Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effects of single-objective management on disturbances in central interior dry forests of British… Leclerc, Marc-Antoine François 2017

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

Item Metadata

Download

Media
24-ubc_2017_september_leclerc_marc-antoine.pdf [ 3.07MB ]
Metadata
JSON: 24-1.0354571.json
JSON-LD: 24-1.0354571-ld.json
RDF/XML (Pretty): 24-1.0354571-rdf.xml
RDF/JSON: 24-1.0354571-rdf.json
Turtle: 24-1.0354571-turtle.txt
N-Triples: 24-1.0354571-rdf-ntriples.txt
Original Record: 24-1.0354571-source.json
Full Text
24-1.0354571-fulltext.txt
Citation
24-1.0354571.ris

Full Text

 THE EFFECTS OF SINGLE-OBJECTIVE MANAGEMENT ON DISTURBANCES IN CENTRAL INTERIOR DRY FORESTS OF BRITISH COLUMBIA by  Marc-Antoine François Leclerc  B.Sc. in Forest Sciences, The University of British Columbia, 2014  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE In THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Forestry)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) August 2017  © Marc-Antoine Leclerc, 2017  ii Abstract  Mule deer are an important game species, and have become the focus of applying a particular silvicultural treatment that enhances habitat while allowing timber harvesting. Mule deer winter range management (MDWRM) involves the proportional removal of trees based on their diameter and abundance resulting in a multilayered, Douglas-fir dominated forest with a clumpy tree distribution. I assessed changes in forest stand attributes brought about by MDWRM through time and how these attributes related to stand susceptibility to the western spruce budworm, Douglas-fir beetle, and wildfire using a randomized complete block single factor mixed-effects model with subsampling. In the short-term, MDWRM significantly changed (p<0.05) forest stand attributes by decreasing sub-canopy tree density and basal area, canopy cover, leaf area index, and increasing large surface fuel load. In the long-term, most attributes recovered to untreated levels. Relative to untreated stands, treated stands maintained a multilayered structure and an abundance of Douglas-fir trees thus their susceptibility to the western spruce budworm did not change through time. In the short-term, a reduction in mature host-tree density lowered susceptibility to the Douglas-fir beetle. With subsequent forest recovery, long-term susceptibility did not differ relative to untreated stands. In the treated stands the likelihood of crown fire was greater shortly after than longer after treatment. This was likely due to more large surface fuels immediately following treatment. In addition, I extrapolated the effects of MDWRM across eligible stands of interior British Columbia in a hypothetical simulation to evaluate the current and forecasted landscape-level fire risk. The forecasted forest under widespread application of MDWRM resulted in a homogenized landscape dominated by low fire risk. Further, widespread application of  iii MDWRM may result in a fire resilient landscape, but with consequences for other ecological processes. The present study contributes to our understanding of the relationship between single-objective management and subsequent stand susceptibility to biotic and abiotic disturbances. I concluded that understanding this relationship should play an important role in responsible resource management.    iv Lay abstract  In the central interior of British Columbia, the effects of mule deer winter range management (MDWRM) on forest characteristics such as the number of trees per area, and canopy cover were assessed through time and linked to the susceptibility to the 3 primary disturbance agents of the region: the western spruce budworm, Douglas-fir beetle, and wildfire. MDWRM did not alter susceptibility to the western spruce budworm. Susceptibility to the Douglas-fir beetle was lower shortly after applying MDWRM yet the long-term susceptibility was similar to unaffected forest stands. Likelihood of crown fire was greatest shortly after MDWRM while lowest in the long-term. Hypothetically applying MDWRM to all eligible stands to influence landscape-level fire risk resulted in a homogenized landscape dominated by low fire risk. While broad-scale application of MDWRM created a fire resilient landscape, potential negative impacts to other ecosystem processes such as biodiversity and insect disturbance might be of concern.    v Preface  This thesis was made possible with the collaboration of my co-advisors, Drs. Allan Carroll and Lori Daniels, who guided me through the development of my research questions, research methods and sampling design, a portion of the analysis in Chapter 2, and provided thorough edits to the manuscript. Dr. Valerie LeMay provided guidance in conducting analysis of Chapter 2. Further, Dr. Sarah Gergel also provided detailed edits and suggestions improving the manuscript. Finally, I was involved in developing my research questions, research methods and sampling design, data collection and data analysis of Chapters 2 and 3 along with the full write up. Chapter 2 is expected to be submitted for scientific publication.    vi Table of contents  Abstract .............................................................................................................................. ii	Lay abstract ...................................................................................................................... iv	Preface ................................................................................................................................ v	Table of contents .............................................................................................................. vi	List of tables...................................................................................................................... ix	List of figures .................................................................................................................... xi	List of abbreviations ...................................................................................................... xiii	Acknowledgements ........................................................................................................ xiv	Dedication ........................................................................................................................ xv	1	 General introduction ................................................................................................ 1	1.1	 Influence of disturbance patterns on forest ecosystems ...................................... 1	1.1.1	 The influence of forest management on disturbance patterns .................... 2	1.2	 Influential abiotic disturbance patterns in the mixed-conifer forests in British Columbia ......................................................................................................................... 3	1.2.1	 General description of dry forests in British Columbia .............................. 3	1.2.2	 The role of fire in the dry forests of British Columbia ............................... 4	1.3	 Biotic disturbance in the dry forests of British Columbia .................................. 5	1.3.1	 The western spruce budworm ..................................................................... 5	1.3.1.1	 Western spruce budworm biology .......................................................... 5	1.3.1.2	 Western spruce budworm outbreaks and stand susceptibility to outbreaks ................................................................................................................. 6	1.3.2	 The Douglas-fir beetle ................................................................................ 7	1.3.2.1	 Douglas-fir beetle biology ...................................................................... 7	1.3.2.2	 Douglas-fir beetle outbreak potential ...................................................... 8	1.3.3	 Interactions between the western spruce budworm and the Douglas-fir beetle ..................................................................................................................... 8	1.4	 Interactions between biotic and abiotic disturbances .......................................... 9	1.4.1	 The interaction between defoliators and fire ............................................ 11	1.4.2	 The interactions between bark beetles and fire ......................................... 13	1.5	 Mule deer .......................................................................................................... 14	1.6	 Potential for land-use changes and climate change to alter mule deer habitat . 15	 vii 1.7	 Management of mule deer habitat ..................................................................... 16	1.7.1	 Potential interactions between mule deer winter range management and disturbance ................................................................................................................ 17	1.8	 Research goals .................................................................................................. 18	1.8.1	 Objective 1: Effects of mule deer winter range management on forest stand attributes and disturbance likelihood ............................................................... 18	1.8.2	 Objective 2: Widespread application of mule deer winter range management on fire risk ............................................................................................ 18	1.9	 Thesis structure overview ................................................................................. 19	2	 Effects of mule deer winter range management on susceptibility to disturbance 20	2.1	 Introduction ....................................................................................................... 20	2.2	 Materials and methods ...................................................................................... 28	2.2.1	 Study area: Knife Creek Alex Fraser Research Forest, BC ...................... 28	2.2.2	 Mule deer winter range experiment at Knife Creek .................................. 30	2.2.2.1	Field sampling: Forest structure and disturbance susceptibility in 2014-2015 ........................................................................................................... 31	2.2.2.2	 Deriving plot-level attributes in 2014-2015 .......................................... 33	2.2.3	 Susceptibility to disturbance ..................................................................... 33	2.2.3.1	 Western spruce budworm susceptibility ............................................... 34	2.2.3.2	 Douglas-fir beetle susceptibility ........................................................... 36	2.2.3.3	 Crown fire likelihood ............................................................................ 37	2.2.4	 Forest recovery within a mule deer management context ........................ 39	2.2.5	 Data analyses ............................................................................................ 39	2.2.5.1	 Forest structure in 2014-2015 ............................................................... 39	2.2.5.2	 Wildfire hazard ..................................................................................... 41	2.3	 Results ............................................................................................................... 41	2.3.1	 Changes in forest structure ........................................................................ 41	2.3.2	 Forest recovery within a mule deer winter range management context ... 44	2.3.3	 Western spruce budworm susceptibility ................................................... 44	2.3.4	 Douglas-fir beetle susceptibility ............................................................... 45	2.3.5	 Crown fire likelihood ................................................................................ 46	2.3.6	 Wildfire hazard ......................................................................................... 46	2.4	 Discussion ......................................................................................................... 49	2.4.1	 Short and long-term effects of winter range management on stand .............  structure ..................................................................................................... 49	2.4.2	 Implications for biotic and abiotic disturbance ......................................... 52	2.4.2.1	 Western spruce budworm susceptibility ............................................... 52	2.4.2.2	 Douglas-fir beetle susceptibility ........................................................... 57	2.4.2.3	 Crown fire likelihood and wildfire hazard ............................................ 59	2.4.3	 Interactions ................................................................................................ 62	2.4.4	 Conclusions ............................................................................................... 68	 viii 3	 The potential effects of widespread mule deer winter range management application on wildfire risk in dry forests of interior British Columbia ................... 69	3.1	 Introduction ....................................................................................................... 69	3.2	 Materials and methods ...................................................................................... 73	3.2.1	 Study area .................................................................................................. 73	3.2.2	 Data and processing .................................................................................. 75	3.2.2.1	 Habitat management zones ................................................................... 75	3.2.2.2	 Vegetation resource inventory .............................................................. 76	3.2.2.3	 Fire risk ................................................................................................. 76	3.2.3	 Landscape management scenarios ............................................................ 77	3.2.3.1	 Landscape scenario 1: Current landscape without MDWRM .............. 78	3.2.3.2	 Landscape scenario 2: Forecasted landscape with widespread MDWRM  ............................................................................................................... 78	3.2.4	 Data analysis ............................................................................................. 79	3.2.4.1	 Abundance and spatial distribution of fire risk classes ......................... 79	3.2.4.2	 Landscape heterogeneity ....................................................................... 80	3.3	 Results ............................................................................................................... 81	3.3.1	 Abundance and spatial distribution of fire risk classes ............................. 81	3.3.2	 Landscape heterogeneity ........................................................................... 84	3.4	 Discussion ......................................................................................................... 84	4	 General conclusions ................................................................................................ 89	4.1	 Key findings ...................................................................................................... 89	4.2	 Implications for forest management ................................................................. 90	4.3	 Future research .................................................................................................. 95	References ........................................................................................................................ 97	Appendices ..................................................................................................................... 115	Appendix A: Western spruce budworm susceptibility index ..................................... 115	Appendix B: Douglas-fir beetle susceptibility ............................................................ 121	Appendix C: Crown fire likelihood ............................................................................ 123	Appendix D: Basal area increment ............................................................................. 124	   ix List of tables  Table 2.1: A summary of mean tree densities (and standard deviation) of the treatment blocks prior to the mule deer winter range management silvicultural treatment in 1984 provided by the Forest Valuation Branch. Densities were measured using variable radius plots (basal area factor 8) for trees ³12.5 cm .....................................................................30 Table 2.2: Comparison of forest structure attributes among plots in the control and treatment areas that were sampled in 2014 and/or 2015 following implementation of mule deer winter range management. Each variable was transformed to meet assumptions of ANOVA. For variables that differed significantly (p-values in italics), means (standard errors) followed by the same superscripts do not differ significantly based on the applied Bonferroni correction (α=0.05) ..........................................................................................43 Table 2.3: Comparison of the mean basal area increment (BAI) per year (mm2/year) of canopy and sub-canopy trees in the control and treatment areas prior (1955-1984) and after (1985-2014) the implementation of mule deer winter range habitat management. BAI was compared using ANOVA and using a BLOM transformation to meet assumptions. For significantly different variables (p-value in italics), determined by a Bonferroni correction, means (standard errors) followed by the same superscripts did not differ significantly (α=0.05) ...............................................................................................44  Table 2.4: Comparison of the number of plots classified as high or moderate probability of infestation by the Douglas-fir beetle in the control and treatment areas following mule deer winter range management. A c2 test revealed that significantly fewer plots had a high probability of infestation in the treatment area in 2015 relative to 2014. Probability  x of infestation by the Douglas-fir beetle was assessed using the regression tree developed by Negrón (1998) ...............................................................................................................45  Table 2.5: The likelihood ratio test values for the variables making up the wildfire hazard model in 2015 following initial model selection. Significantly contributing variables (p-values in italics), were determined using α=0.05 and variables not contributing to the model were removed from subsequent iterations ..............................................................47 Table 2.6: The coefficients, standard error, t- and p-values of the variables in the final wildfire hazard model in 2015 ...........................................................................................48 Table 3.1: Comparison of the spatial metrics used to describe the abundance and spatial arrangement of the 4 fire risk classes in the dry interior forests of British Columbia designated for mule deer winter range management (MDWRM) between the current state of the forest without widespread MDWRM (scenario 1) and a hypothetical application of MDWRM to all eligible forest stands (scenario 2) ............................................................82 Table 3.2: Comparison of landscape heterogeneity (distribution and abundance of fire risk classes) in the dry interior forests of British Columbia designated eligible for mule deer winter range management (MDWRM) between the current state of the forest without widespread MDWRM (scenario 1) and a hypothetical application of MDWRM to all eligible forest stands (scenario 2) ......................................................................................83    xi List of figures  Figure 1.1: The hypothetical interactions among the primary biotic and abiotic disturbance agents in the interior dry forests of British Columbia. ‘Enhancing’ indicates predisposition of a stand to the subsequent disturbance, while a ‘suppressing’ interaction indicates disturbances excluding one another. ‘Balancing’ indicates that one disturbance predisposes a stand to the other disturbance while one disturbance reduces susceptibility of a stand to the other ......................................................................................................... 12 Figure 2.1: The location of the mule deer winter range experimental blocks at the Alex Fraser Research Forest. (A) The Knife Creek Block is located in central British Columbia, Canada. (B) The Alex Fraser Research Forest’s Knife Creek block in central British Columbia. (C) Three paired blocks were established to study the effects of mule deer winter range management. Treated blocks (polygons) with respective control pairs immediately adjacent to the treated blocks ........................................................................29 Figure 2.2: Comparison in the mean (standard errors) likelihood of crown fire occurrence following mule deer winter range management in the treatment area in 2014 and the treatment and control areas in 2015 of Knife Creek. Bars with the same superscript letters did not differ significantly after applying a Bonferroni correction (a=0.05) ....................46 Figure 2.3: The relationships between crown base height of sub-canopy trees (top), large fuels (centre) and 10-hour fuels (bottom) with likelihood of crown fire occurrence ........48 Figure 2.4: Hypothetical interactions between the primary disturbance agents of interior dry forest ecosystems of British Columbia prior to the application of mule deer winter range management (MDWRM). A ‘+’ symbol indicates an enhancing effect, whereby a  xii disturbance predisposes the stand to the subsequent disturbance. A ‘-’ symbol indicates a suppressing effect, where disturbances exclude one another .............................................63 Figure 2.5: Hypothetical interactions between the primary disturbance agents of interior dry forests of British Columbia with the application of mule deer winter range management (MDWRM) and associated Douglas-fir beetle management tactics in the short and long-term. A ‘+’ symbol indicates an enhancing effect, whereby a disturbance predisposes the stand to the subsequent disturbance. A ‘-’ symbol indicates a suppressing effect, where disturbances exclude one another. A ‘0’ indicates that there is no effect, the interaction is neutral. Red arrows with ‘+’ or ‘-’ symbols indicate direct effects of MDWRM on the disturbance, while red ‘+’ or ‘-’ with black arrows indicate cascading effects resulting from MDWRM application .....................................................................65 Figure 3.1: The extent of the mule deer winter range management zones (A) located in central British Columbia (B), Canada (C) .........................................................................74 Figure 3.2: The estimated area covered by different fire risk classes within the mule deer winter range management landscape of interior British Columbia. (A) Estimated area covered by the different fire risk classes in the current state of the forest without MDWRM application (scenario 1). (B) Estimated area covered by different fire risk classes under the hypothetical application of MDWRM to all eligible stands (scenario 2)............................................................................................................................................81    xiii List of abbreviations  AIC- Akaike Information Criterion ANOVA- Analysis of variance BAI- Basal area increment  BC- British Columbia CART-Classification and regression tree CFIS- Crown fire initiation and spread  DBH- Diameter at breast height GIS- Geographic information system IDF- Interior Douglas-fir LAI- Leaf area index MDWRM- Mule deer winter range management VRI- Vegetation resource inventory    xiv Acknowledgements  Many thanks go out to Drs. Allan Carroll, Lori Daniels, Valerie LeMay, and Sarah Gergel for lots of patience, guidance, encouragement, comments, feedback, laughs and for being totally groovy. An immense thank you to Cathy Koot, Ken Day, and Stephanie Ewen from the Alex Research Forest for providing information, feedback, and thought provoking conversations. My field crew: Wesley Brookes, Tyler Dergousoff, Angelica Kaufman, Marie-Eve Leclerc, Anna Weixelman, and Eileen Xu deserve a round of applause for a job well done! Shout out to Ben Tudor for keeping me entertained while up at the research forest and to Mike Tudor for great dish-soap advice and for serving me at badminton. A big thanks to Ingrid Jarvis for EXCELing in the lab. A serious thank you goes out to Gregory Greene, Ian Eddie, and Tanya Gallagher for helping me navigate in ArcMap. To my group of amigos Barbara Wong, Curtis Chance, Yuhao (Bean) Lu, Raphaël Charvardès, Jordan Burke, and Stan Pokorny who took the time to read and provided feedback for this piece of writing. Thanks to the Moss Rock Foundation and NSERC for funding.  Most importantly, a big hand, hugs and kisses go out to my family (Maman, Papa, Marie-Eve, and Socks) for their support, moral and financial, through all the ups and downs of this project. Thank you to my ‘thesis support group’ Bill New, Barbara Wong, Marie-Eve Leclerc, Curtis Chance, Ian Eddie, Stan Pokorny, Jordan Burke, and Jenny Liu for providing ears while I would complain and rant. To all my friends, past and present, for somehow leading me to this point, ‘Hasta la victoria siempre’. And Lindsay Hill.    xv Dedication  À Maman et Papa 1  1 General introduction  1.1 Influence of disturbance patterns on forest ecosystems  Disturbances are vital ecosystem processes that may alter the availability and spatiotemporal distribution of resources (White and Pickett 1985). Disturbance events can influence nutrient cycling by affecting nutrient release rates and the distribution of organic matter (Harmon et al. 1986, Brown et al. 2003, Marañón-Jiménez and Castro 2013). Disturbances may also affect the biodiversity of an ecosystem when an event significantly alters the distribution of resources through changes in forest composition and/or structure, or destroys existing biotic assemblages (e.g. Haggstrom and Kelleyhouse 1996, Sullivan and Sullivan 2001, Lyons et al. 2008, Millington et al. 2011, Latif et al. 2016). Following a fire, tree species assemblages and the ability of species to re-colonize from neighbouring stands through seed dispersal may change based on the number of surviving individuals (Turner et al. 1998, Romme et al. 1998). Further, tree mortality resulting from disturbance may also alter species assemblages and structural changes by influencing tree establishment, or allowing the release of subordinate trees (Daniels et al. 2011). Therefore, disturbances play a key role in promoting heterogeneity throughout succession (Hessburg et al. 2000, Turner 2010, Buma and Wessman 2013) enhancing forest health and maintaining ecosystem services (Franklin et al. 2002, Palik et al. 2002).     2 1.1.1 The influence of forest management on disturbance patterns  Forest management practices have the potential to alter disturbance patterns, and/or create new ones that may affect forest health and ecosystem functions. Managers may apply silvicultural treatments in an attempt to replace or mimic natural disturbances (Radeloff et al. 2000), promote specific habitats (Koot et al. 2015), or convert forested lands for agriculture and grazing (Hessburg et al. 2005). In certain cases, management actions may predispose forests to a shift in disturbance regime. For example, logging and fire suppression have homogenized stands throughout western North America at both stand and landscape scales (Hessburg et al. 2000, Schoennagel et al. 2004). These stands may become denser and/or change species composition resulting in more frequent and severe insect outbreaks (Raffa et al. 2008, Sturtevant et al. 2012), as well as a shift from low-severity to more intense high-severity fires (Covington and Moore 1994, Fulé et al. 1997, Allen et al. 2002, Schoennagel et al. 2004, Daniels 2004, Falk et al. 2007, Amoroso et al. 2011, Charvardès and Daniels 2016, Pogue 2017).   Forest management practices intended to address one specific objective may create or exacerbate other forest health issues (Hessburg et al. 1994, Hessburg et al. 2000). In the case of the western spruce budworm in Colorado, fire suppression, insecticides, and past logging likely created multilayered stands with abundant hosts, providing food loading and ideal conditions for larval dispersal and survival (Fellin and Dewey 1986, Brookes et al. 1987, Swetnam and Lynch 1993, Hadley 1994). In British Columbia, widespread fire suppression and inadequate harvesting rates created an over-abundance of overmature, even-aged lodgepole pine [Pinus contorta var. latifolia] forest leading to an  3 unprecedented mountain pine beetle [Dendrotonus ponderosae Hopkins] epidemic that began in 1999 (Taylor and Carroll 2004, Barclay et al. 2005, Taylor et al. 2006). However, other practices, such as thinning young stands, and treating or removing slash, can help reduce susceptibility to bark beetles (Ross and Daterman 1997). For Douglas-fir beetle [Dendroctonus pseudotsugae Hopkins], fire prevention is also beneficial in reducing potential beetle habitat by lowering the number of burned, and therefore injured and susceptible hosts trees (Furniss 1965, Ross and Daterman 1997).  1.2 Influential abiotic disturbance patterns in the mixed-conifer forests in British Columbia  1.2.1 General description of dry forests in British Columbia  The “dry forest” is a collective term used to describe pine and mixed-conifer forests that are typically open, and patchy in structure with low tree density (Arno 1980, Hessburg et al. 2005, Kaufmann et al. 2007, Binkley et al. 2007). Dry forests are characterized by structural variability across spatial scales (Allen et al. 2002), where most stands are multilayered with clumps of intermediate and small-sized trees with grassy openings, and single or scattered clumps of large, old trees (Maclauchlan and Brooks 2009). In British Columbia, these dry forests are found in the south-central interior, typically experiencing hot summers and cold winters with most precipitation occurring in the late spring and early summer (Hope et al. 1991, Pojar and Meidinger 1991). Dry forests in British Columbia generally occur at elevations ranging from 300-1450m on brunisolic and luvisolic soils (Hope et al. 1991).   4 The tree species composition of dry forests is generally comprised of ponderosa pine [Pinus ponderosa Douglas ex C. Lawson], Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco var. glauca] and western larch [Larix occidentalis Nutt.] stands (Hope et al. 1991). However, other species such as lodgepole pine [Pinus contorta Dougl. var. latifolia Engelm.], trembling aspen [Populus tremuloides Michx.], hybrid white spruce [Picea engelmannii x glauca], and paper birch [Betula papyrifera Marsh.] also occur on wetter sites (Hope et al. 1991, Klinka et al. 2004, Koot et al. 2015). The understory is most often open and composed of herbs, shrubs, and mosses along with regeneration of shade-tolerant species including Douglas-fir and hybrid white spruce (Hope et al. 1991, Klinka et al. 2004).  1.2.2 The role of fire in the dry forests of British Columbia  Fire has played an important historic role in dry forests, determining forest composition and structure while also preventing forest encroachment into adjacent grasslands (Government of British Columbia Ministry of Forests 1996, Schoennagel et al. 2004, Nesbitt 2010, Pogue 2017). However, fire suppression during the previous century has allowed for the development of a dense understory in previously open stands throughout British Columbia (Daniels 2004, Da Silva 2009, Nesbitt 2010, Pogue 2017). Infilling of open stands increases the occurrence and size of high-severity fires in the dry forests (Schoennagel et al. 2004), and as such, these forests may be experiencing fire regimes outside their historic range of variability (Covington and Moore 1994, Allen et al. 2002, Schoennagel et al. 2004, Da Silva 2009, Nesbitt 2010, Marcoux et al. 2015, Pogue 2017).   