UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Incorporating stand-health metrics into monitoring the effects of soil disturbance during logging on… Reid, Anya Martina 2016

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

Item Metadata

Download

Media
24-ubc_2016_september_reid_anya.pdf [ 7.46MB ]
Metadata
JSON: 24-1.0306863.json
JSON-LD: 24-1.0306863-ld.json
RDF/XML (Pretty): 24-1.0306863-rdf.xml
RDF/JSON: 24-1.0306863-rdf.json
Turtle: 24-1.0306863-turtle.txt
N-Triples: 24-1.0306863-rdf-ntriples.txt
Original Record: 24-1.0306863-source.json
Full Text
24-1.0306863-fulltext.txt
Citation
24-1.0306863.ris

Full Text

Incorporating stand-health metrics into monitoring the effects of soil disturbance during logging on long-term forest productivity  by ANYA MARTINA REID  MSc, York University, 2011 BSc, The University of British Columbia, 2009  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in The Faculty of Graduate and Postdoctoral Studies (Forestry)   THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)   July 2016  © Anya Martina Reid, 2016                       ii ABSTRACT Stand volumes at rotation are usually predicted from models based on tree growth, assuming that fast-growing trees are, and will remain, healthy. However, greater-than-expected disease occurrence on dominant trees has been reported in regenerating forests in British Columbia (BC), suggesting that forest-productivity monitoring should include health measurements. In this thesis, I assess growth and health of lodgepole pine at six installations of the Long-Term Soil Productivity (LTSP) project 15 to 20 years after soil-disturbance treatments (organic-matter removal and soil compaction). To determine treatment effects on forest health, 5400 lodgepole pine seedlings at six sites were examined for vigour and pest occurrence. Treatment effects differed among sites, suggesting that the effects of the soil-disturbance treatments on forest health are context-dependent. Larger trees generally had higher frequency of disease, contradicting the assumption that fast-growing trees are healthy. Spectral-reflectance indices of greenness calculated from aerial images correlated with ground-based measures of tree vigour and foliar disease, indicating a role for remote-sensing techniques in forest health monitoring. Topsoil-nutrient content was reduced in the forest-floor removal treatment plots, with foliar phosphorus and potassium concentrations (%) being reduced in the three older sites in the sub-boreal spruce zone. The forest-floor removal treatment was also associated with lower abundance of ectomycorrhizae, but greater abundance Suillus sp., a fungus that is associated with nitrogen-fixing diazatrophic bacteria. The main predictive growth-and-yield model used in BC (called TASS with the interface TIPSY), underestimated tree height (m) but overestimated stand density (stems/ha) at the six sites, suggesting greater-than-predicted tree mortality. Monitoring forest health in conjunction with tree growth after free-to-grow standards are met is recommended for more accurate management of long-term forest productivity.        iii  PREFACE The research program detailed in this thesis was lead by myself with valuable input from my supervisory committee including Cindy Prescott, Sue Grayston, Marty Kranabetter and Bill Chapman. This research program collaborated with The British Columbia Ministry of Forests, Lands, and Natural Resource Operations (MFLNRO) and was conducted on the MFLNRO’s Long-Term Soil Productivity (LTSP) sites. Therefore, data used in this thesis originates both from the long-term monitoring of the LTSP sites by the MFLNRO and from original data I collected during the thesis program. Data on tree growth, foliar nutrients, soil chemistry and soil-physical properties was collected by the MFLNRO. I helped collect and enter the tree growth data for Skulow Lake at year 20. I helped collect soil samples for chemical- and physical-property analysis on the Topley site at year 20. I designed, collected and entered all the data on stand health, tree mortality, canopy closure and spectral reflectance. I summarized, amalgamated and analyzed all the data used in this thesis. I conducted all the writing of this thesis with editing from my supervisory committee and anonymous reviewers, in the case of published work.   LIST OF PUBLICATIONS: Reid, A.M., Chapman, B.K., Kranabetter, J.M., and Prescott, C.E. 2015. Response of lodgepole pine health to soil-disturbance treatments in British Columbia, Canada. Canadian Journal of Forest Research 45:1045-1055. The material in this manuscript is primarily located in Chapter 2 although the data is located throughout the thesis. As the lead author, Anya Reid collected the data, conducted the analysis, and wrote the manuscripts with valuable contributions from the co-authors. Bill K. Chapman was involved in planning, initiating, funding and advising this study.    iv Reid, A.M., Chapman, B.K., and Prescott, C.E. 2016. Comparing lodgepole pine growth and disease occurrence at six Long-Term Soil Productivity (LTSP) sites in British Columbia, Canada. DOI: 10.1139/cjfr-2015-0441. The material in this manuscript is primarily located in Chapter 3. As the lead author, Anya Reid collected the data, conducted the analysis, and wrote the manuscripts with valuable contributions from the co-authors.   Reid, A.M., Chapman, B.K., and Prescott, C.E. 2016. Using spectral-reflectance indices of excess greenness and green chromatic coordinate to assess lodgepole pine stand vigour, mortality and disease occurrence. Forest Ecology and Management 374: 146-153.  The material in this manuscript is primarily located in Chapter 4. As the lead author, Anya Reid collected the data, conducted the analysis, and wrote the manuscripts with valuable contributions from the co-authors.    v TABLE OF CONTENTS ABSTRACT ....................................................................................................................... ii	PREFACE ......................................................................................................................... iii	TABLE OF CONTENTS .................................................................................................. v	LIST OF TABLES .......................................................................................................... vii	LIST OF FIGURES ........................................................................................................... x	LIST OF ABBREVIATIONS ....................................................................................... xiii	ACKNOWLEDGEMENTS ........................................................................................... xiv	Chapter 1. Introduction .................................................................................................... 