RINGBARKING TO MANAGE ARMILLARIA ROOT DISEASE IN THE FORESTS OF CENTRAL BRITISH COLUMBIA by Daniel R. Sklar B.Sc., University of British Columbia, 2015 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN NATURAL RESOURCES AND ENVIRONMENTAL STUDIES (BIOLOGY) UNIVERSITY OF NORTHERN BRITISH COLUMBIA December 202 1 ©Daniel R. Sklar, 2021 ABSTRACT Armillaria root disease, caused by the fungus Armillaria ostoyae (Romagn.) Herink, is an important agent of diversity in the forests of British Columbia. Forest operations can disrupt co-evolved host-pathogen balances, leaving behind a supply of carbohydrate-rich defenceless stumps, allowing for a post-harvest inoculum flush and disease spread. One management option is pre-harvest ringbarking, requiring the removal of a strip of bark, phloem, and cambium around a tree, in an attempt to limit carbohydrate transportation to the roots. In theory, ringbarking should deplete starch levels in roots prior to harvest, restricting A. ostoyae' s energy base, directly limiting the pathogen, and indirectly increasing saprophytic competition and exclusion. I set out to determine if ringbarking Armillaria root disease centres in the forests of central BC prior to harvest would have a short-term influence on A. ostoyae ’s characteristic post-disturbance inoculum flush. My objectives were to quantify A. ostoyae colonization and starch content within the roots of ringbarked and untreated Douglas-fir (Pseudotsuga menziesii var. glauca (Mirb.) Franco (Beissn.) Franco). Thirty-six plots were established across three sites, with ringbarking randomly applied to eighteen plots. Sites were clearcut and sample stumps excavated. Armillaria ostoyae colonization was quantified by scoring disease severity of roots and by measuring percent colonized root lengths. Root samples were quantitatively analyzed for starch content. Logistic mixed models were used to assess disease incidence data, and linear mixed models to assess categorical measures of colonization severity, as well as intensive measures of colonization and starch content. No statistically significant differences were observed in post-harvest A. ostoyae colonization or starch content on roots from ringbarked trees versus untreated plots at any site or overall. My study was unable to confinn that ringbarking trees in Armillaria root disease centres prior to their felling influences starch content or ii colonization within root systems following harvest. My results combined with reviewed literature instead suggest that ringbarking may provide control through a host stress response mechanism. Additional long-term research is required to clarify the mechanisms of control and the effectiveness of treatment. Future work should address challenges surrounding initial disease assessment, treatment methods and timing, sampling, and the quantification of colonization and starch. iii TABLE OF CONTENTS Abstract ii List of Tables vi List of Figures vii Acknowledgement viii Introduction 1 Literature Review Armillaria ostoyae as a pathogen Affected hosts Environmental factors Abiotic conditions Rhizospheric conditions Climate change Ecological roles Natural disturbance interactions Abiotic agents Biotic agents Management in British Columbia Distribution Management challenges Management options Ringbarking as a treatment 5 5 7 9 9 10 11 11 12 12 13 14 14 14 16 17 Methods Site selection Gavin Lake trial Horsefly Lake trial Monte Lake trial Plot establishment Pre-harvest measurements Identification of Armillaria ostoyae Initial disease assessment Ringbarking treatment Root colonization assessment Starch assessment Statistical analyses 20 20 23 25 27 29 31 31 33 34 35 39 41 Results Root colonization Starch content 43 43 44 iv Discussion The influence of ringbarking on Armillaria root disease The influence of ringbarking on starch in roots Research limitations Initial disease assessment Ringbarking treatment Sampling for colonization Timing the trial Starch content 49 49 55 57 57 58 59 59 60 Conclusion and Management Recommendations 62 References 64 v LIST OF TABLES Table 1 - Site conditions in plots at the Gavin Lake, Horsefly Lake, and Monte Lake trials, showing means with standard deviations in parentheses (* different letters indicate statistically significant differences using one-way ANOVA with a Bonferroni comparison, a 22 = 0.05, letters are only shown for models with statistical significance) Table 2 - Estimates from logistic and linear mixed models for the fixed-effect ringbarking treatment, showing response variables including Armillaria ostoyae incidence (logistic) and severity (linear) for (a) analyses of categorical colonization and (b) intensive colonization. Response of (c) starch content (linear) is also provided. Estimates include effect coefficients (Coef.), standard errors (SE), Z-statistics (Z), and /2-values (p). All a = 0.05. 45 LIST OF FIGURES Figure 1 - Three trial sites in southcentral British Columbia, at Gavin Lake, Horsefly Lake, 21 and Monte Lake Figure 2 - Gavin Lake trial map showing the research area with the size and location of ringbarked and untreated plots and the locations of root disease centres 24 Figure 3 - Horsefly Lake trial map showing the research area with the size and location of ringbarked and untreated plots and the locations of Armillaria root disease centres 26 Figure 4 - Monte Lake trial map showing the research area with the size and location of ringbarked and untreated plots and the locations of Armillaria root disease centres 28 Figure 5 - Conceptual diagram of a disease centre with plots setup along a delineated active disease edge (left) and of an example plot and buffer setup showing one sample tree (right). 30 Figure 6 - Timeline of my 24-month trial indicating ringbarking to harvest, harvest to excavation, and excavation to sampling periods at Gavin Lake, Horsefly Lake, and Monte Lake. Showing time periods in between these events at each site 36 Figure 7 - Ringbarking treatment applied using (top left) a chainsaw, (top right) a chainsaw with Log Wizard adapter, and (bottom left) a bowsaw. Example of (bottom right) outer bark and phloem tissues removed during treatment also shown 37 Figure 8 - Examples of intensive examination of root colonization during spring 2018 sampling, showing numbering of primary roots prior to measuring them for healthy and infected lengths to average into stump-level colonization 38 Figure 9 - Spring 2018 categorical assessment of ringbarking showing (left) mean Armillaria ostoyae incidence for all sites, for Gavin Lake, Horsefly Lake, and Monte Lake, and (right) mean Armillaria ostoyae (1-6) severity for all sites, for Gavin Lake, Horsefly Lake, and Monte Lake. Standard deviation bars shown 46 Figure 10 - Spring 2018 intensive assessment of ringbarking showing (left) mean Armillaria ostoyae incidence for all sites, for Gavin Lake, Horsefly Lake, and Monte Lake, and (right) mean Armillaria ostoyae (%) colonization for all sites, for Gavin Lake, Horsefly Lake, and 47 Monte Lake. Standard deviation bars shown Figure 11 - Mean starch content by dry weight within sampled healthy Douglas-fir roots from plots in the untreated control versus the ringbarked treatment for the overall trial, Gavin Lake, Horsefly Lake, and Monte Lake. Standard deviation bars shown 48 vii ACKNOWLEDGEMENT I would like to thank Dr. Kathy Lewis for her expert guidance during every stage of my graduate studies at UNBC. Thanks to all my committee members for dedicating their time and contributing their expertise to this project. Special thanks to David Rusch for proposing the project and for fueling my interest in forest pathology, and to Dr. Che Elkin for his expert assistance with statistics and sampling design. Major funding was provided by the British Columbia Ministry of Forests, Lands, Natural Resource Operations and Rural Development. It has been a pleasure collaborating with you all. I would like to thank Dr. Nicole Sukdeo for her unwavering dedication to helping me with my starch assay and with all my lab work. Throughout this project I have relied on many research assistants and volunteers who tirelessly helped me in the field and in the lab. I would like to thank Allie Golt, Liz Wass, Keaton Freel, Ian Higgins, John-Clare Laxton, and Aaron Zwiebel for their assistance, among others I am sure to have missed. I could not have done it without all of you or without the support of my family and my friends. I would like to thank UNBC’s Enhanced Forestry Lab and Tree Ring Lab as well as UBC’s Alex Fraser Research Forest team and the Gavin Lake Forest Education Centre for their support and assistance, and for making me and my team feel at home. Thanks to West Fraser Mills Ltd. and Gilbert Smith Forest Products Ltd. for their cooperation at Horsefly Lake and Monte Lake. viii INTRODUCTION Root-inhabiting pathogens contribute to diverse and healthy forests by influencing host-tree growth and mortality (van der Kamp 1991, Lewis and Lindgren 2000). This increase in diversity often involves a shift towards less susceptible species, and a reduction in stand volume. Collectively these effects can be significant and can negatively impact economic values in commercial forests. Disruptions to co-evolved host defences, whether natural or human-caused, can provide opportunistic conditions that allow parasitic fungi to thrive (Morrison and Mallett 1996, Lewis and Lindgren 2000). Forest management practices can significantly alter the host-pathogen relationship between root pathogens and commercial tree species, and root disease management practices must account for these effects if maintenance of forest productivity and resilience is a management goal. Armillaria root disease is caused by root-inhabiting fungi in the genus Armillaria. These pathogens are widespread and have broad host ranges (Edmonds et al. 2000), making them difficult to manage. This diverse complex of facultative parasites can persist saprophytically (Campbell 1934, Wargo 1996, Cleary et al. 2008) on energy sources such as root carbohydrates to establish, grow, and spread. Pathogenic Armillaria spp. cause long¬ term changes in forest structure and composition, succession (Lewis and Lindgren 2000), and habitat (Carswell et al. 2021), and can impede reforestation and reduce timber volume and quality (Cleary et al. 2008, Cruickshank et al. 2011, Cruickshank and Filipescu 2012). Host species have evolved chemical and physical defences against infection and colonization by Armillaria spp., and under natural conditions, many hosts are able to tolerate Armillaria root disease by limiting spread of the pathogen. Harvest of infected trees can disrupt co-evolved host defence mechanisms and host-parasite balances, leaving behind defenceless carbohydrate-rich root systems and resulting in an aggressive post-disturbance 1 flush of inoculum and disease spread (Campbell 1934, Leach 1937, 1939, Redfern 1968, Swift 1972, Lewis and Lindgren 2000). In southern British Columbia (BC), and much of western North America, the virulent Armillaria ostoyae (Romagn.) Herink is the species causing Armillaria root disease. Management is challenging due to inconspicuous symptoms and a broad host range. Current options are limited, and while ideal in theory (Edmonds et al. 2000, Lewis and Lindgren 2000, Gonthier and Nicolotti 2013), avoiding harvest of infected stands is not practical where incidence is high across large areas of forest. Inoculum removal can provide effective Armillaria root disease management (Morrison et al. 1988, 2014, Cleary et al. 2013), but post-harvest stump excavation is expensive and can degrade soils. Push-harvesting trees with a large excavator (Chapman et al. 2011, Shaw et al. 2012, Morrison et al. 2014) may be more affordable and cause less soil damage, however the use of either method is limited on steep rocky sites. Planting tolerant or partially resistant tree species is also effective, but options are limited due to A. ostoyae’s host range and the limited number of species both adapted to the site and acceptable for reforestation given mandated stocking standards (Morrison et al. 1992, Chapman and Xiao 2000, Cleary et al. 2008, 2012). Biocontrol tests (inoculation of hosts with saprophytic fungi to increase competition and exclusion) have been attempted (Chapman and Xiao 2000, Chapman et al. 2004), however commercial certification and use of these treatments has not occurred as research is required to verily its effectiveness and economic practicality. Another option for management of Armillaria root disease may be pre-harvest ringbarking, which involves the removal of a strip of bark, phloem, and cambium around each tree (Leach 1939). In theory, this should limit carbohydrate transportation and deplete 2 starch levels in roots prior to harvest, restricting A. ostoyae’s energy base, directly limiting the pathogen, and indirectly increasing saprophytic competition and exclusion (Campbell 1934, Leach 1937, 1939, Chapman and Schellenberg 2015). It is expected that ringbarking should thereby reduce post-harvest inoculum flushing and disease spread, enhancing the survival of regenerating trees. Armillaria ostoyae is an opportunist and requires sufficient energy to overcome host defences. The pathogen attacks when the availability of carbohydrates in host tissues is sufficient to support its energetic needs (Entry et al. 1992). Trees maintain carbohydrate reserves in roots and aboveground tissues (Wiley and Helliker 2012, Mei et al. 2015), mainly in the form of starch (Zimmermann and Brown 1971). After girdling, Miller and Berryman (1986) observed decreased starch levels in tissues below treatment, suggesting that stored carbohydrates were mobilized and used to maintain tree survival, continuing to be depleted without the ability to be replenished. Mei et al. (2015) observed that ringbarking caused carbohydrates stored in roots to be quickly mobilized to maintain root function and aboveground transportation to support continuing photosynthetic efforts in the canopy until complete decline. Studies by Hogberg et al. (2001) and Binkley et al. (2006) suggest roots are able to continue to function for months following ringbarking. Ringbarking to manage Armillaria root disease was developed in Malawi (then Nyasaland) and was used to limit disease in tea plantations following the clearing of native forest (Leach 1937, 1939). Early tests in forestry in Great Britain (Redfern 1968), Zimbabwe (then Rhodesia) (Swift 1970, 1972), and Russia (Sokolov 1964) provided less certain results on the efficacy of the treatment. However, recent research in southern BC indicated 3 approximately 50% less tree mortality in ringbarked versus untreated areas 15 years after harvest (Chapman and Schellenberg 2015). My main goal was to determine if ringbarking Armillaria root disease centres in the forests of central BC prior to harvest would have a short-term influence on A. ostoyae’s characteristic post-disturbance inoculum flush. My objectives were to quantify A ostoyae colonization and starch content within the roots of ringbarked and untreated Douglas-fir (Pseudotsuga menziesii var. glauca (Mirb.) Franco (Beissn.) Franco). My hypotheses were that ringbarking decreased post-harvest A. ostoyae colonization and reduced starch content. The research took place at three ecologically distinct sites in the Cariboo and ThompsonOkanagan regions of central BC. This thesis is comprised of a literature review on the biology and ecology of A. ostoyae, and a review of disease management practices including ringbarking treatment. Research methods and results are described, followed by discussion and recommendations regarding the efficacy of ringbarking to limit spread of A. ostoyae, and some limitations of my study. 