5 Spatial and temporal variability in the frequency and severity of fires influences the susceptibility of dry forests to subsequent disturbance (Hessburg et al. 2000, Kulakowski and Jarvis 2013), as the resulting stand structure and composition can, in turn, affect the fire regime (Bigler et al. 2005). Mixed- and low-severity fires tend to be more common than high-severity fires in dry forests (Marcoux et al. 2013, Chavardès and Daniels 2016, Greene and Daniels 2017). Mixed-severity fires are comprised of both high- and low-severity fire characteristics, varying spatially and temporally, creating a mosaic of even- and uneven-aged stands (Schoennagel et al. 2004, Perry et al. 2011, Marcoux et al. 2013, 2015).  High-severity fires will burn the forest floor, understory vegetation, and into the canopy leading to extensive over- and understory mortality (Schoennagel et al. 2004). These fires are typically associated with forests with abundant ladder fuels with low crown base heights, high stocking density, and sparse understory (Schoennagel et al. 2004, Falk et al. 2007). Typically, low-severity fires only burn the understory, rarely killing over-story trees, and result in open woodlands stands with understory grasses and herbs (Schoennagel et al. 2004).   1.3 Biotic disturbance in the dry forests of British Columbia  1.3.1 The western spruce budworm  1.3.1.1 Western spruce budworm biology The western spruce budworm, Choristoneura occidentalis Freeman (Lepidoptera: Tortricidae), is an important disturbance agent in dry forests and is a widespread, native defoliator in western North America (Brookes et al. 1987, Maclauchlan et al. 2006). The insect preferentially feeds on the current year’s buds and needles, consuming older  6 needles and cones when needed (Fellin and Dewey 1986, Ryerson et al. 2003). The western spruce budworm’s life cycle consists of 6 instars, and eggs are laid on the underside of needles (Brookes et al. 1987). Eggs hatch immediately, and first instar larvae then spin a hibernaculum of silk at the end of branches and overwinter within, and emerge in synchrony with budburst in the spring to begin feeding and development (Brookes et al. 1987). Defoliation by the western spruce budworm lowers available sugar concentrations in affected trees (Clancy et al. 2004), resulting in top kill (Fellin and Dewey 1986) and reduced radial growth (Swetnam et al. 1985). While over-story trees commonly experience decreased growth and/or reduced cone crop, understory trees may die after a large defoliation event (Alfaro et al. 1982, Maclauchlan and Brooks 2009). Consequently, multiple defoliation events may postpone natural regeneration of host trees resulting in more open stands (Fellin and Dewey 1986).  1.3.1.2 Western spruce budworm outbreaks and stand susceptibility to outbreaks Since first being recorded in 1909 in British Columbia (Unger 1983), western spruce budworm populations have undergone periodic outbreaks (Campbell et al. 2006). Population surges appear to coincide with tree stress, particularly during low precipitation (Brookes et al. 1987, Campbell et al. 2006, Flower et al. 2014a). Outbreaks occur on average every 30 years, lasting from 10 to 14 years (Swetnam and Lynch 1993, Ryerson et al. 2003, Campbell et al. 2006), but the severity of outbreaks can subsequently influence both the return interval and duration (Axelson et al. 2015).   7 Stand impacts from a western spruce budworm outbreak are influenced by host abundance, stand structure, adaptive seasonality, and precipitation. Multilayered stands with high absolute and relative host tree densities are highly susceptible, as they increase western spruce budworm survival and successful dispersal events (Fellin and Dewey 1986, Brookes et al. 1987). Although such stands are susceptible, tree-level susceptibility is dependent on growth rate, and synchronicity of larval hatch with bud break (Chen et al. 2001, 2003). Further, site susceptibility may be exacerbated in dry areas where trees are under moisture stress as observed in British Columbia, central Oregon, and Montana (Brookes et al. 1987, Campbell et al. 2006, Flower et al. 2014a). In contrast, outbreaks in New Mexico and Colorado have been associated with wetter periods possibly related to greater foliage availability (Swetnam and Lynch 1993, Ryerson et al. 2003). However, Campbell (1993) suggests that although western spruce budworm survival is greater in warm-dry conditions, extreme drought conditions may be detrimental to survival possibly due to reduced food quality.  1.3.2 The Douglas-fir beetle  1.3.2.1 Douglas-fir beetle biology The Douglas-fir beetle, Dendroctonus pseudotsugae Hopkins (Coleoptera: Curculionidae), is another important biotic disturbance agent in the dry forests of British Columbia (Erickson 1992). It has the ability to cause widespread mortality throughout the range of Douglas-fir in western North America (Furniss and Carolin 1977). The majority of the life cycle of Douglas-fir beetles is spent underneath the bark of the host plant, where eggs are laid and larvae feed exclusively on phloem tissue (Wood 1982,  8 Christiansen et al. 1987). Douglas-fir beetles preferentially attack low vigour (Rudinsky 1962), fire scarred (Furniss 1965, Hood and Bentz 2007), and freshly downed Douglas-fir trees when the beetle is at low densities (Aukema et al. 2016). However, when food is abundant and climate is conducive to survival, populations may experience widespread outbreaks.  1.3.2.2 Douglas-fir beetle outbreak potential Usually found in endemic states (Christiansen et al. 1987), Douglas-fir beetle populations have the ability to increase to outbreak levels and infest well defended host trees (Rudinsky 1962, Rudinsky 1966). Attack on healthy trees by the Douglas-fir beetle results in patches of tree mortality (Rudinsky 1962, Wood 1982, Powers et al. 1999), and is dependent on the abundance of susceptible hosts coupled with high population levels which may build up in weakened and downed trees (Rudinsky 1966, Negrón 1998, Aukema et al. 2016). Susceptibility of trees to bark beetles typically increases during periods of drought (Rudinsky 1966, Hart et al. 2014a). Relative proportion of Douglas-fir basal area and stand density were found to be the best predictors of Douglas-fir beetle damage, where stands comprised mostly of Douglas-fir at high density were more susceptible (Negrón 1998, Negrón et al. 1999, Dodds et al. 2006).  1.3.3 Interactions between the western spruce budworm and the Douglas-fir beetle  On the same host tree species, defoliators and bark beetles have been observed to interact, with	defoliators predisposing host trees to attack by bark beetles (Lessard and Schmidt 1990, Steed et al. 2007). In Scots pine [Pinus sylvestris L.] forests in Scandinavia, trees  9 heavily defoliated by Diprion pini (L.) (Hymenoptera: Diprionidae) were more likely to be attacked by Tomicus piniperda (L.)  (Coleoptera: Scolytinae), whereas trees with little to no defoliation were not (Långström et al. 2001). Similarly, severe defoliation on Scots pine by Bupalus piniara (L.) (Lepidoptera: Geometridae) lead to tree mortality caused by T, piniperda (Cedervind et al. 2003). In western North America, susceptibility of host trees to western spruce budworm is generally a sign that trees may be stressed and also prone to Douglas-fir beetle attacks (Hadley and Veblen 1993, Negrón 1998).  Defoliation of trees acts as a stressor (Wood 1982, Christiansen et al. 1987, Hart et al. 2014a), and may result in a weakened tree without compromising phloem availability resulting in available and susceptible host for Douglas-fir beetles (Marciniak 2015). Conversely, bark beetles may also create conditions that are prone to the western spruce budworm. For example, in the Pacific Northwest, the mountain pine beetle may kill the lodgepole pine component of the canopy, thereby creating gaps, and allowing establishment of Douglas-fir regeneration, which over time may result in a multilayered stand that is susceptible to western spruce budworm outbreak (Wilson et al. 1998).  Interaction between the biotic agents along with abiotic disturbances adds another level of complexity in trying to understand ecosystem processes.  1.4 Interactions between biotic and abiotic disturbances  Ecosystem alterations from disturbances, such as changes in structure and/or resource availability, result in a legacy upon which subsequent disturbance can act (Morrison 1987, Buma 2015). Legacies left behind by disturbances may be a driving force behind changes in ecosystem functions (McCullough et al. 1998, Rhemtulla and Mladenoff  10 2007, Fraterrigo and Rusak 2008). For example, following a blowdown event, salvage logging and fire, forests in the Rocky Mountains of Colorado initially composed of lodgepole pine may become dominated by trembling aspen, likely lowering susceptibility to future fire and windthrow (Buma and Wessman 2012).  The legacies left behind by previous disturbance vary and depend on the type, extent, intensity, and time between different disturbance events (Turner et al. 1998, Romme et al. 1998, Paine et al. 1998). For example, crown fires that burned over a large area in Yellowstone National Park have resulted in forest heterogeneity at multiple scales (Turner and Romme 1994, Schoennagel et al. 2008). Pre-fire conditions and fire severity interact to create a mosaic of burned and unburned areas, resulting in differences in stand age and structure at a broad scale, while also affecting cone seritony, coarse wood arrangement, and understory recovery at finer scales (Schoennagel et al. 2008).   As a result of legacies and changing patterns occurring at multiple scales, determining interactions between disturbances are particularly important: for example, bark beetles and fire (e.g. Bebi et al. 2003, Simard et al. 2011, Kulakowski and Jarvis 2013) or defoliators and fire (e.g. Lynch and Moorcroft 2008). Further, the time between disturbances will also affect the interaction observed, where one disturbance may enhance or reduce the severity of the subsequent disturbance, over the short or long-term, respectively (e.g. Fleming et al. 2002, Sturtevant et al. 2012). Determining the interaction between disturbances is not always straight forward (Parker et al. 2006), especially since effects of one disturbance may last for some time, influencing plant diversity or occurrence of other disturbances (Dupouey et al. 2002, Hart et al. 2014b).  11  Interactions between disturbances are generally described as either linked or synergistic (Paine et al. 1998, Simard et al. 2011, Buma 2015). A linked disturbance changes severity, probability of occurrence, or extent of following disturbances (Simard et al. 2011), such as when bark beetles alter forest structure creating a short-term pulse of fuel creating the possibility for a more intense fire or reducing aerial fuels lowering fire severity in the long-term (Hicke et al. 2012, Donato et al. 2013). Thus, linked disturbances may enhance or suppress conditions for subsequent disturbances. A synergistic interaction between two disturbances occurs when the effect of the combined disturbances is greater than the effects of the individual disturbances (Paine et al. 1998, Buma and Wessman 2011). For example, defoliation by the western spruce budworm predisposes trees to attack by the Douglas-fir beetle potentially leading to changes in species composition and forest structure (Hadley 1994, Marciniak 2015). Novel disturbances may result from disturbance interactions, as impacts on the ecosystem will be different than if there was a single disturbance acting on the system (Buma and Wessman 2011).   1.4.1 The interaction between defoliators and fire  The effect of defoliation on fire susceptibility and severity varies temporally and is often context-dependent. Defoliation by the eastern spruce budworm [Choristoneura fumifera Clemens] is believed to increase surface and ladder fuel loading resulting in greater fire susceptibility (Stocks 1987). Many fires occur 3-9 years following eastern spruce budworm outbreaks, though this ‘window’ varies geographically (Fleming et al. 2002).  12 However, over a longer time period, fires may be less frequent as a result of defoliation (Sturtevant et al. 2012). In British Columbia, the western spruce budworm appears to decrease fire severity immediately after defoliation for the following 5 to 10 years (Lynch and Moorcroft 2008; Figure 1.1). Similarly, using a physics-based fire model, Cohn et al. (2014) found that defoliation by the western spruce budworm decreased the chance of torching and crowning by reducing the fuel load and/or fuel density in tree crowns, thereby making conditions more conducive to surface fires rather than severe crown fires. A decrease in wildfire likelihood was observed following western spruce budworm defoliation in the Pacific Northwest (Meigs et al. 2015), although rapid accumulation of understory and greater amounts of coarse woody debris may, for a short time, lead to an increase in rate of spread and burn severity (Hummel and Agee 2003, Meigs et al. 2016). Alternatively, western spruce budworm outbreaks in Oregon and Montana were not found to affect fire severity (Flower et al. 2014b).   Figure 1.1: The hypothetical interactions among the primary biotic and abiotic disturbance agents in the interior dry forests of British Columbia. ‘Enhancing’ indicates predisposition of a stand to the subsequent disturbance, while a ‘suppressing’ interaction indicates disturbances excluding one another. ‘Balancing’ indicates that one disturbance predisposes a stand to the other disturbance while one disturbance reduces susceptibility of a stand to the other.  13  1.4.2 The interactions between bark beetles and fire  Interactions between fire and several Dendroctonus species have been studied extensively. The spruce beetle [Dendroctonus rufipennis Kirby], mountain pine beetle, and Douglas-fir beetle affect different fuel components at different points during their population cycle (Jenkins et al. 2008). Spruce beetles have not been found to affect fires (Kulakowski and Veblen 2007, Black et al. 2013) unless outbreaks occur in drought conditions (Bebi et al. 2003, Bigler et al. 2005, Hart et al. 2014a). Following mountain pine beetle epidemics, stands have more large woody surface fuels and understory vegetation, but reduced crown base height (due to regeneration) and aerial fuels (Page and Jenkins 2007a). These forest structural changes result in greater probabilities of passive crown fire and torching due to the increased fire intensity (Page and Jenkins 2007b, Simard et al. 2011). In contrast, spread of active crown fire may be reduced in the short-term by fewer aerial fuels which reduce canopy bulk density (Page and Jenkins 2007b, Simard et al. 2011). In the Pacific Northwest, mountain pine beetle activity did not affect wildfire likelihood, but instead had a thinning effect which decreased subsequent burn severity (Meigs et al. 2015, 2016). Following Douglas-fir beetle epidemics, stands have more large woody surface fuels which results in greater burn residency (Donato et al. 2013). Furthermore, lower crown bulk density reduces potential crown fire spread (Donato et al. 2013; Figure 1.1). Douglas-fir beetles also play an important role in post-fire tree mortality (Hood and Bentz 2007). Burned trees that would otherwise survive a fire are more likely than unscathed trees to be selected by bark beetles (Furniss 1965, Hood and Bentz 2007).   14 1.5 Mule deer  Management for mule deer [Odocoileus hemionus Rafinesque] has been a principal objective in dry forest ecosystems of British Columbia, as mule deer are an important game species (Blood 2000). Populations were negatively affected by reductions in winter habitat due to past logging (Armleder et al. 1989, Dawson et al. 2007). During the winter, deer at the periphery of their range are faced with traveling through deep snow and maintaining body temperature in cold conditions (Armleder and Dawson 1992). Winter food limitations result in extensive energetic deficits that compromise reproductive capabilities come spring (Armleder and Dawson 1992). To reduce energetic demands during winter, deer require a multilayered stand that is uneven aged with large, old (>140 years) Douglas-fir trees (Armleder et al. 1986, Armleder and Dawson 1992), with moderate to high canopy cover (Armleder et al. 1994). The canopy cover and multilayered structure intercept snow, reduce snowpack, and provide shelter, therefore lower energy requirements for travel and warmth (Armleder et al. 1986, Armleder and Dawson 1992). While mule deer prefer to feed on shrubs, they must dig through snow to access this during the winter (Waterhouse et al. 1991,1994). Instead, mule deer are forced to mainly feed on the more easily accessible foliage and arboreal lichen from large, old Douglas-fir (Waterhouse et al. 1991, 1994). As such the composition and structure of mule deer winter range is essential in maintaining a positive energy balance (Armleder et al. 1986, 1994). At the northern most extent of the deer’s distribution, maintaining winter range is a contentious issue as the Douglas-fir stands required by the deer to survive are also sought after by the forest industry (Armleder et al. 1998, Dawson et al. 2007).    15 1.6 Potential for land-use changes and climate change to alter mule deer habitat  Timber harvest, driven by economic incentives and the desire to meet production quotas, has created the need to re-assess land-use impacts to mule deer winter range. Past logging used more profitable even-aged management practices such as clearcutting which allowed for quick formation of second-growth forests (Armleder et al. 1989). Intensive harvest on winter range removed large amounts of Douglas-fir with a harvest interval too short for forests to reach late-successional stages required by mule deer, thereby reducing the available winter habitat (Armleder et al. 1989). As a result, a compromise between mule deer habitat needs and timber production was needed. Two harvesting solutions were put forth. The first involved even-aged cuts such as shelterwood, seed trees, or clear cuts which deer did not use until stands had ‘recovered’ (Armleder et al. 1989). Alternatively, an uneven-aged system was proposed with selective tree removal over a period of time (Armleder et al. 1989). Despite costs and potential problems, the uneven-aged system has been utilized most (Armleder et al. 1989, Dawson et al. 2007).   Mule deer winter range may not only be affected by changes in silvicultural practices, but also by a changing climate that alters disturbances regimes (Dale et al. 2001). Climate is anticipated to be more variable with greater fluctuations in temperature and precipitation (Easterling et al. 2000) with varying consequences for fire regimes (e.g. Westerling et al. 2003). The changes in precipitation coupled with earlier snowmelt, warmer springs and summers, is expected to increase fire frequency and extend the wildfire season (Heyerdahl et al. 2002, Westerling et al. 2006). A prolonged wildfire season with more frequent fires may cause dry forests to shift to a non-forested state (Turner et al. 2013) or,  16 in boreal and subalpine forests cause transitions to fuel-limited fire regimes (Westerling et al. 2011).   Similarly, a changing climate will also impact biotic disturbances such as insects (Ayres and Lombardero 2000). Insects are poikilothermic organisms and so warming temperatures are expected to accelerate their development potentially resulting in multiple generations per year (Bentz et al. 2010) as well as increase overwintering survival (Volney and Fleming 2007, Raffa et al. 2015) leading to greater outbreak frequencies (Murdock et al. 2013). Further, a changing climate will affect synchrony of insect emergence causing changes in geographical distribution (Jepsen et al. 2011, Weed et al. 2013), and possibly increasing disturbance extent (Volney and Fleming 2007). Such changes may result in novel disturbances and ecosystem states (Buma and Wessman 2011, 2013).  1.7 Management of mule deer habitat  In order to ensure habitat in a changing environment, mule deer winter range management (MDWRM) involves harvesting a low volume of timber, proportionate to abundance and diameter class of trees (Armleder et al. 1986, Dawson et al. 2007). Additionally, trees are removed in small groups, or as single trees, at minimum intervals of 30 years (Armleder and Dawson 1992, Dawson et al. 2007). This silvicultural treatment provides a multilayered, Douglas-fir dominated stand, with a clumpy spatial pattern of trees that maximizes litterfall and canopy cover for snow interception (Armleder et al. 1998). This treatment removes many small diameter trees thus creating  17 gaps for Douglas-fir regeneration (Dawson et al. 2007, Koot et al. 2015). Mule deer appear to use such treated areas in winter as frequently as untreated areas, as measured by the number of tracks found after snowfall (Armleder et al. 1998). The demographic impacts of MDWRM have not yet been assessed, however. In addition to incorporating harvesting to create habitat, MDWRM utilizes Douglas-fir beetle sanitation management tactics which include piling and burning logs >20cm, or application of anti-aggregation pheromones on downed logs to limit infestation (Armleder and Thomson 1984, Armleder et al. 1986, Dawson et al. 2007).  1.7.1 Potential interactions between mule deer winter range management and disturbance  The relationships between disturbances are highly complex (e.g. Buma and Wessman 2011, Hicke et al. 2012) and depend on stand attributes and climatic conditions (Hart et al. 2014a, b). Furthermore, silvicultural management may alter susceptibility to one or more disturbances and disrupt otherwise established interactions. Silvicultural impacts resulting from single-objective goals for disturbance management may be variable and are system dependent. Susceptibility to disturbance can be reliably determined using forest stand attributes as indicators of likely changes to disturbance severity expected following management actions (Hicks Jr. et al. 1987, Kulakowski and Jarvis 2011). Determining susceptibility may provide a way to quantify the effects of silvicultural management on disturbances while also revealing potential trade-offs (Hicks Jr. et al. 1987). The intent of my research is to quantify the changes in disturbance susceptibility resulting from silvicultural management while highlighting potential management trade-offs.   18  1.8 Research goals  1.8.1 Objective 1: Effects of mule deer winter range management on forest stand attributes and disturbance likelihood  The silvicultural treatment used in MDWRM serves to enhance mule deer winter range as well as allow timber extraction (Armleder et al. 1986, Dawson et al. 2007). The direction and duration of MDWRM effects on disturbances is unclear, however. My primary objectives were to quantify short- and long-term changes in forest stand attributes and to relate these changing attributes to stand susceptibility to 3 disturbances: western spruce budworm, Douglas-fir beetle, and wildfire. To do so I use long-term experimental forest attribute data that were collected in the field and then inputted into empirically-derived susceptibility models. I also developed a conceptual model of the hypothetical disturbance interactions in dry forests to highlight potential trade-offs or synergies resulting from applying MDWRM and its two associated Douglas-fir beetle management sanitation methods in the short and long-term.  1.8.2 Objective 2: Widespread application of mule deer winter range management on fire risk  Mule deer winter range management provides timber harvesting opportunities in otherwise restricted areas while simultaneously creating and enhancing habitat conditions for mule deer. As such, timber companies are interested in the potential prospects of incorporating and applying MDWRM to the landscape. However, the effects of widespread application of this silvicultural treatment remain unknown. Thus, my second objective was conducting a thought-experiment to extrapolate the potential effects of  19 widespread MDWRM on landscape composition and structure as it relates to fire risk using geographic information systems (GIS) and FRAGSTATS. To do so, I combined vegetation and fire risk data, and compared the fire risk of the current forest relative to the forecasted forest under MDWRM.   1.9 Thesis structure overview  In chapter 2, I assess the short and long-term changes in forest stand attributes. I then relate these attributes to susceptibility to disturbance using published, empirically-derived statistical relationships. Further, I develop a conceptual model illustrating disturbance interactions in dry forests to assess synergies and/or trade-offs resulting from MDWRM.  In chapter 3, I use geographic information systems (GIS) in combination with FRAGSTATS to assess the projected change in fire risk, using landscape composition and structure, following the implementation of widespread MDWRM.  In chapter 4, I summarize the main results of this study. Additionally, I briefly discuss the implications of applying MDWRM in the context of a changing climate and general problems with application of single-objective management.    20 2 Effects of mule deer winter range management on susceptibility to disturbance  2.1 Introduction  Sustainable forest management requires balancing increasing timber harvest and maintenance of species habitat. This can be a trade-off as one aspect often negatively affects the other (Rees 2003). For example, harvesting to maintain or enhance habitat may impose unacceptable logistic and timber removal restrictions (e.g. Armleder et al. 1989, Koot et al. 2015, Dellasalla et al. 2015, Sutherland et al. 2016), whereas unrestricted removal of timber may alter forest structure to the detriment of many wildlife species (Haggstrom and Kelleyhouse 1996, Sullivan and Sullivan 2001, Thompson et al. 2003, Perkins and Hunter 2006, Millington et al. 2011, O’Donnell et al. 2015). Furthermore, ecological effects of forest management may not be immediately obvious and can influence long-term forest structure and composition affecting ecological processes (Palik et al. 2002, Franklin et al. 2002), and compromising stand resilience to subsequent disturbance (e.g. Swetnam and Lynch 1993, Hadley and Veblen 1993, Radeloff et al. 2000, Allen et al. 2002, Hessburg et al. 2000, Hessburg et al. 2005, Taylor et al. 2006, Raffa et al. 2008).   The potential for forest management tactics to have lagged impacts on ecosystem processes, like disturbances, creates additional challenges for sustainable forestry. Management activities that do not facilitate or emulate natural disturbances may inadvertently create homogeneous landscapes prone to unprecedented disturbance severity. This is exemplified by the contribution of fire suppression to (i) the surplus of host trees for the recent mountain pine beetle outbreak [Dendroctonus ponderosae  21 Hopkins] (Taylor and Carroll 2004, Barclay et al. 2005, Taylor et al. 2006), and (ii) fuel loading for extensive fires in pine and mixed-conifer forests (Fulé et al. 1997, Hessburg et al. 2000, Hessburg et al. 2005). Moreover, management efforts that do not consider longer time scales may further exacerbate future disturbance events. For example, land clearing during early European settlement of North America followed by intensive fire suppression efforts have resulted in forests prone to defoliator outbreaks (Hadley and Veblen 1993, Swetnam and Lynch 1993). Similarly, timber extraction in the boreal mixed-wood forest has resulted in larger areas of early seral forest with greater flammability and wildfire risk (James et al. 2011).  A forest disturbance is a discrete event that alters resource and substrate availability in time and/or space (White and Pickett 1985), and can be associated with a biotic agent, such as a herbivorous insect, or abiotic agent, like wildfire. Within the interior dry forests of British Columbia (BC), Canada, the western spruce budworm [Choristoneura occidentalis Freeman] is one of the most significant biotic disturbance agents (Erickson 1992, Axelson et al. 2015). It is a defoliator native to western North America that feeds primarily on Douglas-fir (Fellin and Dewey 1986, Brookes et al. 1987, Ryerson et al. 2003). Feeding can result in reduced radial growth (Swetnam et al. 1985), top kill (Fellin and Dewey 1986) and even tree mortality (Brookes et al. 1987). The western spruce budworm undergoes episodic outbreaks within mature Douglas-fir forests (Unger 1983) that tend to coincide with periods of low precipitation (Brookes et al. 1987, Campbell et al. 2006, Flower et al. 2014a). Outbreaks occur at 30-year intervals and can last from 10 to 14 years (Swetnam and Lynch 1993, Ryerson et al. 2003, Campbell et al. 2006), but  22 the interval and duration will vary with the severity of the outbreak (Axelson et al. 2015). A stand’s susceptibility is directly related to the density of available hosts, and is also influenced by the vertical complexity of the stand (Fellin and Dewey 1986, Brookes et al. 1987). Dense, multilayered stands with a high proportion of host trees are very susceptible to outbreaks, especially under moisture stress (Brookes et al. 1987). Extreme drought conditions, however, are likely to negatively impact western spruce budworm survival (Campbell 1993).  The Douglas-fir beetle [Dendroctonus pseudotsugae Hopkins] is another important herbivorous insect within interior dry forests of BC (Erickson 1992). It is a bark beetle that feeds, reproduces, and spends most of its life cycle under the bark of Douglas-fir trees (Rudinsky 1962, Wood 1982, Aukema et al. 2016). The normative state of Douglas-fir beetle populations is the low density, endemic phase, characterized by attacks on defensively compromised hosts, windthrown trees, large branches, and stumps (Schmitz and Gibson 1996, Humphreys 2000). When conditions for beetle survival are good and there is an abundance of compromised hosts, as in drought (Hart et al. 2014a), beetles can rapidly increase in number and switch their attacks to large, well-defended hosts (Rudinsky 1962, Christiansen et al. 1987). The overall susceptibility of a stand to outbreak is related to the basal area and density of Douglas-fir trees (Negrón 1998), as well as the distribution and abundance of defensively impaired hosts (Dodds et al. 2006).   Biotic disturbance agents can leave behind a legacy, such as a change in stand structure, composition, or stand health, upon which subsequent disturbance agents may act (Buma  23 2015). The western spruce budworm and Douglas-fir beetle act in a synergistic manner, where moderate defoliation by the former acts as a stressor that predisposes trees to attack by the latter (Hadley and Veblen 1993, Marciniak 2015). Alternatively, Douglas-fir beetle attacks may increase susceptibility of stands to the western spruce budworm by creating gaps that promote the release of understory trees as has been suggested for the mountain pine beetle and its removal of overstory pine trees (Wilson et al. 1998).  