-2020 and 1961-1990 cli atic reference ","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/hasType":[{"value":"Thesis\/Dissertation","type":"literal","lang":"en"}],"http:\/\/vivoweb.org\/ontology\/core#dateIssued":[{"value":"2023-11","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/isShownAt":[{"value":"10.14288\/1.0435560","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/language":[{"value":"eng","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#degreeDiscipline":[{"value":"Forestry","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/provider":[{"value":"Vancouver : University of British Columbia Library","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/publisher":[{"value":"University of British Columbia","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/rights":[{"value":"Attribution-NonCommercial-NoDerivatives 4.0 International","type":"literal","lang":"*"}],"https:\/\/open.library.ubc.ca\/terms#rightsURI":[{"value":"http:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/","type":"literal","lang":"*"}],"https:\/\/open.library.ubc.ca\/terms#scholarLevel":[{"value":"Graduate","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/contributor":[{"value":"Aitken, Sally N.","type":"literal","lang":"en"},{"value":"Rieseberg, Loren H.","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/title":[{"value":"Genomics of adaptation in interior spruce to past, present, and future climates of western Canada","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/type":[{"value":"Text","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#identifierURI":[{"value":"http:\/\/hdl.handle.net\/2429\/85601","type":"literal","lang":"en"}]}}