4 LITERATURE REVIEW This literature review covers the biology and ecology of Armillaria ostoyae, including its hosts, and environmental and climatic factors that contribute to disease development. It then reviews A. ostoyae' s ecological roles and interactions with abiotic and biotic natural disturbances, followed by management of the root disease in BC including distribution, challenges, and options. The review finishes with an overview of ringbarking as a potential treatment. Armillaria ostoyae as a pathogen Armillaria root disease is caused by fungal pathogens in the genus Armillaria, a complex of diverse species that at one time were all referred to as Armillaria mellea (Gonthier and Nicolotti 2013). Armillaria mellea was later separated into distinct species due to sexual incompatibility and morphological differences. In BC, the causal pathogen of root disease in conifers is the virulent A. ostoyae. Armillaria ostoyae is a facultative parasite, capable of persisting in a saprophytic phase for decades on decaying tissues, waiting for living tissues rich in carbohydrates which allow for a parasitic phase of aggressive growth and spread (Campbell 1934, Wargo 1996, Desprez-Loustau et al. 2006, Cleary et al. 2008). Upon contact with living roots via rhizomorphs or root-to-root contact, A. ostoyae attempts to penetrate the bark and move through to the cambium (Leach 1937, Edmonds et al. 2000, Cleary et al. 2008, 2012, Gonthier and Nicolotti 2013). As it feeds through the roots, up the root collar, and into the bole of the tree, the fungus causes the death of infected tissues, eventually girdling and killing its host, leading once again to the pathogen’s saprophytic phase (Cleary et al. 2012). Armillaria ostoyae in its saprophytic phase causes a white rot of wood tissues and is able to decompose all components of plant cell walls, including cellulose and lignin (Campbell 5 1931, Entry et al. 1993). Utilization of wood as a substrate by the pathogen is enhanced by carbohydrates such as starch and sugars, and relatively higher levels of carbohydrates in the bark of susceptible species under attack enhance A. ostoyae ’s ability to overcome host defences (Entry et al. 1992). Due to its relative susceptibility to host defences, A. ostoyae is an opportunistic pathogen. It can be widespread yet cause minimal mortality, acting as a natural thinning agent, removing stressed trees and increasing landscape-level diversity and vigour, and therefore resistance to the disease (van der Kamp 1991, 1993, Morrison and Mallett 1996, Lewis and Lindgren 2000). The pathogen’s ability to spread and infect new hosts, and subsequently disease incidence and severity, depends on a range of local conditions which influence its behaviour (Shaw and Loopstra 1988, Smith et al. 1994, Dettman and van der Kamp 2001, Holdenrieder et al. 2004). Major factors of influence include abiotic and biotic environmental considerations, such as climatic (temperature, precipitation) and soil conditions (soil type, saturation, nutrient conditions), and rhizospheric interactions with local biota. Virulent Armillaria spp. are thought to be poorly rhizomorphogenic, dispersing primarily via root contact with inoculum (Gonthier and Nicolotti 2013). Armillaria ostoyae in BC has been observed to spread less than 1 m per year (van der Kamp 1993, Peet et al. 1996, Morrison 2011). Root contact may instead be the primary means of spread into mature hosts with extensive roots, while rhizomorphs may be more involved in infection of young trees with limited root systems (Morrison 2011, Gonthier and Nicolotti 2013). While there is evidence of dispersal and new disease centre initiation via spores (Cleary et al. 2008, Cruickshank and Filipescu 2012, Gonthier and Nicolotti 2013), infection by spores is thought 6 to play a minimal role in disease spread relative to direct contact between the pathogen and host roots. Once trees become infected, symptoms can include reduced growth, chlorotic and thinning foliage, production of stress cones, basal resinosis, healed-over lesions, and yellow¬ white root and butt rot depending on the extent of colonization (Edmonds et al. 2000, Morrison et al. 2000, Cleary et al. 2008, Morrison 2011, Gonthier and Nicolotti 2013). Armillaria root disease is difficult to distinguish from other stressors using above ground symptoms alone. Signs confirming diagnosis include white mycelial fans under the bark of roots and the root collar, dark shoe-string-like rhizomorphs in the soil, and pale-brown honey mushrooms around the base of trees (Edmonds et al. 2000, Morrison et al. 2000, Cleary et al. 2008, Morrison 2011, Gonthier and Nicolotti 2013). Affected hosts Armillaria ostoyae is a generalist, able to attack a broad range of host species across forests in various stages of succession. Disease susceptibility is a function of host species, age, size, and vigour, as well as site specific conditions and inoculum levels (Edmonds et al. 2000, Lewis and Lindgren 2000, Morrison 2011, Cleary et al. 2012). In BC, A. ostoyae more aggressively attacks softwood versus hardwood species, though western redcedar {Thuja plicata Donn ex D. Don) and western larch (Larix occidentalis Nutt.) exhibit partial tolerance or resistance relative to highly susceptible Douglas-fir, subalpine fir (Abies lasiocarpa (Hook.) Nutt.), and spruce (Picea spp.), or moderately susceptible western hemlock (Tsuga heterophylla (Raf.) Sarg.) and pine (Pinus spp.) (Morrison et al. 1992, Cleary et al. 2008, British Columbia Ministry of Forests, Lands, Natural Resources and Rural Development 2018). Some hardwood species appear to exhibit a degree of tolerance to the 7 pathogen, and continue to grow despite infection (Schafer 1971, Cruickshank and Jaquish 2014). Effective resistance to A. ostoyae is a carbohydrate-intensive effort for hosts (Entry 1992, Cleary 2007, Cleary et aL 2012). Younger host trees with limited starch reserves are unable to respond quickly to attack, and young roots cannot produce a necrophylactic periderm with phellem capable of blocking the pathogen’s penetration to internal tissues (Robinson and Morrison 2001, Morrison 2011, Cleary et al. 2012). As trees mature and roots extend, thicken, and begin to overlap, contact with A. ostoyae is more likely and the pathogen’s inoculum potential is enhanced (Morrison 2011). However, starch reserves increase as trees mature, and larger mature trees can allocate additional resources and successfully defend themselves against infection despite increased contact with the pathogen. The point of root contact with inoculum is also important, and tap roots, lateral roots, and root collars vary in susceptibility to infection (Robinson and Morrison 2001). Host resistance to A. ostoyae comes in the form of pre-existing defensive resins, toxins, barriers, and enzymes, along with attack-induced defences (Cruickshank and Jaquish 2014). These include chemical and structural responses to attack, such as the production of defensive phenols (Entry 1992, Robinson and Morrison 2001) or the development of protective lignified tissues in the bark of roots (Cleary et al. 2012). Resistance appears to be related to a host’s ability to rapidly blocks, ostoyae’ s attempts at colonization (Cleary 2007). Upon attack, western redcedar and western larch, and other partially resistant species quickly produce a necrophylactic periderm with multiple bands of phellem that serve as structural barriers to protect internal root tissues (Cleary et al. 2012). Douglas-fir and more susceptible species respond more slowly to infection, and produce more penetrable defensive 8 tissues, enhancing the pathogen’s ability to spread from external lesions to internal tissues (Robinson and Morrison 2001, Robinson et al. 2004, Cleary et al. 2012). Entry et al. (1992) found that species more susceptible to A. ostoyae produced lower amounts of phenolics in the bark of roots and had higher concentrations of sugars in these tissues when under attack, providing the pathogen with an enhanced energy base relative to less prone species. Environmental factors Abiotic conditions Climate is an important determinant of disease incidence and severity. In BC, A ostoyae is most common at low elevations, and in sites with moderate temperatures and precipitation levels. It is widespread across the southern interior (aside from the Chilcotin Plateau, David Rusch, personal communications, July 15, 2021) and is found only on drier sites on the coast (Cruickshank et al. 1997, Morrison 2011). The pathogen is susceptible to frost and is less aggressive during winter (Cleary et al. 2012). Disease incidence decreases with increasing latitude and A. ostoyae is most common on low elevation south-facing slopes (David Rusch, personal communications, July 15, 2021). Armillaria ostoyae can tolerate a wide range of soil conditions. Incidence is related to sand content and water retention, with the pathogen being less common in finer soils with higher clay contents (Cruickshank et al. 1997, Mallett and Maynard 1998, Desprez-Loustau et al. 2006). Cruickshank et al. (1997) observed that soils which experience periodic saturation and drying also limit A. ostoyae. Periods of saturation and anoxia reduce the viability of inoculum in roots and inhibit parasitic spread, while periods of drying limit rhizomorph growth in soils and saprophytic spread (Cruickshank et al. 1997). 9 Imbalanced or deficient soil nutrient conditions stress host defence systems and can provide opportunistic conditions for parasites such as A. ostoyae. Incidence of infection of lodgepole pine (Pinus contorta Douglas ex Loudon) and black spruce (Picea mariana (Mill.) Britton, Stems & Poggenb.) was higher in soils with lower ammonium levels in the organic layers (Mallett and Maynard 1998) and increasing soil phosphorus levels (Wiensczyk et al. 1997) respectively. Shields and Hobbs (1979) noted increased colonization on Douglas-fir in Idaho in nitrogen-deficient and acidic soils. One explanation for these observations is that hosts growing in stressful soil conditions may extend their roots to increase resource uptake, which would increase the probability of contact with A. ostoyae (Wiensczyk et al. 1997, Popoola and Fox 2003, Desprez-Loustau et al. 2006). Rhizospheric conditions Armillaria ostoyae competes for resources with other biota in the rhizosphere, and natural biological controls exist that limit or exclude the pathogen. Delong et al. (2002) suggested that plant growth promoting rhizobacteria (PGPR) inhibit A. ostoyae and promote host resistance. Pseudomonas fluorescens Migula and other pseudomonas PGPR have shown antagonism towards various plant pathogens, and Delong et al. (2002) observed a trend of decreasing A. ostoyae growth with increasing PGPR in culture. Stand-level disease resistance may be enhanced with moderate nutrient and mesic soil conditions that promote PGPR (DeLong et al. 2002). Hosts such as paper birch (Betula papyrifera Marsh.) are associated with a rhizosphere more suitable for PGPR populations, including higher phenols and lower sugar and ethanol contents in the bark of roots, along with lower soil pH levels (DeLong et al. 2002). This may explain partial resistance or tolerance to the disease, versus more 10 susceptible softwood species such as Douglas-fir. Planting paper birch and other partially resistant or tolerant trees may promote stand-level resistance while reducing A. ostoyae’ s virulence (Gerlach et al. 1997, DeLong et al. 2002). Climate change Rapidly changing climatic conditions exert pressure on co-evolved host-pathogen equilibria. Conditions may become advantageous for A ostoyae, with increasing winter temperatures allowing for longer growth through the year, and the defence systems of mal adapted hosts may be inhibited by increasingly hot and dry summers (Ayres and Lombardero 2000, Harvell et al. 2002, Dukes et al. 2009, Klopfenstein et al. 2009, Heineman et al. 2010, Sturrock et al. 2011, Sturrock 2012). Climate-induced stress inhibits host growth and reproduction and induces changes in root systems, altering carbohydrate and nutrient conditions, and influencing A. ostoyae’s energy base (Popoola and Fox 2003, Desprez- Loustau et al. 2006, Klopfenstein et al. 2009). Ecological roles Armillaria ostoyae slowly and continuously diversifies the physical and biological landscape, creating forests that are distinct from those free of its influence (van der Kamp 1991), for example, by killing susceptible trees and causing them to be replaced with more resistant species. This is due to co-evolved host-parasite equilibria that exert natural controls over the pathogen, increasing landscape-level disease resistance through time (van der Kamp 1991, Lewis and Lindgren 2000, Lundquist 2000). Further, the pathogen removes overstory species, creating gaps and reducing competition, and promotes understory expansion and the establishment of shade-intolerant trees as it thins out stressed hosts (van der Kamp 1991, Lewis and Lindgren 2000, Cruickshank et al. 2009). Tree mortality caused by A. ostoyae 11 adds nutrients to the soil and causes an increase in sunlight reaching the forest floor, providing resources to more vigorous neighbouring trees. In these ways, A. ostoyae creates structurally and biologically diverse habitats, including for wildlife such as ungulates (Carswell et al. 2021) that prefer mixed-open forests to browse, and for birds and cavity¬ nesting wildlife that rely on dead-standing trees (van der Kamp 1991). Natural disturbance interactions Abiotic agents Interactions among abiotic and biotic agents influence the relationship between pathogens and their hosts. Agents that stress trees can induce chemical changes in host roots, increasing carbohydrate levels which favours A ostoyae (Entry et al. 1992, Popoola and Fox 2003). Clinton et al. (1993) observed that drought-induced stress inhibited the ability of hosts to defend against Armillaria root disease in the dry soils of North Carolina’s oak (Quercus spp.) forests, and Desprez-Loustau et al. (2006) noted stress due to water rationing increased Armillaria root disease related mortality in nursery seedlings. Floods can stress both sides of host-pathogen relationships, creating anaerobic soil conditions that inhibit A. ostoyae (Cruickshank et al. 1 997), or displacing soils and damaging and exposing tree roots to infection (Baughman et al. 2017). Wildfires also interact with A. ostoyae. The pathogen influences fire severity through its impact on forest structure and fuel availability (Parker et al. 2006). Depending on severity, fire can result in damaged and susceptible hosts adjacent to fire-killed infected trees with high inoculum loads. Blodgett and Lundquist (2007) observed short-term post-fire increases in A. ostoyae inoculum in the roots of damaged hosts with inhibited vigour in the ponderosa pine (Pinus ponderosa Lawson & C. Lawson) forests of North Dakota. 