In addition to biotic disturbance agents, fire can also influence the dynamics of subsequent disturbance, especially given its spatial and temporal variability (Kulakowski and Jarvis 2013). The severity of a fire is affected by a number of characteristics including tree density, ladder fuels, and presence or absence of understory plants (Allen et al. 2002, Schoennagel et al. 2004, Falk et al. 2007). Low-severity fires tend to burn the understory causing minimal over-story tree mortality, typically creating uneven-aged clumpy stands with greater structural complexity resulting in conditions conducive to more surface fires (Arno 1980, Allen et al. 2002, Schoennagel et al. 2004, Hessburg et al. 2005, Binkley et al. 2007). Conversely, high-severity fires often result in high over-story tree mortality producing even-aged stands that are less structurally complex and are subsequently prone to further high-severity fires (Schoennagel et al. 2004). Mixed-severity fires comprise a spatial and temporal mosaic of low- and high-severity fires producing patches of even- and uneven-aged stands creating heterogeneous structure at multiple scales (Schoennagel et al. 2004, Perry et al. 2011).    24 Abiotic and biotic disturbances also interact. Western spruce budworm defoliation events may reduce the severity of subsequent fires (Lynch and Moorcroft 2008) by feeding on foliage thereby reducing fuel load and/or fuel density and lowering vertical and horizontal fire propagation through the forest canopy (Cohn et al. 2014, Gavin et al. 2013). Further, defoliation results in a reduction in potential ladder fuels given that most mortality during an infestation occurs among small-diameter trees (Alfaro et al. 1982, Maclauchlan and Brooks 2009, Sturtevant et al. 2012). However, in the longer term there is also evidence that budworm outbreaks may increase coarse woody debris by causing tree mortality, thereby potentially influencing fire intensity and severity (Hummel and Agee 2003). The impacts of Douglas-fir beetle infestation on forest structure and subsequent fire are more nuanced as effects are dependent on the system and time since an outbreak (Jenkins et al. 2008). In stands undergoing a Douglas-fir beetle outbreak there tends to be an increase in the fine surface fuel load (<7.6cm diameter) resulting in an increase in the rate of spread of a fire as well as flame length (Jenkins et al. 2008, Hicke et al. 2012). In post-epidemic stands, Douglas-fir beetle activity increases the amount of large surface fuels also positively contributing to fire rate of spread and flame length (Hicke et al. 2012), while creating gaps in the forest canopy by reducing aerial fuels (Jenkins et al. 2008, Donato et al. 2013). The reduction in aerial fuels and creation of a discontinuous canopy would suggest a reduction in fire severity by inhibiting the occurrence of high-severity crown fires (Simard et al. 2011, Donato et al. 2013).   While biotic disturbance agents tend to affect fire over a shorter time period by changing fuel availability, fire regimes tend to affect biotic disturbance agents over the long-term  25 by influencing stand structure. Low-severity fires tend to burn more frequently to produce stands with complex vertical and horizontal structure (Allen et al. 2002, Schoennagel et al. 2004, Binkley et al. 2007, Beaty et al. 2007); conditions that are beneficial to the western spruce budworm given the greater food availability and increased inter-tree dispersal success associated with a more stratified stand structure (Fellin and Dewey 1986, Brookes et al. 1987). Also, low-severity fires typically do not tend to kill big, old trees leaving high-quality hosts for the Douglas-fir beetle (Negrón 1998, Aukema et al. 2016). Furthermore, in the short-term such fires may injure the old trees creating a pulse of preferred available host for the bark beetle (Furniss 1965, Hood and Bentz 2007). High-severity fires, by contrast, burn both surface and crown fuels, creating even-aged forests that are not vertically complex (Schoennagel et al. 2004), thereby negatively affecting the western spruce budworm (McCullough et al. 1998). In the short-term following a high-severity fire, a pulse of dead trees may be utilized by the Douglas-fir beetle (personal observation); however, in the longer term the stand structure resulting from a high-severity fire may not be suitable for the Douglas-fir beetle as there may be fewer large, mature trees (Humphreys 2000), although these denser stands may be less vigorous (Negrón 1998).  Disturbances and their interactions have helped shape the composition and structure of the mixed-conifer dry forests of interior BC (Pojar and Medinger 1991). The dry forests are mainly composed of Douglas-fir [Pseudotsuga menziesii (Mirb.) Franco], occurring in the over- and understory (Hope et al. 1991). On moister sites, lodgepole pine [Pinus contorta Dougl. ex Loud. var. latifolia Engelm.], trembling aspen [Populus tremuloides  26 Michx.], hybrid white spruce [Picea engelmannii x glauca], and paper birch [Betula papyrifera Marsh.] may occur (Klinka et al. 2004, Koot et al. 2015). Forests are typically open and are structurally heterogeneous (Allen et al. 2002), with trees occurring alone or in clumps (Maclauchlan and Brooks 2009). Forest composition and structure are the product of interactions among the 3 primary disturbances of this forest type: fire, the western spruce budworm and the Douglas-fir beetle (Hope et al. 1991, Erickson 1992, Alfaro et al. 1982, Powers et al. 1999, Maclauchlan and Brooks 2009, Axelson et al. 2015).  The mixed-conifer dry forests in the central interior of BC have become the home of competing wildlife habitat and timber management objectives. Mule deer [Odocoileus hemionus Rafinesque] is an ungulate species that depends upon Douglas-fir stands as critical winter habitat in BC (Armleder and Dawson 1992, Armleder et al. 1994). However, Douglas-fir is a very valuable timber species that is preferred for harvest (Armleder et al. 1998, Dawson et al. 2007). Mule deer utilize multilayered old Douglas-fir forests (>140 years) with moderate to high crown cover as winter range (Armleder and Dawson 1992, Armleder et al. 1994). There they feed on Douglas-fir foliage and arboreal lichen that drop to the ground due to wind and snow breakage (Waterhouse et al. 1991, 1994). Survival of mule deer in BC is dependent on this winter habitat because it provides snow interception, thermal cover and high quality food (Armleder and Dawson 1992).   27 Conservation of mule deer populations has become a high priority management objective in BC where it is an important game species (Blood 2000, Lehmkuhl et al. 2001). Populations were believed to be declining due to winter habitat reduction from past forest management practices such as clearcutting (Armleder et al. 1989). In an attempt to increase mule deer populations, silvicultural treatments to create conditions that emulate critical aspects of mule deer winter range were devised (Dawson et al. 2007). Treatments involve the proportional removal of trees based on their diameter and abundance with a focus on removing small-diameter trees to create gaps and fashion a multilayered, uneven-aged stand with a clumpy tree distribution over the long-term (Armleder et al. 1986, Armleder and Dawson 1992, Dawson et al. 2007). Given that silvicultural treatments to improve mule deer winter range require the selective harvest of valuable Douglas-fir, it has the potential to satisfy both the objectives of wildlife habitat conservation and timber management. However, direct impacts of treatment on susceptibility of stands to independent and interacting disturbances by western spruce budworm, Douglas-fir beetle and fire have not been considered.  In this study, I quantified the impacts of mule deer winter range management (MDWRM) on short- and long-term changes in forest stand attributes, and then using these attributes I assessed the susceptibility of the forest to three disturbances: the western spruce budworm, Douglas-fir beetle and wildfire. Disturbance susceptibility was evaluated using empirically-derived, published models. Also, I developed models depicting disturbance interactions in dry forests to assess trade-offs and synergies that may result from MDWRM. I hypothesized that in the short-term, the removal of sub-canopy trees will  28 reduce host availability and suitability to the western spruce budworm and Douglas-fir beetle and decrease sub-canopy ladder fuels thereby lowering crown fire likelihood. In the long-term however, recovery of understory trees within gaps will increase host availability and suitability to the western spruce budworm, and increase ladder fuels and crown fire susceptibility, whereas, retention of large, old trees will result in abundant hosts for the Douglas-fir beetle. Therefore, I predicted that changes to forest stand attributes resulting from MDWRM will lower forest susceptibility to western spruce budworm, Douglas-fir beetle and wildfire in the short-term while susceptibility to the budworm, beetle and wildfire will be greater in the long term.  2.2 Materials and methods  2.2.1 Study area: Knife Creek Alex Fraser Research Forest, BC  The Knife Creek block of the University of British Columbia’s Alex Fraser Research Forest (52o03’N, 121o53’W) comprises 3487 hectares (ha) of rolling terrain located in the Cariboo Forest Region, 15 kilometers southeast of Williams Lake British Columbia (Armleder et al. 1994, Day 1997; Figure 2.1). Most slope aspects are south and west facing with elevation ranging from 700 to 1000 m (Armleder et al. 1994). The area is characterized by a mean annual precipitation of 569 mm and temperature of 4.7oC with mean minimum and maximum summer temperatures of 7.3 and 22.1oC respectively (Wang et al. 2012). The frost-free period lasts 104 days from May to September with 277 mm of precipitation falling during this period (Wang et al. 2012). The soils are mainly composed of well-drained Brunisols and Luvisols (Hope et al. 1991).  Knife Creek is located in a fire-dominated uneven-aged dry forest with the late successional species  29 being Douglas-fir generally forming small closed-canopy patches interspersed among large opening (Klinka et al. 2004).                            Figure 2.1: The location of the mule deer winter range experimental blocks at the Alex Fraser Research Forest. (A) The Knife Creek Block is located in central British Columbia, Canada. (B) The Alex Fraser Research Forest’s Knife Creek block in central British Columbia. (C) Three paired blocks were established to study the effects of mule deer winter range management. Treated blocks (polygons) with respective control pairs immediately adjacent to the treated blocks.  Historically, frequent fires, management for food-plants and animals by First Nations, various logging practices, fire suppression, and cattle ranching have shaped the forest’s composition and structure (Klinka et al. 2004, Day 2007). Following the establishment of the Alex Fraser Research Forest in 1987, little harvesting was done at Knife Creek although salvage logging methods were employed to address Douglas-fir beetle activity along with pre-commercial thinning and brushing (Day 2007). Harvesting of mountain  30 pine beetle (Dendroctonus ponderosae) affected trees began in 1998 and ceased in 2004 (Day 2007).  2.2.2 Mule deer winter range experiment at Knife Creek  In 1983, prior to the establishment of the Alex Fraser Research Forest, part of the Knife Creek block was dedicated to mule deer winter range habitat research (Figure 2.1; Armleder and Thomson 1984). A controlled experiment was established with three treatment (33.3, 35.7 and 17.3ha) and corresponding control areas (27.9, 30.4, and 16.7ha) (Table 2.1, Figure 2.1c). Silvicultural treatments intended to improve mule deer winter range were applied in 1984 and again to the same areas in 2014. In 1984, 16% of volume across all merchantable diameter classes (DBH ³12.5 cm) was removed in proportion to their abundance by harvesting in clumps while leaving trees DBH<12.5 cm, to create a multilayered stand and promote thermal and security cover (Armleder and Thomson 1984, Armleder and Dawson 1992, Armleder et al. 1998). In 2014, the low volume harvest was repeated removing 12-22% of volume per area (Koot et al. 2015). Harvesting in the second treatment served to maintain a multilayered structure and was adjusted to emphasize gap creation to increase forage for mule deer (Armleder and Dawson 1992), and address declining tree vigour (Koot et al. 2015). The second treatment involved harvesting across all diameter classes paired with an emphasis on removal of smaller diameter trees (Koot et al. 2015).   31  Table 2.1: A summary of mean tree densities (and standard deviation) of the treatment blocks prior to the mule deer winter range management silvicultural treatment in 1984 provided by the Forest Valuation Branch. Densities were measured using variable radius plots (basal area factor 8) for trees ³12.5 cm.  Block Number of plots Area (ha) Density pre-harvest 1984 (stems/ha) 222 13 35.7 472.5 (278.2) 229 13 33.3 372.6 (281.0) 232 8 17.3 356.9 (255.5)  Permanent sample plots were established throughout the control and treatment blocks in 1984 (Armleder and Thomson 1984, Koot et al. 2015). The sample plots were located at 100m intervals along 15 systematically placed transects creating a grid (Armleder and Thomson 1984). The plots were sampled in 1984 and in 2013 prior to the treatments to assess forest stand attributes. I sampled the treatment area in 2014 and 2015, before and after silvicultural treatment 2, respectively, and the control area in 2015 only, to address questions regarding changes in forest stand attributes and disturbance susceptibility through time.  2.2.2.1 Field sampling: Forest structure and disturbance susceptibility in 2014-2015  Several biophysical attributes were measured at the centre of the permanent plots:  location in UTM coordinates, elevation in metres above sea-level (m.a.s.l.), slope aspect and angle in degrees. To quantify changes in forest structure, I used the N-tree sampling method (Jonsson et al. 1992) to assess the treatment area in 2014 and 2015 and the control area in 2015. From the centre of each permanent plot, I measured the distance to the 10 closest canopy (emergent and dominant height classes, Oliver and Larson 1996)  32 and 10 closest sub-canopy (intermediate and suppressed height classes) trees. In the treated area, different trees were measured in 2015 as some of the trees measured in 2014 were harvested. I recorded species, state (dead or alive), DBH, and canopy class. The height and crown base height of the sampled canopy and sub-canopy trees were taken in 2015. The crown base height was defined as the lowest living or dead branches, where a fire could spread up the tree crown (Scott and Reinhardt 2007). The number of regenerating trees (height>30cm, DBH<12.5 cm) was counted and the total percent cover of understory plants including living and dead shrubs, herbs and grasses, was estimated in fixed area plots (Marshall and LeMay 2005) with a 3.99m radius representing 0.005 ha. I also recorded height of the tallest individual shrub.  Crown cover and fuels were quantified using the permanent plot centres. To quantify crown cover, I photographed the canopy using a 35 mm digital camera fitted with a fish-eye lens with a 180o field of view (Bréda 2003). Photographs were taken ca. 1m above the ground. I measured the depth of leaf litter and duff layers at plot centre and used the line-intercept method to quantify dead woody surface-fuel loads along randomly-oriented, 30 m transects (Brown et al. 1982). Downed wood intersecting each transect was measured using calipers and stratified by size class (0-<0.6, 0.6-<2.5, 2.5-7.6, >7.6 cm). Wood in 0-<0.6 cm and 0.6-<2.5 cm size classes was tallied along the first 1m and 2 m of the transect. For each piece of wood in the 2.5-7.6 cm and >7.6 cm classes, I recorded the location along the transect, species, and differentiated logs resulting from harvesting in 2014 based on the presence of bark, branches and needles and cut log ends.   33 At every second plot in each control and treatment area, an increment core was taken from each sampled tree in 2015. A single core was removed from each tree while staying at or below 30 cm of the base of the tree. A tree was cored multiple times to ensure that cores intercepted pith, or based on arched rings, were very near pith.  2.2.2.2 Deriving plot-level attributes in 2014-2015 I compared 16 forest stand attributes of the treatment area in 2014 and 2015 (n=34) with the control (n=33) to determine differences through time (Table 2.2). Canopy and sub-canopy tree densities per hectare were calculated using a negative binomial distribution to account for the clumped tree pattern (Lessard et al. 2002) in both control and treated areas. Counts of regenerating trees in the 50 m2 plots were scaled to one hectare. Basal areas of the canopy and sub-canopy strata were calculated as the average tree basal area per stratum multiplied by the density of each stratum.  2.2.3 Susceptibility to disturbance  I used the biophysical attributes and information derived from the increment cores to calculate disturbance susceptibility indices at the plot-level in the treatment area for 2014 and 2015 (n=34), and the control area for 2015 (n=33). The susceptibility indices for each disturbance were subsequently compared before and after the silvicultural treatment in 2014 and relative to the control plots to quantify the effects of MDWRM on disturbance susceptibility, as described below.   34 2.2.3.1 Western spruce budworm susceptibility I used the plot-level susceptibility index developed by Wulf and Carlson (1985; Appendix A) to assess susceptibility to western spruce budworm. The index incorporates a total of 9 variables: 6 that represent plot-level forest composition, structure, and host-tree availability, 2 variables representing climate attributes affecting outbreaks, and 1 variable representing landscape-level composition. Each variable was subsequently converted to weighted index values based on empirical relationships and when multiplied together provide an overall susceptibility index, ranging from 0 to 100 (Wulf and Carlson 1985; Appendix A). Indices were converted to a susceptibility ranking based on the overall values grouped into three categories: low (overall susceptibility index = 0-20), moderate (21-50), and high (51-100) (Appendix A).  Four variables were used to determine plot-level host tree availability based on structure and composition data (Appendix A). Stand structure was represented by the coefficient of variation of the DBH (CVDBH) of the sampled trees at the permanent plots (Wulf and Carlson 1985; Appendix A). Total percent crown cover of all species, percent host crown cover, and percent late-successional host cover were derived from hemispherical photographs by calculating ‘site openness’ using the program Gap Analyzer Version 2 (Frazer et al. 1999). ‘Site openness’ accounted for both topographic shading and proportion of open sky (Frazer et al. 1999) and was used to assess percent total crown, percent host, and percent late-successional host crown cover by subtracting ‘site openness’ from 100%. Percent host crown cover was the ratio of host crown cover and total crown cover, multiplied by 100. Percent late-successional host cover was the ratio of  35 late-successional host crown cover and total crown cover multiplied by 100. Douglas-fir made-up ³95% of trees in each plot, therefore crown cover of host trees, late-successional host crown cover and total crown cover values were nearly the same and percent host crown cover and late-successional host crown cover were >95% in each plot.   The final two variables representing plot-level host tree availability were ‘maturity’ and ‘vigour’. Maturity was derived using the increment cores extracted from 696 trees. Increment cores were prepared and mounted according to Stokes and Smiley (1968), measured in Coorecorder (Larsson 2011a) and cross-dated in CDendro (Larsson 2011b) against a master chronology provided by Daniels (2004). For cores that did not intercept the pith (n=553), the number of missed rings (mean=7.67; standard deviation= 9.77) was estimated from geometric measurements of the inner-most rings (Norton et al. 1987, Duncan 1989). Tree age was the difference between the calendar year of the outer-most ring with complete latewood (e.g. 2014 for living trees) and the year of the pith, plus one, plus the number of years for trees to grow to coring height (Daniels 2004). Maturity was calculated as the mean of the basal-area-weighted age of the host trees in each plot (Wulf and Carlson 1985; Appendix A). Plot ‘vigour’ was determined by comparing the plot-level basal areas relative to the maximum basal areas observed in the control plots in 2015 (Wulf and Carlson 1985; Appendix A). Since <30% of host trees in each plot were affected by insects or diseases at the time of sampling, all plots were considered free of biotic stress (Wulf and Carlson 1985).   36 Climatic and landscape composition variables were determined based on the location, forest type at the plot-level, and in the forest surrounding the study area (Wulf and Carlson 1985; Appendix A). Since the Knife Creek study area encompassed 3487 ha with little variation in elevation and aspect, the climatic and landscape composition variables and their index values were considered constant across all plots. In Wulf and Carlson’s (1985) model, site climate and regional climate determine the habitat type or late-successional plant community and the degree of continentality (Appendix A). Site climate was classified as a warm dry Douglas-fir habitat and regional climate was dry. Finally, host-type continuity characterized the proportion of available host in the 400 ha surrounding the study area (Wulf and Carlson 1985; Appendix A). The Knife Creek study area was surrounded predominantly by Douglas-fir-dominated forests.  2.2.3.2 Douglas-fir beetle susceptibility I used the probability of infestation classification tree and potential mortality regression model created by Negrón (1998) to assess plot-level susceptibility to Douglas-fir beetle. The probability of infestation classification tree is an empirically-derived model that uses the percentage of basal area occupied by Douglas-fir and tree density to classify stand susceptibility to Douglas-fir beetle (Negrón 1998; Appendix B). The classification tree results in 4 classes with probabilities, expressed as a percentage, ranging from 0 to 100 corresponding to low, moderate, high, and very high probabilities of infestation (Appendix B).  Further, the potential mortality of Douglas-fir (m2/ha) that could result from an infestation was calculated according to Negrón (1998) (Appendix B). I used this linear equation to determine plot-level potential mortality by using the calculated basal  37 area occupied by Douglas-fir at each plot and subsequently determining the potential basal area of Douglas-fir killed by Douglas-fir beetle at each plot.  2.2.3.3 Crown fire likelihood I used Crown Fire Initiation and Spread (CFIS) software developed by Alexander et al. (2006) to determine the likelihood of crown fire occurrence. The software uses a multiple logistic regression model to generate a probability of crown fire occurrence expressed as a percent ranging from 0-100%, where a crown fire is most likely when probabilities exceed 50%, while a surface fire is most likely below the 50% threshold (Alexander et al. 2006).   The four input variables are the fuel strata gap, estimated fine fuel moisture, wind speed and surface fuel consumption (Appendix C). The first three variables represent the fire environment, and the final variable is a descriptor of fire behaviour (Alexander et al. 2006, Cruz et al. 2004). The fuel strata gap for each plot was calculated as the difference between the sub-canopy tree crown base height and the height of the tallest shrub (Cruz et al. 2004). The fuel strata gap was assumed to have remained the same from 2014 to 2015. Fine fuel moisture was estimated from an empirically-derived model including 3 variables describing the physical environment at each plot and 4 fire weather variables representing the study area (Alexander et al. 2006). Slope angle and aspect were measured in the field and the degree of shading (³ 51% crown closure <50%) was determined from the hemispheric photographs. Fire weather variables were calculated from the 2000-2012 records for Knife Creek (station number: 225, lat: 52.04972, long: - 38 121.87383, elevation: 821m, 1999-2012; BC Wildfire Service 2014). In the Cariboo Fire Zone, 76% of lightning ignitions occur on days with high or extreme fire danger (BC Wildfire Service, unpubl. data). To represent conditions when risk of wildfire is greatest, I selected the 21 days in July and August (2000-2012) when fire danger was high or extreme and averaged values for temperature (24.6 oC), relative humidity (30.4%), and wind speed at 10m above the ground (9.33m/s). For the final fire weather variable, I used August as the month to represent conditions most conducive to fire.   Dead-woody surface fuel load was calculated at the plot-level by combining the load (kg/m2) of 1-, 10-, 100-, and 1000-hour fuels corresponding to the downed woody surface fuel diameter classes of 0-<0.6, 0.6-<2.5, 2.5-7.6, and >7.6 cm respectively. Dead-woody fuel load for each surface fuel diameter class was calculated by multiplying the specific wood density for Douglas-fir of 440kg/m3 (Gonzalez 1990) by the number of individuals in each diameter class that intersected their respective transect lengths. One and 10-hour fuel load classes were calculated individually while 100- and 1000-hour fuel load classes were combined to address large dead woody surface fuels. One and 10-hour fuels were assumed to have remained the same from 2014 to 2015.  The fire behaviour descriptor used in the CFIS model, surface fuel consumption, is a class variable that serves to describe the amount of available fuel consumed during flaming combustion serving as a proxy for flame height, depth, and length which influence the heat reaching aerial fuels (Cruz et al. 2004). Surface fuel consumption is a class variable with 3 options: <1kg/m2, 1-2 kg/m2 and >2 kg/m2.  Consumption of surface  39 woody fuels during prescribed burns in dry forests of BC ranged from 25-45% (Daniels et al. 2014, R. Kubian, Parks Canada, unpubl. data), and wildfires tend to burn under more extreme fire weather conditions than prescribed fires (e.g. Natural Resources Conservation Service 2009), therefore, I estimated that 50 % of the calculated surface woody fuels would be consumed. As a result, calculated fuel loads were multiplied by 0.5 and I assigned the resulting values into one of the fuel classes.  2.2.4 Forest recovery within a mule deer management context  I assessed forest recovery following the silvicultural treatment in 1984 using basal area increments (BAI) calculated for the 696 cored trees. Basal area increments (Biondi and Qeadan 2008) were derived from the cross-dated ring widths described above (Appendix D). Forest recovery was determined by comparing the mean BAI 30 years prior (1955-1984) and mean BAI 30 years after the implementation (1985-2014) of MDWRM in both the control and treatment areas to identify a silvicultural effect. The period of 30 years was used as this is the minimum harvest rotation length under MDWRM (Dawson et al. 2007), the interval used at Knife Creek.  2.2.5 Data analyses  2.2.5.1 Forest structure in 2014-2015 I used a randomized complete block single-factor mixed-effects model with subsampling to compare forest stand attributes between the control area and with the treatment area before and after the silvicultural treatment in 2014. This same model was used to assess recovery of the forest following the 1984 silvicultural treatment. The randomized  40 complete block accounted for the division between control and treatment areas, thus accounting for variation within each area to ensure observed differences were due to the silvicultural treatment. Within the treatment and control areas, the permanent plots served as subsamples to test for differences among treatment levels. The silvicultural treatment was the single factor tested when comparing forest structure and consisted of 3 levels: treatment area in 2014, treatment area in 2015, and control area in 2015; for 6 variables, there were only 2 levels since measurements were not made in 2014. The silvicultural treatment was also the single factor tested to assess the recovery of the forest, as measured by BAI, and consisted of 4 levels: treatment and control areas 30 years prior to implementation of MDWRM (1955-1984) and treatment and control areas 30 years after the implementation of MDWRM (1985-2014). I analyzed canopy and sub-canopy tree BAI separately.   An analysis of variance (ANOVA) was conducted on each of the 18 variables (16 forest attributes and basal area increment of canopy and sub-canopy trees) followed by pair-wise analyses using least squares means and a Bonferroni correction to determine any differences between silvicultural treatment levels. Variables were transformed using log base 10, square root and BLOM transformations (Beasley et al. 2009) as required to meet assumptions of ANOVA. All analyses were conducted in R 3.2.2 (R Core Team 2015) using Type I sum of squares, and residual maximum likelihood. This analysis was also conducted on western spruce budworm susceptibility, Douglas-fir beetle potential mortality, and likelihood of crown fire occurrence. Additionally, I conducted a Chi- 41 squared test to determine the change in number of susceptible plots to Douglas-fir beetle resulting from MDWRM.  2.2.5.2 Wildfire hazard A mixed-effects multiple linear regression was used to identify forest stand attributes associated with high likelihood of crown fire occurrence. Two groups were created to account for the variation within fire types (surface or crown). The threshold value used to determine fire type was 50%, where above this value a fire was classified as a crown fire and below, a surface fire (Alexander et al. 2006). I developed linear models incorporating forward, backward and stepwise selection procedures using the Akaike Information Criterion (AIC) for initial selection of variables. Following initial model selection procedures, model parsimony was iteratively determined using likelihood ratio tests, while ensuring assumptions were met. I selected the best model and determined the relationship of each included variable with likelihood of crown fire occurrence.  2.3 Results  2.3.1 Changes in forest structure  Seven of the 16 forest stand attributes differed significantly between the control (2015) and treatment areas (2014 and 2015) before and after the silvicultural treatment (Table 2.2). The 2014 silvicultural treatment significantly decreased canopy cover, and sub-canopy tree density and basal area (Table 2.2). Canopy tree density differed significantly, but the subsequent pair-wise analysis revealed no differences among treatment levels suggesting a decrease in absolute density while means did not differ among time periods  42 (Table 2.2). Differences in canopy cover, and sub-canopy tree density as well as basal area were not significant between the control area in 2015 and the treatment area in 2014, suggesting that the forest recovered over a 30-year interval since the 1984 harvest (Table 2.2).   The leaf area index (LAI) between the control (2015) and treatment areas (2014 and 2015) all significantly differed from each other. The treatment area in 2015 had the lowest LAI while the control area had the greatest LAI (Table 2.2). Large fuels in the treatment area in 2014 were significantly different from both the control and treatment areas in 2015 (Table 2.2). Large fuels in the treatment and control areas in 2015 did not differ significantly. Leaf litter depth was significantly different in the treatment area in 2015 from the control area in 2015 (Table 2.2).  