1	1.1 Background ............................................................................................................... 1	1.1.1 Forest health .................................................................................................................... 1	1.1.2 Monitoring forest productivity ........................................................................................ 4	1.1.3 Management effects of forest health ............................................................................... 8	1.2 Thesis objectives and organization ......................................................................... 10	1.3 Study sites and species ............................................................................................ 11	1.4 Tables and figures ................................................................................................... 13	Chapter 2. Effect of soil disturbance on stand health .................................................. 18	2.1 Synopsis .................................................................................................................. 18	2.2 Introduction ............................................................................................................. 18	2.3 Methods................................................................................................................... 19	2.4 Results ..................................................................................................................... 22	2.5 Discussion ............................................................................................................... 24	2.6 Tables and figures ................................................................................................... 27	Chapter 3. Relationships between tree growth and stand health ................................ 31	3.1 Synopsis .................................................................................................................. 31	3.2 Introduction ............................................................................................................. 31	3.3 Methods................................................................................................................... 32	3.4 Results ..................................................................................................................... 34	3.5 Discussion ............................................................................................................... 34	3.6 Tables and figures ................................................................................................... 38	Chapter 4. Spectral reflectance and stand health ......................................................... 43	4.1 Synopsis .................................................................................................................. 43	4.2 Introduction ............................................................................................................. 43	4.3 Methods................................................................................................................... 45	4.4 Results ..................................................................................................................... 47	4.5 Discussion ............................................................................................................... 49	4.6 Tables and figures ................................................................................................... 51	Chapter 5. Nutrient contribution of the forest floor ..................................................... 59	5.1 Synopsis .................................................................................................................. 59	5.2 Introduction ............................................................................................................. 59	 vi 5.3 Methods................................................................................................................... 62	5.4 Results ..................................................................................................................... 66	5.5 Discussion ............................................................................................................... 67	5.6 Tables and figures ................................................................................................... 69	Chapter 6. Conclusion and synthesis ............................................................................. 78	6.1 Synthesis ................................................................................................................. 80	6.2 Future research directions ....................................................................................... 90	Bibliography ..................................................................................................................... 92	Appendix A: Data tables ............................................................................................... 106	  vii  LIST OF TABLES Table 1.1. Nitrogen amount (kg/ha) and percent (%) distribution within ecosystem components in Douglas-fir stands that are 36, 47 and 450 years old in Washington and Oregon…………………………………………………………………………….......13 Table 1.2. Nutrients pools (kg/ha) and percentages calculated in each ecosystem components of a 450-year-old Douglas-fir forest in Oregon, USA (Cole and Rapp 1980).……………………………………………….…………………………………13 Table 1.3. Black Pines, Dairy Creek, and O’Connor Lake sites were located in the Interior Douglas Fir (IDF) biogeoclimactic ecosystem classification (BEC) zone. Log Lake, Skulow Lake and Topley sites were located in the Sub Boreal Spruce (SBS) BEC zone. Information at each site is provided: latitude (˚), longitude (˚), elevation (m), year planted, mean annual temperature (MAT), mean annual precipitation (MAP; Wang et al. 2012), soil texture, and dominant soil classification (Berch et al. 2010).………………………………………………………………….………………14 Table 2.1. Average number and percentage of healthy trees, dead or dying trees, tree with high or medium foliar-disease severity, trees with galls from western gall rust, and root-disease symptoms at each site. Three sites are in the Interior Douglas Fir (IDF) zone and three are in the Sub Boreal Spruce (SBS) zone. Percentages of healthy and dead or dying trees are calculated from the total number of trees planted per site (900) and percentages of disease occurrence are calculated from the number of living trees per site.