12 In the longer-term, depending on fire severity, post-fire conditions may naturally inhibit A. ostoyae, as nutrient-enriched soils and release from competition promote host vigour and resistance. A study on Douglas-fir in Oregon suggested that elevated cation levels in post-bum soils inhibited the pathogen while enhancing antagonism by saprophytes such as Trichoderma fungus (Reaves et al. 1990). Fire suppression can inhibit such natural controls, leaving behind stressed and overly dense forests of susceptible trees (Parker et al. 2006). Biotic agents Armillaria ostoyae can exist in disease complexes with root-dwelling fungi such Coniferiporia weirii (Murrill) L.W. Zhou & Y.C. Dai (formerly Phellinus weirii or P. sidphurascens, causing laminated root disease), Pleterobasidion annosum s.L (causing annosus root disease) (Edmonds et al. 2000, Gonthier and Nicolotti 2013, Goheen and Hansen 1993), or Leptographium wageneri (W.B. Kendr.) M.J. Wingf. (causing black stain root disease) (David Rusch, personal communications, July 15, 2021). Interactions between A. ostoyae and above-ground pathogens can also occur. Kulhavy et al. (1984) found that blister rusts in the western white pine (Pinus monticola Dougl. ex D. Don) forests of Idaho predisposed trees to Armillaria root disease, which subsequently predisposed hosts to mountain pine beetle (Dendroctonus ponderosae Hopkins). They also noted that A. ostoyae can act to limit insect infestation, restricting host phloem development and inhibiting brood survivability. Armillaria ostoyae predisposes trees to attack by some insects by stressing hosts and inhibiting defences. Infection of lodgepole pine in Utah limited defensive resin production and made trees more susceptible to infestation by mountain pine beetle, a primary beetle that is able to attack healthy trees but has to build up population levels on stressed trees (Tkacz 13 and Schmitz 1986). Population levels of secondary beetle species, that normally attack only stressed trees, can be positively affected by root pathogens as well. For example, Armillaria root disease enhances susceptibility of fir (Abies spp.) to Scolytus ventralis LeConte (Goheen and Hansen 1993). Management in British Columbia Distribution Armillaria ostoyae is prevalent throughout southern BC, where it is distributed across all biogeoclimatic (BEC) zones, with especially high incidence in many mature Interior- Cedar-Hemlock (ICH) forests (Cruickshank et al. 2009, Morrison 2011). The pathogen is found throughout the interior as far north as McBride, and along the southern coast of the mainland up to approximately Bella Coola, including across Vancouver Island (Henigman et al. 1999, British Columbia Ministry of Forests, Lands, Natural Resource Operations and Rural Development 2018). Management challenges Armillaria ostoyae presents a significant challenge for commercial forestry. Silviculture operations alter forest structure, composition, and succession, and can disrupt host-pathogen equilibria. Clearcut harvest can enhance the severity of Armillaria root disease when infected trees are cut, leaving behind colonized stumps with carbohydrate-rich roots and no defence response which causes a flush of inoculum and rapid disease spread (Leach 1937, 1939, Lewis and Lindgren 2000). Partial cutting exacerbates this, enhancing disease spread via root contact between colonized felled and retained trees (Morrison et al. 2001). Further, planting after harvest using susceptible species often preferred for commercial 14 reforestation can provide A. ostoyae with closely spaced trees with little resistance to infection (Lewis and Lindgren 2000), increasing mortality in subsequent rotations. Precommercial thinning and other stand management practices also influence disease severity and spread. Thinning out and removing hardwoods to enhance softwood productivity can enhance spread between felled and retained trees, reducing stand-level resistance to the disease (Cruickshank et al. 1997, Baleshta et al. 2005, 2015). Baleshta et al. (2005) observed increased Armillaria root disease-related mortality (from approximately 0.2% to 2.0% over 3 years) due to removal of paper birch from commercial stands. Retaining paper birch and other partially resistant or tolerant species can increase stand productivity and offset host mortality (Simard et al. 2005, Baleshta et al. 2015). Armillaria ostoyae can cause significant growth reductions and tree mortality (up to 1 - 2% or more annually according to Morrison and Pellow 1994) and inhibit reforestation and future harvests (Bloomberg and Morrison 1989, Peet et al. 1996, Chapman et al. 2004, 2011, Cleary et al. 2008, 2012). Morrison’s (201 1) study on managed Douglas-fir in BC noted evidence of A. ostoyae within 5 years of planting, as rhizomorphs growing from harvested stumps began to contact nearby host tissues. Infection escalated at approximately 1 5 years post-harvest, as the increasing overlap of host roots made contact with inoculum more likely. Impacts of Armillaria root disease on timber production can be significant. Cruickshank et al. (2011) observed up to a 27% loss in volume per stem across 8 sites in southern interior BC, depending on host age and duration of infection. Taylor (1986) put volume losses due to A. ostoyae in BC’s forests at over 500,000 m3/year, while provincial losses of up to 3.8 million m3/year have been suggested when considering all root pathogens (Morrison et al. 1992). Timber quality and value can also be impacted as the pathogen 15 influences tree growth, height, stem taper, and overall form, and causes warp and knots in the wood along with undesirable changes in annual ring widths (Cruickshank 2010, Cruickshank and Filipescu 2012). Management options Options for management of A ostoyae are limited. Inoculum and host root removal can be effective but requires intensive effort to remove small pieces of root from the soil (Chapman and Xiao 2000, Chapman et al. 2004, 2011). Post-harvest stump excavation is expensive and can cause soil compaction. While push-falling during harvest (Chapman et al. 2011, Shaw et al. 2012, Morrison et al. 2014) can be less disruptive, either method of inoculum removal is not suitable on steep or rocky terrain or in areas with silts and clays prone to soil compaction (David Rusch, personal communications, July 15, 2021). Planting partially resistant or tolerant tree species can be effective, though is limited by A. ostoyae’s extensive host range. Planting a mix of species such as paper birch, western redcedar, and western larch, though often not preferred commercially or acceptable given mandated stocking standards, can improve stand-level disease resistance (Chapman and Xiao 2000, Cleary et al. 2008, 2012, Morrison et al. 2014). Inoculation of colonized trees or stumps with antagonistic saprophytes to increase competition and exclusion has also been evaluated. Observations of the fungus Hypholoma fasciculare (Huds.) Kumm, in culture and in the field suggest its potential for A ostoyae management in BC (Chapman and Xiao 2000, Chapman et al. 2004), although commercial certification and use of these treatments has not occurred as more research is required to verify effectiveness and economic practicality as a biocontrol. Prescribed fires may also be an option, as Reaves et al. (1990) observed increased incidence of saprophytic Trichoderma, 16 a fungus that is antagonistic to Armillaria spp., in post -bum soils of ponderosa pine forests in Oregon. However, A. ostoyae can escape natural or prescribed fires as it inhabits roots deep in the soil (Parker et al. 2006). Ringbarking as a treatment Pre-harvest ringbarking, or the removal of a strip of bark, phloem, and cambium around a tree, is also a possible management option. Ringbarking is different from girdling as the goal is to remove the outer bark and phloem layers with minimal disturbance to the xylem. In theory, this should limit the transportation of carbohydrates from a tree’s canopy down through its phloem to reduce starch content in roots prior to tree mortality (Leach 1937, 1939). If ringbarking is applied with sufficient time prior to harvest (1 year or more), it may directly limit intensification of A. ostoyae in colonized root systems, and indirectly increase competition and exclusion of the parasite by saprophytic antagonists better suited to these conditions. This should effectively limit post-harvest inoculum flushing and disease spread, enhancing the survivability of regenerating host species (Leach 1939, Chapman and Schellenberg 2015). In theory, ringbarking should effectively limit A. ostoyae" s opportunity as a facultative parasite. The pathogen requires a sufficient source of carbohydrates in host roots to support its efforts in overcoming defences (Entry et al. 1992). Host trees store carbohydrates, primarily as starch (Edmonds et al. 2000), in roots and aboveground tissues that can be called upon in times of need, including during times of stress (Wiley and Helliker 2012, Mei et al. 2015). Ringbarking has been shown to cause the mobilization and depletion of starch below treatment in the trunk and roots for use by the host to support survival, while completely inhibiting host ability to replenish these resources (Miller and Berryman 1986, 17 Mei et al. 2015). Carbohydrates mobilized by hosts to maintain metabolic processes will permanently deplete root starches available to A ostoyae as an energy base, while allowing trees to continue photosynthesis during decline (Hogberg et al. 2001, Binkley et al. 2006). This treatment minimizes impacts to timber quality by allowing trees to survive until harvest. Ringbarking was developed by Leach (1937, 1939) to manage Armillaria root disease (Armillaria heimmii Pegler or Armillaria fuscipes Petch, then referred to as A. mellea) in Malawian (then Nyasaland) tea plantations after observing that the fungus relied on carbohydrates or sugars to establish and compete in culture. Swift (1970, 1972) also observed the pathogen behaving as a facultative parasite in culture, able to persist on sawdust but requiring carbohydrates or sugars to compete with more saprophytic antagonists. Though short-term success was seen using Leach’s method of ringbarking to manage Armillaria root disease in tropical tea gardens, Swift (1970, 1972) noted that recurrence of infection was often observed. Swift’s studies in tropical pine plantations in Zimbabwe, along with Sokolov’s (1964) research in the forests of Russia, suggested that while ringbarking prevented colonization of uninfected stumps, treatment instead enhanced disease severity in infected roots. Redfern (1968) observed increased inoculum in the roots of ringbarked stumps in a temperate hardwood plantation in Great Britain. He proposed that ringbarking might affect management by reducing the pathogen’s energy base through stressing hosts and speeding up colonization, suggesting that this effect may be short-term, reversing as starch in treated roots declines to the point where it becomes a limiting factor, effectively inhibiting A. ostoyae’s growth and reducing inoculum potential by the time of planting (Redfern 1968). To support this, Redfern evaluated root systems from his trial as food bases for A. mellea by 18 incubating colonized segments and observing their ability to produce rhizomorphs. Root systems from his ringbarked treatment appeared to be less viable energy sources for the pathogen to spread from (39% of root systems from ringbarked trees produced rhizomorphs versus 90% of roots systems from untreated trees, Redfern 1968). Chapman and Schellenberg (2015) evaluated the use of ringbarking to manage A. ostoyae across three sites in southern BC. Combining data across three similar Douglas-fir dominant sites with qualitatively high levels of infection, the authors observed a significant effect on the mortality of regenerating hosts 15 years following harvest (0.8% annual mortality in treated areas versus 1.8% in untreated controls). Mortality was observed within small two-metre plots centred on old cut stumps and was calculated as the proportion of live and dead trees that were likely killed by Armillaria root disease. 19 METHODS Site selection Selection of sites was limited to stands within the Cariboo and Thompson-Okanagan regions of southern interior BC that were planned for clearcut harvest during the fall and winter of 2016 - 2017 (Figure 1). I focused on BC’s Interior Cedar-Hemlock, Sub-boreal Spruce, and Interior Douglas-fir BEC zones (Pojar et al. 1987) known to be highly susceptible to Armillaria root disease. Digital maps with overlain forest harvest plans and BEC zones, along with suggestions from industry and government personnel were used (DataBC 2016) to select candidate areas. Aerial imagery was reviewed (Google Earth 20 1 6) within candidate sites to locate areas of apparently disturbed forest within otherwise continuous cover. Potential sites were visited to verify the presence of Armillaria root disease and to determine if the site had identifiable root disease centres as indicated by distinct brushy openings with coniferous mortality and relatively healthy closed-canopy coniferous forests beyond (van der Kamp 1993). The presence of white mycelial fans under the bark of infected trees was used to confirm A. ostoyae as the causal agent. Root disease centres in selected sites were mapped using GPS. Three sites were chosen (Figure 1) that represented a range of site conditions and characteristics, as well as Armillaria root disease incidence and severity (Table 1). 