43 Table 2.2: Comparison of forest structure attributes among plots in the control and treatment areas that were sampled in 2014 and/or 2015 following implementation of mule deer winter range management. Each variable was transformed to meet assumptions of ANOVA. For variables that differed significantly (p-values in italics), means (standard errors) followed by the same superscripts do not differ significantly based on the applied Bonferroni correction (α=0.05).  Attribute Transformation Control 2015 (n=33) Treatment 2014 (n=34) Treatment 2015 (n=34) p-value Canopy tree density (stems/ha) Logarithm 179.1 (28.6)a 202.8 (28.5)a 141.4 (28.5)a 0.038 Sub-canopy tree density (stems/ha) BLOM 331.3 (63.7)a 253.0 (63.0)a 174.7 (63.0)b 0.001 Regeneration density (stems/ha) Square root 1026.6 (395.8) 2016.0 (393.6) 1422.1 (393.6) 0.066 DBH coefficient of variation BLOM 2.852 (0.005) 2.846 (0.005) 2.847 (0.005) 0.539 Leaf Area Index BLOM 1.389 (0.035)a 1.282 (0.034)b 1.084 (0.034)c <0.001 Mean plot DBH BLOM 28.5 (0.8) 28.6 (0.8) 30.47 (0.8) 0.076 Canopy tree basal area (m2/ha) Logarithm 22.52 (2.51) 27.06 (2.47) 21.92 (2.47) 0.3673 Sub-canopy tree basal area (m2/ha) BLOM 10.67 (2.14)a 5.97 (2.10)a 4.52 (2.10)b 0.007 Sub-canopy tree canopy base height (m) Logarithm 2.8 (0.5) - 3.3 (0.5) 0.171 Canopy cover (%) BLOM 73.53 (0.94)a 71.45 (0.93)a 65.12 (0.93)b <0.001 Understory cover (%) Logarithm 32 (4.3) - 31 (4.2) 0.858 100- and 1000-hour fuels (kg/m2) BLOM 2.35 (0.46)a 0.84 (0.46)b 2.14 (0.46)a 0.001 10-hour fuels (kg/m2) BLOM 0.48 (0.09) - 0.55 (0.09) 0.783 1-hour fuels (kg/m2) Square root 0.24 (0.03) - 0.27 (0.03) 0.444 Leaf litter depth (cm) Square root 0.7 (0.1)a - 1.1 (0.1)b 0.008 Duff depth (cm) BLOM 1.0 (0.2) - 1.2 (0.2) 0.378  44 2.3.2 Forest recovery within a mule deer winter range management context  Mean BAI of sub-canopy trees was significantly greater following the implementation of MDWRM (F3,60: 4.47, P: 0.007; Table 2.3), suggesting that remaining sub-canopy trees benefited from the silvicultural treatments. Furthermore, mean sub-canopy tree BAI did not differ in the control area before or after the implementation of MDWRM suggesting a treatment effect (Table 2.3). The mean BAI of canopy trees did not differ before or after implementation of MDWRM (F3,60: 0.09, P: 0.967; Table 2.3).  Table 2.3: Comparison of the mean basal area increment (BAI) per year (mm2/year) of canopy and sub-canopy trees in the control and treatment areas prior (1955-1984) and after (1985-2014) the implementation of mule deer winter range habitat management. BAI was compared using ANOVA and using a BLOM transformation to meet assumptions. For significantly different variables (p-value in italics), determined by a Bonferroni correction, means (standard errors) followed by the same superscripts did not differ significantly (α=0.05).    2.3.3 Western spruce budworm susceptibility  The treated and control areas had a moderate western spruce budworm susceptibility ranking. Susceptibility to western spruce budworm did not differ among any of the treatment levels (F2,96: 1.59, P: 0.210). Mean (± standard errors) susceptibility values were 44.0 ±1.8, 40.1 ±1.8, 41.6 ±1.8 for the treatment area in 2014, and 2015, and control area in 2015 respectively.   Canopy position          Mean BAI in control (n=16)          Mean BAI in   treatment (n=17) p-value 1955-1984 1985-2014 1955-1984 1985-2014  Canopy  1314.5 (186.0) 1007.1 (186.0) 1024.7 (180.4) 1022.7 (180.4) 0.967 Sub-canopy 248.5 (22.2)a 227.2 (22.2)a 215.9 (21.5)a 332.1 (21.5)b 0.007  45 2.3.4 Douglas-fir beetle susceptibility  Douglas-fir beetle susceptibility was either moderate or high at all plots since the relative basal area of Douglas-fir always exceeded 89.4%. The reduction in density of host trees with DBH³12.5 cm to 292 trees/ha or less changed the probability of infestation from high to moderate in 15 of 34 sites in the treatment area (c2 1=13.8, P<0.0001; Table 2.4). Twenty-eight of 34 plots in the treatment area in 2014 had a high probability of infestation suggesting recovery of host tree density above the 292 trees/ha threshold, and resulting in a similar number of high susceptibility plots to that of the control area in 2015 (Table 2.4). Despite a reduction in tree density, the potential mortality from Douglas-fir beetle did not differ significantly among the 2015 control or treatment in 2014 or 2015 areas (F2,96 : 1.09, P: 0.339), reflecting the high relative basal area occupied by Douglas-fir in the study area. Potential reduction in mean basal area due to predicted Douglas-fir beetle-caused mortality averaged 14.56 ±1.18, 11.99 ±1.18, and 11.88 ±1.19 m2/ha for the treatment area in 2014, and 2015, and the control area in 2015, respectively.  Table 2.4: Comparison of the number of plots classified as high or moderate probability of infestation by the Douglas-fir beetle in the control and treatment areas following mule deer winter range management. A c2 test revealed that significantly fewer plots had a high probability of infestation in the treatment area in 2015 relative to 2014. Probability of infestation by the Douglas-fir beetle was assessed using the regression tree developed by Negrón (1998).    Probability of infestation  Number of plots Control 2015 Treatment 2014 Treatment 2015  Moderate (42%)  6 6 21     High (71%)  27  28  13    46 2.3.5 Crown fire likelihood  Likelihood of crown fire occurrence differed significantly among treatment levels (F2,96: 5.57, P: 0.005; Figure 2.2). The treatment area in 2014 had a significantly lower likelihood of crown fire occurrence relative to the control area in 2015 (Figure 2.2). The treatment area in 2015 did not significantly differ from the treatment in area 2014 or the control in area 2015 (Figure 2.2).                   Figure 2.2: Comparison in the mean (standard errors) likelihood of crown fire occurrence following mule deer winter range management in the treatment area in 2014 and the treatment and control areas in 2015 of Knife Creek. Bars with the same superscript letters did not differ significantly after applying a Bonferroni correction (a=0.05).  2.3.6 Wildfire hazard  Crown base height of sub-canopy trees, 10-hour fuels and large fuels were found to be the strongest predictors of the likelihood of crown fire occurrence (Tables 2.5, 2.6). All model selection methods converged onto a single model with an AIC of 290.08 comprising 6 variables: litter depth, sub-canopy tree density, canopy cover, crown base height of sub-canopy trees, 10-hour fuels and large fuels.  Subsequent likelihood ratio tests revealed that litter depth, canopy cover and sub-canopy tree density were not strong predictors of likelihood of crown fire occurrence and were removed from the model  47 iteratively (Table 2.5). In the final model the likelihood of crown fire occurrence was inversely related to the crown base height of sub-canopy trees and positively related to the abundance of 10-hour and large fuels (Table 2.6; Figure 2.3). Likelihood of crown fire occurrence declined 4% for every meter increase in crown base height of sub-canopy trees (Figure 2.3). For every unit increase of fuel in kg/m2, likelihood of crown fire occurrence also increased by 7% and 3% for 10-hour and large fuels respectively (Figure 2.3).  Table 2.5: The likelihood ratio test values for the variables making up the wildfire hazard model in 2015 following initial model selection. Significantly contributing variables (p-values in italics), were determined using α=0.05 and variables not contributing to the model were removed from subsequent iterations.  Model iteration Variable Likelihood ratio value p-value 1  Litter depth 2.10 0.148 Sub-canopy tree density 5.31 0.021 Canopy cover 4.49 0.034 Crown base height of sub-canopy trees 30.47 <.001 10 hour fuels 10.41 0.001 Large fuels 28.09 <.001 2    Canopy cover 3.61 0.058 Sub-canopy tree density 5.11 0.024 Crown base height of sub-canopy trees 29.50 <.001 10 hour fuels 12.02 <.001 Large fuels 27.60 <.001 3  Sub-canopy tree density 3.67 0.055 Crown base height of sub-canopy trees 28.67 <.001 10 hour fuels 9.18 0.003 Large fuels 24.82 <.001          48 Table 2.6: The coefficients, standard error, t- and p-values of the variables in the final wildfire hazard model in 2015.   Variable  Coefficient Standard Error t-value p-value Intercept 54.84 23.96 2.288 0.026 Crown base height of sub-canopy trees -4.406 0.7361 -5.986 0.000 10 hour fuels 7.143 2.347 3.043 0.003 Large fuels 2.944 0.5381 5.471 0.000                       Figure 2.3: The relationships between crown base height of sub-canopy trees (top), large fuels (centre) and 10-hour fuels (bottom) with likelihood of crown fire occurrence.   49 2.4 Discussion  2.4.1 Short and long-term effects of winter range management on stand structure  Silvicultural tactics intended to improve winter habitat by means of mule deer winter range management significantly altered dry forest stand structure in the short and long-term. The silvicultural treatment in 2014 resulted in a short-term reduction in sub-canopy tree density and basal area, canopy cover, LAI, and an increase in large surface fuels. The observed changes in measured forest stand attributes at Knife Creek are due to the silvicultural treatment as the area was not affected by any other stand-level disturbance between the field sampling done in 2014 and 2015. To enhance vertical complexity and create gaps (Koot et al. 2015), the second entry of the silvicultural treatment combined the retention of large, old canopy Douglas-fir, while removing a larger proportion of sub-canopy trees (DBH³12.5cm) and leaving the regeneration (height>30cm, DBH<12.5 cm) intact. The significant decrease in sub-canopy tree density and basal area, while maintaining a constant density of regeneration and a relatively unchanged canopy layer through time, is expected to maintain the multilayered stand required by mule deer (Armleder and Dawson 1992, Dawson et al. 2007).   The silvicultural treatment also reduced canopy cover and LAI by creating canopy gaps allowing for Douglas-fir regeneration (Koot et al. 2015), and development of a greater shrub understory providing deer with a preferred source of food (Waterhouse et al. 1991, 1994). However, in the short-term, a more open canopy may be detrimental to the mule deer due to increased energetic expenditure from reduced snow interception (Armleder and Dawson 1992). Additionally, the silvicultural treatment increased the large woody  50 surface fuel load, a common impact of mechanical treatments (Agee and Skinner 2005, Stephens and Moghaddas 2005, Vaillant et al. 2015). The large woody surface fuel was used to help reduce soil compaction during the silvicultural treatment (K. Day, personal communication) and had not yet been removed to meet the required Douglas-fir beetle sanitation objectives. Mule deer winter range management requires the piling and burning of large logs to reduce subsequent susceptibility to the Douglas-fir beetle (Armleder and Thomson 1984, Dawson et al. 2007). The opening of the canopy also has potential repercussions on the curing of the dead woody surface fuels and surface fire intensity. Greater understory cover in conjunction with a larger initial abundance of dead woody surface fuels could lead to more intense surface fires (Vaillant et al. 2015).   Though no data were collected following the silvicultural treatment in 1984, inferences on forest recovery can be made by comparing forest stand attributes of the treatment area in 2014 to the control area in 2015. As prescribed for MDWRM, the silvicultural treatment in 1984 removed 16% of volume (Armleder et al. 1998) in proportion to the abundance of each diameter class in all merchantable size classes (DBH³12.5 cm; Armleder and Thomson 1984). Further, trees were removed to leave a clumpy tree distribution post-treatment (Armleder et al. 1986). Similarly, the silvicultural treatment that occurred in 2014, on average, removed 18% of volume proportionally across all merchantable diameter classes to create a clumpy tree distribution with the additional removal of small diameter trees (Koot et al. 2015). Given the parallels between the two silvicultural treatments, it is reasonable to assume that the 1984 treatment impacts were similar to the impacts measured before and after the silvicultural treatment in 2014.  51  Although the lack of difference between certain forest stand attributes and increased BAI of sub-canopy trees strongly indicate forest recovery following silvicultural treatment, I cannot entirely eliminate the possibility that forest stand attributes differed between the control and treatment areas prior to the treatment in 1984. The forest stand attributes that were affected by the silvicultural treatment but did not differ between the treatment area in 2014 and the control area in 2015 represent forest recovery from 1984 to 2014. The treated forest was likely able to recover over the 30-year period between 1984 and 2014 as demonstrated by the lack of significant difference in tree densities, sub-canopy tree basal area, and canopy cover in the 30-year-old treatment versus the current control. Also, the BAI of the sub-canopy trees in the treated area significantly differed from the BAI of sub-canopy trees in the control area following the implementation of MDWRM further suggesting forest recovery from 1984 to 2014.   Not all forest stand attributes were similar in the treatment area in 2014 relative to the control area in 2015. Leaf area index was lower in the 30-year-old treatment relative to the recent control suggesting an absolute reduction in leaf area due to the clumpy tree removal pattern and gap creation required by the ungulate species. Furthermore, the lower abundance of large surface fuels in the treatment area in 2014 was likely due to the requirement to meet Douglas-fir beetle sanitation measures (Armleder and Thomson 1984, Dawson et al. 2007), as large logs left on site in Douglas-fir forests tend to decay slowly (Sollins 1982, Spies et al. 1988). Finally, the canopy tree basal area, regeneration density, and DBH coefficient of variation remained unchanged over time, meeting the  52 objectives of MDWRM in favouring the retention of big, old Douglas-fir trees, and maintenance of a multilayered forest structure (Dawson et al. 2007). Therefore, removal of no more than 20% of volume across all diameter classes, and cutting at an interval of 30 years or more would permit timber extraction, creation of mule deer winter range (Dawson et al. 2007), and allow enough time for the forest to recover while mimicking forest structures already common to the landscape (Maclauchlan and Brooks 2009).  2.4.2 Implications for biotic and abiotic disturbance  2.4.2.1 Western spruce budworm susceptibility Contrary to my prediction, MDWRM did not affect short or long-term susceptibility to western spruce budworm. Somewhat surprisingly, MDWRM did not alter seven of the nine variables used to determine susceptibility to western spruce budworm. The composition and structure of the stands in Knife Creek did not vary, as measured by total percent crown cover, percent host crown cover, percent late-successional host cover, and the coefficient of variation of DBH. Further, due to the small size of the study area, the coarse-scale variables such as site climate, regional climate, and surrounding host continuity remained constant.  Maturity and vigour were the only variables that were altered due to MDWRM. However, these variables did not vary enough among treatment levels to influence overall susceptibility of individual plots or treatments.   In addition to no effect on susceptibility, MDWRM did not affect the overall availability of food for western spruce budworm. Mule deer primarily feed on old Douglas-fir needles during the winter (Waterhouse et al. 1991, 1994), and stands are therefore  53 managed to be mainly composed of Douglas-fir trees (Armleder et al. 1986). In fact, MDWRM is actually expected to enhance Douglas-fir regeneration (Dawson et al. 2007, Koot et al. 2015). Therefore, the change in canopy cover resulting from MDWRM was not great enough to decrease susceptibility to western spruce budworm because sufficient canopy cover is needed to intercept snow reducing the deer’s energetic requirements (Armleder and Dawson 1992). Further, MDWRM requires the maintenance of a multilayered stand and so the vertical complexity, as measured by the DBH coefficient of variation, did not change to ensure deer have thermal and security cover (Armleder and Dawson 1992). Unfortunately, the regenerating host layer at Knife Creek was not captured by the susceptibility index I used, but the presence of regeneration adds an additional canopy layer, and may allow for greater western spruce budworm population build-up due to more foliage resulting in stands with greater outbreak potential (Magnussen et al. 2004, Maclauchlan et al. 2006).Therefore, conditions created by MDWRM of abundant food, and complex vertical structure may enhance western spruce budworm outbreak potential by increasing survival and successful dispersal events (Fellin and Dewey 1986, Brookes et al. 1987, Swetnam and Lynch 1993).   Although the maturity and vigour variables did vary, the variation was not great enough to influence overall plot or treatment susceptibility to the western spruce budworm. The variation in maturity was likely a result of creating a clumpy tree distribution along with the proportional removal of trees based on their abundance (Armleder et al. 1986, Dawson et al. 2007). The heterogeneous removal of trees likely caused some variation among plots because lone single large trees could be removed along with different sized  54 clumps of variably sized trees (Armleder and Thomson 1984, Armleder et al. 1986, Dawson et al. 2007). Therefore, maturity, calculated as the mean basal-area-weighed age, was dependent on the age and size of the remaining trees. Similarly, variation in vigour, measured as the proportion of basal area present relative to the maximum basal area a plot could support, was likely also due the creation of a clumpy tree distribution coupled with the proportional removal of trees with the additional requirement of meeting a residual basal area target (Dawson et al. 2007, Koot et al. 2015).   To reduce susceptibility to the western spruce budworm, winter range management may need to be more aggressive in terms of volume removal across all diameter classes both greater and less than 12.5cm DBH. Alternatively, the silvicultural treatments applied on winter range may emulate frequencies and stand structures that would result from defoliation events. Proportionally removing trees with 12.5 cm³ DBH based on their abundances results in greater removal of smaller diameter trees, in effect creating stand structures that would be similar to stands affected by budworm outbreaks as most mortality occurs in smaller diameter trees (Alfaro et al. 1982, Fellin and Dewey 1986, Anderson et al. 1987, Alfaro and Maclauchlan 1992, Swetnam and Lynch 1993, Hadley and Veblen 1993, Maclauchlan and Brooks 2009). Additionally, the harvesting interval of 30 years also appears to fall within with the range of frequencies at which western spruce budworm outbreaks occur in dry forests (Campbell et al. 2006, Alfaro et al. 2014, Axelson et al. 2015).     55 Mule deer winter range management may indirectly reduce susceptibility to western spruce budworm through its effects on the growth of trees. Sub-canopy tree basal area increment was greater in the treated area, suggesting greater vigour, thereby potentially reducing susceptibility to western spruce budworm as the insect prefers stressed hosts (Redak and Cates 1984, Brookes et al. 1987), and more vigorous hosts are better able to recover from defoliation (Clancy et al. 2004). However, Knife Creek is located in Douglas-fir dominated, multilayered forest that has historically been particularly prone to western spruce budworm defoliation, and enhancing Douglas-fir regeneration to meet MDWRM objectives may result in greater susceptibility to defoliation (Maclauchlan et al. 2006, Maclauchlan and Brooks 2009), despite larvae landing on more vigorous host trees. Additionally, more vigorous trees produce more foliage, suggesting that the individual tree in question would likely be able to survive multiple defoliation events, potentially sustaining populations and even building them up through time (Magnussen et al. 2004, Maclauchlan et al. 2006, Maclauchlan and Brooks 2009), especially with a harvesting interval of 30 years or more (Dawson et al. 2007).   The classification of the treated and control areas as being moderately susceptible to western spruce budworm is not consistent with the widespread and severe outbreak that is currently occurring in the region (Axelson et al. 2015). The discordance suggests that forest stand attributes are not the only factors contributing to the outbreaks and likely are not limiting outbreak extent and severity. Climatic events may play an important role in influencing susceptibility of the forest to the western spruce budworm (Flower et al. 2014a) and can allow for population synchrony (e.g. Peltonen et al. 2002, Haynes et al.  56 2014) influencing outbreak extent and severity (Axelson et al. 2015). Forest structure and composition therefore provide the framework of predisposing stands to widespread and severe outbreaks. However, something more is required to trigger an outbreak. Conducive climatic conditions could therefore act as the necessary trigger to ‘set-off’ an outbreak by allowing populations to become large enough to persist over time and in suboptimal habitats (Raffa et al. 2008).  While the susceptibility of a forest stand to western spruce budworm as quantified by the model developed by Wulf and Carlson (1985) characterizes composition, structure, and available host within a forest stand, additional variables not included may contribute to susceptibility. Emphasis in the Wulf and Carlson (1985) susceptibility model is placed on the proportion of host and the vertical structure of the stand which are both key to budworm survival and larval dispersal (Fellin and Dewey 1986, Brookes et al. 1987). The model also quantifies tree stress as preferential feeding tends to occur on stressed hosts (Redak and Cates 1984, Brookes et al. 1987), and greater growth rates appear to be associated with budworm resistance (Chen et al. 2001). Additionally, the model includes climatic variables which also influence host stress, and quantifies host composition in the surrounding area. However, my research highlights the potential of including more variables to improve predictions of stand susceptibility. Accounting for climatic events such as droughts would be highly beneficial since budworm outbreaks tend to occur with drier conditions (Campbell et al. 2006, Flower et al. 2014a). Additionally, synchrony of larval emergence and bud burst also affect susceptibility with later bud burst relative to feeding resulting in greater resistance (Chen et al. 2001, 2003). Asynchrony between  57 larval feeding and bud burst may even be detrimental to larval fitness (Chen et al. 2003), therefore, greatly influencing tree and stand-level susceptibility.  2.4.2.2 Douglas-fir beetle susceptibility Douglas-fir beetle susceptibility was lower in the treatment area in the short-term as predicted. In the long-term, the treatment area’s susceptibility was not greater, as predicted, but ‘recovered’ to a level similar to the control area. The percentage of basal area occupied by Douglas-fir was the same in treated and control areas. Therefore, the greater number of plots with a moderate susceptibility to Douglas-fir beetle in the treatment area in 2015, relative to the treatment area in 2014, was due to the reduction in mature host tree (12.5cm ³ DBH) density to or below 292 trees/ha. As a result of forest recovery after the silvicultural treatment in 1984, most plots in the treatment area in 2014 were ranked as high susceptibility. Density reduction may therefore be effective in preventing bark beetle infestations (Fettig et al. 2007, 2014), especially with the removal of large host trees (Temperli et al. 2014).  In addition to affecting Douglas-fir beetle susceptibility, MDWRM also resulted in greater sub-canopy tree growth, suggesting greater vigour, and lower stress, in this stratum.  Overall, a more vigorous stand is expected with canopy trees maintaining steady growth and greater sub-canopy tree growth. Fewer stressed trees would suggest fewer beetle infestations as Douglas-fir beetle preferentially attack compromised hosts (Furniss 1965, Negrón 1998, Aukema et al. 2016). Further, a more vigorous stand would have a higher probability of keeping a resident beetle population in an endemic state given that  58 greater beetle densities are required to attack and overwhelm healthy trees (Raffa et al. 1993, Boone et al. 2011).   Currently, the Cariboo Forest Region is also being affected by a Douglas-fir beetle outbreak (Axelson 2014). The predictions from Negrón’s (1998) susceptibility model appear consistent with the ongoing widespread epidemic. Further, forest stand attributes, particularly the density of mature trees in Douglas-fir dominated stands, is a limiting factor to Douglas-fir beetle susceptibility. However, as mentioned above, a trigger is required to initiate an outbreak by local beetles. A tree stressor such as western spruce budworm defoliation (Hadley and Veblen 1993, Marciniak 2015), root pathogens (e.g. Hoffman 2004), fire (Furniss 1965), drought as observed in spruce beetle (Hart et al. 2014a) or recently downed trees from windthrow or logging activities (Rudinsky 1966, Aukema et al. 2016) would be necessary to build-up a large enough population to outbreak and attack healthy trees (Rudinsky 1962, 1966).   Bark beetle outbreaks can be devastating, affecting timber supply and carbon sequestration (Kurz et al. 2008 a,b), making it important to accurately assess contributing factors to a stand’s susceptibility to outbreaks to reduce ecological and economic impacts (Hicks Jr. et al. 1987). The Negrón (1998) model first accounts for the proportion of host in the stand. Naturally, a low proportion of host would suggest a low probability of infestation. Afterwards, tree density is incorporated into the model and may be a variable that indirectly accounts for reduced growth (Negrón 1998), therefore being an indirect measure of tree stress.  Although the Negrón (1998) model quantifies the available, live  59 standing host, my research highlights the need to include large coarse wood in future Douglas-fir beetle susceptibility models. The estimation of the probability of infestation model that I used did not account for the changes in large coarse wood with viable phloem on the ground due to the mule deer winter range treatment in 2015, which can serve an important role in sustaining and building up the beetle populations (Rudinsky 1962, Christiansen et al. 1987, Aukema et al. 2016). Winter range management does incorporate active Douglas-fir beetle management through piling and burning (Armleder and Thomson 1984) or the use of anti-aggregation pheromones (Dawson et al. 2007); however, the length of time to which logs are potentially exposed to infestation is not specified and could advantage beetle population growth. Removal of downed logs may help control Douglas-fir beetle but may consequently influence nutrient release back into the system (Brown et al. 2003, Marañón-Jiménez and Castro 2013).  2.4.2.3 Crown fire likelihood and wildfire hazard Contrary to my predictions, likelihood of crown fire occurrence was greater in the treatment area in the short-term and lower in the long-term. The observed increase in crown fire likelihood in the treatment area in 2015 relative to the treatment area in 2014 is likely explained by a change in the surface fuel abundance since estimated fine fuel moisture, wind speed, and the fuel strata gap remained constant. However, the likelihood of crown fire occurrence in the treatment and control areas in 2015 did not differ suggesting that the control area already had abundant surface fuels. The piling and burning of logs following the silvicultural treatment in 1984 to manage for Douglas-fir beetle (Armleder and Thomson 1984) likely changed the probability of crown fire  60 occurrence in the treatment in 2014 by altering available surface fuels. Despite the recovery of forest stand attributes, such as infilling of ladder fuels in the long-term, the difference in surface fuels abundance had a greater influence on likelihood of crown fire. However, only logs >20 cm were removed following the silvicultural treatment in 1984 (Armleder and Thomson 1984, Armleder et al. 1986), whilst smaller logs were left on site. The leftover logs may still contribute to the fuel hazard as the decay state of these logs can affect torching and spot fires (Brown et al. 1991, Knapp et al. 2005).   As outlined above, the removal of ladder fuels may reduce crown fire likelihood (Agee and Skinner 2005, Fiedler et al. 2010, Vaillant et al. 2015), but the resultant increase in surface fuels (Agee and Skinner 2005, Vaillant et al. 2015), may ultimately exacerbate fire intensity (Stephens 1998). As fire intensity is related to flame length (Byram 1959, Rothermel and Deeming 1980), potential for crowning may be enhanced without further fuel mitigation efforts (Van Wagner 1977), such as increasing the fuel strata gap or removing surface fuels. Thus, the timely removal of large surface fuels is essential if MDWRM is to incorporate fuel management, otherwise in the short-term, stands that undergo silvicultural treatment would have a similar likelihood of crown fire occurrence to unmanaged stands.  The change in fuel components that influenced a flame’s ability to travel up into a tree crown due to MDWRM were assessed but fire spread in a dry forest fuel complex was not assessed. Although, fire suppression in dry forests has changed the vegetation structure resulting in altered fire behaviour (Hessburg et al. 2000, Hessburg et al. 2005,  61 Beaty and Taylor 2007, Falk et al. 2007, Stephens et al. 2008) that may be similar to an even-aged closed canopy stand, my research highlights the need to assess fire behaviour in forests composed of variably sized clumps of trees of varying ages that are not necessarily uniformly distributed (Arno 1980, Allen et al. 2002, Schoennagel et al. 2004, Binkley et al. 2007, Perry et al. 2011). The likelihood of crown fire occurrence model, from Crown Fire Initiation and Spread (CFIS), focuses on the probability that a surface fire will transition to a crown fire and was developed using even-aged, closed canopy forest types (Cruz et al. 2004). Although such a model provides useful insights, even-aged closed canopy forests may result in different fire behaviour than in more open Douglas-fir forest types (e.g. Brown and Bevins 1986, Cruz et al. 2004) due to different canopy fuel characteristics (e.g. Cruz et al. 2003).   The silvicultural treatment used for MDWRM coupled with additional efforts to increase the sub-canopy tree crown base height could simultaneously benefit the ungulate species and be used as a fuel treatment. The maintenance of a high crown base height helps ensure a larger fuel strata gap reducing wildfire hazard. Fuel strata gaps less than 2.0 m result in high incidence of crown fire, while gaps ³7 m have a much lower incidence of crown fire (Cruz et al. 2004).  Further, since removal of large dead woody surface fuels is already incorporated into MDWRM as part of the Douglas-fir beetle sanitation objective (Dawson et al. 2007) it would be a matter of ensuring prompt removal of these fuels. The sanitation objective would therefore reduce susceptibility to the Douglas-fir beetle and wildfire. Effective wildfire management would incorporate management of both crown base height and surface fuels because altering only one aspect may not be operationally  62 feasible. For example, to reduce crown fire likelihood to ~30% the crown base height of all sub-canopy trees would need to be ~8m. Alternatively, even if all large and 10-hour fuels were removed from a site there would still be a likelihood of crown fire of ~45%.  2.4.3 Interactions  To examine the independent and interacting effects of MDWRM on disturbance processes over the short and long-term, I developed qualitative models illustrating potential trade-offs and synergies. The initial model consists of a historic baseline in the absence of MDWRM depicting the effect of one disturbance on another through structural alterations (Figure 2.4). Generally, high-severity fires are expected to negatively affect future insect infestations by removing host trees, while low-severity fires created forest structures prone to infestations (McCullough et al. 1998, Schoennagel et al. 2004, Hood and Bentz 2007). Infestation by the western spruce budworm or Douglas-fir beetle would result in forest structures prone to low-severity fires (Lynch and Moorcroft 2008, Donato et al. 2013). Each individual disturbance is expected to create stand structures that are later prone to that particular disturbance (Arno 1980, Brookes et al. 1987, Covington and Moore 1994, Allen et al. 2002, Schoennagel et al. 2004), except for the Douglas-fir beetle, which kills available and susceptible host, removing the large, old trees it would preferentially feed on (Humphreys 2000, Aukema et al. 2016).       63  Figure 2.4: The hypothetical interactions among the primary biotic and abiotic disturbance agents of interior dry forest ecosystems of British Columbia prior to the application of mule deer winter range management (MDWRM). A ‘+’ symbol indicates an enhancing effect, whereby a disturbance predisposes the stand to the subsequent disturbance. A ‘-’ symbol indicates a suppressing effect, where disturbances exclude one another.  The implementation of MDWRM is expected to result in an unbalanced network of disturbance interactions, although the removal or retention of large surface fuels and the time elapsed since treatment may affect the suite of interactions in different ways (Figure 2.5). Separate scenarios for the removal or retention of large surface fuels were created because, as mentioned above, Douglas-fir beetle sanitation objectives under MDWRM dictate that large downed wood may be treated and left behind or removed (Armleder and Thomson 1984, Dawson et al. 2007). Different downed wood management options result  64 in different disturbance interactions in the short and long-term. The removal of the large surface fuels in the short-term (Figure 2.5) is expected to lower the potential for crown fires, suggesting ignitions would result in surface fires possibly reducing the stand’s susceptibility to the western spruce budworm by burning the saplings and regenerating trees (Schoennagel et al. 2004). Further, susceptibility to the western spruce budworm may decline because surface fires create host for the Douglas-fir beetle, which kills large, old trees reducing the vertical complexity of the stand (Furniss 1965, Fellin and Dewey 1986, Humphreys 2000, Hood and Bentz 2007). However, the removal of large surface fuel lowers the amount of available material that may be infested by the Douglas-fir beetle, thereby offsetting a potential increase in susceptibility (Humphreys 2000, Aukema et al. 2016). The removal of large surface fuels shortly after MDWRM appears to create stands predisposed to surface fires with variable consequences for subsequent Douglas-fir beetle attack while lowering susceptibility to the western spruce budworm and crown fires.  65   Figure 2.5: The hypothetical interactions between the primary disturbance agents of interior dry forests of British Columbia with the application of mule deer winter range management (MDWRM) and associated Douglas-fir beetle management tactics in the short- and long-term. A ‘+’ symbol indicates an enhancing, whereby a disturbance predisposes the stand to the subsequent disturbance. A ‘-’ symbol indicates a suppressing effect, where disturbance exclude one another. A ‘0’ indicates that there is no effect, the interaction is neutral. Red arrows and ‘+’ or ‘-’ symbols indicate direct effects of MDWRM on the disturbance, while red ‘+’ or ‘-’ symbols with black arrows indicate cascading effects resulting from MDWRM application. 66  Shortly after the implementation of MDWRM, leaving the large surface fuel creates an interaction cascade favouring the occurrence of crown fires, and infestation by the Douglas-fir beetle (Figure 2.5). Leaving large surface fuels on the ground results in greater susceptibility to crown fires. Further, without the immediate application of anti-aggregation pheromones, the large surface fuels left behind by MDWRM would potentially increase susceptibility to the Douglas-fir beetle (Rudinsky 1962, Humphreys 2000, Aukema et al. 2016). Under this scenario where crown fires are favoured, the result may be an exclusion of the other main disturbance agents. However, in the absence of ignition and any efforts to mitigate Douglas-fir beetle impacts, an infestation of large-diameter trees may result in lowered susceptibility to crown fires, and defoliation by the western spruce budworm.  In the long-term, MDWRM with the removal of large surface fuels, has the potential to lower susceptibility to crown fires (Figure 2.5). In the absence of crown fires, ignitions would likely result in surface fires that would create a complex vertical and horizontal stand structure (Allen et al. 2002, Hessburg et al. 2005, Binkley et al. 2007) that is susceptible to the western spruce budworm (Fellin and Dewey 1986, Brookes et al. 1987, Maclauchlan and Brooks 2009). Additionally, western spruce budworm activity may also predispose trees to attack by the Douglas-fir beetle (Marciniak 2015). Further, fire-injured trees would be suitable hosts for the Douglas-fir beetle (Furniss 1965, Hood and Bentz 2007). The combined effects of surface fires, defoliation by the western spruce budworm, and infestations by the Douglas-fir beetle will likely result in stands with a more open canopy, inhibiting potential crown fire spread (Jenkins et al. 2008, Donato et  67 al. 2013, Cohn et al. 2014). However, leaving large surface fuel following MDWRM in the long-term favours a system dominated by crown fires with lower susceptibility to surface fires (Figure 2.5). With more crown fires expected in the system, the susceptibility to both the western spruce budworm and Douglas-fir beetle will be lowered due to fewer available host and altered stand structure (McCullough et al. 1998).   Application of MDWRM appears to culminate in an unbalanced disturbance interaction network resulting in shifting susceptibility to disturbance that depends on time since treatment and large surface fuel removal. Fire will be omnipresent under MDWRM; however, the severity will depend on whether large surface fuels are left on site thereby altering the susceptibility of the stand to other disturbance agents through a cascade of interactions. Low-severity fires are expected to favour greater insect activity while high-severity fires may exclude insects. Further, the frequency of crown and surface fires may be altered depending on time since treatment and whether large surface fuels are removed. Moreover, given the potential for prolonged fire seasons creating conditions more conducive to lightning ignitions that could result in higher severity fires (Easterling et al. 2000, Westerling et al. 2006, Lutz et al. 2009, Flannigan et al. 2013, Woolford et al. 2014), forest managers may focus on the removal of large surface fuels in order to mitigate future high-severity fires. However, such an application of MDWRM may later compromise susceptibility to insects, especially with the potential for ‘hotter’ droughts (Allen et al. 2015, Millar and Stephenson 2015). Therefore, managers need to be aware that the implementation of MDWRM may lower susceptibility to some disturbance agents at certain points in time while simultaneously increasing susceptibility to others.   68 2.4.4 Conclusions  Mule deer winter range management does alter forest stand attributes in the short-term by reducing tree density, sub-canopy tree basal area, canopy cover, LAI and increasing the number of large surface fuels on the ground. Over the long-term, MDWRM allows for the recovery of original canopy cover, tree density, and sub-canopy tree basal area, and likely lowers the number of larger surface fuels on the ground. Additionally, BAI of sub-canopy trees increases following the implementation of MDWRM.   The changes in forest stand attributes had differential effects on biotic and abiotic disturbance over time. Susceptibility to the western spruce budworm remained unchanged in the short and long-term through continued maintenance of abundant host trees and multilayered stand structure. Susceptibility to the Douglas-fir beetle was lower in the short-term due to fewer available mature host trees; however, over the long-term forest recovery facilitated beetle susceptibility increases to reach levels seen in the control. Likelihood of crown fire in the short-term was greater with more large surface fuels following treatment, but lower in the long-term with removal of the large surface fuels to meet Douglas-fir beetle sanitation requirements. The silvicultural treatment applied for MDWRM coupled with a focus on increasing the crown base height of sub-canopy trees could simultaneously benefit mule deer and be utilized as a fuel mitigation treatment. Finally, implementation of MDWRM is expected to affect the network of disturbance interactions in the short and long-term. Cascading effects on disturbance interactions due to the application of MDWRM will vary depending on the decision to remove large surface fuels or not, and time since treatment.   69 3 The potential effects of widespread mule deer winter range management application on wildfire risk in dry forests of interior British Columbia  3.1 Introduction  Ecological processes are influenced by a landscape’s composition and configuration. Landscape composition is an indicator of patch diversity and abundance, while landscape configuration refers to the location of the patches relative to each other (Turner 1989, Dunning et al. 1992). Both of these landscape characteristics can directly determine the extent and severity of disturbances such as wildfires or insect outbreaks (Turner et al. 1989, O’Neill et al. 1992, Raffa et al. 2008, Hansen et al. 2016, Seidl et al. 2016b). Disturbance extent is dependent on the proportion of susceptible area on the landscape. Theoretically when landscapes are composed of >60% susceptible area, disturbances spread easily whereas disturbance spread is constrained when susceptible area is less than 20% (Turner et al. 1989).   At broad spatial scales, landscapes that are configured with high degree of connectivity between susceptible patches can facilitate the spread of disturbance, increasing the extent of the affected area (Turner 1989, Turner et al. 1989, O’Neill et al. 1992, Seidl et al. 2016c). For example, widespread fire suppression in dry forests has increased tree densities over large areas and has led to more severe fires (Hessburg et al. 2000, Schoennagel et al. 2004, Hessburg et al. 2005, Falk et al. 2007), and fires have been further exacerbated by increased connectivity among areas with high fuel loads leading to greater potential for ignition and subsequent spread (Turner and Romme 1994, Beverly et al. 2009). Bark beetle outbreaks also tend to be more widespread and severe in better  70 connected landscapes composed of large areas of susceptible host trees (Taylor and Carroll 2004, Raffa et al. 2008, Hansen et al. 2016). Therefore, management activities that alter composition and configuration of forest landscapes may negatively affect the ability of forest ecosystems to recover and maintain processes and structures following disturbance (Peterson et al. 1998, Gunderson 2000).  In addition to affecting disturbances, landscape composition and configuration influences species composition and biotic communities through direct and indirect impacts to habitat and resource availability (Dunning et al. 1992, Dormann et al. 2007, Prugh et al. 2008). Directly, a landscape with a greater diversity of patches results in greater habitat and resource availability (Dunning et al. 1992) suggesting greater biodiversity (Tscharntke et al. 2012). Indirectly, landscape composition and configuration can influence community composition by affecting successful colonization events by dispersing individuals (Gustafson and Gardner 1996, King and With 2002). Further, composition and configuration of the landscape is particularly important for weak dispersers as higher abundance or an aggregation of suitable patches is necessary for successful colonization (King and With 2002).   Landscape composition and configuration may be altered due to single-objective management activities applied over broad scales. Single-objective management focuses on achieving a single goal perhaps with unforeseen effects. In pursuit of achieving the objective, managers may apply tactics that affect the patch diversity and connectivity of a landscape, altering or exacerbating disturbances (Hessburg et al. 2005, Barclay et al.  71 2005, Safranyik et al. 2010, Ciesla 2015). For example, managers may wish to increase timber yield by planting a single species over a large area without necessarily considering that in doing so, landscape connectivity increases and landscape patch diversity decreases resulting in a more disturbance prone landscape (Hessburg et al. 2000, Taylor and Carroll 2004). Further, under single-objective management reducing patch diversity may affect available habitat for endangered species. For example, past forest management has reduced habitat for caribou [Ranger tarandus caribou Gmelin] (Cumming 1992). Simultaneously, fire suppression has resulted in a more homogeneous landscape with fewer food resource patches for caribou (Euler et al. 1976). Similarly, fire suppression, and subsequent infilling, has also negatively affected gopher tortoise [Gopherus Polyphemus Daudin] habitat by lowering the abundance of and connectivity between the open, dry forest patches used for burrows and foraging, adversely affecting populations (McRae et al. 1981, Eubanks et al. 2003, Jones and Dorr 2004). In essence, a drawback of single-objective management is its oversimplified approach, focusing on the maintenance of a single part as opposed to viewing the entire system and working towards maintaining the system (Benson 2012).  Mule deer are an important game species and have become an important focus of management (Blood 2000, Dawson et al. 2007). Mule deer preferentially choose to overwinter in old forests (>140 years) with moderate to high canopy cover (Armleder et al. 1994). Its winter survival and subsequent reproduction is dependent on older, multilayered, uneven-aged forests with large Douglas-fir trees (Armleder and Dawson 1992) for two reasons. First, such forest structure reduces their energetic requirements by  72 providing thermal and security cover (Armleder and Dawson 1992) and snow interception during winter (Armleder et al. 1994). Second, large old Douglas-fir trees provide winter forage in the form of needles and arboreal lichen (Waterhouse et al. 1991, 1994).   In British Columbia, a silvicultural treatment was developed to achieve the required species composition and forest structure needed by mule deer. The treatment involves harvesting 15-20% of volume as single and/or small groups of trees (12.5 cm ³DBH) (Dawson et al. 2007). Additionally, larger trees are typically retained while a greater number of smaller diameter trees are removed; essentially proportionally removing different sized trees based on their abundance (Armleder et al. 1986, Dawson et al. 2007). This silvicultural treatment is used to create a multilayered, uneven-aged stand with large, old Douglas-fir trees to enhance the habitat and food requirements of mule deer. To ensure that the habitat and food requirements are maintained, a re-entry period of 30 or more years has been incorporated into the treatment schedule (Armleder et al. 1986, Dawson et al. 2007).  In the interior dry forests of British Columbia (BC) mule deer winter range management (MDWRM) may lead to landscape homogenization by altering landscape composition and configuration. Mule deer winter range management creates and enhances habitat conditions for deer, and provides timber harvesting opportunities in areas that are otherwise restricted, sparking considerable interest in widespread application by companies. Further, in chapter 2, I demonstrated that modification of individual stands to  73 improve their suitability for mule deer winter range also reduced the likelihood of crown fire occurrence. Therefore, I initiated a study using geographical information systems (GIS) and FRAGSTATS to determine the effects of widespread application of MDWRM on landscape composition and configuration by comparing the current and forecasted fire risk of dry forest dominated landscapes in British Columbia. I evaluated the hypothesis that widespread application of mule deer management will affect the fire risk in dry forest landscapes. I predicted that widespread application would result in a more homogenized landscape with lower fire risk.  3.2 Materials and methods  3.2.1 Study area  The Cariboo region of British Columbia is located in the central interior plateau of the province, east of the Coastal mountains and adjacent to the Fraser Basin and Thompson Plateau (Figure 3.1; Pojar and Meidinger 1991). Much of the Cariboo region is classified as part of the Interior Douglas-fir (IDF) biogeoclimatic zone (Hope et al. 1991). Generally dry, with a mean annual precipitation of 434 mm, the Cariboo region receives 19 and 34% of its precipitation in the spring and summer respectively (Wang et al. 2016). Mean temperatures vary from 16.7oC to -6.6oC during the warmest and coldest months respectively (Wang et al. 2016).   A large portion of the Cariboo region is composed of dry forests dominated by Douglas-fir. Such forests typically exhibit low tree densities occurring in small clumps of varying ages (Arno 1980, Hessburg et al. 2005, Binkley et al. 2007, Kaufmann et al. 2007,  74 Maclauchlan and Brooks 2009). Low elevation dry forests generally have open canopies dominated by Douglas-fir, while moister and higher elevation forests are more diverse including other species such as lodgepole pine, trembling aspen, white spruce, and paper birch (Hope et al. 1991). Fire plays an important role determining species composition. Frequent surface fires often create Douglas-fir dominated stands, while more severe fires can promote the establishment of lodgepole pine within Douglas-fir forests (Hope et al. 1991). The understory generally has few shrubs and herbs, very little moss, and Douglas-fir regeneration is found in the understory at varying densities (Hope et al. 1991).                                 Figure 3.1: The extent of the mule deer winter range management zones (A) located in central British Columbia (B), Canada (C)    75  3.2.2 Data and processing  To determine the potential effects of widespread MDWRM, I assembled vector polygon data comprising mule deer winter range habitat management zones, vegetation resource inventory (VRI) and fire risk (BC Ministry of Forests Lands and Natural Resource Operations 2012, 2015a, 2016) as described below.  3.2.2.1 Habitat management zones Habitat management zones delineate the MDWRM polygons in British Columbia (BC Ministry of Forests Lands and Natural Resource Operations 2015a). The management zones are determined using the criteria outlined by the General Wildlife Measures for mule deer winter ranges (BC Ministry of Environment 2007). These criteria include an individual site’s ability to ensure enhanced Douglas-fir regeneration, an eventual species composition of 80% Douglas-fir or more, residual basal area of 29 m2/ha for trees DBH³12.5 cm with 55% of this basal area occupied by large trees (DBH >37.5 cm) to maintain high canopy cover, as well as the ability to implement harvesting to achieve the desired multilayered, clumpy structure required by mule deer (Armleder et al. 1986, Dawson et al. 2007). In addition, potential habitat management zones must be situated away from commercial forestry activities to help maintain security cover and reduce predation and/or harassment, aiding in mule deer survival (Armleder et al. 1986, Dawson et al. 2007). Harvesting activities in habitat management zones are further restricted so as not to affect established mule deer trails including ridges as these topographic features are preferentially used by mule deer during the winter because of reduced snow depth from sun and wind (Armleder et al. 1986).  76  3.2.2.2 Vegetation resource inventory Vegetation resource inventory is derived from aerial photography combined with sample plots to determine the type of vegetation cover and corresponding forest attributes for individual forest stands (Sandvoss et al. 2005), where, a stand is defined as a group of trees that share certain characteristics that distinguish this group from the surrounding area including composition, age, and arrangement (Haddon 1988). The VRI data consist of multiple forest attributes such as species composition (with up to six species identified in each stand), canopy closure, and biogeoclimatic zone (BC Ministry of Forests Lands and Natural Resource Operations 2016). The data are periodically updated to reflect changes due to forest harvesting and natural disturbances. I used the VRI data to characterize the habitat management zones and specifically selected for polygons that were part of the Interior Douglas-fir (IDF) zone, since mule deer management also occurs in different biogeoclimatic zones (Dawson et al. 2007). Polygons in the IDF were generally dominated (>50%) by Douglas-fir, however those with a second dominant species were still included because under such conditions, harvesting of non-Douglas-fir trees is done to create and/or maintain conditions for Douglas-fir regeneration (Dawson et al. 2007).   3.2.2.3 Fire risk Fire risk vector polygons were developed by the BC Ministry of Forests Lands and Natural Resource Operations (2015b) to identify areas of high risk of wildfires and improve landscape planning in the province. The fire risk vector polygons combined  77 information on ignition probability, the fuel complex, topography and weather to account for fire occurrence, intensity and spotting. From these attributes, they derive three variables: fire density, head fire intensity under extreme conditions, and fire spotting impact (BC Ministry of Forests Lands and Natural Resource Operations 2015b). Fire density was determined using fire start densities and fire perimeters dating back to 1950 (BC Ministry of Forests Lands and Natural Resource Operations 2015b). Head fire intensity was determined using the 90th percentile fire weather index value to characterize fire intensity under extreme conditions. Finally, spotting impact, the ability of wind-carried embers to ignite a fire outside the main perimeter (Scott and Reinhardt 2007), was assessed up to 2 km from a point, with a reduction in impact with increasing distance (BC Ministry of Forests Lands and Natural Resource Operations 2015b). The three variables are combined using a weighted average, attributing greatest importance to head fire intensity, then fire density and finally spotting, to determine fire risk (BC Ministry of Forests Lands and Natural Resource Operations 2015b). Fire risk is divided into 4 classes: low, moderate, high and severe (BC Ministry of Forests Lands and Natural Resource Operations 2012).  3.2.3 Landscape management scenarios  Given that the primary impact of MDWRM on stands is altered crown fire occurrence (see Chapter 2), I quantified the spatial arrangement and abundance of fire risk classes to determine the effects of widespread application of MDWRM over landscapes. The spatial arrangement and abundance of fire risk classes were considered indicators of landscape structure and composition. To quantify the landscape effects of MDWRM, I imported the  78 vector polygon data described above into ArcMapTM 10.3.1 (ESRI 2015) and then I derived two scenarios.  3.2.3.1 Landscape scenario 1: Current landscape without MDWRM Scenario 1 described the fire risk of the landscape without widespread application of MDWRM in eligible stands, while scenario 2 described the long-term projected fire risk of a landscape in which all eligible stands underwent widespread application of MDWRM. In scenario 1, I used the mule deer habitat management zones layer to delineate and restrict the area of interest to the Douglas-fir dominated forests with suitable forest structures. I combined the VRI layer with the management zones later to select forests that were suitable habitat in the Interior Douglas-fir (IDF) zone, which were combined with fire risk. The combined IDF habitat management zones and fire risk layers served to represent the forest’s current vegetative and fire risk state representing the landscape without widespread application of MDWRM (scenario 1).   3.2.3.2 Landscape scenario 2: Forecasted landscape with widespread MDWRM To reflect long-term changes in fire risk associated with MDWRM (scenario 2), I derived an additional layer from scenario 1, under the following assumptions. Fire risk is the chance of a fire occurring based on fuel hazard, ignitions and weather (Kulakowski and Jarvis 2011). First, I assumed the decrease in fuel hazard experienced at Knife Creek due to MDWRM (see Chapter 2) as measured by a decrease in likelihood of crown fire occurrence, could be applied to all dry Douglas-fir forests within the mule deer management landscape. I then assumed that the other contributors to fire risk, ignitions  79 and weather, remained constant in order to isolate the changes in fuel hazard. A reduced fuel hazard resulted in a decrease in crown fire likelihood (Chapter 2) and was therefore assumed to result in a decrease in fire risk since other aspects remained constant. The new scenario 2 layer lowered the level of fire risk of all areas by one class to reflect a decrease in fuel hazard (i.e. high fire risk was changed to moderate fire risk), however sites classified as low fire risk remained as low fire risk. Finally, I assumed that the forests in the MDWRM landscape had not been affected by stand-replacing disturbances, including harvesting, since the VRI data was last collected.  3.2.4 Data analysis  I assessed landscape changes resulting from widespread MDWRM application by comparing the current forest’s fire risk (scenario 1) to the forecasted forest’s fire risk (scenario 2) using FRAGSTATS v.4.2 (McGarigal and Ene 2013). I first rasterized the layers with a cell size of 720 x 720 m which were then converted to ASCII files to be read in FRAGSTATS. The cell dimensions I selected best fit the polygon shapes without adding any unnecessary raster cells. From the rasterized data, changes in the abundance and distribution of the fire risk classes on the current and forecasted landscapes were assessed.  3.2.4.1 Abundance and spatial distribution of fire risk classes I assessed the area, aggregation, and connectivity of each fire risk class to determine changes in spatial arrangement resulting from MDWRM. I calculated the total area and percent of the landscape occupied, in addition to ‘clumpiness’ and patch cohesion indices  80 of each fire risk class. The clumpiness index quantifies the degree to which focal class patches are aggregated, solely focusing on configuration without including changes in focal class area (McGarigal 2014). Clumpiness values range from -1 to 1; where, 1 indicates an aggregated distribution, -1 indicates a disaggregated focal class and a value around 0 suggests a random distribution (McGarigal 2014). The patch cohesion index ranges from 0-100, and indicates the area occupied by and the degree to which a focal class is physically connected, with higher values suggesting greater coverage and better connectivity (McGarigal 2014).  3.2.4.2 Landscape heterogeneity I assessed landscape heterogeneity of the current and forecasted forests by quantifying the distribution and abundance of the fire risk classes using Shannon’s evenness and contagion indices (McGarigal 2014). Shannon’s evenness index, ranging from 0 to 1, indicates whether areas covered by different focal classes are similar such that values closer to 1 suggest an even distribution among focal classes (McGarigal 2014). Contagion describes the landscape’s tendency to aggregate (values ranging from 0 to 100) by looking at the spatial distribution of each class and occurrence of particular patch classes being next to one another (McGarigal 2014). Values closer to 100 indicate an aggregated landscape while values near 0 a disaggregated landscape (McGarigal 2014).      81 3.3 Results  3.3.1 Abundance and spatial distribution of fire risk classes  In the current landscape, the severe fire risk was the most aggregated and the third most abundant class on the landscape (Table 3.1). In contrast, moderate fire risk occupied the most area in the current landscape followed by low fire risk and were also the best-connected classes (scenario 1; Table 3.1). Widespread application of MDWRM tactics (scenario 2) increased the amount of the landscape considered to be at low and high fire risk by 40% and 6% respectively, while reducing the amount of the landscape considered to be at moderate fire risk by 1% (Table 3.1; Figure 3.2). Further, widespread MDWRM reduced the patch cohesion of the moderate fire risk class by ~5%, but increased high fire risk patch cohesion by ~ 5% with low fire risk remaining unchanged (Table 3.1; Figure 3.2). Finally, aggregation of low and moderate fire risk classes were reduced by 15% and 2% respectively, under scenario 2 (Table 3.1). However, under scenario 2, aggregation of high fire risk classes increased by 8% (Table 3.1).              82                               Figure 3.2: The estimated area covered by different fire risk classes within the mule deer winter range management landscape of interior British Columbia. (A) Estimated area covered by the different fire risk classes in the current state of the forest without MDWRM application (scenario 1). (B) Estimated area covered by different fire risk classes under the hypothetical application of MDWRM to all eligible stands (scenario 2). 83 Table 3.1: Comparison of the spatial metrics used to describe the abundance and spatial arrangement of the 4 fire risk classes in the dry interior forests of British Columbia designated for mule deer winter range management (MDWRM) between the current state of the forest without widespread MDWRM (scenario 1) and a hypothetical application of MDWRM to all eligible forest stands (scenario 2).  Scenario Fire risk Total area (ha) Percent of landscape (%) Patch cohesion index (%) Clumpiness index  Current state of forest without widespread MDWRM Low 136,339.2 19 93.68 0.72 Moderate 147,018.2 21 93.74 0.74 High 96,370.6 14 88.58 0.72 Severe 103,628.2 15 93.35 0.80             Hypothetical application of MDWRM to eligible forest stands Low 283,357.4 59 94.89 0.57 Moderate 96,370.6 20 88.58 0.72 High 103,628.2 21 93.35 0.80  84  3.3.2 Landscape heterogeneity  Simulated application of widespread MDWRM reduced Shannon’s evenness index by 11% and increased the contagion index by 12% (Table 3.2). Together, both metrics indicated that the forecasted forest has a skewed class distribution and is more aggregated relative to the current forest (Table 3.2; Figure 3.2).   