20 Kakwa Provincial Park Grande Cache Vanderhoof Prince George Edso Hinton McBride Wells Jasper Cariboo Mountains Quesnel Valemount Provincrai Park FRASER PLATEAU Wi liams Lake 100 Mile House Guide Giaaw Natural Par* Revelstoke Monte Lilloo^t Ashcroft Kamloo^ Nakusp Vernon Stan Valley Provincial Park Whistler liver Kelowna Garibaldi Pruvmdai Park Peachland Powell River \e Is (j Summerland Penticton Squamish nay Pfir Sechelt Parksville l Albernl Nanaimo Hope Vancouver CANADA Chilliwack Abbotsford Ladysmith Duncan Sidney Victoria EfoOntfe. EmL HERE, Gamin. FAQ NOA.A, V3GS, ERA, NHjQm. Six m Im NllAA, Em, l.StiSx Im Cnada Pm ( 25 50 km UNITED STATES Figure 1 - Three trial sites in southcentral British Columbia, at Gavin Lake, Horsefly Lake, and Monte Lake. 21 Table 1 - Site conditions in plots at the Gavin Lake, Horsefly Lake, and Monte Lake trials, showing means with standard deviations in parentheses (* different letters indicate statistically significant differences using one-way ANOVA with a Bonferroni comparison, a = 0.05, letters are only shown for models with statistical significance). BEC zone Gavin Lake SBSdwl Horsefly Lake IHCwk2 Monte Lake IDFdk2 Forest cover Dominant (co-dominant) species Douglas-fir (trembling aspen, hybrid spruce) Douglas-fir (western redcedar, western hemlock, paper birch) Douglas-fir (ponderosa pine, trembling aspen, paper birch) Soil description Coarse, sandy, shallow Coarse, sandy, shallow Coarse, sandy, deep 34(12) 35(17) 40(15) Live overstory density (stems per ha)* 591A(260) 1086b (239) 486a(203) Live understory density (stems per ha)* 995a(418) 4550b(1768) 500A(304) Dead overstory density (stems per ha)* 501a(161) 355a(164) 180B (85) Dead understory density (stems per ha)* 656a (423) 433A(252) 133B(105) Overstory dead (%)* 47a(12) 25b (11) 28b (11) Understory dead (%)* 40a(19) 9b (4) 23b(22) Overstory deciduous (%)* 8a(6) 19b (11) 5a(7) Understory deciduous (%)* 66a (18) 2b (3) 2b(4) Qualitative Armillaria root disease assessment Heavy, single large discrete centre Heavy-moderate, dispersed Moderate-low, small discrete centre Prism basal area (m2/ha) 22 Gavin Lake trial The Gavin Lake trial was located within the Cariboo Forest District, approximately 70 km northeast of Williams Lake, BC (Figure 1), on a planned 18.4 ha clearcut (Figure 2). LIDAR height class data was used to identify a large open and severe root disease centre of approximately 1 1.9 ha. Mortality of mature Douglas-fir within the disease centre was almost complete, aside from a notable area of relatively healthy conifers within the clearing. Overall tree mortality in plots at Gavin Lake was higher compared with the other two sites (Table 1). The Gavin Lake site was in the Sub-boreal Spruce BEC zone (SBSdwl) (Table 1). Soils were coarse and sandy, with areas of rocky outcrops and shallower soils. Forest cover was dominantly Douglas-fir, with a trembling aspen (.Populus tremuloides Michx.) deciduous component and scattered hybrid spruce (Picea x.). Understory cover within the Armillaria root disease centre was heavier in deciduous regeneration versus the other two sites (Table 1), being brushy and heavily occupied by Rubus and Rosa spp., with mosses dominating the understory in the closed canopy forest outside of the root disease centre. 23 Plot Legend Radius (m) Treated? 10.92 Y 17.34 Y 11.25 N 12.81 N 5 10.68 Y 6 12.42 N 7 8.89 N 8 10.1 Y 9 11.12 N 10 7.54 Y Y 10.18 11 7.32 N 12 1 16-O3Pk>t_TreatmentBuffer 0Landmgs 2 3 4 Roads •Highway •on-nary —Secondary unimproved •Deaavawj Temporary 2m contour interval Dem 177 boundary WTP — SkidTrails | Root Rot (DRusch) ArmiBada pne;nu* Root Rot (Hoitam) SU 1 Qvr-i ana gpfie; nui 18.361 0T&Tertt>su6 SU 2 - sU3a Gap NP Access ^Wetlands —Streams 0.38 Ring-barked Plots Nil-treated Plots Armillaria Ringbarking Trials Gavin Lake Research Site UBC Alex Fraser Research Forest CB 177 Cutblock Area 99.6 ha Ring-barked Area 0.20 ha Nil-treated Area 0.20 ha Cariboo Forest District UTM10N NAD83 Base map created and provided by UBC Alex Fraser Research Forest 1:3,000 100 0 100 Meters Modified by Daniel Sklar, October 2016 Figure 2 - Gavin Lake trial map showing the research area with the size and location of ringbarked and untreated plots and the locations of root disease centres. 24 Horsefly Lake trial The Horsefly Lake trial was also located in the Cariboo Forest District, approximately 120 km northeast of Williams Lake, BC (Figure 1), on a planned 91.9 ha clearcut (Figure 3). The site was characterized by severe Armillaria root disease infection with heavy but dispersed and scattered mortality, in contrast to the single large open Armillaria root disease centre at Gavin Lake. Overall tree mortality in plots at Horsefly Lake was lower versus the other two trial locations (T able 1 ). The Horsefly Lake site was in the Interior Cedar-Hemlock BEC zone (ICHwk2) (Table 1). Soils were shallow, coarse and sandy, with areas of clay and steep exposed bedrock that had to be avoided. Forest cover was dominantly Douglas-fir, with a heavy coniferous component of western redcedar and scattered western hemlock. The percentage of the canopy comprised of deciduous species was larger at this site versus the others (Table 1), predominantly paper birch. Hybrid spruce, lodgepole pine, and trembling aspen were also observed. Overall canopy cover was higher than in the other sites and as a result, understory species were predominantly mosses. 25 X Plot A - c Radius (m) Treated? 8.35 Y 1 13.15 N 2 7.59 Y 3 4 7.95 Y 9.6 N 5 9.91 N 6 17.87 N 7 8 12.22 N 9 10.66 Y 10 13.07 N 11 11.6 Y 10.42 Y 12 Legend Ring-barked Plots Nil-treated Plots Roads CP366 Rasarv** Armtllana Ringbarking Trials Hen Ingram Research Site West Fraser A200 17 CP35504 Canboo Forest District Cutblock Area 91.9 he Research Area 6.0 ha Ring-barked Area 0.17 ha Nil-treated Area 0.31 ha UTM10N NAD83 Base map created and provided by West Fraser, modified by Daniel Sklar 1:2,500 100 Meters Figure 3 - Horsefly Lake trial map showing the research area with the size and location of ringbarked and untreated plots and the locations of Armillaria root disease centres. 26 Monte Lake trial The Monte Lake trial was located within the Kamloops Forest district of the Thompson Okanagan region, approximately 60 km southeast of Kamloops, BC (Figure 1), on a 9-ha planned clearcut (Figure 4). Armillaria root disease incidence and severity on site was moderate relative to the other two sites, and disease centres were small and discrete. Overall tree mortality in plots at Monte Lake was lower than at Gavin Lake but higher than at Horsefly Lake (Table 1). The Monte Lake site was in the Interior Douglas-fir BEC zone (IDFdk2) (Table 1). Soils on site were coarse and sandy, and relatively deep in comparison with the other trial locations. Douglas-fir was the dominant species with some ponderosa pine and a deciduous component of trembling aspen and paper birch. Forest canopy was open and clumpy, promoting an understory of heavy brush. 27 Plot Radius (m) Treated? 10.92 Y 17.34 Y 2 3 11.25 N 12.81 N 4 5 10.68 Y 6 12.42 N 8.89 N 7 10.1 Y 8 9 11.12 N 10 7.54 Y 11 10.18 Y 7.32 N 12 1 Legend • Armillaria Centers Ring-barked Plots O Nil-treated Plots Research Area CP904-13 Cutblock Armillaria Ringbarking Trials Monte Lake Research Site Cutblock Area 9.01 ha Research Area 4.7 ha Gilbert Smith CP904-13 Ring-barked Area 0.27 ha Kamloops Forest District Nil-treated Area 0.20 ha UTM11N NAD83 Base map created and provided by Gilbert Smith, modified by Daniel Sklar 100 Meters 1:2,500 Figure 4 - Monte Lake trial map showing the research area with the size and location of ringbarked and untreated plots and the locations of Armillaria root disease centres. 28 Plot establishment Twelve plots were established along edges of identified Armillaria root disease centres at each site, to target areas where the pathogen was more active in order to maximize uniformity of disease severity and inoculum potential. I attempted to locate plots in areas of similar species composition and tree density. Sites also had to be accessible by an excavator without causing much soil disturbance. Areas of forest cover dominated by Douglas-fir were chosen for this trial due to a high susceptibility to Armillaria root disease (Morrison et al. 1992, Cleary et al. 2008, British Columbia Ministry of Forests, Lands, Natural Resources and Rural Development 2018). Plots were approximately centered in between two or more dead standing Douglas-fir with signs of Armillaria root disease. Plot size was variable in order to include 15 live Douglas-fir larger than 12.5 cm DBH (diameter at breast height, 1.3 m), with each of these sample trees being at most 10 m from a standing live or dead stem exhibiting A. ostoyae fans (Figure 5). A 5 m buffer, treated but not sampled, was added to each plot to reduce contact between the roots of sample trees and surrounding untreated trees. Plot locations were recorded using a GPS. 29 Figure 5 - Conceptual diagram of a disease centre with plots setup along a delineated active disease edge (left) and of an example plot and buffer setup showing one sample tree (right). 30 Pre-harvest measurements Sample trees were tagged at their base, and tree height (using Haglof Vertex IV), DBH, and signs or symptoms of Armillaria root disease were recorded. Tree height and diameter was also recorded for live deciduous trees larger than 12.5 cm DBH within plot areas. Stem counts of live and dead trees in all plots were carried out, grouping trees into layers by height and DBH. Total plot basal areas were measured using wedge prisms, and individual sample tree basal areas were calculated from the DBH measurements. Dominant vegetation (species and incidence), soil (texture and classification), and topography (slope and aspect) were recorded at each plot. Identification of Armillaria ostoyae One infected root section was sampled from each plot and taken to the laboratory. Small mycelial fragments were removed under a laminar-flow hood using flame-sterilized tools and were plated for culturing on malt agar (30 g malt, 15 g agar, 1 L distilled water). Colonies were allowed to incubate in the dark at ambient temperature for several weeks and when developed were subcultured to obtain single isolates. MoBio PowerSoil DNA Isolation Kits (Manufactured as “DNeasy PowerSoil Pro” after acquisition of MoBio by Qiagen in 2016; Carlsbad, CA) were used for DNA extractions. Rhizomorph samples were taken from isolates in culture and were transferred to empty tubes for preliminary processing. Two 2 mm autoclaved zircon beads were added to each tube before agitating rhizomorphs tissues using a Bullet Blender (Model BBX24, Next Advance, Troy, NY, USA) at a setting of 4 - 5 for 5 minutes. The lysis buffer and bead matrix for the kit was then added to the tube for lysis in the Bullet Blender at a setting of 4 31 for 10 minutes. DNA isolation and purification followed the manufacturer’s directions. Purified DNA extracts (eluted in 10 mM Tris buffer, pH 7) were stored at -20 °C. Identification was via DNA PCR analysis using two species specific primer sequences, AoF700 and 18S-rev (McLaughlin and Hsiang 2010). This primer pair amplifies a segment of the 5’ intergenic spacer 2 (IGS2) region directly preceding the 18S gene in Armillaria spp. (McLaughlin and Hsiang 2010). Size differences between amplicons distinguished A. ostoyae from Armillaria calvescens Berube & Dessur., Armillaria mellea (Vahl) P. Kumm., A. sinapina Berube & Dessur., and Armillaria gallica Marxm. & Romagn. in McLaughlin and Hsiang’s protocol (2010). A DNA extract from an A sinapina sporocarp from Dr. Chow Lee’s (UNBC) voucher collection was used to provide a comparison to the cultured specimens. For the PCR assay 50 pl reaction volumes were used, with 1 ng of purified template DNA. The PCR contained 200nM (final concentration) of primers AoF700 and 18S-rev (McLaughlin and Hsiang 2010), containing 1 ul of 10 pM primer working stock of both AoF700 and 18S-rev, along with 0. 1 mg/mL BSA (final concentration) to bind PCR inhibitors and 20 pl of 5 Prime Hot Master Mix (2.5X stock concentration). Negative controls with no template DNA were run as controls to check for DNA contamination of PCR reagents. The thermal cycle profile for IGS2 amplification was as follows: • Step 1: 94 °C for 2 minutes • Step 2: 94 °C for 30 seconds • Step 3: 58 °C for 45 seconds • Step 4: 72 °C for 90 seconds 32 • Step 5: repeat step 2 - 5 34 times • Step 6: 72 °C for 10 minutes • Step 7: final hold at 4 °C PCR products were stored -20 °C for longer terms. Discrimination of A. ostoyae from A. sinapina was by size comparison of amplicons using agarose gel electrophoresis. Amplicons were electrophoresed on a 0.5X TBE, 1% agarose gel with a size standard. Armillaria ostoyae amplicons were expected to be approximately 731 bp in size and A. sinapina amplicons approximately 622 bp. Initial disease assessment Trap-strakes were used in an attempt to assess Armillaria root disease at each site (Mallett and Hiratsuka 1985). Trap stakes were 30 - 40 cm long and cut from freshly cut subalpine fir saplings (5 - 10 cm DBH) then driven with bark intact 20 - 25 cm into the soil on 5 x 5 m grids centered over plots. Setup took place during spring 2016. Where possible, subsets of 3 - 5 stakes were removed pre-harvest during fall 2016 and post-harvest during spring 2017. Stakes were examined in the field and lab and the presence of Armillaria spp. mycelium and rhizomorphs were recorded. Tissue samples were collected and sent to a lab at UBC for culturing and identification via DNA-based species-specific PCR primers (McLaughlin and Hsiang 2010) using the same methods as described in the previous section. While many stakes at Gavin Lake and Horsefly Lake were colonized both pre- and post¬ harvest, no colonization was observed on stakes at Monte Lake. The majority of trapped samples were rhizomorph tissues identified as Armillaria sinapina rather than A. ostoyae, though A. ostoyae was identified from mycelial tissues trapped on at least two stakes at Gavin Lake. 33 Surveys for A. ostoyae fruiting bodies were carried out during the fall of 2016 (the last week of September through the first week of October). Existing sample plots were examined in quadrants. In the case of already harvested areas, transects were run in adjacent undisturbed forested areas. No A. ostoyae mushrooms were observed which may be due to seasonal conditions or the narrow observation period. Ringbarking treatment Ringbarking occurred June 2016 (Figure 6). I treated half of the plots at each site, randomly selected, for a total of 1 8 plots. A total area of approximately 0.2 ha was treated at each site. All live trees larger than 7.5 cm DBH within plots and buffers were ringbarked. At all sites, treatment was applied at 0.3 - 1.3 m height and removed a single strip of bark, phloem, and cambial tissues 10 - 15 cm wide around each tree (Figure 7). Care was taken to minimize damage to xylem tissues, identified by changes in sawdust colour or moisture, in order to restrict carbohydrate transportation to the roots without influencing the movement of water and nutrients upwards to the canopy. Ringbarking at Monte Lake was done using a chainsaw adapted with a Log Wizard (Log Wizard 2016). Due to safety concerns and a lack of control over xylem contact, the other sites were treated using hand tools such as bowsaws and hand chainsaws to make cuts and chisels and hammers to remove bark. All sites were clearcut between fall 2016 and winter 2017 (Figure 6). Limited control over licensee harvest plans resulted in earlier harvest than planned at Monte Lake and Horsefly Lake, which were meant to follow the same timeline as Gavin Lake. Root excavations were delayed by wildfire during the summer of 2017 and finally undertaken in fall 2017. Contractors employed a Hitachi EX60 at the Gavin Lake and Horsefly Lake sites, and a Volvo EC220DL at the Monte Lake site. Both excavators were fitted with hydraulic 34 bucket-thumb setups. I attempted to have a minimum of 10 sample stumps from each plot removed with minimal damage, directing excavator operators to invert each sample to expose the roots. Root colonization assessment Post-harvest A. ostoyae colonization was assessed during fall 2017, but was rushed due to snowfall, therefore observations were repeated in spring 2018 (Figure 6). I used categories of colonization extent by rating each accessible and intact sample stump on a 0 - 6 scale of severity based on number of colonized primary roots observed (0 = no accessible roots colonized to 6 = all accessible roots colonized). Post-harvest colonization was intensively examined during spring 2018 (Figure 8) on up to 2 root systems in each of the 0 6 categories (as determined in spring 2018) per plot. Lengths of healthy and infected root sections for up to 5 accessible and intact primary roots (roots arising directly from the stump) from each sample were measured and averaged into stump-level percent colonization. 