Table 3.2: Comparison of landscape heterogeneity (distribution and abundance of fire risk classes) in the dry interior forests of British Columbia designated eligible for mule deer winter range management (MDWRM) between the current state of the forest without widespread MDWRM (scenario 1) and a hypothetical application of MDWRM to all eligible forest stands (scenario 2).  Scenario Shannon’s evenness index Contagion index Current state of forest without widespread MDWRM   0.99  39.9 Hypothetical application of MDWRM to eligible forest stands  0.88 44.7  3.4 Discussion  As expected, MDWRM lowered the long-term likelihood of crown fire and led to a more homogeneous landscape when applied over all eligible areas within the interior Douglas-fir forests of BC by creating a large contiguous area with low fire risk. Widespread homogeneous areas of low fire risk may desirable with regards to wildfire management (Seidl et al. 2016a). When low susceptibility patches occupy a large proportion of the landscape, this likely reduces fire spread and extent, as well as susceptibility to infrequent large and severe fire events (Turner 1989, Turner et al. 1989, O’Neill et al. 1992).    85 Changes in habitat connectivity and aggregation were associated with the increased homogeneity arising from the simulated widespread MDWRM that could likely alter fire dynamics. The tendency for increased cohesion and clumpiness among high fire risk habitats suggests that if an ignition were to occur there, spread could be enhanced (O’Neill et al. 1992, Ryu et al. 2007, O’Donnell et al. 2011). However, due to the widespread application of MDWRM, high-risk patches were contained within an otherwise low fire risk landscape, suggesting limited fire spread even in the event of successful ignition (Gonzalez et al. 2008).   In contrast, over broad spatial scales, homogeneous landscapes can be more susceptible to other types of disturbance (Hessburg et al. 2000, Taylor and Carroll 2004, Raffa et al. 2008). For example, increased landscape homogeneity has been associated with increased host-tree availability causing exacerbated outbreaks by several primary bark beetle species (Taylor and Carroll 2004, Hansen et al. 2016, Seidl et al. 2016b). Further, a more homogeneous forest landscape may be less resilient, and thereby more prone to widespread disturbance exacerbated by climate extremes (Easterling et al. 2000, Williams and Liebhold 2002, Peltonen et al. 2002, Haynes et al. 2013, Allstadt et al. 2015), such as drought-induced insect outbreaks (Raffa et al. 2008, Bentz et al. 2010, Hart et al. 2014a, Flower et al. 2014a), and elevated fire ignition frequency (Lutz et al. 2009, Woolford et al. 2014). Interestingly, in my research a reduction in crown fire likelihood was not detected until 30 years post treatment (Chapter 2). Therefore, any negative effects on landscape uniformity may be attenuated with the gradual imposition of widespread MDWRM.  86  Increased landscape homogeneity as a consequence of widespread MDWRM has implications for potential spread and extent of biotic disturbances specific to dry interior Douglas-fir forests. In these forests, the western spruce budworm and Douglas-fir beetle are major biotic disturbance agents (Erickson 1992, Axelson et al. 2015). Although MDWRM-related changes to stand structure are not expected to alter the long-term likelihood of infestation by either insect at that scale (Chapter 2), widespread application of MDWRM to all eligible stands in the absence of fire will favour a landscape dominated by old Douglas-fir forests with a complex vertical structure and clumpy tree distribution (Armleder et al. 1986, Dawson et al. 2007) which may facilitate outbreaks by both the western spruce budworm and Douglas-fir beetle (Fellin and Dewey 1986, Brookes et al. 1987, Swetnam and Lynch 1993, Powers et al. 1999, Negrón et al. 2001, Dodds et al. 2006). Consequently, a homogeneous fire resilient landscape may in fact facilitate widespread insect outbreaks through the creation of extensive, contiguous susceptible hosts (Turner 1989, Turner et al. 1989, Aukema et al. 2006, Raffa et al. 2008, Hansen et al. 2016) with consequences for the disturbance interactions in the ecosystem. Exclusion of high-severity fires and promotion, to some extent, of low-severity fires may ultimately create conditions prone to infestation (see Figure 1.1). In summary, MDWRM at the landscape scale may be an effective means of limiting future wildfire impacts yet may create conditions conducive to other disturbance agents.  Reduction of landscape heterogeneity seen with widespread MDWRM has implications to habitat diversity which extends beyond impacts to natural disturbances. Removal of  87 logs, and large diameter branches to lower the risk of Douglas-fir beetle infestation (Armleder and Thomson 1984, Dawson et al. 2007) and fire (Chapter 2), is an important aspect of management for mule deer (Armleder et al. 1986). However, widespread removal of coarse woody debris may alter habitat structure (Harmon et al. 1986, Brown et al. 2003) with negative effects across trophic levels (Horn and Hanula 2008) leading to reduced species diversity (Tscharntke et al. 2012). For example, removal of coarse woody debris in the southeastern United States resulted in lower insect diversity reducing available prey for multiple bark-foraging and gleaning bird species (Horn and Hanula 2008). Further, presence of coarse woody debris in south-central British Columbia affected abundance of voles and their predators (Sullivan and Sullivan 2012). Additionally, absence of coarse woody debris may negatively affect insects, amphibians, reptiles, and mammals by reducing available feeding and reproductive sites (Harmon et al. 1986, Brown et al. 2003). Furthermore, landscape configuration becomes critical in determining successful dispersal when suitable habitat is limited, especially for weak dispersers (King and With 2002). Therefore, even though some species may benefit from widespread application of MDWRM (Dawson et al. 2007), it may create trade-offs. Despite creating large areas of suitable habitat for mule deer (and several other species) the resulting unsuitability for other species may lead to an overall reduction in landscape-level biodiversity.   Resource management efforts that favour a single objective often do so to the detriment of other objectives (Bennett et al. 2009, Turner et al. 2013). For example, increasing human habitation of the wildland-urban interface has resulted in widespread fire  88 suppression to protect homes in many forested systems. This in turn has caused a build-up of fuels, creating a homogeneous fuel complex with potential to support larger and more severe fires (Fulé et al. 1997, Hessburg et al. 2000, Schoennagel et al. 2004, Hessburg et al. 2005). Similarly, extensive forest fire suppression aimed at preserving long-term timber supply resulted in a dramatic increase in susceptible hosts to the mountain pine beetle (Taylor and Carroll 2004), and resulted in an unprecedented outbreak (Safranyik et al. 2010). Although widespread application of MDWRM may result in a more fire resilient landscape it may be detrimental to other ecosystem services (Foley et al. 2005), and even create large areas prone to other disturbances (Hessburg et al. 2000). The cases mentioned above highlight a possible trade-off, where, meeting a single objective may potentially create a less resilient landscape. It is therefore important to quantify the interactions between multiple services and objectives to provide a better understanding of management effects at multiple spatial scales (Kremen 2005).    89 4 General conclusions  My research highlights the need to understand the unintended yet detrimental effects of single-objective management on disturbances and how the resulting changes in forest structure and composition may differentially affect a variety of other ecosystem services and ecosystem processes.   4.1 Key findings  I demonstrated in chapter 2 that treatments associated with MDWRM altered forest stand attributes in the short and long-term with differential effects on disturbance likelihood. Mule deer winter range management maintained abundant Douglas-fir canopy cover, and a multilayered structure (Dawson et al. 2007, Koot et al. 2015), leaving the susceptibility to the western spruce budworm unchanged over the short- and long-term. Short-term reduction in density of mature host-trees lowered susceptibility to the Douglas-fir beetle, but susceptibility in the long-term was similar to untreated stands. A short-term increase in the likelihood of crown fire occurrence was likely due to the greater amount of large surface fuels remaining whereas the long-term decrease was likely linked to removal of large downed wood as recommended by MDWRM to control Douglas-fir beetle (Armleder and Thomson 1984, Dawson et al. 2007). The interactions and feedbacks occurring among the 3 primary disturbances of the interior dry forests will be altered with the implementation of MDWRM. The application of MDWRM also includes strategies that serve to limit Douglas-fir beetle infestations such as treating downed logs with anti-aggregation pheromones, or piling and burning the logs (Armleder et al. 1986, Dawson et al. 2007). The first method may either favour greater occurrence of more severe fires by  90 applying anti-aggregation pheromones and leaving logs onsite, whereas the latter may favour, in the long-term, increased insect activity. Therefore, application of MDWRM is expected to result in susceptibility to different disturbance agents at different times depending on whether downed logs are removed.   In chapter 3, I found that widespread application of MDWRM created a more aggregated and homogenized landscape. Such management resulted in a greater areal extent of the low fire risk class while also resulting in greater aggregation of the high fire risk class. The homogenized landscape may be more fire resilient yet have negative implications on biodiversity and insect disturbances. Some species may suffer reductions by dispersal limitations which reduce colonization, whereas a landscape composed of a single widespread host-tree species, with similar structure, may be subject to widespread insect outbreaks.   4.2 Implications for forest management  Single-objective management may have negative ecological and economic implications for forests. Single-objective management could change disturbance severity. For example, increases to the fuel distribution and abundance brought about by fire suppression in the dry forests of North America has resulted in shift from less severe surface fires to more severe crown fires (Fulé et al. 1997, Allen et al. 2002, Schoennagel et al. 2004, Binkley et al. 2007). Further, implementing any objective in a way that does not reflect historic disturbance regimes may create a homogeneous landscape, composed of a single species and/or an abundance of same-sized trees thus creating conditions  91 conducive to widespread, unprecedented disturbance negatively affecting the timber supply (Raffa et al. 2008, Safranyik et al. 2010). Therefore, a deviation from historical disturbances may result in compromised resilience to subsequent disturbance with potential ecological and economic impacts (Safranyik et al. 2010, Churchill et al. 2013, Wu and Kim 2013, Dymond et al. 2014, Boucher et al. 2017).   The silvicultural treatment used in MDWRM, while originally developed to balance habitat enhancement and timber extraction, also affected disturbance likelihood with consequences in a changing climate. Mule deer winter range management addresses large downed logs by either piling and burning, or by treating piles with anti-aggregation pheromones to limit Douglas-fir beetle infestations (Dawson et al. 2007). Despite being reasonable Douglas-fir beetle management options, both approaches must be implemented in a timely manner following harvest. However, treating downed logs with anti-aggregation pheromones does not lower the surface fuel load, as the treated logs are left onsite, increasing stand susceptibility to crown fire. To simultaneously address susceptibility to both crown fire and Douglas-fir beetle, both piling and burning of downed logs would be required (Armleder and Thomson 1984, Dawson et al. 2007; Chapter 2). The removal of the downed logs (or not) will also determine stand susceptibility to future disturbance, favouring greater insect activity or higher severity fires. However, MDWRM implementation on a large scale, with timely removal of downed logs could create large block of low fire risk resulting in a fire resilient landscape (Chapter 3; Seidl et al. 2016a). Such an approach would protect values at risk in the event  92 of a longer fire season that is expected to have more frequent and severe fires (Westerling et al. 2006, Flannigan et al. 2013, Jolly et al. 2015).   In a changing climate, MDWRM may favour conditions prone to herbivorous insects. Mule deer winter range management does not reduce available foliage for the western spruce budworm yet strives towards creating favourable conditions for western spruce budworm survival and dispersal by enhancing a multilayered forest structure (Fellin and Dewey 1986, Brookes et al. 1987, Dawson et al. 2007). Further, a harvest interval exceeding 30 years may create a window of opportunity for western spruce budworm outbreaks since mean outbreak occurrence is roughly 30 years (Campbell et al. 2006, Axelson et al. 2015). Similarly, harvesting at a 30-year interval or more would allow for forest recovery and a subsequent return to ‘high’ Douglas-fir beetle susceptibility. These conditions could be further exacerbated with more variable temperature and precipitation (Easterling et al. 2000), and the prevalence of more severe and prolonged drought (Williams et al. 2013, Allen et al. 2015). The moisture stress resulting from drought could create a large pulse of available host trees for the western spruce budworm and Douglas-fir beetle as both insects preferentially feed on less vigorous host trees (Redak and Cates 1984, Rudinsky 1962, Millar and Stephenson 2015). Moreover, warmer and drier conditions that favour insect development and survival may also allow for longer outbreaks and persistence in suboptimal habitats (Raffa et al. 2008, Bentz et al. 2010). The possibility of widespread insect outbreaks is of particular concern with large-scale application of MDWRM.   93 Managing for a single objective runs the risk of being short-sighted with unforeseen lagged effects and surprises (e.g. Hadley and Veblen 1993, Hessburg et al. 2000, Taylor and Carroll 2004), and such examples abound in forested ecosystems worldwide. For example, increased agricultural production in Australia required clearing of forest, providing short-term benefit. Over the long-term, however, forest evapotranspiration responsible for maintaining low ground water levels was lowered, thus no longer preventing salinization of the soil (Rodriguez et al. 2006). Tree removal raised ground water levels and salts to eventually move to the surface resulting in unproductive lands (Rodriguez et al. 2006). Historical management in Bavaria, Germany resulted in extensive tree mortality due to Ips typographus in 2009 (Ciesla 2015). Under a policy of ‘no management’ established in the 1970s, severe storms dating back to the 1980s resulted in large amounts of windthrow, building-up populations of I. typographus, and eventually leading to an outbreak (Ciesla 2015). Activities such as salvage logging, to reduce economic loss following a disturbance event, may also have unintended long-term structural effects on the affected ecosystem. Salvage logging may increase the cumulative disturbance severity while also decreasing the number of standing dead and downed trees potentially leading to an altered successional trajectory through a change in stand structure as witnessed in northeastern Minnesota (D’Amato et al. 2011).  The short-sightedness of single-objective management is a result of placing greater emphasis on a particular goal while failing to consider other potential effects. Caribou and timber production provide two strong examples of this phenomenon. For example, single species management as opposed to managing for whole ecosystem processes leads  94 to a simplified view of ecosystems (Benson 2012). In the case of dwindling caribou populations, fire suppression was used to create late-successional lichen patches to support short-term lichen supply (Cumming 1992). However, over the long-term, fire suppression homogenized the forest and reduced the overall number of lichen patches available to caribou (Euler et al. 1976, Cumming 1992). Alternatively, timber production tactics that solely focus on maintaining live, similar-sized trees in fully stocked stands may create less resilient ecosystems (Franklin et al. 2002) leading to severe and widespread insect outbreaks (Raffa et al. 2008). Therefore, forest management should strive to create or maintain conditions that resemble low, moderate, or mixed disturbance frequency and/or severity as these may be beneficial in enhancing biodiversity while minimizing the reduction in harvestable timber (Thom and Seidl 2016). Managing in such a manner may result in greater variability across spatial and temporal scales, resulting in structures that better mimic the effects of natural disturbance (e.g. Churchill et al. 2013). Maintaining ecosystem processes by increasing biodiversity, structural diversity and planning at a scale of thousands of hectares as opposed to single stands should be a priority for managers while also being aware of potential conflicts and trade-offs among objectives (Seidl et al. 2016a).  The potential for lagged effects and surprises resulting from single-objective management also applies to MDWRM. At the stand level, a focus on creating a multilayered Douglas-fir dominated stand coupled with a harvest interval of 30 years or more may create suitable winter range for mule deer, however, may in the short-term result in greater susceptibility to crown fires and in the long-term favour western spruce budworm and/or  95 Douglas-fir beetle activity (Chapter 2). Further, the potential for altered interactions among disturbances may lead to stand structures and/or compositions undesirable to mule deer (Chapter 2). Additionally, widespread MDWRM application may, in the long-term, result in a fire resilient landscape yet could be prone to insect outbreaks (Chapter 3). Managers should be aware of this risk in creating a homogeneous landscape composed of stands with a single host-tree species, multilayered structure, and large old trees.   4.3 Future research  My research addressed the effects of changes in forest stand attributes on disturbance likelihood due to MDWRM at a fine scale and in only one forest type. This creates the need to assess other forest types where MDWRM might be applied. Further, in my research, forest stand attribute measurements were not taken immediately following the first treatment in 1984. Therefore, a similar experiment should be conducted elsewhere in interior dry forests. This would allow forest stand attribute measurements to be taken before and after the first and second silvicultural treatments, and then compared to a control thus confirming that forest recovery did in fact occur within a 30-year period. Additionally, re-running the likelihood of crown fire occurrence in a model conceived to accurately depict fire effects in interior dry forests will provide managers with a clearer idea of the effects of MDWRM on wildfire. Finally, and perhaps most importantly, while habitat usage in treated and untreated areas has been quantified (Armleder et al. 1998), the direct effects of MDWRM on mule deer populations have not. Thus far, MDWRM is only assumed to be beneficial for mule deer and is lacking empirical confirmation. Thus, one would ask: what is the effect of MDWRM on mule deer population dynamics? The  96 answer to the above question may have major implications in terms of the harvesting restrictions currently in place in the interior dry forests of BC.   More broadly, assessing landscape disturbance likelihood and susceptibility through network analysis could be used to determine a landscape’s ‘disturbance connectivity’ as well as identify ‘gateway’ stands that would promote or reduce disturbance spread. Similar to O’Neill et al. (1992), application of an epidemiological model on the landscape could identify changes to potential disturbance extent resulting from application of MDWRM or other management tactics. Further, multiple scenarios using modified epidemiological models (e.g. Munz et al. 2009 for the concept) could be generated to determine the effects of applying different silvicultural tactics on the proportion of susceptible area and subsequent disturbance spread across the landscape.    97 References  Agee, J.K., and Skinner C.N. 2005. Basic principles of forest fuel reduction treatments. Forest Ecology and Management 211 (1-2): 83-96   Alfaro, R.I., Van Sickle, G.A., Thomson, A.J., and Wegwitz, E. 1982. Tree mortality and radial growth losses caused by the western spruce budworm in a Douglas-fir stand in British Columbia. Canadian Journal of Forest Research 12 (4): 780-787  Alfaro, R.I., and Maclauchlan L.E. 1992. A method to calculate the losses caused by western spruce budworm in uneven-aged Douglas-fir forests of British Columbia. Forest Ecology and Management 55 (1-4): 295-313  Alfaro, R.I., Berg, J., and Axelson, J. 2014. Periodicity of western spruce budworm in southern British Columbia, Canada. Forest Ecology and Management 315: 72-79  Alexander, M.E., Cruz, M.G., and Lopes, A.M.G. 2006. CFIS: A software tool for simulating crown fire initiation and spread. Proceedings of 5th international conference on forest fire research, 27-30 November 2006. Amsterdam, The Netherlands pp. 1-13  Allen, C.D., Savage, M., Falk, D.A., Suckling, K.F., Swetnam, T.W., Schulke, T., Stacey, P.B., Morgan, P., Hoffman, M., and Klingel, J.T. 2002. Ecological restoration of southwestern ponderosa pine ecosystems: A broad perspective. Ecological Applications 12: 1418-1433  Allen, C.D., Breshears, D.D., and McDowell, N.G. 2015. On underestimation of global vulnerability to tree mortality and forest die-off from hotter drought in the Anthropocene. Ecosphere 6 (8): 129  Allstadt, A.J., Liebhold, A.M., Johnson, D.M., Davis, R.E., and Haynes, K.J. 2015. Temporal variation in the synchrony of weather and its consequences for spatiotemporal population dynamics. Ecology 96 (11): 2935-2946  Amoroso, M.M., Daniels, L.D., Bataineh, M., and Andison, D.W. 2011. Evidence of mixed-severity fires in the foothills of the Rocky Mountains of west-central Alberta, Canada. Forest Ecology and Management 262 (12): 2240-2249  Anderson, L., Carlson, C.E., and Wakimoto, R.H. 1987. Forest fire frequency and western spruce budworm outbreaks in western Montana. Forest Ecology and Management 22 (3-4): 251-260  Armleder, H.M., and Thomson R. 1984. Experiment harvesting of dry-belt Douglas-fir on a mule deer winter range at Knife Creek: working plan and progress report. Research Branch, Ministry of Forests. WHR-17. Victoria, B.C.  Armleder, H.M., and Dawson, R.J., Thomson, R.N. 1986. Handbook for timber and mule deer management co-ordination on winter ranges in the Cariboo forest region. Land management handbook number 13. BC ministry of forests. ISSN 0229-1622  Armleder, H.M., Leckenby D.A., Freddy, D.J., and Hicks, L.L. 1989. Integrated management of timber and deer: Interior forests of western North America. USDA. Forest Service. General Technical Report PNW-GTR-227 pp 1-23  Armleder, H.M., and Dawson, R.J. 1992. Logging on mule deer winter range: An integrated management approach. Forestry Chronicle 68 (1): 132-137  Armleder, H.M., Waterhouse, M.J., Keisker, D.G., and Dawson, R.J. 1994. Winter habitat use by mule deer in the central interior of British Columbia. Canadian Journal of Zoology 72 (10): 1721-1725  98  Armleder, H.M., Waterhouse, M.J., Dawson, R., and Iverson, K. 1998. Mule deer response to low-volume partial cutting on winter ranges in central interior British Columbia. B.C. Min. For., Victoria, B.C. Research Report 16.  Arno, S. 1980. Forest fire history in the northern Rockies. Journal of Forestry 78 (8): 460-465  Aukema, B.H., Carroll, A.L., Zhu, J, Raffa, K.F., Sickley T.A., and Taylor S.W. 2006. Landscape level analysis of mountain pine beetle in British Columbia Canada: spatiotemporal development and spatial synchrony within the present outbreak. Ecography 29 (3): 427-441  Aukema, B.H., McKee, F.R., Wytrykush, D.L., Carroll, A.L. 2016. Population dynamics and epidemiology of four species of Dendroctonus (Coleoptera: Curculionidae): 100 years since J.M. Swaine. Canadian Entomologist 148 (s1): S82-S110  Axelson, J.N. 2014. A guide to managing Douglas-fir beetles on private property. B.C. Ministry of Forests, Lands and Natural Resource Operations. Retrieved from https://www.for.gov.bc.ca/dqu/DouglasFirBeetles_factsheet_FLNRO_Web.pdf [accessed November 28 2016]  Axelson, J.N., Smith, D.J., Daniels, L.D., Alfaro, R.I. 2015. Multicentury reconstruction of western spruce budworm outbreaks in central British Columbia, Canada. Forest Ecology and Management 335: 235-248  Ayres, M.P., and Lombardero M.J. 2000. Assessing the consequences of global chance for forest disturbance from herbivores and pathogens. The Science of the Total Environment 262 (3): 263-286  Barclay, H.J., Li, C., Benson, L., Taylor, S., and Shore, T. 2005. Effects of fire return rates on traversability of lodgepole pine forests for mountain pine beetle (Coleoptera: Scolytidae) and the use of patch metrics to estimate traversability. Canadian Entomologist 137 (5): 566-583  BC Ministry of Environment. 2007. Amended-order #U-5-001, U-5-002 and U-5-003- Ungulate winter ranges Cariboo and Chilcotin land use plan, shallow and moderate snowpack. Available from http://www.env.gov.bc.ca/wld/documents/wha/Amendment_ShallowModerate_Feb07_Ord.pdf [accessed November 1 2016]  BC Ministry of Forests Lands and Natural Resource Operations. (2012). Wildfire hazard assessment and abatement. B.C. Ministry of Forests, Wildfire Management Branch. URL. ftp://ftp.for.gov.bc.ca/HPR/external/!publish/GIS/FuelHazFireRisk/ [accessed June 6th 2016]   BC Ministry of Forests Lands and Natural Resource Operations (2015a). Mule deer winter range habitat management zones-Cariboo region. B.C. Ministry of Forests, Regional Operations-Cariboo. URL  https://catalogue.data.gov.bc.ca/dataset/mule-deer-winter-range-habitat-management-zones-cariboo-region [accessed May 17, 2016]   BC Ministry of Forests Lands and Natural Resource Operations (2015b). Provincial Strategic Threat Analysis: 2015 Wildfire Threat Analysis Component. BC Wildfire Service pp. 1-33  BC Ministry of Forests Lands and Natural Resource Operations (2016). VRI- Forest Vegetation Composite Polygons and Rank 1 Layer. B.C. Ministry of Forests, Forest Analysis and Inventory Branch. URL  https://catalogue.data.gov.bc.ca/datasetvri-forest-vegetation-composite-polygons-and-rank-1-layer  [accessed May 17, 2016]   BC Wildfire Service. 1950-2013. Fire ignition data: Cariboo 100 mile house fire zone. Cariboo Fire Centre. Unpublished data provided by E., Meyer, Superintendent, Fire Weather. BC Ministry of Lands and Natural Resource Operation, Victoria, BC.   99 BC Wildfire Service. 2014. British Columbia weather stations. URL http://apps.gov.bc.ca/pub/dmf-viewer/?siteid=8461315703751178746&ll=52.3,-122&z=7 [accessed September 6 2016]  Beasley, T.M., Erickson, S., Allison, D.B. 2009. Rank-based inverse normal transformations are increasingly used, but are they merited? Behavior Genetics 39 (5): 580-595  Beaty, M.R., and Taylor, A.H. 2007. Fire disturbance and forest structure in old-growth mixed conifer forests in the northern Sierra Nevada, California. Journal of Vegetation Science 18 (6): 879-890  Bebi, P., Kulakowski, D., Veblen, T.T. 2003. Interactions between fire and spruce beetles in a subalpine Rocky mountain forest landscape. Ecology 84 (2): 362-371  Bennett, E.M., Peterson, G.D., and Gordon, L.J. 2009. Understanding relationships among multiple ecosystem services. Ecology Letters 12 (12): 1394-1404  Benson, M.H. 2012. Intelligent Tinkering: The Endangered Species Act and resilience. Ecology and Society 17 (4): 28  Bentz, B.J., Régnière, J., Fettig, C.J., Hansen, E.M., Hayes, J.L., Hicke, J.A., Kelsey, R.G., Negrón, J.F., and Seybold, S.J. 2010 Climate change and bark beetles of the western United States and Canada: direct and indirect effects. BioScience 60 (8): 602-613  Beverly, J.L., Herd, E.P.K., and Conner, J.C.R. 2009. Modeling fire susceptibility in west central Alberta, Canada. Forest Ecology and Management 258 (7): 1465-1478  Bigler, C., Kulakowski, D., Veblen, T.T. 2005. Multiple disturbance interactions and drought influence fire severity in Rocky mountain subalpine forests. Ecology 86 (11): 3018-3029  Binkley, D., Sisk, T., Chambers, C., Springer, J., Block, W. 2007. The role of old-growth forests in frequent fire landscapes. Ecology and Society 12: 18  Biondi, F., and Qeadan, F. 2008. A theory-driven approach to tree-ring standardization: defining the biological trend from expected basal area increment. Tree-Ring Research 64 (2): 81-96  Black, S.H., Kulakowski, D., Noon, B.R., DellaSala, D.A. 2013. Do bark beetle outbreaks increase wildfire risks in the central U.S. Rocky Mountains? Implications from recent research. Natural Areas Journal 33 (1): 59-65  Blackford, D. 2004. Management guide for western spruce budworm. Forest Health Protection and State Forestry Organizations USDA Forest Service 6.1 pp 1- 11  Blood, D.A. 2000. Mule and black-tailed deer in British Columbia, Ecology, Conservation and Management. B.C. Minist. Environ., Lands and Parks, Wildl. Branch. pp. 6  Boone, C.K., Aukema, B.H., Bohlmann, J., Carroll, A.L., and Raffa, K.F. 2011. Efficacy of tree defense physiology varies with bark beetle population density: a basis for positive feedback in eruptive species. Canadian Journal of Forest Research 41 (6): 1174-1188  Boucher, Y., Perrault-Hébert, M., Fournier, R., Drapeau, P., and Auger, I. 2017. Cumulative patterns of logging and fire (1940-2009): Consequences on the structure of the eastern Canadian boreal forest. Landscape Ecology 32 (2): 361-375  Bousfield W.E., Wulf, N.W., and Carlson, C.E. 1986. Rating susceptibility of stands to western spruce budworm: users guide and documentation to SBW-hazard USDA Forest Service Northern Region Report 96-7   100 Bréda, NJ-J. Ground-based measurements of Leaf Area Index: A review of methods, instruments and current controversies. Journal of Experimental Botany 54 (392): 2403-2417  Brookes, M.H., Campbell, R.W., Colbert, J.J., Mitchell, R.G., and Stark, R.W. 1987.  Western spruce budworm. United States Department of Agriculture Forest Service Cooperative State Research Service Technical Bulletin Number 1694 pp 1-198  Brown, J.K., Oberheu, R.D., and Johnston, C.M. 1982. Handbook for inventorying surface fuels and biomass in the interior West. United States Department of Agriculture Forest Service Intermountain Forest and Range Experiment Station Ogden, UT 84401 General Technical Report INT-129 pp 1-52   Brown, J.K, and Bevins, C.D. 1986. Surface fuel loadings and predicted fire behavior for vegetation types in the northern Rocky Mountains. USDA Forest Service, Intermountain Research Station, Research Note INT-358, pp 9.  Brown, J., K., Reinhardt, E., D., and Fischer, W., C. 1991.  Predicting duff and woody fuel consumption in northern Idaho prescribed fires. Forest Sciences 37 (6): 1550-1566  Brown, J.K., Reinhardt, E.D., and Kramer, K.A. 2003. Coarse woody debris: managing benefits and fire hazard in the recovering forest. United States Department of Agriculture. Forest Service. Rocky Mountain Research Station. General Technical Report RMRS-GTR-105   Buma, B., and Wessman, C.A. 2011. Disturbance interactions can impact resilience mechanisms of forests. Ecosphere 2 (5): UNSP 64  Buma, B., and Wessman, C.A. 2012. Differential species responses to compounded perturbations and implications for landscape heterogeneity and resilience. Forest Ecology and Management 266: 25-33  Buma, B., and Wessman, C.A. 2013. Forest resilience, climate change and opportunities for adaptation: a specific case of a general problem. Forest Ecology and Management 306: 216-225  Buma, B. 2015. Disturbance interactions: characterization, prediction, and the potential for cascading effects. Ecosphere 6 (4): 70  Byram, G.M. 1959. Combustion of forest fuels In: Forest fire: control and use. Davis, K.P (ed.). McGraw-Hill, New York, pp. 61-89.  Campbell, R.W. 1993. Population dynamics of the major North American needle-eating budworms. Research Paper PNW-RP-463. Portland, OR: U.S. Department of Agriculture, Forest Pacific Northwest Research Station pp. 1-222  Campbell, R., Smith, D.J., Arsenault, A. 2006. Multicentury history of western spruce budworm outbreaks in interior Douglas-fir forests near Kamloops, British Columbia. Canadian Journal of Forest Research 36 (7): 1758-1769  Cedervind, J., Pettersson, M., and Långström, B. 2003. Attack dynamics of the pine shoot beetle, Tomicus piniperda (Col.; Scolytinae) in Scots pine stands defoliated by Bupalus piniaria (Lep.; Geometridae). Agrilcultural and Forest Entomology 5 (3): 253-261  Chavardès, R.D., and Daniels, L.D. 2016. Altered mixed-severity fire regime has homogenised montane forests of Jasper National Park. International Journal of Wildland Fire 25 (4): 433-444  Chen, Z., Kolb, T.E., and Clancy, K.M. 2001. Mechanisms of Douglas-fir resistance to western spruce budworm defoliation: Bud burst phenology, photosynthetic compensation and growth rate. Tree physiology 21 (16): 1159-1169   101 Chen, Z., Clancy, K.M., and Kolb, T.E. 2003. Variation in budburst phenology of Douglas-fir related to western spruce budworm (Lepidoptera: Tortricidae) fitness. Journal of Economic Entomology 92 (2): 377-387  Christiansen, E., Waring, R.H., and Berryman A., A.1987. Resistance of conifers to bark beetle attack: Searching for general relationships. Forest Ecology and Management 22 (1-2): 89-106  Churchill, D.J., Larson, A.J., Dahgreen, M.C., Franklin, J.F., Hessburg, P.F., and Lutz, J.A. 2013. Restoring forest resilience: from reference spatial patterns to silvicultural prescriptions and monitoring. Forest Ecology and Management 291: 442-457  Ciesla, W.M. 2015. The role of human activities on forest insect outbreaks worldwide. International Forestry Review 17 (3): 269-281  Clancy, K.M., Chen, Z., and Kolb, T.E. 2004. Foliar nutrients and induced susceptibility: Genetic mechanisms of Douglas-fir resistance to western spruce budworm defoliation. Canadian Journal of Forest Research 34 (4): 939-949  Cohn, G.M., Parsons, R.A., Heyerdahl, E.K., Gavin, D.G., Flower, A. 2014. Simulated western spruce budworm defoliation decreases torching and crowning potential: A sensitivity analysis using a physics-based fire model. International Journal of Wildland Fire 23 (5): 709-720  Covington, W.W. and Moore, M.M. 1994. Southwestern ponderosa forest structure: Changes since Euro-American settlement.  Journal of Forestry 92 (1): 39-47  Cruz, M.G., Alexander, M.E., and Wakimoto, R.H. 2003. Assessing canopy fuel stratum characteristics in crown fire prone fuel types of western North America. International Journal of Wildland Fire 12 (1): 39-50  Cruz, M.G., Alexander, M.E., Wakimoto R.H. 2004. Modeling the likelihood of crown fire occurrence in conifer forest stands. Forest Science 50 (5): 640-658  Cruz, M.G. and Alexander, M.E. 2006. Crown Fire Initiation and Spread system description]. Available from https://www.frames.gov/catalog/7374 [accessed September 6 2016]  Cumming, H.G. 1992. Woodland caribou: facts for forest managers. Forestry Chronicle 68 (4): 481-489  D’Amato, A.W., Fraver, S., Palik, B.J., Bradford, J.B., and Patty, L. 2011. Singular and interactive effects of blowdown, salvage logging, and wildfire in sub-boreal pine systems. Forest Ecology and Management 262 (11): 2070-2078  Dale, V.H., Joyce, L.A., McNulty, S., Neilson, R.P., Ayres, M.P., Flannigan, M.D., Hanson, P.J., Irland, L.C., Lugo, A.E., Peterson, C.J., Simberloff, D., Swanson, F.J., Stocks, B.J., and Wotton B.M. 2001. Climate change and forest disturbances. Bioscience 51 (9): 723-734  Daniels, L.D. 2004. Climate and Fire: A case study of the Cariboo Forest, British Columbia. Pp. 235-246 in Taylor, L., J. Zelnik, S. Cadwallader, and B. Hughes (editors). Proceedings of the Mixed Severity Fire Regimes: Ecology and Management Conference, Washington State University, WA.  Daniels, L.D., Maertens, T.B., Stan, A.B., McCloskey, S.P.J., Cochrane, J.D., and Gray, R.W. 2011. Direct and indirect impacts of climate change on forests: three case studies from British Columbia. Canadian Journal of Plant Pathology 33 (2): 108-116  Daniels, L., Erasmus, H., Forge, B., Greene, G., Lavallee, S. 2014. Monitoring fire intensity and severity: 2013 prescribed burn at the Westside Unit, Vaseux_Bighorn National Wildlife Area. Report to the Canadian Wildlife Service.   102 Da Silva, E. 2009. Wildfire history and its relationship with top-down and bottom-up controls in the Joseph and Gold Creek watersheds, Kootenay Mountains, British Columbia. M.Sc. Thesis. University of Guelph, Faculty of Graduate Studies. Guelph, ON, Canada 107p.  Dawson, R.J., Armleder, H.M., Bings, B.A., and Peel, D.E. 2007. Management strategy for mule deer winter ranges in the Cariboo-Chilcotin-Part1a: Management Plan for shallow and moderate snowpack zones. B.C. Ministry of Forests and Range, Research Branch, Victoria, B.C. Land Management Handbook 60. Available from http://www.for.gov.bc.ca/hfd/pubs/lmh.htm [accessed December 15, 2015]  Day, K. 1997. Interior Douglas-fir and selection management at the UBC/Alex Fraser Research Forest: Forest management for mule deer winter range For: IUFRO Uneven-aged Silviculture Symposium Corvallis, Oregon September 15-19, 1997 pp 1-11  Day, K. 2007. History of the UBC Alex Fraser Research Forest 1987-2007. Available from http://afrf.forestry.ubc.ca/files/2012/03/HistoryAFRF.pdf [accessed December 17, 2014]  Dellasala, D.A., Baker, R., Heiken, D., Frissell, C.A., Karr, J.R., Nelson, S.K., Noon, B.R., Olson, D., and Strittholt, J. 2015. Building on two decades of ecosystem management and biodiversity conservation under the northwest forest plan, USA. Forests 6 (9): 3326-3352  Dodds, K.J., Garman, S.L., Ross, D.W. 2006. Landscape analyses of Douglas-fir beetle populations in northern Idaho. Forest Ecology and Management 231 (1-3): 119-130  Donato, D.C., Harvey, B.J., Romme, W.H., Simard, M., Turner, M.G. 2013. Bark beetle effects on fuel profiles across a range of stand structures in Douglas-fir forests of Greater Yellowstone. Ecological Applications 23 (1): 3-20  Dormann, C.F., Schweiger, O., Augenstein, I., Bailey, D., Billeter, R., de Blust, G., DeFilippi, R., Frenzel, M., Hendrickx, F., Herzog, F., Klotz, S., Liira, J., Maelfait, J-P., Schmidt, T., Speelmans, M., van Wingerden, W.K.R.E., Zobel, M. 2007. Effects of landscape structure and land-use intensity on similarity of plant and animal communities. Global Ecology and Biogeography 16 (6): 774-787  Duncan, R.P. 1989. An evaluation of errors in tree age estimates based on increment cores in Kahikatea (Dacrycarpus dacrydioides). New Zealand Natural Sciences 16: 31-37  Dunning, J.B., Danielson, B.J., Pulliam, H.R. 1992. Ecological processes that affect populations in complex landscapes. Oikos 65 (1): 169-175  Dupouey, J.L, Dambrine, E., Laffite, D., Moanes, C. 2002. Irreversible impact of past land use on forest soils and biodiversity. Ecology 83 (11): 2978-2984  Dymond, C.C., Tedder, S., Spittlehouse, D.L., Raymer, B. Hopkins, K., McCallion, K., and Sandland, J. 2014. Diversifying managed forests to increase resilience. Canadian Journal of Forest Research 44 (10): 1196-1205  Easterling, D.R., and Meehl, G.A., Parmesan, C., Changnon S.A., Karl, T.R., and Mearns, L.O. 2000. Climate extremes: Observations, modelling and impacts. Science 289 (5487): 2068-2074  Erickson, R.D. 1992. History of important forest insects in the Cariboo Region 1913-1991. For. Can. FIDS Rpt. 21-14, p.48  ESRI. 2015. ArcMap 10.3.1  Eubanks, J.O., Michener, W.K., and Guyer, C. 2003. Patterns of movement and burrow use in a population of gopher tortoise (Gopherus polyphemus). Herpetologica 59 (3): 311-321   103 Euler, D.L., Snyder, B., and Timmerman, H.R. 1976. Woodland caribou and plant communities on the Slate Islands, Lake Superior. Can. Field-Nat. 90: 17-21  Falk, D.A., Miller, C., McKenzie, D., and Black, A.E. 2007. Cross-scale analysis of fire regimes. Ecosystems 10 (5): 809-823  Fellin, D.G., and Dewey, J.E. 1986. Western spruce budworm. U.S Department of Agriculture Forest Service. Forest Insect & Disease Leaflet 53  Fettig, C.J., Klepzig, K.D., Billings, R.F., Munson, A.S., Nebeker, T.E., Negròn J.F., and Nowak, J.T. 2007. The effectiveness of vegetation management practices for prevention and control of bark beetle infestations in coniferous forests of the western and southern United States. Forest Ecology and Management 238 (1-3): 24-53  Fettig, C.J., Gibson, K.E., Munson, A.S., and Negròn J.F. 2014. Cultural practices for prevention and mitigation of mountain pine beetle infestations. Forest Science 60 (3):450-463  Fiedler, C.E., Metlen, K.L., and Dodson, E.K. Restoration treatment effects on stand structure, tree growth, and fire hazard in a ponderosa pine/Douglas-fir forest in Montana. 2010. Forest Science 56 (1): 18-31  Flannigan, M., Cantin, A.S., de Groot, W.J., Wotton, M., Newbery, A., and Gowman, L.M. 2013. Global wildland fire season severity in the 21st century. Forest Ecology and Management 294 (SI): 54-61  Fleming, R.A., Candau, J-N., McAlpine, R.S. 2002. Landscape-scale analysis of interactions between insect defoliation and forest fire in central Canada. Climate Change 55 (1-2): 251-272  Flower, A., Gavin, D.G., Heyerdahl, E.K., Parsons, R.A., and Cohn, G.M. 2014a. Drought-triggered western spruce budworm outbreaks in the interior Pacific northwest: A multi-century dendrochronological record. Forest Ecology and Management 324: 16-27  Flower, A., Gavin, D.G., Heyerdahl, E.K., Parsons, R.A., and Cohn, G.M. 2014b. Western spruce budworm outbreaks did not increase fire risk over the last three centuries: A dendrochronological analysis of inter-disturbance synergism. PLoS ONE 9 (12): 1-25  Foley, J.A., DeFries, R., Asner, G.P., Barford, C., Bonan, G., Carpenter, S.R., Chapin, F.S., Coe, M.T., Daily, G.C., Gibbs, H.K., Helkowski, J.H., Holloway, T., Howard, E.A., Kucharik, C.J., Monfreda, C., Patz, J.A., Prentice, I.C., Ramankutty, N., and Snyder, P.K. 2005. Global consequences of land use. Science 309 (5734): 570-574  Franklin, J.F., Spies, T.A., Van Pelt, R., Carey, A.B., Thornburgh, D.A., Berg, D.R., Lindenmayer, D.B., Harmon, M.E., Keeton, W.S., Shaw, D.C., Bible, K., Chen, J. 2002. Disturbances and structural development of natural forest ecosystems with silvicultural implications, using Douglas-fir forests as an example. Forest Ecology and Management 155 (1-3): 399-423  Fraterrigo, J.M., and Rusak, J.A. 2008. Disturbance-driven changes in the variability of ecological patterns and processes. Ecological Letters 11 (7): 756-770  Frazer, G.W., Canham, C.D., and Lertzman, K.P. 1999. Gap Light Analyzer (GLA), Version 2.0: Imaging software to extract canopy structure and gap light transmission indices from true-colour fisheye photographs, users manual and program documentation. Copyright © 1999: Simon Fraser University, Burnaby, British Columbia, and the Institute of Ecosystem Studies, Millbrook, New York.  Fulé, P.Z., Covington, W.W., Moore, M.M.1997. Determining reference conditions for ecosystem management of southwestern ponderosa pine forests. Ecological Applications 7 (3): 895-908   104 Furniss, M.M. 1965. Susceptibility of fire-injured Douglas-fir to bark beetle attack in southern Idaho. Journal of Forestry 63 (1): 8-11  Furniss, R.L., and Carolin, V.M. 1977. Western forest insects. Miscellaneous Publication 1339. United States Department of Agriculture-Forest Service, Pacific Northwest Forest and Range Research Station, Portland, Oregon, United States of America.  Furniss, M.M., McGregor, M.D., Foiles, M.W., and Partridge, A.D. 1979. Chronology and characteristics of a Douglas-fir beetle outbreak in northern Idaho. USDA Forest Service General Technical Report INT-59. Intermountain Forest and Range Experiment Station, Ogden, UT. Pp 19.  Furniss, M.M., Livingston, R.L., and McGregor, M.D. 1981. Development of a stand susceptibility classification for Douglas-fir beetle. In: Hedden, R.L., Barras, S.J., Coster, J.E. (Tech. Coords.), Hazard-Rating Systems in Forest Insect Pest Management: Symposium Proceedings. USDA Forest Service Washington Office General Technical Report WO-27, pp.115-128  Gavin, D., Flower, A., Cohn. G., Heyerdahl, E., and Parsons, R. 2013. Interactions of insects, fire and climate on fuel loads and fire behavior in mixed conifer forest. Final Report for JFSP Project 09-1-096-5. Available from https://www.firescience.gov/projects/09-1-06-5/project/09-1-06-5_final_report.pdf [accessed October 2, 2016]  Government of British Columbia, Ministry of Forests. 1996. A PDF providing basic information on the Ecology of the Interior Douglas-fir. Available from http://www.for.gov.bc.ca/hfd/pubs/docs/bro/bro47.pdf [accessed November 19, 2014]  Gonzalez, J.S. 1990. Wood density of Canadian tree species. Forestry Canada, Northwest Region, Northern Forestry Centre, Edmonton, Alberta. Information Report NOR-X-315. p. 130.  Gonzalez, J.R., del Barrio, G., Duguy, B. 2008. Assessing functional landscape connectivity for disturbance propagation on regional scales-a cost-surface model approach applied to surface fire spread. Ecological Modelling 211 (1-2): 121-141  Greene, G.A., and Daniels, L.D. 2017. Spatial interpolation and mean fire interval analyses quantify historical mixed-severity fire regimes. International Journal of Wildland Fire 26 (2): 136-147  Gunderson, L.H. 2000. Ecological resilience- in theory and application. Annual Review of Ecology and Systematics 31 (1): 425-439  Gustafson, E.J., and Gardner, R.H. 1996. The effect of landscape heterogeneity on the probability of patch colonization. Ecology 77 (1): 94-107  Haddon, B.D. (editor). 1988. Forest inventory terms in Canada. Part II. Glossary. Canadian Forest Services, Petawawa National Forestry Institute, Chalk River Ont. Available from https://www.for.gov.bc.ca/hfd/library/documents/glossary/glossary.pdf [accessed November 21, 2016]  Hadley, K.S., and Veblen, T.T. 1993. Stand response to western spruce budworm and Douglas-fir beetle outbreaks, Colorado Front Range. Canadian Journal of Forest Research 23 (3): 479-491  Hadley, K.S. 1994. The role of disturbance, topography, and forest structure in the development of a montane forest landscape. Bulletin of the Torrey Botanical Club 121 (1): 47-61  Haggstrom, D.A., and Kelleyhouse, D.G. 1996. Silviculture and wildlife relationships in the boreal forest of Interior Alaska. The Forest Chronicle 72 (1): 59-62   105 Hansen, W.D., Chapin III, F.S., Naughton, H.T., Rupp, T.S., and Verbyla, D.  2016. Forest-landscape structure mediates effects of a spruce bark beetle (Dendroctonus rufipennis) outbreak on subsequent likelihood of burning in Alaskan boreal forest. Forest Ecology and Management 369: 38-46  Harmon, M.E., Franklin, J.F., Swanson, F.J., Sollins, P., Gregory, S.V., Lattin, J.D., Anderson, N.H., Cline, S.P., Aumen, N.G., Sedell, J.R., Lienkaemper, G.W., Cromack Jr., K., and Cummins, K.W. 1986. Ecology of coarse woody debris in temperate ecosystems. Advances in Ecological Research 15: 133-302  Hart, S.J., Veblen, T.T., Eisenhart, K., Jarvis, D., and Kulakowski D. 2014a. Drought induces spruce beetle (Dendroctonus rufipennis) outbreaks across northwestern Colorado. Ecology 95 (4): 930-939  Hart, S.J., Veblen, T.T., Kulakowski, D. 2014b. Do tree and stand-level attributes determine susceptibility  of spruce-fir forests to spruce beetle outbreaks in the early 21st century? Forest Ecology and Management 318 (15): 44-53  Haynes, K.J., Bjørnstad, O.N., and Allstadt, A.J. 2013. Geographical variation in the spatial synchrony of a forest-defoliating insect: isolation of environmental and spatial drivers. Proceedings of the Royal Society B-Biological Sciences 280 (1753): 20130112  Haynes, K.J., Allstadt, A.J., and Klimetzek, D. 2014. Forest defoliator outbreaks under climate change: effects on the frequency and severity of outbreaks of five pine insect pests. Global Change Biology 20 (6): 2004-2018  Hessburg, P.F., Mitchell, R.G., and Filip, G.M. 1994. Historical and current roles of insects and pathogens in eastern Oregon and Washington forested landscapes. USDA Forest Service, Pacific Northwest Research Station, General Technical Report PNW-GTR-327.  Hessburg, P.F., Smith, B., G., Salter, R., B., Ottmar, R., D., and Alvarado, E. 2000. Recent changes (1930s-1990s) in spatial patterns interior northwest forests, USA. Forest Ecology and Management 136 (1-3): 53-83  Hessburg, P.F., Agee, J.K., and Franklin, J.F. 2005. Dry forests and wildland fires of the inland northwest USA: Contrasting the landscape ecology of the pre-settlement and modern eras. Forest Ecology and Management 211 (1-2): 117-139  Heyerdahl, E.M., Brubaker, L.B., and Agee, J.K. 2002. Annual and decadal climate forcing of historical fire regimes in the interior Pacific Northwest, USA. Holocene 12 (5): 597-604  Hicke, J.A., Johnson, M.C., Hayes, J.L., and Preisler, H.K. 2012. Effects of bark beetle-caused tree mortality on wildfire.  Forest Ecology and Management 271 (1): 81-90  Hicks Jr., R.R. Coster, J.E., Mason, G.N. 1987. Forest insect hazard rating. Journal of Forestry 85 (10): 20-26 Available from http://scholarworks.sfasu.edu/forestry/346   Hoffman, J.T. 2004. Management guide for Schweinitzii root and butt rot. USDA US Forest Service p3.  Hood, S., and Bentz, B. 2007. Predicting postfire Douglas-fir beetle attacks and tree mortality in the northern Rocky Mountains. Canadian Journal of Forest Research 37 (6): 1058-1069  Hope, G.D., Mitchell W.R., Lloyd D.A., Erickson W.R., Harper W.L., and Wikeem B.M. 1991. Chapter 10: Interior Douglas-fir Zone. In Ecosystems of British Columbia. D. Meidinger and J. Pojar (ed.) B.C. Ministry of Forests, Research Branch, pp. 153-165.  Horn, S., and Hanula, J.L. 2008. Relationship of coarse woody debris to arthropod availability for Red-Cockaded Woodpeckers and other bark-foraging birds on Loblolly Pine boles. Journal of Entomological Science 43 (2): 153-168  106  Hummel, S., and Agee, J.K. 2003. Western spruce budworm defoliation effects on forest structure and potential fire behaviour. Northwest Science 77 (2): 159-169  Humphreys, N. 2000. Douglas-fir beetle in British Columbia. Forest Pest Leaflet 14. Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, Victoria, British Columbia, Canada  James, P.M.A, Fortin, M.-J., Sturtevant, B.R., Fall, A., and Kneeshaw, D. 2011. Modelling spatial interactions among fire, spruce budworm, and logging in the Boreal forest. Ecosystems 14 (1): 60-75  Jenkins, M.J., Hebertson, E., Page, W., and Jorgensen, C.A. 2008. Bark beetles, fuels, fires and implications for forest management in the Intermountain West. Forest Ecology and Management 254 (1): 16-34  Jepsen, J.U., Kapari, L., Hagen. S.B., Schott, T., Vindstad, O.P.L., Nilssen, A.C., and Ims, R.A. 2011. Rapid northwards expansion of a forest insect pest attributed to spring phenology matching with sub-Arctic birch. Global Change Biology 17 (6): 2071-2083  Jolly, W.M., Cochrane, M.A., Freeborn, P.H., Holden, Z.A., Brown, T.J., Williamson, G.J., and Bowman, D.M.J.S. 2015. Climate-induced variation in global wildfire danger from 1979 to 2013. Nature Communications 6: 7537  Jones, J.C., and Dorr, B. 2004. Habitat associations of gopher tortoise burrows on industrial timberlands. Wildlife Society Bulletin 32 (2): 456-464  Jonsson, B., Holm, S., and Kallur, H. 1992. A forest inventory method based on density-adapted circular plot size. Scandinavian Journal of Forest Research 7 (1-4): 405-421  Kaufmann, M., Binkley, D., Fulé, P., Z., Johnson, M., Stephens, S., L., Swetnam, T., W. 2007. Defining old-growth for fire-adapted forest of the western United States. Ecology and Society 12 (2): 15  King, A.W., and With, K.A. 2002. Dispersal success on spatially structure landscapes: when do spatial pattern and dispersal behaviour really matter? Ecological Modelling 147 (1): 23-39  Klinka, K., Varga, P., Thretheway, C., Koot, C., Rau, M., Macku, J., and Kusbah, T. 2004 Site units of the University of British Columbia Alex Fraser Research Forest. UBC Alex Fraser Research Forest ©  Knapp, E.E., Keeley, J.E., Ballenger, E.A., and Brennan, T.J. 2005. Fuel reduction and coarse woody debris dynamics with early season and late season prescribed fire in a Sierra Nevada mixed conifer forest. Forest Ecology and Management 208 (1-3): 383-397  Koot, C., Day, K., Ewen, S., and Skea, D. 2015. Harvesting on mule deer winter range under general wildlife measures for shallow and moderate snowpack zones: Approach and lessons learned following a second harvest entry after 30 years. Technical report prepared for Ministry of Forests, Lands and Natural Resource Operations, Cariboo Region.  Kremen, C. 2005. Managing ecosystem services: what do we need to know about their ecology? Ecology Letters 8 (5): 468-479  Kulakowski, D., and Veblen, T.T. 2007. Effects of prior disturbances on the extent and severity of wildfire in Colorado subalpine forests. Ecology 88 (3): 759-769  Kulakowski, D., and Jarvis, D. 2011. The influence of mountain pine beetle outbreaks and drought on severe wildfires in northwestern Colorado and southern Wyoming: A look at the past century. Forest Ecology and Management 262 (9): 1686-1696   107 Kulakowski, D., and Jarvis, D. 2013. Low severity fires increase susceptibility of lodgepole pine to mountain pine beetle outbreaks in Colorado. Forest Ecology and Management 289 (1): 544-550  Kurz, W.A., Dymond, C.C., Stinson, G., Rampley, G.J., Neilson, E.T., Carroll, A.L., Ebata, T., and Safranyik, L. 2008a. Mountain pine beetle and forest carbon feedback to climate change. Nature 452 (7190): 987-990  Kurz, W.A., Stinson, G., Rampley, G.J., Dymond, C.C., and Neilson, E.T. 2008b. Risk of natural disturbances makes future contribution of Canada’s forests to the global carbon cycle highly uncertain. Proceedings of the National Academy of Sciences of the United States of America 105 (5): 1551-1555  Långström, B., Annila, E., Hellqvist, C., Varama, M., and Niemelä, P. 2001. Tree mortality, needle biomass recovery and growth losses in Scots pine following defoliation by Diprion pini (L.) and subsequent attack by Tomicus piniperda (L.). Scandinavian Journal of Forest Research 16 (4): 342-353  Larsson, L. 2011a. CooRecorder program of the CDendro package version 7.8. Cybis Elektronik & Data AB, Saltsjöbaden, Sweden.  Larsson, L. 2011b. CDendro program of the CDendro package version 7.8. Cybis Elektronik & Data AB, Saltsjöbaden, Sweden.  Latif, Q.S., Sanderlin, J.S., Saab, V.A., Block, W.M., and Dudley, J.G. 2016. Avian relationships with wildfire at two dry forest locations with different historical fire regimes. Ecosphere 7 (5): e01346  Lehmkuhl, J.F., Kie, J.G., Bender, L.C., Servheen, G., and Nyberg, H. 2001. Evaluating the effects of ecosystem management alternatives on elk, mule deer, and white-tailed deer in the interior Columbia River basin, USA. Forest Ecology and Management 153 (1-3): 89-104  Lessard, E.D., and Schmid, J.M. 1990. Emergence, attack densities, and host relationships for the Douglas-fir beetle (Dendroctonus pseudotsugae Hopkins) in northern Colorado. Great Basin Naturalist 50 (4): 333-338  Lessard, V.C., Drummer, T.D., and Reed, D.D. 2002. Precision of density estimates from fixed-radius plots compared to N-tree distance sampling. Forest Science 48 (1): 1-6  Lutz, J.A., van Wagtendonk, J.W., Thode, A.E., Miller, J.D., and Franklin, J.F. 2009. Climate, lightning ignitions, and fire severity in Yosemite National Park, California, USA. International Journal of Wildland Fire 18 (7): 765-774  Lyons, A.L., Gaines, W.L., Lehmkuhl, J.F., and Harrod, R.J. 2008. Short-term effects of fire and fire surrogate treatments on foraging tree selection by cavity-nesting birds in dry forests of central Washington. Forest Ecology and Management 255 (8-9): 3203-3211  Lynch, H.J., and Moorcroft, P.R. 2008. A spatiotemporal Ripley’s K-function to analyze interactions between spruce budworm and fire in British Columbia, Canada. Canadian Journal of Forest Research 38 (12): 3112-3119  Maclauchlan, L.E., Brooks, J.E., and Hodge, J.C. 2006. Analysis of historic western spruce budworm defoliation in south central British Columbia. Forest Ecology and Management 226 (1-3): 351-356  Maclauchlan LE., and Brooks J.E. 2009. Influence of past forestry practices on western spruce budworm defoliation and associated impacts in southern British Columbia. BC Journal of Ecosystems and Management 10 (2): 37-49  Magnussen, S., Boudewyn, P., and Alfaro, R. 2004. Spatial prediction of the onset of spruce budworm defoliation. The Forest Chronicle 80 (4): 485-494  108  Marañòn-Jiménez, S., and Castro, J. 2012. Effect of decomposing post-fire coarse woody debris on soil fertility and nutrient availability in a Mediterranean ecosystem. Biogeochemistry 112 (1-3): 519-535  Marciniak, A.R. 2015. Shifting disturbances in a warming environment: the western spruce budworm and Douglas-fir beetle in British Columbia. M.Sc. Thesis. University of British Columbia, Faculty of Forestry. Vancouver, BC, Canada. p80.  Marcoux, H.M., Gergel, S.E., and Daniels, L.D. 2013 Mixed-severity fire regimes: How well are they represented by existing fire-regime classification systems? Canadian Journal of Forest Research 43 (7): 658-668  Marcoux, H.M., Daniels, L.D., Gergel, S.E., Da Silva, E., Gedalof, Z., and Hessburg, P.F. 2015. Differentiating mixed- and high-severity fire regimes in mixed-conifer forests of the Canadian Cordillera. Forest Ecology and Management 341: 45-58  Marshall, P., and LeMay, V. 2005. Forest Measurements: Forest inventory In: Forestry handbook for British Columbia fifth edition edited by: Susan Watts Ph.D. RPF and Lynne Tolland MSc. pp 584-591  McCullough, D.G., Werner, R.A., and Neumann, D. 1998. Fire and insects in northern and boreal forest ecosystems of North America. Annual Review of Entomology 43: 107-127  McGarigal K and Ene E. 2013. FRAGSTATS: A spatial pattern analysis program for categorical maps version 4.2  McGarigal K. 2014. FRAGSTATS help manual.  McRae, W.A., Landers, J.L., and Garner, J.A. 1981. Movement patterns and home range of the gopher tortoise. The American Midland Naturalist 106 (1): 165-179  Meigs, G.W., Campbell, J.L., Zald, H.S.J., Bailey, J.D., Shaw, D.C., and Kennedy, R.E. 2015. Does  wildfire likelihood increase following insect outbreaks in conifer forests? Ecosphere 6 (7): 118  Meigs, G.W., Zald, H.S.J., Campbell, J.L., Keeton, W.S., and Kennedy, R.E. 2016. Do insect outbreaks reduce the severity of subsequent forest fires? Environmental Research Letters 11 (4): 045008  Millar, C.I., and Stephenson, N.L. 2015. Temperate forest health in an era of emerging megadisturbance. Science 349 (6250): 823-826  Millington, J.D.A., Walters, M.B., Matonis, M.S., Laurent, E.J., Hall, K.R., and Liu, J. 2011. Combined long-term effects of variable tree regeneration and timber management on forest songbirds and timber production. Forest Ecology and Management 262 (5): 718-729  Morrison D. 1987. Chapter 9: Landscape restoration in response to previous disturbance. In Landscape heterogeneity and disturbance. 1st edition. Turner, M.G., editor. Springer-Verlag, New York, USA. pp.159-172  Munz, P., Hudea, I, Imad, J., and Smith? R.J. 2009. Chapter 4: When zombies attack!: Mathematical modelling of an outbreak of zombie infection. In: Infectious disease modelling research progress. Editors: J.M. Tchuenche and C. Chiyaka, Nova Science Publishers, Inc. pp.133-150  Murdock, T.Q., Taylor, S.W., Flower, A., Mehlenbacher, A., Montenegro, A., Zwiers, F.W., Alfaro, R., and Spittlehouse, D.L. 2013. Pest outbreak distribution and forest management impacts in a changing climate in British Columbia. Environmental Science & Policy 26: 75-89   109 Natural Resources Conservation Service. 2009. Prescribed burning Iowa job sheet conservation practice 338. United States Department of Agriculture.  Negrón, J.F. 1998. Probability of infestation and extent of mortality associated with the Douglas-fir beetle in the Colorado front range. Forest Ecology and Management 107 (1-3): 71-85  Negrón, J.F., Schaupp, W.C., Gibson, K.E., Anhold, J., Hansen, D., Thier, R., and Macettinni, P. 1999. Estimating extent of mortality associated with the Douglas-fir beetle in the central and northern Rockies. Western Journal of Applied Forestry 14 (3): 121-128  Negrón, J.F., Anhold, J.A., and Munson, A.S. 2001. Within-stand spatial distribution of tree mortality caused by the Douglas-fir beetle (Coleoptera: Scolytidae). Environmental Entomology 30 (2): 215-224  Nesbitt, J.H. 2010. Quantifying forest fire variability using tree rings Nelson, British Columbia 1700-Present. M.Sc. Thesis. University of British Columbia, Faculty of Arts. Vancouver, BC, Canada, 85p.  Norton, D., A., Palmer, J., G., and Ogden, J. 1987. Dendroecological studies in New Zealand 1. An evaluation of tree age estimates based on increment cores. New Zealand Journal of Botany 25 (3): 373-383  O’Donnell, A.J., Boer, M.M., McCaw, W.L., and Grierson, P.F. 2011. Vegetation and landscape connectivity control wildfire intervals in unmanaged semi-arid shrublands and woodlands in Australia. Journal of Biogeography 38 (1): 112-124  O’Donnell, K., M., Thompson III, F., R., and Semlitsch, R., D. 2015. Prescribed fire and timber harvest effects on terrestrial salamander abundance, detectability and microhabitat use. The Journal of Wildlife Management 79 (5): 766-775  Oliver, C.D., and Larson, B.C. 1996. Chapter 5: Overview of stand development patterns. In Forest stand dynamics update edition. John Wiley & Sons, Inc. New York, Chichester, Brisbane, Toronto, Singapore. pp.145-170  O’Neill, RV., Gardner, R.H., Turner, M.G., and Romme, W.H. 1992. Epidemiology theory and disturbance spread on landscapes. Landscape Ecology 7 (1):19-26  Page, W.G., and Jenkins, M.J. 2007a. Mountain pine beetle-induced changes to selected lodgepole pine complexes within the Intermountain region. Forest Science 53 (4): 507-518  Page, W.G., and Jenkins, M.J. 2007b. Predicted fire behavior in selected mountain pine beetle-infested lodgepole pine. Forest Science 53 (6): 662-674  Paine, R.T., Tegner, M.J., and Johnson, E.A. 1998. Compounded perturbations yield ecological surprises. Ecosystems 1 (6): 535-545  Palik, B.J., Mitchell, R.J., and Hiers, J.K. 2002. Modeling silviculture after natural disturbance to sustain biodiversity in the longleaf pine (Pinus palustris) ecosystem: balancing complexity and implementation 155 (1-3): 347-356  Parker, T.J., Clancy, K.M., Mathiasen, R.L. 2006. Interactions among fire, insects and pathogens in coniferous forests of the interior western United States and Canada. Agricultural and Forest Entomology 8 (3): 167-189  Peltonen, M., Liebhold, A.M., Bjønstad, O.N., and Williams D.W. 2002. Spatial synchrony in forest insect outbreaks: roles of regional stochasticity and dispersal. Ecology 83 (11): 3120-3129  Perkins J.W., and Hunter JR., M.L. 2006. Effects of riparian timber management on amphibians in Maine. The Journal of Wildlife Management 70 (3): 657-670  110  Perry, D.A., Hessburg, P.F., Skinner, C.N., Spies, T.A., Stephens, S.L., Taylor, A.H., Franklin, J.F., McComb, B., and Riegel, G. 2011. The ecology of mixed severity fire regimes in Washington, Oregon, and northern California. Forest Ecology and Management 262 (5): 703-717  Peterson, G., Allen, C.R., and Holling, C.S. 1998. Ecological resilience, biodiversity, and scale. Ecosystems 1 (1): 6-18  Pojar, J., and Meidinger, D. 1991. Chapter 2: British Columbia: The environmental setting. In Ecosystems of British Columbia. D. Meidinger and J. Pojar (ed.) B.C. Ministry of Forests, Research Branch pp. 39-68  Pogue, A.M. 2017. Humans, climate and an ignitions-limited fire regime at Vaseux Lake. M.Sc. Thesis. University of British Columbia, Faculty of Forestry. Vancouver, BC, Canada p60.  