35 JUN 2016 NOV 2016 APR. 2016 SEP. 2016 FEB. JUL. 2017 2017 Ringbarking to harvest Harvest to excavation Excavation to sampling Figure 6 - Timeline of my 24-month trial indicating ringbarking to harvest, harvest to excavation, and excavation to sampling periods at Gavin Lake, Horsefly Lake, and Monte Lake. Showing time periods in between these events at each site. 36 Figure 7 - Ringbarking treatment applied using (top left) a chainsaw, (top right) a chainsaw with Log Wizard adapter, and (bottom left) a bowsaw. Example of (bottom right) outer bark and phloem tissues removed during treatment also shown. 37 Figure 8 - Examples of intensive examination of root colonization during spring 2018 sampling, showing numbering of primary roots prior to measuring them for healthy and infected lengths to average into stump-level colonization. 38 Starch assessment Sampling for starch content took place during fall 2017 after harvest and excavations were completed (Figure 6). Healthy and infected roots from trees in treated and untreated plots at each site were assessed. Sections (5-15 cm in length) of accessible primary and secondary roots with diameters between 5 and 10 cm were collected from sample stumps using a bowsaw. A total of 62 roots were assessed for starch content across the trial distributed among sites as follows: • Gavin Lake, one sample from each plot, with an even number of treated and untreated samples • Horsefly Lake, 23 samples from 10 plots, 14 from treated plots and 9 from untreated plots • Monte Lake, 27 samples from all 12 plots, 15 from treated plots and 12 from untreated plots Root sections were cut crosswise to expose a fresh surface and a sapwood sample was removed from each using a drill and a bit that was cleaned between samples with 70% ethanol to avoid cross-contamination. An assay to quantify starch content in roots was developed based on Smith and Zeeman (2006) and Chow and Landhausser (2018). Starches were solubilized using heat and depolymerized into glucose using amyloglucosidase before being assayed using a hexokinase-glucose 6-phosphate dehydrogenase reagent. In preparation, 20 mg drilled sapwood chips were oven-dried in uncapped 2 ml flip-top polypropylene Eppendorf tubs at 60 - 65 °C for 48 hours. Wood chips were powdered after drying by adding two 3.2 mm 39 stainless steels beads to tubes and pulverizing in a Bullet Blender (BBX-24) at a setting of 10 (maximum intensity) for three intervals of one minute. Beads were removed using a magnet. Low molar-weight sugars were extracted by adding 1.5 ml of 80% ethanol to each tube and mixing gently before placing in a 60 °C water bath for three minutes. Samples were then centrifuged at 10,000 G for five minutes and the supernatant decanted, repeating until supernatants were colourless. Pellets were dried overnight at 60 °C and each tube’s weight including pellet was recorded to account for sapwood loss during ethanol extraction. Pellets were resuspended in 1 ml of distilled deionized water, and, using a wide bore pipette tip, 0.6 ml of sapwood homogenate was transferred to a screw-cap polypropylene Eppendorf tube. Three successive transfers of 200 pL were used, vortexing briefly in between, to allow for transfer of approximately equal quantities of homogenate. Samples were then gelatinized by incubating tubes in boiling water for one hour, vortexing every five minutes. Starch depolymerization into glucose used five units of alpha-amyloglucosidase (Sigma A7095, powdered amyloglucosidase from Aspergillus nigerAAzgh. with < 0.02% free glucose at approximately 30 - 60 units/mg) added to 0.6 mL of 200 mM NaCH3COO buffer at a pH of 5.2 - 5.5. A working stock of enzyme in buffer was prepared at a concentration of 200 units/mL enzyme and 25 pL of stock added to each tube. Samples were incubated with 80 RPM shaking for 24 hours at 50 °C. A control without starch was run with every set of samples to account for possible background glucose contamination in the enzyme stock, setup in a 2 ml Eppendorf tube with 0.6 mL distilled deionized water, 0.6 mL of the 200 mM NaCH3COO buffer, and five units of alpha-amyloglucosidase at the same working stock concentration as the non-control samples. All tubes were then centrifuged at 14,000 G for 10 40 minute and supernatants were transferred to clean 1.5 ml Eppendorf tubes before being assayed with the hexokinase and G6PDH kit. My assay used 96-well plates (polystyrene, flat-bottom wells). For each sample, triplicates of 10 pl subsamples were pipetted with 100 pl of assay reagent. Multiple blanks were included for each sample, pipetting 10 pl of subsample with 100 pl of distilled deionized water. Plates were incubated for 30 minutes at 37 °C and read at 340 nm using a microplate reader. Standard curves (0 - 1 mg/ml glucose range, 6-7 different concentrations) for were generated for every 20 - 30 samples assayed, and for each time a new hexokinase-G6PDH working stock was made. Statistical analyses Data were analyzed in STATA 14 (StataCorp 2015). To address zero-inflation, Armillaria ostoyae colonization datasets were split into disease incidence (all non-zero values represented as 1) and severity of infected (all zero values removed) for analysis with mixed models. Models for each dataset compared stumps in ringbarked versus untreated plots at each site and overall (all a = 0.05). Logistic mixed models were used to assess disease incidence, aside from for the Horsefly Lake categorical data, which exhibited too little variation to run the site-level model (all ringbarked samples were infected). Linear mixed models were used to assess fall 2017 and spring 2018 categorical measures of colonization (1-6) severity, as well as spring 2018 intensive measures of (%) colonization. Linear mixed models were used to examine for differences in starch content within healthy roots, aside from for the Gavin Lake data which represented too small of a sample to run the site-level model. 41 All mixed models were fitted using maximum likelihood estimation and assumed a standard normal distribution for the fixed effects. Plot was included as a random effect, as well as site in trial-wide (all site) models. Pre-harvest stand structure metrics were included as covariates in preliminary analyses. This included pre-harvest sample tree DBH, height, and a variable indicating the presence of signs or symptoms of Armillaria root disease. These analyses indicated a lack of statistical significance of these covariates in most models. As such, covariates were excluded from all mixed models reported on in my thesis, and analyses only includes the fixed effect ringbarking treatment, reflecting the hypotheses being tested. A two-tailed paired t-test was used to examine for differences between 2017 and 2018 categorical severity data. Two-way ANOVAs were employed to identify interactions between site and treatment in categorical and intensive colonization data, and in starch data. These tests analyzed overall datasets prior to splitting into incidence and severity for analysis via mixed models. 42 RESULTS Root colonization Armillaria ostoyae was tentatively visually identified as the pathogen causing Armillaria root disease across my trial. DNA PCR analysis of mycelial tissues sampled from a subset of infected root systems within my trial plots confirmed A. ostoyae as the causal pathogen at each site. Armillaria ostoyae colonization severity in fall 2017 based on categories of colonization was not significantly different from spring 2018 (t368 = 1.49, p = .14), therefore these results focus on 2018 data. Averaged data from this assessment indicated a possible trend of lower A ostoyae incidence and severity in stumps in the ringbarked plots at Gavin Lake, and higher incidence in the ringbarked plots at Monte Lake (Figure 9). No statistically significant differences in incidence were observed on roots from treated versus untreated plots at Gavin Lake, Horsefly Lake, Monte Lake, or overall (Table 2a, Figure 9). No statistically significant differences in severity (1-6) were observed on roots from treated versus untreated plots at Gavin Lake, Horsefly Lake, Monte Lake, or overall (Table 2a, Figure 9). An interaction between site and treatment was identified in categorical colonization data (F2,38i = 6.78, p = .001). Averaged data from my intensive assessment of incidence and percent colonization in spring 2018 indicated a possible trend of increased A. ostoyae (%) colonization in stumps in the ringbarked plots at Monte Lake (Figure 10). No statistically significant differences were observed in measures of incidence on roots from treated versus untreated plots at Gavin Lake, Horsefly Lake, Monte Lake, or overall (Table 2b, Figure 10). No statistically significant differences were observed in measures of (%) colonization on roots from treated versus untreated plots at Gavin Lake, Horsefly Lake, Monte Lake, or overall (Table 2b, 43 Figure 10). No interactions between site and treatment were identified in intensive colonization data (F2, 245 = 1 .24, p = .29). Starch content Starch content in healthy root samples ranged from 0 - 0.26% of dry weight across the trial (Figure 1 1). Values were close to or below the detectability of the developed assay. Due to assay detectability limitations, assessment was discontinued, and infected root samples were not assessed for comparison alongside healthy samples. No statistically significant differences in starch content in healthy roots were noted between trees in ringbarked and untreated plots at Gavin Lake, Horsefly Lake, Monte Lake, or across the overall trial (Table 2c, Figure 1 1). No interactions between sites and treatments were identified in these data (F2.56 = 0. 19, p = .83). 44 Table 2 - Estimates from logistic and linear mixed models for the fixed-effect ringbarking treatment, showing response variables including Armillaria ostoyae incidence (logistic) and severity (linear) for (a) analyses of categorical colonization and (b) intensive colonization. Response of (c) starch content (linear) is also provided. Estimates include effect coefficients (Coef.), standard errors (SE), Z-statistics (Z), and /^-values (p). All a = 0.05. Site All Gavin Horsefly Coef. 0.589 -0.524 SE 0.550 0.667 Z 1.07 -0.79 IL .28 .43 - - - - Monte All Gavin Severity (1-6) Horsefly Monte All Gavin Incidence (0-1) Horsefly (b) Monte Intensive A. ostoyae All Colonization Gavin Severity (°/o colonization) Horsefly Monte All (c) Gavin Starch Starch (% dry weight) Horsefly Content Monte 0.826 -0.093 -0.619 0.097 0.307 -0.110 -0.300 0.482 -0.266 6.719 -2.319 8.766 15.52 -0.003 0.847 0.176 0.350 0.243 0.333 0.304 0.492 0.638 0.503 4.792 8.409 9.097 8.339 0.021 0.98 -0.53 -1.77 0.40 0.92 -0.36 -0.61 0.76 -0.53 1.40 -0.28 0.96 1.86 -0.13 .33 .60 .08 .69 .36 .72 .54 .45 .60 .16 .78 .34 .06 .89 - - - - 0.013 -0.011 0.041 0.032 0.32 -0.34 .75 .73 Variable (a) Categorical A. ostoyae Colonization Incidence (0-1) 45 All Sites Control (n= 188) All Sites Ringbark (n= 199) Gavin Lake Control (n = 66) Ringbark (n = 63) Control (n = 49) Ringbark (n= 39) Horsefly Lake Ringbark (n= 65) Monte Lake Control (n = 66) Ringbark(n= 151) Gavin Lake Horsefly Lake Control (n = 56) Control (n= 132) Control (n = 5 1) Ringbark (n= 65) Monte Lake Ringbark (n= 71) Control (n = 32) Ringbaik (n = 47) Figure 9 - Spring 2018 categorical assessment of ringbarking showing (left) mean Armillaria ostoyae incidence for all sites, for Gavin Lake, Horsefly Lake, and Monte Lake, and (right) mean Armillaria ostoyae (1-6) severity for all sites, for Gavin Lake, Horsefly Lake, and Monte Lake. Standard deviation bars shown. 46 All Sites All Sites 100 90 80 70 60 50 40 30 20 10 0 Control (n= 120) Control (n = 94) Ringbark (n= 131) Gavin Lake Ringbaik (n= 100) Gavin Lake 100 90 80 70 60 50 40 30 20 10 0 Control (n = 46) Control (n = 36) Ring bark (n= 44) Ringbark (n= 32) Horsefly Lake Horsefly Lake 100 90 80 70 60 50 40 30 20 10 Control (n = 39) Ringbark (n = 42) Control (n = 32) Ringbaik (n = 37) Monte Lake Monte Lake 100 90 80 70 60 50 40 30 20 10 0 Control (n = 35) Ringbaik (n= 45) Control (n = 26) Ringbark (n= 31) Figure 10 - Spring 2018 intensive assessment of ringbarking showing (left) mean Armillaria ostoyae incidence for all sites, for Gavin Lake, Horsefly Lake, and Monte Lake, and (right) mean Armillaria ostoyae (%) colonization for all sites, for Gavin Lake, Horsefly Lake, and Monte Lake. Standard deviation bars shown. 47 All Sites Gavin Lake 0.22 t 0.2 30.18 0.22 t 0.2 •• >0.18 T £ 0.16 0.16 7.0.14 Q 0.12 ^0.14 5 0.12 £ 0.1 f8 0.08 " * 0.06 •• 0.04 0.02 0 o o •• £ 0.1 •• O 1 1 Control (n = 35) Ringbark (n= 27) J- 0.06 0.04 * 0.02 “ 0 J Control (n = 6) Horsefly Lake 0.22 0.2 3 0.18 0.16 0.14 0.12 Q 0.1 0.08 0.06 E 0.04 0.02 0 -0.02 T i o.o8 Ringbark (n = 6) Monte Lake 0.22 T 0.2 T 30.18 g 0.16 I ^,0.14 Q 0.12 £ 0.1 " o Control (n = 15) Ringbark (n = 12) 0.08 a o.o6 •• 0.04 •• * 0.020 Control (n= 14) Ringbark (n = 9) Figure 11 - Mean starch content by dry weight within sampled healthy Douglas-fir roots from plots in the untreated control versus the ringbarked treatment for the overall trial, Gavin Lake, Horsefly Lake, and Monte Lake. Standard deviation bars shown. 