Powers, J.S., Sollins, P., Harmon, M.E., and Jones, J.A. 1999. Plant-pest interactions in time and space: a Douglas-fir bark beetle outbreak as a case study. Landscape Ecology 14 (2): 105-120  Prugh, L.R., Hodges, K.E., Sinclair, A.R.E., and Brashares, J.S. 2008. Effect of habitat area and isolation on fragmented animal populations. Proceedings of the National Academy of Sciences of the United States of America 105 (52): 20770-20775  R Core Team. 2015. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.  Radeloff, V.C., Mladenoff, D.J., and Boyce, M.S. 2000. Effects of interacting disturbances on landscape patterns: Budworm defoliation and salvage logging. Ecological Application 10 (1): 233-247  Raffa, KF., Phillips, T.W., and Salom S.M. 1993. Strategies and mechanisms of host colonization by bark beetles. In: Beetle-pathogen interactions in conifer forests. Edited by: Schowalter T.D. and Filip, G.M. Academic Press, Harcourt Brace & Co. London, UK. Pp 103-128   Raffa, K.F., Aukema, B.H,. Bentz, B.J., Carroll, A.L., Hicke, J.A., Turner, M.G., and Romme, W.H. 2008. Cross-scale drivers of natural disturbance prone to anthropogenic amplification: The dynamics of bark beetle eruptions. Bioscience 58 (6): 501-517  Raffa, K.F., Aukema, B.H., Bentz, B.J., Carroll, A.L., Hicke, J.A., and Kolb, T.E. 2015. Chapter 10: Responses of tree-killing bark beetles to a changing climate In: Climate change and insect pests. Björk man, C. H., and Niemelä, P. (eds.).173-201 pp.  Redak, R.A., and Cates, R.G. 1984. Douglas-fir (Pseudotsuga menziesii)-spruce budworm (Choristoneura occidentalis) interactions: the effect of nutrition, chemical defenses, tissue phenology, and tree physical parameters on budworm success. Oecologia 62 (1): 61-67  Rees, W.E. 2003. Economic development and environmental protection: an ecological economics persperctive. Environmental Monitoring and Assessment 86 (1-2): 29-45  Rhemtulla, J.M., and Mladenoff, D.J. 2007. Why history matters in landscape ecology. Landscape Ecology 22: 1-3  Rodriguez, J.P., Beard Jr., T.D., Bennett, E.M., Cumming, G.S., Cork, S.J., Agard, J., Dobson, A.P., and Peterson, G.D. 2006. Trade-offs across space, time, and ecosystem services. Ecology and Society 11 (1): 28  Romme, W.H., Everham, E.H., Frelich, L.E., Moritz, M.A., Sparks, R.E. 1998. Are large infrequent disturbances qualitatively different from small, infrequent disturbances? Ecosystems 1 (6): 524-534   111 Ross, D.W., and Daterman, G.E. 1997 Integrating pheromone and silvicultural methods for managing the Douglas-fir beetle. In: Integrating cultural tactics into the management of bark beetle and reforestation pests. USDA Forest Service General Report NE-236 pp 135-145  Rothermel, R.C., and Deeming, J.E. 1980. Measuring and interpreting fire behavior for correlation with fire effects. General Technical Report INT-GTR-93. Ogden, UT: USDA Forest Service, Intermountain Forest and Range Experiment Station, p3.  Rudinsky, J.A. 1962. Ecology of Scolytidae. Annual Review of Entomology 7 (1): 327-348  Rudinsky, J.A. 1966. Host selection and invasion by Douglas-fir beetle, Dendroctonus pseudotsugae Hopkins, in coastal Douglas-fir forests. Canadian Entomologist 98 (1): 98-111  Ryerson, D.E., Swetnam, T.W., and Lynch, A.M. 2003. A tree-ring reconstruction of western spruce budworm outbreaks in the San Juan mountains, Colorado, USA. Canadian Journal of Forest Research 33 (6): 1010-1028  Ryu. S-R., Chen, J., Zheng, D., Lacroix, J.J. 2007. Relating surface fire spread to landscape structure: an application of FARSITE in a managed forest landscape. Landscape and Urban Planning 83 (4): 275-283  Safranyik, L., Carroll, A.L., Régnière, J., Langor, D.W., Riel, W.G., Shore, T.L., Peter, B., Cooke, B.J., Nealis, V.G., and Taylor, SW. 2010. Potential for range expansion of mountain pine beetle into the boreal forest of North America. The Canadian Entomologist 142 (5): 415-442  Sandvoss, M., McClymont, B., Farnden, C. 2005. A user’s guide to the vegetation resources inventory. Timberline Forest Inventory Consultants Ltd.  Schmitz, R.F, and Gibson, K.E. 1996. Douglas-fir beetle. Forest Insect & Disease Leaflet 5. United States Department of Agriculture, Forest Service, Washington, District of Columbia, United States of America.  Schoennagel, T., Veblen, T.T., and Romme, W.H. 2004. The interaction of fire, fuels, climate across Rocky mountain forests. Bioscience 54 (7): 661-676  Schoennagel, T., Smithwick, E.A.H., and Turner, M.G. 2008. Landscape heterogeneity following large fires: insights from Yellowstone National Park, USA. International Journal of Wildland Fire 17 (6): 742-753  Scott, J.H., and Reinhardt, E.D. 2007, compilers. FireWords Version 1.0: Fire Science Glossary A website providing the definition for crown base height and spotting. Available from: http://www.fs.fed.us/fmi [online] [Accessed June 7th 2015]  Seidl, R., Spies, T.A., Peterson, D.L., Stephens. S.L., and Hicke, J.A. 2016a. Searching for resilience: addressing the impacts of changing disturbance regimes on forest ecosystem services. Journal of Applied Ecology 53 (1): 120-129  Seidl, R., Müller, J., Hothorn, T., Bässler, C., Heurich, M., and Kautz, M. 2016b. Small beetle, large-scale drivers: how regional and landscape factors affect outbreaks of the European spruce bark beetle. Journal of Applied Ecology 53 (2): 530-540  Seidl, R., Donato, D.C., Raffa, K.F., and Turner, M.G. 2016c. Spatial variability in tree regeneration after wildfire delays and dampens future bark beetle outbreaks. Proceedings of the National Academy of Sciences of the United States of America 113 (46): 13075-13080  Simard, M., Romme, W.H., Griffin, J.M., and Turner, M.G. 2011. Do mountain pine beetle outbreaks change the probability of active crown fire in lodgepole pine forests? Ecological Monographs 81 (1): 3-24   112 Sollins, P. 1982. Input and decay of coarse woody debris in coniferous stands in western Oregon and Washington. Canadian Journal of Forest Research 12 (1): 18-28  Spies, T.A., Franklin, J.F., and Thomas, T.B. 1988. Coarse woody debris in Douglas-fir forests of western Oregon and Washington. Ecology 69 (6): 1689-1702  Steed, B., Sturdevant, N., and Gibson, K. 2007. Establishment Report: Permanent plots to monitor Douglas-fir mortality due to western spruce budworm and Douglas-fir beetle interactions. USDA Forest Service, Forest Health Protection, Northern Region, Missoula, MT Numbered report 07-09  Stephens, S.L. 1998. Evaluation of the effects of silvicultural and fuels treatment on potential fire behaviour in Sierra-Nevada mixed-conifer forests. Forest Ecology and Management 105 (1-3): 21-35  Stephens, S.L., and Moghaddas J.J. 2005. Experimental fuel treatment impacts on forest structure, potential fire behavior, and predicted tree mortality in a California mixed conifer forest. Forest Ecology and Management 215 (1-3): 21-36  Stephens, S.L., Fry, D.L., and Franco-Vizcaino, E. 2008. Wildfire and spatial patterns in forests in Northwestern Mexico: the United States wishes it had similar fire problems. Ecology and Society 13 (2):10  Stocks, B.J. 1987. Fire potential in the spruce budworm-damaged forests of Ontario. The Forestry Chronicle 63 (1): 8-14  Stokes, M.A., and Smiley, T.L. 1968 An introduction to tree-ring dating. Chicago: University of Chicago Press: pp. 73  Sturtevant, B.R., Miranda, B.R., Shinneman, D.J., Gustafson, E.J., Wolter, P.T. 2012. Comparing modern and presettlement forest dynamics of a subboreal wilderness: Does spruce budworm enhance fire risk? Ecological Applications 22 (4): 1278-1296  Sullivan, T.P., and Sullivan D.S. 2001. Influence of variable retention harvests on forest ecosystems. II. Diversity and population dynamics of small mammals. Journal of Applied Ecology 38 (6): 1234-1252  Sullivan, T.P., and Sullivan, D.S. 2012. Woody debris, voles, and trees: Influence of habitat structures (piles and windrows) on long-tailed vole populations and feeding damage. Forest Ecology and Management 263: 189-198  Sutherland, I.J., Bennett, E.M., and Gergel, S.E. 2016. Recovery trends for multiple ecosystem services reveal non-linear responses and long-term tradeoffs from temperate forest harvesting. Forest Ecology and Management 374: 61-70  Swetnam, T.W., Thompson, M.A., Sutherland, E.K. 1985. Using dendrochronology to measure radial growth of defoliated trees. United States Department of Agriculture Agricultural Handbook No. 639.  Swetnam, T.W., and Lynch, A.M. 1993. Multicentury, regional-scale pattern of western spruce budworm outbreaks. Ecological Monographs 63 (4): 399-424  Taylor, S.W., and Carroll, A.L. 2004. Disturbance, forest age, and mountain pine beetle outbreak dynamics in BC: A historical perspective. In: Mountain Pine Beetle Symposium: Challenges and Solutions. October 30-31, 2003, Kelowna, British Columbia. T.L. Shore, J.E. Brooks, and J.E. Stone (editors). Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, Information Report BC-X-399, Victoria BC. pp 41-51  Taylor S.W., Carroll A.L., Alfaro R.I., and Safranyik L. 2006. Chapter 2: Forest, climate and mountain pine beetle outbreak dynamics in western Canada In: The mountain pine beetle: A synthesis of biology,  113 management, and impacts on lodgepole pine edited by: Les Safranyik and Bill Wilson Natural Resources Canada, Canadian Forest Service, Pacific Forestry centre Victoria BC: 67-94   Temperli, C., Hart, S.J., Veblen, T.T., Kulakowski, D., Hicks, H.J., and Andrus, R. 2014. Are density reduction treatments effective at managing for resistance or resilience to spruce beetle disturbance in the southern Rocky Mountains? Forest Ecology and Management 334: 53-63  Thom, D., and Seidl, R. 2016. Natural disturbance impacts on ecosystem services and biodiversity in temperate and boreal forests. Biological Reviews 91 (3): 760-781  Thompson, I., D., Baker, J., A., and Ter-Mikaelian, M. 2003. A review of the long-term effects of post-harvest silviculture on vertebrate wildlife, and predictive models, with an emphasis on boreal forests in Ontario, Canada. Forest Ecology and Management 177 (1-3): 441-469  Tscharntke, T., Tylianakis, J.M., Rand, T.A., Didham, R.K., Fahrig, L., Batáry P., Bengtsson, J., Clough, Y., Crist, T.O., Dormann, C.F., Ewers, R.M., Fründ, J., Holt, R.D., Holzschuh, A., Klein, A.M., Lindenmayer, D., Scherber, C., Sodhi, N., Steffan-Dewenter, I., Thies, C., van der Putten, W.H., and Westphal, C. 2012. Landscape moderation of biodiversity patterns and processes- eight hypotheses. Biological Reviews 87 (3): 661-685  Turner, M.G. 1989. Landscape ecology: the effect of pattern on process. Annual Review of Ecology and Systematics 20: 171-197  Turner, M.G., Gardner, R.H., Dale, V.H., and O’Neill, R.V. 1989. Predicting the spread of disturbance across heterogeneous landscapes. Oikos 55 (1): 121-129  Turner, M.G., and Romme, W.H. 1994. Landscape dynamics in crown fire ecosystems. Landscape Ecology 9 (1): 59-77  Turner, M.G., Baker, W.L., Peterson, C.J., Peet, R.K. 1998. Factors influencing succession: Lessons from large, infrequent natural disturbances. Ecosystems 1 (6): 511-523  Turner, M.G. 2010. Disturbance and landscape dynamics in a changing world. Ecology 91 (10): 2833-2849  Turner, M.G., Donato, D.C., and Romme, W.H. 2013. Consequences of spatial heterogeneity for ecosystem services in changing forest landscapes: priorities for future research. Landscape Ecology 28 (6): 1081-1097  Unger, L.S. 1983. Spruce budworms. Pest Leaflet. Victoria, BC: Environment Canada, Canadian Forestry Service, Pacific Forest Research Centre. Cat. No. Fo 29-6/31-1995E pp 1-4  Vaillant, N., M., Noonan-Wright, E., K., Reiner, A., L., Ewell, C., M., Rau, B., M., Fites-Kaufman, J., and Dailey, S., N. 2015. Fuel accumulation and forest structure change following hazardous fuel reduction treatments throughout California. International Journal of Wildland Fire 24 (3): 361-371  Van Wagner, C.E. 1977. Conditions for the start and spread of crown fire. Canadian Journal of Forest Research 7 (1): 23-34  Volney, W.J.A., and Fleming, R.A. 2007. Spruce budworm (Choristoneura spp.) biotype reactions to forest and climate characteristics. Global Change Biology 13 (8): 1630-1643  Wang, T., Hamann, A., Spittlehouse, D.L., and Murdock, T. 2012. ClimateWNA- High-resolution spatial climate data for western North America. Journal of Applied Meteorology and Climatology 51 (1): 16-29.   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   114 Waterhouse, M.J., Armleder, H.M., Dawson, R.J. 1991. Forage litterfall in Douglas-fir forests in the central interior of British Columbia. Research note 108 B.C. Ministry of Forests, Victoria B.C  Waterhouse, M.J., Armleder H.M., and Dawson R.J. 1994. Winter food habits of mule deer in the central interior of British Columbia. Research note 113, B.C Ministry of Forests, Victoria B.C.   Weed, A.S., Ayres, M.P., and Hicke, J.A. 2013. Consequences of climate change for biotic disturbances in North American forests. Ecological Monographs 83 (4): 441-470  Westerling, A.L., Gershunov, A., Brown, T.J., Cayan, D.R., and Dettinger, M.D. 2003. Climate and wildfire in the western United States. Bulletin of the American Meteorological Society 84 (5): 595-604   Westerling, A.L., Hidalgo, H.G., Cayan, D.R., and Swetnam, T.W. 2006. Warming and earlier spring increase western U.S. forest wildfire activity. Science 313 (5789): 940-943  Westerling, A.L., Turner, M.G., Smithwick, E.A.H., Romme, W.H., and Ryan, M.G. 2011. Continued warming could transform Greater Yellowstone fire regimes by mid-21st century. Proceedings of the National Academy of Sciences of the United States of America 108 (32): 13165-13170  White, P.S., Pickett, S.T.A. 1985. Natural disturbance and patch dynamics: an introduction. In: Pickett STA, White PS, editors. The ecology of natural disturbance and patch dynamics. New York: Academic. p 3–13.  Williams, D.W. and Liebhold, A.M. 2002. Climate change and the outbreak ranges of two North American bark beetles. Agricultural and Forest Entomology 4 (2): 87-99  Williams, A.P., Allen, C.D., Macalady, A.K., Griffin, D., Woodhouse, C.A., Meko, D.M., Swetnam, T.W., Rauscher, S.A., Seager, R., Grissino-Mayer, H.D., Dean, J.S., Cook, E.R., Gangodagamage, C., Cai, M., and McDowell N.G. 2013. Temperature as a potent driver of regional forest drought stress and tree mortality. Nature Climate Change 3 (3): 292-297  Wilson, J.S., Isaac, E.S., and Gara, R.I. 1998. Impacts of mountain pine beetle (Dendroctonus ponderosae) (Col. Scolytidae) infestation on future landscape susceptibility to western spruce budworm (Choristoneura occidentalis) (Lep. Tortricidae) in north central Washington. Journal of Applied Entomology 122 (5): 239-245  Wood, S.L. 1982. The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. Great Basin Naturalist Memoirs No. 6. Provo, UT.  Woolford, D.G., Dean, C.B., Martell, D.L., Cao, J., and Wotton, B.M. 2014. Lightning-caused forest fire risk in Northwestern Ontario, Canada, is increasing and associated with anomalies in fire weather. Envrionmetrics 25 (6): 406-416  Wu, T. and Kim, Y-S. 2013. Pricing ecosystem resilience in frequent-fire ponderosa pine forests. Forest Policy and Economics 27: 8-12  Wulf, N.W., Carlson C.E. 1985. Chapter 6: Rating Stand Hazard to Western Spruce Budworm, 6.6 Generalized indexing model. In: Managing trees and stands susceptible to western spruce budworm. Tech. Bull. 1695. Washington, DC: U.S. Department of Agriculture, Forest Service, Cooperative State Research Service pp. 51-54.    115  Appendices  Appendix A: Western spruce budworm susceptibility index  Wulf and Carlson (1985) developed a model that is currently used by the US Forest Service (Blackford 2004) to assess forest susceptibility to western spruce budworm in western Montana, central Idaho (Wulf and Carlson 1985) and in Washington (Wilson et al. 1998). The model incorporates 9 variables that are each converted to a weighted index and when multiplied together provide a susceptibility index, that ranges from 0-100, with corresponding susceptibility rankings. Variables 1-6 were calculated at the plot-level using equations adapted from Wulf and Carlson (1985) and Bousfield et al. (1986); variables 7-9 serve to represent the study area, as follows.    1. Stand structure = coefficient of variation of DBH (CVDBH)  Originally, Wulf and Carlson (1985) used tree height (m); however, I used the DBH of trees ³12.5cm since this was consistently measured for all trees. Additionally, instead of using fixed area plots, I used the N-tree design, and calculated weighted means and weighted standard deviations for plot-specific densities of sub-canopy and canopy trees. The coefficient of variation of DBH (CVDBH) was calculated as:   !"#$% = (()#*+,)((./01*+,) ∗ 100          (Eq. 1)  where, wSDDBH is the weighted standard deviation and wMeanDBH is the weighted mean of tree DBHs. Since trees were sampled using the N-tree method rather than fixed area plots, weighted means were calculated as:  56789#$% = 6789#$%:01 ∗ ;<79=>?@:01 + 6789#$%BCD ∗ ;<79=>?@BCD    (Eq. 2)  where, MeanDBHcan was the mean DBH of the 10 measured canopy trees; rDensitycan was the density of canopy trees (ha-1) relative to the total plot density (sum of canopy and sub-canopy trees, ha-1) expressed as a proportion; MeanDBHsub was the mean DBH of the 10 measured sub-canopy trees; and rDensitysub was the density of sub-canopy trees (ha-1) relative to the total plot density expressed as a proportion. Weighted standard deviations were calculated as:  5E<#$% = (FGH IJK(./01*+, LMNJOM P (QRS IJK(./01*+, LMNJOMTUMT       (Eq. 3)   116 where, xi was the measured DBH of each tree, wsub was the relative density of sub-canopy trees divided by 10; wcan was the relative density of canopy trees divided by 10; and, N=20, the total number of sub-canopy and canopy trees sampled in each plot. The calculated coefficient of variation was then converted to the corresponding index value using the following table:  Coefficient of variation Index value 0-10 0.9 11-20 1.1 21-30 1.3 31-40 1.5 41+ 1.7  2. Total percent crown cover = total percent crown cover of all tree species  Wulf and Carlson (1985) referred to this variable as ‘density’ and used allometric equations to estimate crown cover from tree density (Bousfield et al. 1986). I measured crown cover (%) of trees using hemispheric photographs (see Methods).  Total percent crown cover Index value 1-20 0.8 21-40 1.1 41-60 1.3 61-80 1.4 81-100 1.5 100+ 1.6   3. Percent host crown cover= percentage of crown cover occupied by host  VWX(1	:XZ/W[\F]VWX(1	:XZ/W]\]R^ ∗ 100          (Eq. 4)  Percent host crown cover refers to the relative canopy cover of hosts (%), including Douglas-fir, grand fir, subalpine fir, white fir, Engelmann spruce, and western larch (Wulf and Carlson 1985). Like total percent crown cover, allometric equations were originally used to estimate crown cover from tree densities (Bousfield et al. 1986). I measured crown cover (%) using hemispheric photographs.  Percent host cover Index value 1-10 0.1 11-20 0.3 21-30 0.5 31-40 0.8 41-50 1.0  117 51-60 1.3 61-70 1.6 71-80 1.8 81-90 2.1 91-100 2.4   4. Late-successional host crown cover=   :WX(1	:XZ/W^R]_	FGQQ_FFJ\SR^	[\F]:WX(1	:XZ/W]\]R^ ∗ 100         (Eq. 5)  Wulf and Carlson (1985) referred to this variable as ‘percent climax host cover’. It is the relative canopy cover of late-successional hosts (%), which vary by habitat-type. ‘Major’ (dominant) late-successional host in various habitat-types include: Douglas-fir, grand fir, subalpine fir, white fir, and Engelmann spruce. Originally, allometric equations were used to estimate crown cover from tree densities (Bousfield et al. 1986) and calculate cover of late-successional hosts relative to all trees (%). I measured crown cover (%) using hemispheric photographs.  Percent late-successional host cover Index value 0-10 0.6 11-20 1.0 21-30 1.3 31-40 1.6 41-50 1.8 51-60 2.0 61-70 2.1 71-80 2.2 81-90 2.3 91-100 2.4   5. Maturity= mean basal-area-weighted age of host trees  Maturity was calculated as the mean of the basal-area-weighted age of the host trees in each plot. The basal-area-weighted age of individual trees was calculated using the following equation (adapted from Wulf and Carlson 1985):  `8=8a − 8;78 − 57>cℎ?7e	8c7fW// = 8c7fW// ∗ $g]h__∗i_SFJ]j	QRS	Q^RFFSQRS	Q^RFFkXf0l	$g ∗ 4.356      (Eq. 6)  where, agetree is the age in years of each cored tree; BAtree is the basal area (m2) of each tree; densitycan class is the plot-level density of the canopy class (sub-canopy or canopy) of each tree; ncan class is the number of trees in that canopy class that were aged in the plot;  118 the ratio of densitycan class/ncan class is the correction factor for the N-tree design to scale basal area to a hectare; Total BA for each plot is the sum of the mean basal area of each canopy class (canopy or sub-canopy) multiplied by the density of that canopy class; and, 4.356 is the conversion from metric to imperial units to match Wulf and Carlson’s (1985) scaled index.  For the 200 trees that were harvested in 2014 as part of the silvicultural treatment, and therefore could not be cored in 2015, age was estimated using a regression describing the relationship between age, DBH, and crown class from the 337 trees in the control area for which age was determined from tree rings (R2= 0.308, p-value<0.0001 and the standard error of the estimate was 1.3 years) was as follows:  arcst(uc7) = 1.2913 + 0.5723 ∗ arcst<`y + 0.5819 ∗ {;r59{a8== −0.3884 ∗ arcst<`y ∗ {;r59{a8==          (Eq. 7)  where, log10 (Age) is the logarithm of age in years; log10DBH is the logarithm of the DBH in cm; crownclass is a categorical variable distinguishing between canopy and sub-canopy trees; and, log10DBH*crownclass is the interaction between log10DBH and crownclass. Each predicted age was back-transformed to estimate tree age in years.   Mean basal area weighted host age Index value 1-25 0.3 26-50 1.0 51-75 1.1 76-150 1.2 150+ 1.3   6. Vigour= relative basal area  Wulf and Carlson (1985) referred to this variable as ‘relative stand density’ although it was derived from basal areas. Originally, the average maximum basal area was determined from stocking level assessments of forest inventory and was specific to sub-region and habitat-type group (Wulf and Carlson 1985). I calculated the maximum basal area from the control plots, which are considered fully stocked (Pers. Comm. K. Day, Manager, Alex Fraser Research Forest). Relative basal area was calculated as:   |7a8?>}7	~8=8a	8;78 = kXf0l	$ggZ/W0/	Ä0IÅÄCÄ	$g ∗ 100      (Eq. 8)  where, Total BA (m2) was calculated from the weighted basal area of the sub-canopy and canopy trees as described for ‘maturity’; average maximum BA (m2) was the mean plot-level basal area from the upper quartile of the control 2015 plots. The index values below  119 are for stands with low or no biotic stress; index values for stands stressed by biological agents were not relevant in my study area and are not provided.   Relative basal area Index value 1-40 1.5 41-60 1.3 61-80 1.1 81+ 0.9   7. Site climate describes the habitat-type and potential late-successional plant community for each plot.  Habitat type Index value Cold subalpine fir, timberline  0 Cool moist spruce; cool moist subalpine fir  0.6 Warm, wet grand fir; western red cedar, western hemlock; warm, wet subalpine fir  1.0 Cold Douglas-fir; cold grand fir; cool, dry spruce; cool dry subalpine fir  1.2 Moist grand fir; warm, moist spruce; warm, moist subalpine fir  1.3 Mesic Douglas-fir; dry grand fir; warm, mesic spruce; warm, dry subalpine fir  1.4 Warm, dry Douglas-fir 1.5   8. Regional climate describes the degree of maritime influence on the climate based on location.  Wulf and Carlson (1985) classified sites in the interior of British Columbia as an index value of 1.0 for relatively mesic climate with a maritime influence or as an index value of 1.1 for relatively dry climates with a continental influence.  Region (locations in the USA)  Index value  120 St. Joe, Clearwater, westside Lolo, Selway district in Nezperce, Colville  1.0 Flathead, Nezperce (without Selway district), Wallowa-Whitman, Umatilla, Malheur, Ochoco, Okanogan, Wenatchee, Boise, Payette  1.1   9. Surrounding host continuity describes the forests surrounding the study area.   Percent host type in surrounding 400 ha Character of nearby stands Index value 0-15 Nonhost 0.6 0-15 Host 0.8 16-25 Nonhost 1.0 16-25 Host 1.2 26-75 Nonhost 1.4 26-75 Host 1.5 76-100 Nonhost 1.6 76-100 Host 1.7    10. Overall susceptibility to western spruce budworm= relative habitat suitability  The overall susceptibility index values calculated by multiplying the index values for the nine variables described above. The susceptibility index ranges from 0 to 100, representing low to high habitat suitability for western spruce budworm.   11. Susceptibility ranking is based on the susceptibility of each plot to western spruce budworm.  Susceptibility index Susceptibility ranking 0-20 Low 21-50 Moderate 51-100 High      121 Appendix B: Douglas-fir beetle susceptibility  The probability of infestation model and potential mortality regression were developed in the Colorado Front Range (Negrón 1998). The probability of infestation model was developed using Classification and Regression Trees (CARTs). Potential predictor variables were those known to influence stand susceptibility to Douglas-fire beetle (Furniss et al. 1979, 1981). Three potential models were presented and cross-validated (Negrón 1998).  I selected the probability of infestation model that used percent basal area occupied by Douglas-fir and density of trees.  Negrón’s (1998) original model used the density of trees with DBH³12.7 cm (5 inches); however, I used a DBH threshold of 12.5 cm corresponding to merchantable trees in my study area. The modified infestation model was used to assess the probability of infestation (%) for each plot which were classified as low (29%), moderate (41%), high (71%), and very high (75%) (after Negrón 1998):               Negrón’s (1998) potential mortality regression model estimates the basal area (m2/ha) killed by the Douglas-fir beetle if an infestation were to occur. Using data from 200 plots sampled following an infestation in the Front Range of Colorado, a linear regression model was derived using step-wise selection with 7 potential predictors (Negrón 1998). The model with initial basal area occupied by Douglas-fir best predicted Douglas-fir mortality based on AIC and a prediction R2 (Negrón 1998) and was as follows:   Çr?79?>8a	Ér;?8a>?@#XCl0BKÑÅW = 1.36 + 0.49 ∗ `8=8a	8;78#XCl0BKÑÅW     (Eq. 9)  122 where, potential mortalityDouglas-fir is the basal area (m2/ha) estimated to be affected by Douglas-fir beetle at the plot-level; and, Basal areaDouglas-fir is the basal area (m2/ha) of Douglas-fir at each plot.     123  Appendix C: Crown fire likelihood  Crown Fire Initiation and Spread (CFIS) is software that serves to simulate crown fire behavior and can be used to help make fire management decisions by analyzing the effects of fuel treatments on crown fire behaviour (Cruz and Alexander 2006). The CFIS contains 4 models that have been tested against experimental and wildfire observations (Cruz and Alexander 2006). The likelihood of crown fire occurrence (Cruz et al. 2004) is a logistic model included in the software. This particular model incorporated the fuel strata gap, estimated fine fuel moisture, 10m wind speed, and estimated surface fuel consumption to predict the onset of crowning (Cruz and Alexander 2006). The logistic model applies to conifer forests and level terrain as slope angle is not included (Cruz et al. 2004). The likelihood of crown fire occurrence can therefore be applied to Knife Creek as it is dominated by Douglas-fir and the terrain is subdued.  The Crown Fire Initiation and Spread (CFIS) software determines the likelihood of crown fire occurrence using logistic regression (Alexander et al. 2006, Cruz et al. 2004).  Described as a probability, the likelihood of crown fire occurrence is calculated as (Cruz et al. 2004):  Ç = /Ö(Ü)sP/Ö(Ü)         (Eq. 10)  where, P is the likelihood of crown fire occurrence; and, g(x) is the corresponding logit equation (Cruz et al. 2004):  c á = 4.236 − 0.710 ∗ àEâI − 0.331 ∗ äàà6I  + 0.357 ∗ ãåç< + Eà!I         (Eq. 11)  where, FSGx is the fuel strata gap (m) measured at the plot level; EFFMx is the estimated fine fuel moisture for each plot representing moisture in fine, dead surface fuels. It is calculated based on topography and degree of shading, measured at each plot, and fire weather variable (air temperature, humidity, time of year and day) for the period of study.  SFCx is a categorical variable representing surface fuel consumption at each plot. Three broad classes were defined using 2 dummy variables: SFC<1.0kg/m2 [D1=1, D2=0]; 1.0< SFC <2.0 kg/m2 [D1=0, D2=1]; SFC> 2.0 kg/m2 [D1=0, D2=0], each with coefficients of 0, -4.613, and -1.856 corresponding to the three broad classes respectively (Cruz et al. 2004). A categorical variable with 3 broad classes was incorporated to address the difficulty in estimating fuel consumed in flaming combustion (Cruz et al. 2004).         124 Appendix D: Basal area increment  I used ring widths of 696 Douglas-fir increment cores to derive basal area increment (BAI) to assess forest recovery following the silvicultural treatment done in 1984. For each core, I summed the ring widths (mm) from the pith to the outer ring. Starting from the outer-most ring (2014 on living trees) and moving towards the pith, basal area (mm2) in each year was calculated as:  `8=8a	8;78	é/0W	f = è(;8e>ê=é/0W	f)ë          (Eq. 12)  where, radiusyear t is the sum of the ring widths from pith to the outer ring formed in year t. Annual basal area increment (BAI, mm2) were calculated as the difference in basal area between consecutive years, starting from the outer-most ring:  `8=8a	8;78	>9{;7É79?	é/0W	f = 	`8=8a	8;78é/0W	f − `8=8a	8;78é/0W	fKs ∗ =í               (Eq. 13)  where, sf is the scaling factor that adjusts the BAI values to acknowledge that rings do not always form concentric circles around the pith (Norton et al. 1987). It was calculated as:  =í = (#Å0Ä/f/W	0f	:XWÅ1	ì/ÅìfK	ë(D0Wî	fìÅ:î1/BB)/ë	)CÄ	XÑ	WÅ1	(ÅñfìB	ÑWXÄ	óÅfì	fX	XCf/W	WÅ1                     (Eq. 14)  where, diameter at coring height (cm) was measured on each tree; bark thickness was 4.5± 1.9 cm for 33 Douglas-fir trees in the study area (M-A. Leclerc, unpublished data). The numerator in the equation represents the mean radius inside the bark of the tree. The sum of ring widths from the pith to the outer ring represents the measured radius inside the bark. For trees with asymmetrical growth the mean radius can be longer (shorter) than the measured radius, yielding scaling factor greater (less) than 1.0.       

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0354571/manifest

Comment

Related Items