48 DISCUSSION The influence of ringbarking on Armillaria root disease The results of my trial do not support the hypothesis that ringbarking infected forested areas prior to harvest limits the intensification of Armillaria root disease by reducing starch content in root systems before host mortality. Treatment did not have a significant influence on A. ostoyae colonization or starch content within roots of sampled stumps in my study. My data show higher mean A. ostoyae incidence and colonization of treated stumps at Monte Lake which, while not statistically significant, suggests that treatment may provide some control through an alternate mechanism, host stress response. Host stress from injury during ringbarking may enable rapid expansion of the fungus by causing a decline in vigour of the host due to a hindered ability to mobilize and utilize carbohydrates to produce defensive compounds. The rapid expansion and accelerated use of root-based resources by the fungus as a result of reduced host defences would lead to faster decline of inoculum potential in treated trees compared to untreated trees. Successful host resistance to A. ostoyae is a carbohydrate-intensive effort (Entry et al. 1992, Cleary 2007, Cleary et al. 2012). Host defence is likely inhibited by stress from ringbarking, increasing susceptibility to the pathogen as the tree reprioritizes and mobilizes carbohydrates below the treatment to deal with injury, while cutting off any ability for stores to be replenished from above. Studies by Mei et al. (2015) and Miller and Berryman (1986) have demonstrated such a host stress response to girdling and ringbarking, observing the reprioritization, mobilization, and depletion of carbohydrates stores below the treatment to support maintenance and survival of the tree in the face of injury. Leach (1937, 1939) proposed ringbarking in Malawi (then Nyasaland) to manage Armillaria root disease in tropical evergreen hardwood stands being converted to tea 49 plantations or Tung orchards (Vernicia fordei Hemsl. and Vernicia montana Lour.), suggesting that treatment should limit the transportation of carbohydrates from the canopy downward to reduce starch content in root systems prior to mortality. If treatment was applied with sufficient time prior to harvest, it should directly limit the intensification of Armillaria root disease in colonized root systems or should indirectly increase competition and exclusion of the parasite by saprophytic antagonists. Ringbarking through this mechanism would limit post-disturbance inoculum flush and disease spread and enhance the survivability of regenerating host species (Leach 1937, 1939, Chapman and Schellenberg 2015). Leach observed the root systems of a small random sample of Muula trees (Parinari curatellifolia Planch, ex Benth., then referred to as Parinari mobola) and a comparison was made between trees which were felled versus those killed via ringbarking (1937). His short¬ term results indicated a significantly lower percentage of roots infected with Armillaria root disease in samples from ringbarked versus non-treated stumps (17/18 ringbarked trees had no infection on root systems versus 3/24 felled trees). Leach’s (1937) proposal was grounded in the characterization of pathogenic Armillaria spp. as facultative parasites. The ability to overcome or avoid host defences and infect living root tissues provides these fungi with a competitive edge over saprophytes, allowing for an active and aggressive parasitic phase (Campbell 1934, Leach 1937, 1939, Redfern 1968, Swift 1970, Chapman and Xiao 2000, Chapman et al. 2004). Facultative parasites are able to infect and kill living hosts and are then able to persist in dead hosts long¬ term in a quiescent saprophytic phase, sourcing carbohydrates from cellulose and other constituents in decaying tissues. When tree roots are already dead and depleted of carbohydrates, or when saprophytes better adapted to colonizing decaying tissues arrive first, 50 potentially facilitated through a treatment such as ringbarking, facultative parasites may lose their competitive edge and might be inhibited in their ability to infect hosts (Leach 1937, 1939). Chapman and Schellenberg (2015) supported Leach’s theory, suggesting starch reduction was the mechanism responsible for the results of their evaluation of ringbarking in the forests of southern BC. Combining data across three similar Douglas-fir dominant sites with qualitatively high levels of infection, the authors observed a significant decrease at the trial level in the mortality of planted trees 15 years following harvest and ringbarking (0.8% annual mortality in treated areas versus 1.8% in untreated controls). Chapman and Schellenberg’s (2015) long-term observations would also support a reinterpretation of the mechanism behind control through ringbarking. If the mechanism of control was instead host stress response to treatment, a long-term decrease in mortality following treatment would be expected due to a more rapid inoculum flush followed by decline in inoculum before planted tree roots came into contact with roots of colonized trees. My study directly examined differences in inoculum volume of treated trees, whereas Chapman and Schellenberg (2015) examined a stand-level response to treatment that could be due to several different drivers of inoculum potential. Disease reduction due to rapid inoculum flush, followed by relatively rapid decline of inoculum potential in treated trees may be a potential mechanism of control. Ringbarking trees prior to harvest had no significant influence on A. ostoyae colonization within roots in comparison to control stumps, however I did observe a trend of increased mean incidence and mean severity of colonization following ringbarking at Monte Lake. Ringbarking at Monte Lake was carried out using a chainsaw fitted with a Log Wizard (Log Wizard 2016), 51 versus the other two trial locations which were treated by hand. Control over xylem contact was difficult with a chainsaw and it is likely that treatment at Monte Lake was more stressful to hosts versus at Gavin Lake or Horsefly Lake, possibly producing a more detectable inoculum flush effect. An inoculum flush effect following ringbarking was observed by Redfern (1968) in the south of England in Great Britain (the United Kingdom). That study found significantly higher inoculum levels in the roots of stumps from trees ringbarked 2 years prior to harvest versus those left untreated (90% infection in treated areas versus 80% in untreated controls). Redfern’s research took place in a mature temperate hardwood plantation (Quercus robur L. and Quercus petraea (Matt.) Liebl.) and root systems were observed 5 years following harvest, recording the presence or absence of Armillaria spp. on 4 primary roots of 5 stumps in each of 4 replicate plots per treatment. He found that root systems from treated trees were in further stages of decay and suggested that ringbarking might affect management by reducing the pathogen’s energy base through speeding up colonization and causing root die off prior to the spread of infection. To support this, Redfern evaluated root systems from his trial as food bases for A. mellea by incubating colonized segments and observing their ability to produce rhizomorphs. Root systems from his ringbarked treatment appeared to be less viable energy sources for the pathogen to spread from (39% of root systems from ringbarked trees produced rhizomorphs versus 90% of roots systems from untreated trees, Redfern 1968). Other trials to evaluate ringbarking as a control for Armillaria root disease may support an inoculum flush effect following treatment. Sokolov’s (1964) research in mixed stands (Picea spp., Betula spp., Larix spp., Populus spp.) in the Taiga zone of Russia 52 observed the roots of 38 infected trees ringbarked 3 months prior and found that treatment enhanced colonization of already infected roots in the short-term. Sokolov concluded that ringbarking depleted the starch in roots too slowly to inhibit colonization by the pathogen in infected hosts, however his results could also be explained by the stress caused by ringbarking causing an inoculum flush, especially given the short time between treatment and observation in his study. Differences in site conditions, Armillaria spp., and host species could explain discrepancies in results between my study and trials such as Leach’s (1937, 1939). Armillaria root disease around the world is caused by a complex of several Armillaria spp. varying widely in degree of pathogenicity and other characteristics. Many of these species were referred to early on as A. mellea and were later taxonomically separated due to sexual incompatibility and morphological differences, among other dissimilarities. Early research on A. mellea across Africa (Leach 1937, 1939; Swift 1970, 1972) may refer to what is now identified as A. heimmii or A. fuscipes, among a number of other species. Research by Sokolov (1964) in Russia on A. mellea may have also included species such as Armillaria lutea, while Redfern’s (1968) work in Europe might have involved additional species such as A. gallica, A. ostoyae, and Armillaria tabescens (Scop.) Emel. Limited information is available on the relative pathogenicity of this diverse array of species. Armillaria ostoyae in BC is known to be highly virulent and thought to be poorly rhizomorphogenic relative to other species, dispersing primarily via root contact with inoculum, while less aggressive species (or phases) may favour rhizomorphs to increase contact with host tissues (Morrison 201 1, Gonthier and Nicolotti 2013). Redfern (1968) observed rhizomorphs as the primary source of infection in his studies in southern England, 53 suggesting the pathogen causing Armillaria root disease in his trials may have exhibited moderate virulence versus A. ostoyae. Such differences in virulence or in method of infection could provide dissimilar results. Armillaria ostoyae" s phase and virulence, and subsequently its incidence and severity, depend on a range of local ecological conditions which also influence the defensive response of its hosts (Shaw and Loopstra 1988, Smith et al. 1994, Dettman and van der Kamp 2001, Holdenrieder et al. 2004). In my research, I attempted to account for broad ecological differences in A. ostoyae" s virulence by examining 3 sites within 3 distinct ecological zones. Finer-scale ecological differences between or within sites were not accounted for, for example different host species such as deciduous trees and other conifers were not considered. Varying host susceptibility to disease is another factor to consider when making comparisons, and is a function of tree species, age, size, and vigour, as well as site specific conditions and inoculum levels (Edmonds et al. 2000, Lewis and Lindgren 2000, Morrison 2011, Cleary et al. 2012). Armillaria ostoyae in BC is a generalist, able to attack a broad range of species across forests in various stages of succession. Here, the pathogen more aggressively attacks softwood versus hardwood species (Morrison et al. 1992, Cleary et al. 2008, British Columbia Ministry of Forests, Lands, Natural Resource Operations and Rural Development 2018), which may be partially related to competition and natural biological controls in the rhizosphere (Schafer 1971, Cruickshank and Jaquish 2014). This contrasts with Redfern’s (1968) study, where hardwoods were determined as the primary base of infection. Varying susceptibility is also importantly related to the host species’ ability to produce defensive phenols (Entry et al. 1992) and necrophylactic tissues (Cleary 2007, 54 Cleary et al. 2012) to inhibit and compartmentalize the pathogen. These defense responses are linked to carbohydrate dynamics within a tree, which varies by species and may be responsible for differences in host ability to defend against A ostoyae (Entry et al. 1992). The influence of ringbarking on starch in roots Starch is the primary way that plants store carbohydrates, being an essential source of energy for trees (Smith and Zeeman 2006). Ringbarking removes a tree’s phloem tissues and in theory should limit the transportation of carbohydrates from the canopy downwards to reduce levels in root systems prior to mortality. The results of my trial do not support this hypothesis. Starch contents in my samples were at or near the detectability levels of the developed assay, which lacked the required sensitivity to detect differences in starch levels between treated and control samples. Smith and Zeeman (2006) provide possible explanations for inhibited starch detection, including accidental discarding or incomplete digestion of starches, or an incomplete assay. The assay developed was tested and optimized, and I attempted to maximize recovery using an accessible starch control and a range of homogenization, boiling, and incubation times. The primary issue was likely very low starch contents in my samples. Reports of starch levels in plant tissues are highly variable due to discrepancies in methodology, and direct comparisons between outside laboratories should be made with caution (Quentin et al. 2015). My starch results are within the lower range of levels reported in literature for tissues sampled from conifers under stress, such as in Saffell et al.’s (2014) study on carbohydrate dynamics in Douglas-fir infected by Swiss needle cast (Nothophaeocryptopus gaeumannii (T. Rohde) Videira, C. Nakash., U. Braun & Crous). 55 Saffell et al. observed that diseased trees mobilized carbohydrate supplies (including the majority of starch storages) in the trunk to maintain levels in stressed crowns. Severely diseased trees appeared to be devoted to providing carbohydrates to this task over other processes, sacrificing growth and defences in other tissues such as the trunk and roots. Reduced carbohydrates below ground may subsequently influence fungal relationships (Luoma and Eberhart 2006, Saffell et al. 2014). My starch results are also within the lower range of reports from Wiley et al.’s (2016) study on lodgepole pine following mountain pine beetle attack. Wiley et al. observed significant depletion of starch in roots following attack, although total NSC in roots was not significantly depleted until after death. Wiley et al. (2016) concluded that attack caused the immediate local mobilization of carbohydrates, suggesting that long distance transfer to support defence was minimal, and decreases in roots were likely a result of heterotrophy by endophytes rather than directly from partial girdling via beetle attack. A host stress response that reprioritizes carbohydrates between organs may influence the susceptibility of tissues to attack (Wiley et al. 2016). Observations of low starch levels in my study may be due to seasonal carbohydrate dynamics, with trees directing energy into growth. In general, conifers begin to rely on their starch reserves at the start of the growing season and continue to deplete stores until dormancy of growth in the fall, when they begin to replenish starches once again (Edmonds et al. 2000, Saffell et al. 2014, Mei et al. 2015). Starch samples in my study were collected during the fall, at the end of growing season, when reserves may have been mostly depleted, and replenishment having not yet begun. Low starch levels in the case of ringbarked trees may have also been due to stressed hosts mobilizing reserves and redirecting carbohydrate 56 priorities away from replenishing storages to defending against injury (Miller and Berryman 1986, Mei et aL 2015), further lending to suggestion of a host defence response mechanism. The location from which samples were taken from root systems may also have had a role in observations of low starch in my study. Mei et al. (2015) notes the complex structure of root orders, with larger woody roots being devoted to carbohydrate transportation and storage, while finer roots are focused on nutrient uptake from the soil. My study focused on a diameter range to attempt to assess large storage roots, however, these efforts may have been ineffective. Mei et al. (2015) suggests that diameter does not accurately distinguish roots with different functions. The effects of harvest and storage on starch reserves in roots of sample trees likely had an influence in my study. Starch samples were collected 7-12 months following harvest and samples were stored prior to processing. Starch decline in this period may be an important reason for observations of low starch levels in roots. Wiley et al. (2016) discusses how continuing heterotrophic activity following tree death may cause the final decline of starch reserves, as pathogens and saprophytes such as endophytes consume carbohydrates until depletion. Roots in my study colonized by A. ostoyae likely continued to be depleted by the pathogen following harvest and during storage before processing. Research limitations Initial disease assessment Armillaria root disease is difficult to distinguish from other stressors using above¬ ground symptoms alone, limiting the usefulness of aerial and ground surveys. Undertaking a pre-treatment or pre-harvest disease assessment is labour intensive depending on the level of accuracy required. Accurate diagnosis requires identification of mycelial fans under the bark 57 of the root collar and roots. Exposing and sampling roots injures the roots and therefore may influence disease colonization and starch content results. I attempted to use trap-stakes to assess disease levels at each site. Guided by Mallet and Hiratsuka’s (1985) work, freshly cut stakes were driven with bark intact into the soil using grids centered over each plot. Subsets of stakes were removed between treatment and harvest, and again following harvest, and were examined for A. ostoyae. While many stakes were colonized, the majority of samples were identified as the saprophyte A. sinapina and this method was not continued. Armillaria sinapina has strong saprophytic tendencies relative to A. ostoyae, which may indicate cut stakes degraded before any contact was made. A. sinapina is also known to rely on a dense network of rhizomorphs within the soil to reach and infect tissues, whereas A. ostoyae primarily infects new hosts via root contacts with colonized hosts. It is possible that few (if any) of my trap-stakes were in contact with tree roots. An improved design would ensure that stakes were fresh cut from Douglas-fir on site and were immediately driven into the ground using grids surrounding infected and uninfected trees. Along with an increased sample size, these improvements would maximize the possibility of root contact and the successful trapping of A. ostoyae. Ringbarking treatment Ringbarking at the first site was done with a modified chainsaw, but due to safety concerns and lack of control over xylem contact, the other sites were ringbarked using hand tools. Though care was taken to minimize damage to xylem, it was not always possible to avoid, and treatment may have both restricted carbohydrate transport to roots and prohibited water and nutrient movement upward to the canopy, effectively killing host trees. Variability in injury caused by treatment may have influenced the results of my study. 58 Sampling for colonization It was difficult to accurately assess complex, buried, and broken root systems. Sampling methods could be improved by quantifying colonization as a percentage of total root surface area, rather than considering roots as two-dimensional by measuring lengths with any colonization. Reducing root and mycelium damage during excavations and improving access to stumps would be of considerable help. This would however exponentially increase already-prohibitive costs due to additional time required and would potentially increase site and soil degradation. Root breakage is likely exacerbated by heavy colonization, potentially causing an underestimation of colonization on heavily diseased trees in my study. An error adjustment to correct for increasing underestimation with heavier colonization could be applied to analyses. Timing the trial In my study, trees were harvested 3-8 months after treatment and roots were excavated for sampling 8-13 months post-harvest, with a second round of observations occurring 16-21 months after harvest. This contrasts with studies where ringbarking was applied at least 1 - 2 years prior to felling (Leach 1939, Sokolov 1964, Redfern 1968, Swift 1970, Chapman and Schellenberg 2015). Control of treatment, harvest and sample timing were limited due to contractor, licensee, and weather constraints. This included delays due to road access and operational restrictions caused by wildfires during the summer of 2017, followed by winter conditions with snowfall which prohibited access to excavated root systems for examination. Variability in treatment, harvest, and sample timing across the three sites may have influenced my results. 59 The time between treatment and harvest was likely an issue in my study and may explain differences in results between my trial and other research. Leach (1939) suggests that ringbarking depletes carbohydrates in tree roots slowly and notes that proper timing is essential for the treatment’s success. Both A. ostoyae and Douglas-fir are slow growing organisms, taking considerable time to respond to treatment and to each other in a detectable way. The time between treatment and sampling, or between harvest and sampling, may have also been too short in my study to allow for a detectable effect. In a study on carbohydrate dynamics in lodgepole pine that was partially girdled by mountain pine beetle attack, significant influences on starch in roots were only observed after approximately 5-6 months post beetle attack (Wiley et al. 2016). A modification of this design for future use could increase and test a range of times between treatment, harvest, excavation, and sampling. Starch content Starch content results were limited by study design. Sampling occurred in the fall and was complicated by an early snowfall burying excavated root systems. Small sections of healthy and infected roots were cut from primary and secondary roots, however damage to root systems reduced the population of roots to sample. Hand excavating roots for sampling prior to felling would reduce damage and avoid the influence of harvest on starch reserves. Technologies such as high-performance liquid and gas chromatography could be used to quantify very low starch contents. Research could benefit from a larger sample size and an expanded protocol designed to capture the influence of ringbarking across the wide diversity of non-structural carbohydrates (NSC) acting within host trees. An improved design would account for the spatial and temporal complexity of carbohydrate transportation and storage throughout trees and their root systems. Sampling should assess multiple replicates of 60 healthy and infected sections from various sizes and locations of roots in all sample systems at different times. Controls could be taken from untreated trees or from treated stumps left in the ground. Samples should be returned to the lab for processing and proper storage as soon as possible. 61 CONCLUSION AND MANAGEMENT RECOMMENDATIONS Harvest of infected forests disrupts natural balances and biological controls and can provide A. ostoyae with an enhanced energy base. Ringbarking host trees prior to harvest should limit starch in roots and according to one theory, inhibit the pathogen’s spread following disturbance (Leach 1937, 1939, Chapman and Schellenberg 2015). My study was unable to confirm that ringbarking trees in A. ostoyae centres prior to their felling influences starch content or colonization within root systems following harvest. My results combined with reviewed literature instead suggest that ringbarking may provide control through an alternate mechanism, host stress response (Redfern 1968). Observations of increases in mean inoculum levels in stumps of treated trees at one of my sites, while not statistically significant, weakly suggest that host stress from ringbarking may enable rapid expansion of the fungus by causing a decline in vigour of the host due to a hindered ability to mobilize and utilize carbohydrates to produce defences. Host defence is likely inhibited by ringbarking as the tree reprioritizes carbohydrates below the injury to maintain its survival, being unable to replenish reserves (Miller and Berryman 1986, Mei et al. 2015). Rapid expansion and accelerated use of root-based resources by the fungus might be a short-term effect that would reverse as the fungus’ food base becomes depleted, leading to a faster decline in inoculum potential in treated trees. If ringbarking does provide Armillaria root disease control by either lowering the starch content in roots of infected trees or by provoking rapid colonization due to host stress followed by inoculum decline, the balance between stressing hosts and preserving commercial timber values would be a challenge. Provoking a stress response through ringbarking further provides a vector for the introduction of additional fungi or insects. If the 62 host stress mechanism does provide control, other methods of provoking this response to treat Armillaria root disease could be explored. Ringbarking as an Armillaria root disease treatment during my research was labourintensive, time-consuming, and expensive to apply, challenges that would have to be overcome for commercial use. In my trial, we ringbarked hundreds of trees across each site, with each plot taking 2-3 hours, and each site taking up to a week. Contractor costs for chainsaw operators were high, and treatment by hand was slow and required multiple workers. Though efficiency gains could be made with experience and with specially designed tools, scaling up to a large-scale for use in commercial forestry is likely limited in practicality. If ringbarking is undertaken, treatment should be complete around the tree, as in my study. Lateral movement of carbohydrates has been demonstrated and partial girdling may have variable results (Miller and Berryman 1986). Height of treatment may have an important influence as well, and lower treatments may prove more effective, as carbohydrates below the point of ringbarking will be mobilized in a continued effort to support roots (Miller and Berryman 1986, Mei et al. 2015). A cost-benefit analysis comparing ringbarking to other management options such as inoculum removal or the planting of resistant and tolerant species should be undertaken before considering implementation. Additional long-term research is required to clarify the mechanisms of control behind the treatment. Future work should address challenges surrounding initial disease assessment, treatment methods and timing, sampling, and the quantification of colonization and starch. 63 REFERENCES Ayres, M.P., and M.J. Lombardero. 2000. Assessing the consequences of global change for forest disturbance from herbivores and pathogens. Science of the Total Environment 262: 263-286. Baleshta, K.E., S.W. Simard, R.D. Guy, and C.P. Chanway. 2005. Reducing paper birch density increases Douglas-fir growth rate and Armillaria root disease incidence in southern interior British Columbia. Forest Ecology and Management 208: 1-13. Baleshta, K.E., S.W. Simard, and W.J. Roach. 2015. Effects of thinning paper birch on conifer productivity and understory plant diversity. Scandinavian Journal of Forest Research 30: 699-709. Baughman, B.M., E. Forester, and P. Leader. 2017. Trees and Woodlands Flooding effects on trees. Prepared for the University of Minnesota. Available from https://extension.umn.edu/planting-and-growing-guides/how-manage-flood-damagetrees. Binkley, D., J.L. Stape, E.N. Takahashi, and M.G. Ryan. 2006. Tree-girdling to separate root and heterotrophic respiration in two Eucalyptus stands in Brazil. Oecologia 148: 447454. Bloomberg, W.J., and D.J. Morrison. 1989. Relationship of Growth Reduction in DouglasFir to Infection by Armillaria Root Disease in Southeastern British Columbia Canada. Phytopathology 79: 482^487. British Columbia Ministry of Forests, Lands, Natural Resources and Rural Development. 2018. Managing Root Disease in British Columbia. Available from https://www2.gov.bc.ca/assets/gov/farming-natural-resources-andindustry/forestry/forest-health/forest-health-docs/root-diseasedocs/rootdiseaseguidebookjune20 1 8_4.pdf. Campbell, W. G. 1931. The chemistry of the white rots of wood: The effect on wood substance of Armillaria mellea (VahL) Fr., Polyporus hispidus (Bull.) Fr., and Stereum hirsutum Fr. Biochemical Journal 25: 2023-2027. Campbell, A.H. 1934. Zone Lines in Plant Tissues: II. The Black Lines Formed by Armillaria mellea (Vahl) Quel. Annals of Applied Biology 21: 1-22. Carswell, B.M., R.V. Rea, D. Rusch, and C.J. Johnson. 2021. The influence of the root diseases Armillaria solidipes and Inonotus sulphurascens on the distribution of mule deer during winter. Forestry: An International Journal of Forest Research 94: 492501. Chapman, B., and B. Schellenberg. 2015. An exploration of ringbarking to reduce the severity of Armillaria root disease in logged areas in British Columbia. Canadian Journal of Forest Research 45: 1803-1805. 64 Chapman, B., and G. Xiao. 2000. Inoculation of stumps with Hypholoma fasciculare as a possible means to control armillaria root disease. Canadian Journal of Botany 1877: 129-134. Chapman, B., G. Xiao, and S. Myers. 2004. Early results from field trials using Hypholoma fasciculare to reduce Armillaria ostoyae root disease. Canadian Journal of Botany 82: 962-969. Chapman, W.K., B. Schellenberg, and T.A. Newsome. 2011. Assessment of Armillaria root disease infection in stands in south-central British Columbia with varying levels of overstory retention, with and without pushover logging. Canadian Journal of Forest Research 41: 1598-1605. Chow, P.A.K.S., and S.M. Landhausser. 2018. A method for routine measurements of total sugar and starch content in woody plant tissues. Tree Physiology 24: 1 129-1 136. Cleary, M., B. van der Kamp, and D. Morrison. 2008. British Columbia’s Southern Interior Forests: Armillaria Root Disease Stand Establishment Decision Aid. BC Journal of Ecosystems and Management 9: 60-65. Cleary, M.R. 2007. Host responses in Douglas-fir, western hemlock and western redcedar to infection by Armillaria ostoyae and Armillaria sinapina. Doctoral dissertation, University of British Columbia. Cleary, M.R., N. Arhipova, D.J. Morrison, I.M. Thomsen, R.N. Sturrock, R. Vasaitis, T. Gaitnieks, and J. Stenlid. 2013. Stump removal to control root disease in Canada and Scandinavia: A synthesis of results from long-term trials. Forest Ecology and Management 290: 5-14. Cleary, M.R., B.J. van der Kamp, and D.J. Morrison. 2012. Effects of wounding and fungal infection with Armillaria ostoyae in three conifer species. II. Host response to the pathogen. Forest Pathology 42: 109-123. Clinton, B.D., L.R. Boring, and W.T. Swank. 1993. Canopy Gap Characteristics and Drought Influences in Oak Forests of the Coweeta Basin. Ecology 74: 1551-1558. Cruickshank, M.G., D.J. Morrison, and Z.K. Punja. 1997. Incidence of Armillaria species in precommercial thinning stumps and spread of Armillaria ostoyae to adjacent Douglas-fir trees. Canadian Journal of Forest Research 27: 481-490. Cruickshank, M.G. 2010. Effect of Armillaria root disease on quality and value of green Douglas-fir lumber. The Forestry Chronicle 86: 263-266. Cruickshank, M.G., and C.N. Filipescu. 2012. Allometries of coarse tree, stem, and crown measures in Douglas-fir are altered by Armillaria root disease. Botany 90: 711-721. 65 Cruickshank, M.G., and B. Jaquish. 2014. Resistance and tolerance in juvenile interior Douglas-fir trees Pseudotsuga menziesii var. glauca artificially inoculated with Armillaria ostoyae. Forest Pathology 44: 362-371. Cruickshank, M.G., D.J. Morrison, and A. Lalumiere. 2009. The interaction between competition in interior Douglas-fir plantations and disease caused by Armillaria ostoyae in British Columbia. Forest Ecology and Management 257: 443-452. Cruickshank, M.G., D.J. Morrison, and A. Lalumiere. 2011. Site, plot, and individual tree yield reduction of interior Douglas-fir associated with non-lethal infection by Armillaria root disease in southern British Columbia. Forest Ecology and Management 261: 297-307. DeLong, R.L., K.J. Lewis, S.W. Simard, and S. Gibson. 2002. Fluorescent pseudomonad population sizes baited from soils under pure birch, pure Douglas-fir, and mixed forest stands and their antagonism toward Armillaria ostoyae in vitro. Canadian Journal of Forest Research 32: 2146-2159. Desprez-Loustau, M.L., B. Mar^ais, L.M. Nageleisen, D. Piou, and A. Vannini. 2006. Interactive effects of drought and pathogens in forest trees. Annals of Forest Science 63: 597-612. Dettman, J.R., and B.J. van der Kamp. 2001. The population structure of Armillaria ostoyae and Armillaria sinapina in the central interior of British Columbia. Canadian Journal of Botany 79: 600-61 1. Dukes, J.S., J. Pontius, D. Orwig, J.R. Gamas, V.L. Rodgers, N. Brazee, B. Cooke, K. Theoharides, E.E. Stange, R. Harrington, J. Ehrenfeld, J. Gurevitch, M. Lerdau, K. Stinson, R. Wick, and M. Ayres. 2009. Responses of insect pests, pathogens, and invasive plant species to climate change in the forests of northeastern North America: What can we predict? Canadian Journal of Forest Research 39: 231-248. Dumas, M.T. 2012. Effect of site preparation on Armillaria ostoyae infection and growth of spruce after partial cutting in a Boreal Mixedwood Forest. Forestry Chronicle 88: 622-625. Edmonds, R.L., J.K. Agee, and R.I. Gara. 2000. Forest Health and Protection. McGraw-Hill. Entry, J.A., N.E. Martin, R.G. Kelsey, and K. Cromack Jr. 1992. Chemical constituents in root bark of five species of western conifer saplings and infection by Armillaria ostoyae. Phytopathology 82: 393-397. Entry, J. A., P. Donnelly, and K. Cromack Jr. 1993. Effect of nitrogen and carbon sources on lignin and cellulose degradation by Armillaria ostoyae. European Journal of Forest Pathology 23: 129-137. 66 Gerlach, J.P., P.B. Reich, K. Puettmann, and T. Baker. 1997. Species, diversity, and density affect tree seedling mortality from Armillaria root rot. Canadian Journal of Forest Research 27: 1509-1512. Goheen, D., and E. Hansen. 1993. Effects of Pathogens and Bark Beetles on Forests. In Beetle-Pathogen Interactions in Conifer Forests. Edited by T.D. Schowalter and G.M. Filip. Gonthier, P., and G. Nicolotti. 2013. Infectious forest diseases. CABI. Hansen, J., and I.B. Moller. 1975. Percolation of starch and soluble carbohydrates from plant tissue for quantitative determination with anthrone. Analytical Biochemistry 68: 8794. Harvell, C.D., C.E. Mitchell, J.R. Ward, S. Altizer, A.P. Dobson, R.S. Ostfeld, and M.D. Samuel. 2002. Climate warming and disease risks for terrestrial and marine biota. Science 296: 2158-2162. Heineman, J.L., D.L. Sachs, W.J. Mather, and S.W. Simard. 2010. Investigating the influence of climate, site, location, and treatment factors on damage to young lodgepole pine in southern British Columbia. Canadian Journal of Forest Research 40: 1109-1127. Henigman, J., T. Ebata, E. Allen, J. Holt, A. Pollard, and others. 1999. Field Guide to Forest Damage in British Columbia. British Columbia Ministry of Forests. Hogberg, P., A. Nordgren, N. Buchmann, A.F.S. Taylor, A. Ekblad, M.N. Hogberg, G. Nyberg, M. Ottoson-Ldfvenious, and D.J. Read. 2001. Large-scale forest girdling shows that current photosynthesis drives soil respiration. Nature 41 1: 789-792. Holdenrieder, O., M. Pautasso, P.J. Weisberg, and D. Lonsdale. 2004. Tree diseases and landscape processes: The challenge of landscape pathology. Trends in Ecology and Evolution 19: 446-452. van der Kamp, B.J. 1991. Pathogens as agents of diversity in forested landscapes. Forestry Chronicle 67: 353-354. van der Kamp, B.J. 1993. Rate of spread of Armillaria ostoyae in the central interior of British Columbia. Canadian Journal of Forest Research 23: 1239-1241. Klopfenstein, N.B., M.-S. Kim, J.W. Hanna, B.A. Richardson, and J.E. Lundquist. 2009. Approaches to predicting potential impacts of climate change on forest disease: an example with Armillaria root disease. In Res. Pap. RMRS-RP-76. Fort Collins, CO: US Department of Agriculture, Forest Service, Rocky Mountain Research Station. Kulhavy, D.L., A.D. Partridge, and R.W. Stark. 1984. Root diseases and blister rust associated with bark beetles (Coleoptera: Scolytidae) in western white pine in Idaho. Environmental Entomology 13: 813-817. 67 Leach, R. 1937. Observations on the parasitism and control of Armillaria mellea. Proceedings of the Royal Society of London. Series B, Biological Sciences 121 : 561— 573. Leach, R. 1939. Biological control and ecology of Armillaria mellea (vahl) fr. Transactions of the British Mycological Society 23: 320-329. Lewis, K.J., and B.S. Lindgren. 2000. A conceptual model of biotic disturbance ecology in the central interior of BC: how forest management can turn Dr. Jekyll into Mr. Hyde. The Forestry Chronicle 76: 433-443. Lundquist, J.E. 2000. Disturbances on Canopy Gap Size. Forest Science 46: 356-362. Luoma, D.L., and J.L. Eberhart. 2006. Are differences in the ectomycorrhiza community correlated with Swiss Needle Cast Severity?. In Swiss Needle Cast Cooperative Annual Report 2006: 60-64. Oregon State University. Mallett, K.I., and D.G. Maynard. 1998. Armillaria root disease, stand characteristics, and soil properties in young lodgepole pine. Forest Ecology and Management 105: 37-44. Mallett, K.I. 1992. Armillaria Root Rot in the Canadian Prairie Provinces. In Forestry Canada, Northwest Region, Northern Forestry Centre, Edmonton, Alberta. Information Report NOR-X-329. Mallett, K.I., and Y. Hiratsuka. 1985. The" trap-log" method to survey the distribution of Armillaria mellea in forest soils. Canadian Journal of Forest Research 15: 11911193. Marshall, J.D., and R.H. Waring. 1985. Predicting fine root production and turnover by monitoring root starch and soil temperature. Canadian Journal of Forest Research 15: 791-800. Mei, L., Y. Xiong, J. Gu, Z. Wang, and D. Guo. 2015. Whole-tree dynamics of nonstructural carbohydrate and nitrogen pools across different seasons and in response to girdling in two temperate trees. Oencologia 177: 333-344. Miller, R.H., and A.A. Berryman. 1986. Carbohydrate allocation and mountain pine beetle attack in girdled lodgepole pines. Canadian Journal of Forest Research 16: 1036- 1040. Morrison, D., and K. Mallett. 1996. Silvicultural management of Armillaria root disease in western Canadian forests. Canadian Journal of Plant Pathology 18: 194-199. Morrison, D.J. 2011. Epidemiology of Armillaria root disease in Douglas-fir plantations in the cedar-hemlock zone of the southern interior of British Columbia. Forest Pathology 41: 31-40. 68 Morrison, D.J., M.G. Cruickshank, and A. Lalumiere. 2014. Control of laminated and Armillaria root diseases by stump removal and tree species mixtures: amount and cause of mortality and impact on yield after 40 years. Forest Ecology and Management 319: 75-98. Morrison, D.J., H. Merler, and D.J. Norris. 1992. Detection, recognition and management of Armillaria and Phellinus root diseases in the southern interior of British Columbia. FRDA report No. 179. Morrison, D.J., and K.W. Pellow. 1994. Development of Armillaria root disease in a 25year-old Douglas-fir plantation. In Proceedings of the Eighth International Conference on Root and Butt Rots, August 9-16, 1993, Wik, Sweden and Haikko, Finland. Morrison, D.J., K.W. Pellow, A.F.L. Nemec, D.J. Norris, and P. Semenoff. 2001. Effects of selective cutting on the epidemiology of armillaria root disease in the southern interior of British Columbia. Canadian journal of forest research 31: 59-70. Morrison, D.J., K.W. Pellow, D.J. Norris, and A.F.L. Nemec. 2000. Visible versus actual incidence of Armillaria root disease in juvenile coniferous stands in the southern interior of British Columbia. Canadian Journal of Forest Research 30: 405-414. Morrison, D.J., G.M. Wallis, and L.C. Weir. 1988. Control of Armillaria and Phellinus root diseases: 20-year results from the Skimikin stump removal experiment. Prepared for the Canadian Forestry Service, Pacific Forestry Centre. Victoria, BC. Parker, T.J., K.M. Clancy, and R.L. Mathiasen. 2006. Interactions among fire, insects and pathogens in coniferous forests of the interior western United States and Canada. Agricultural and Forest Entomology 8: 167-189. Peet, F.G., D.J. Morrison, and K.W. Pellow. 1996. Rate of spread of Armillaria ostoyae in two Douglas-fir plantations in the southern interior of British Columbia. Canadian Journal of Forest Research 26: 148-151. Popoola, T.O.S., and R.T. V Fox. 2003. Effect of water stress on infection by species of honey fungus [Armillaria mellea and A. gallica). Arboricultural Journal 27: 139-154. Quentin, A.G., E.A. Pinkard, M.G. Ryan, D.T. Tissue, L.S. Baggett, H.D. Adams, P. Maillard, J. Marchand, S.M. Landhausser, A. Lacointe, and Y. Gibon. 2015. Nonstructural carbohydrates in woody plants compared among laboratories. Tree Physiology 35: 1146-1165. Reaves, J.L., C.G. Shaw, III, and J.E. Mayfield. 1990. The effects of Trichoderma spp. isolated from burned and non-bumed forest soils on the growth and development of Armillaria ostoyae in culture. Northwest Science 64: 39^44. Redfern, D.B. 1968. The ecology of Armillaria mellea in Britain biological control. Annals of Botany 33: 293-300. 69 Robinson, R.M., and D.J. Morrison. 2001. Lesion formation and host response to infection by Armillaria ostoyae in the roots of western larch and Douglas-fir. Forest Pathology 31: 371-385. Robinson, R.M., D.J. Morrison, and G.D. Jensen. 2004. Necrophylactic periderm formation in the roots of western larch and Douglas-fir trees infected with Armillaria ostoyae. II. The response to the pathogen. Forest Pathology 34: 1 19-129. Saffell, B. J., F.C. Meinzer, D.R. Woodruff, D.C. Shaw, S.L. Voelker, B. Lachenbruch, and K. Falk. 2014. Seasonal carbohydrate dynamics and growth in Douglas-fir trees experiencing chronic, fungal-mediated reduction in functional leaf area. Tree Physiology 34: 218-228. Schafer, J.F. 1971. Tolerance to plant disease. Annual Review of Phytopathology 9: 235252. Shaw, C.G., D.W. Omdal, A. Ramsey-Kroll, and L.F. Roth. 2012. Inoculum reduction measures to manage Armillaria root disease in a severely infected stand of ponderosa pine in south-central Washington: 35-year results. Western Journal of Applied Forestry 27: 25-29. Shaw III, C.G., and E.M. Loopstra. 1988. Identification and pathogenicity of some Alaskan isolates of Armillaria. Phytopathology 78: 971-974. Shields Jr, W.J., and S.D. Hobbs. 1979. Soil nutrient levels and pH associated with Armillariella mellea on conifers in northern Idaho. Canadian Journal of Forest Research 9: 45-48. Simard, S.W., S.M. Hagerman, D.L. Sachs, J.L. Heineman, and W.J. Mather. 2005. Conifer growth, Armillaria ostoyae root disease, and plant diversity responses to broadleaf competition reduction in mixed forests of southern interior British Columbia. Canadian Journal of Forest Research 35: 843-859. Smith, A.M., and S.C. Zeeman. 2006. Quantification of starch in plant tissues. Nature Protocols 1: 1342-1345. Smith, M.L., J.N. Bruhn, and J.B. Anderson. 1994. Relatedness and spatial distribution of Armillaria genets infecting red pine seedlings. Phytopathology 84: 822-829. Sokolov, D. V. 1964. Root rot caused by Armillaria mellea and its control. Canada Department of Forestry, Translation 37. Sturrock, R. 2012. Climate change and forest diseases: using today’s knowledge to address future challenges. Forest Systems 21: 329-336. Sturrock, R.N., S.J. Frankel, A. V. Brown, P.E. Hennon, J.T. Kliejunas, K.J. Lewis, J.J. Worrall, and A.J. Woods. 2011. Climate change and forest diseases. Plant Pathology 60: 133-149. 70 Swift, M.J. 1970. Armillaria mellea (Vahl ex Fries) Kummer in Central Africa: studies on substrate colonisation relating to the mechanism of biological control by ring¬ barking. In Root diseases and soil-bome pathogens. 150-152. University of California Press Berkeley. Swift, M.J. 1972. The Ecology of Armillaria mellea in the Indigenous and Exotic Woodlands of Rhodesia. Forestry Oxford 45: 67-86. Taylor, S.P. 1986. Forest Insect and Disease Impacts in Timber Supply Areas. British Columbia Ministry of Forests, Protection Branch, Pest Management. 6: 40-1 10. Wargo, P.M. 1975. Estimating Starch Content in Roots of Deciduous Trees - A Visual Technique. USDA Forest Service Research Paper NE-313: 1-6. Wargo, P.M. 1996. Consequences of environmental stress on oak: predisposition to pathogens. Annales des Sciences Forestieres 53: 359-368. Wiensczyk, a M., M.T. Dumas, and R.N. Irwin. 1997. Infection Levels in Black Spruce Plantations as a Function of Environmental Factors. Canadian Journal of Forest Research 27: 1630-1634. Wiley, E., and B. Helliker. 2012. A re-evaluation of carbon storage in trees lends greater support for carbon limitation to growth. New Phytologist 195: 285-298. Wiley, E., B.J. Rogers, R. Hodgkinson, and S.M. Landhausser. 2016. Nonstructural carbohydrate dynamics of lodgepole pine dying from mountain pine beetle attack. New Phytologist 209: 550-562. Zimmermann, M.H., and C.L. Brown. 1971. Trees: structure and function. New York, USE. Springer-Verlag. 71