……………………………………………..…………………………………27 Table 3.1. Tree growth in relation to tree health based on treatment-plot averages.………………………………………………………………………………38 Table 3.2. Tree growth in relation to tree health based on random-group averages.………………………………………………………………………………39 Table 4.1. Significant relationships among tree-health and spectral-reflectance variables calculated at each site separately.………………………………………..……………51 Table 4.2. The percent of variability in EGtree, EGplot, GCCtree and GCCplot explained by healthy vigour, mortality, foliar-disease occurrence, and root-disease occurrence calculated using all 54 plots. Total amount of variability is summarized at the bottom.…………………………………………………………………………..….…51 Table 5.1. Foliar-nutrient concentrations indicating adequate (green), slight to moderate deficiency (yellow), moderate to severe deficiency (orange) and severe deficiency (red) for tree growth (Brockley 2001).…………………………………………..……69 Table 5.2. Mean and standard deviation values of canopy closure calculated at each site. Sorted from highest to lowest canopy-closure values…………………….…………. 69 Table 5.3. The amount (kg/ha) of carbon (C), nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg) and sulfur (S) in the forest floor during the pre-harvest soil survey in the forest-floor removal plots. Darker shading indicates higher numbers……………………………………………………………………………….70  viii Table 5.4. The proportion (%) of carbon (C), nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg) and sulfur (S) removed in the forest floor relative to the topsoil in the forest-floor removal plots. Darker shading indicates higher numbers.…………………………………………………………………………...… 70 Table 5.5. Foliar-nutrient concentrations (%) calculated on a dry mass basis with colour coding for nutrient deficiency ratings (Table 3). When nutrients did indicate deficiencies, there was no evidence for dilution effects (no dilution = ND)…………70 Table 5.6. Effects of organic-matter removal (OMR; forest-floor removal plots versus bole-only harvest plots), soil compaction (COMP), and site on topsoil-nutrient content (kg/ha). Analyses were conducted with plot averages within each BEC zone. Bold font indicates significance considered at p < 0.05……………………………………..…..71 Table 5.7. Effects of organic-matter removal (OMR; forest-floor removal plots versus bole-only harvest plots), soil compaction (COMP), and site on foliar-nutrient concentration (%). Analyses were conducted with plot averages within each BEC zone. Bold font indicates significance considered at p < 0.05.……………………… 71 Table 5.8. Effects of organic-matter removal (OMR; forest-floor removal plots versus bole-only harvest plots), soil compaction (COMP), and site on fungi relative abundance. Analyses were conducted with plot averages within each BEC zone. Bold font indicates significance considered at p < 0.05.………………………..………… 72 Table 5.9. Correlation-coefficient (r) values for the relationships of soil-nutrient content (kg/ha) with tree growth and health variables are summarized. Green shading indicates correlation-coefficient values above zero and red shading indicates correlation-coefficient values below zero. Relationships were tested within each site to determine consistency among sites. ………………………………………………….……….…73 Table 5.10. Correlation-coefficient (r) values for the relationships of foliar-nutrient concentration (%) with tree growth and health variables are summarized. Green shading indicates correlation-coefficient values above zero and red shading indicates correlation-coefficient values below zero. Relationships were tested within each site to determine consistency or differences among sites…………………………........……74 Table 5.11. Correlation-coefficient (r) values for the relationships of fungal-species relative abundance with tree growth and health variables are summarized. Green shading indicates correlation-coefficient values above zero and red shading indicates correlation-coefficient values below zero. Relationships were tested within each site to determine consistency or differences among sites.……………………………...……75 Table 6.1. Average response ratio for tree height, total stem volume, healthy vigour, dead or dying trees and total-tree disease calculated from all eight-treatment plots. Values greater than one indicate an increase after treatments and values less than one indicate a decrease after treatments relative to bole-only harvesting with no compaction.…………………………………………………………………………...82 Table 6.2. Data of pre-harvest topsoil-nutrient content (kg/ha), current topsoil-nutrient content (kg/ha), and organic-matter properties summarized for each of the six sites...84  ix Table 6.3. Site-index values for Douglas-fir (Fd), white spruce (Sx) and lodgepole pine (Pl) based on pre-harvest data collected at each site.…………………………………87 Table A1. Stand-health metrics of the number of healthy trees, dead or dying trees, trees with foliar disease, trees with western gall rust occurrence and trees with symptoms of root disease within each treatment plot at the Interior Douglas-Fir (IDF) and the Sub-Boreal Spruce (SBS) zones. Treatments consist of bole-only harvesting (OM1), whole-tree harvesting (OM2), and whole-tree harvesting plus forest-floor removal (OM3). Soil-compaction treatments consist of no soil compaction (C0), light soil compaction (C1), and heavy soil compaction (C2).…………………………………………106-107 Table A2. Averaged mineral-soil fine-fraction bulk density (g/cm3) and total-soil carbon (kg/ha) within each treatment plot at the Interior Douglas-Fir (IDF) and Sub-Boreal Spruce (SBS) zones. Organic-matter removal treatments consist of bole-only harvesting (OM1), whole-tree harvesting (OM2), and whole-tree harvesting plus forest-floor removal (OM3). Soil-compaction treatments consist of no soil compaction (C0), light soil compaction (C1), and heavy soil compaction (C2)……….……108-109   x LIST OF FIGURES Figure 1.1. Relative importance of nutrients and microclimate (moisture and temperature) on stand productivity changes with time through stand development (Thiffault et al. 2011).…………………………………………...……………………15 Figure 1.2. Map of Biogeoclimactic Ecosystem Classification (BEC) zones in British Columbia with approximate locations of the six long-term soil productivity sites used in this study marked with black crosses. Sites named Topley, Skulow Lake and Log Lake are located in the Sub Boreal Spruce (SBS) zone and sites named Black Pines, Dairy Creek and O’Connor Lake are located in the Interior Douglas Fir (IDF) zone.……………………………………………………………………..……………16 Figure 1.3. Plot design of the Long-Term Soil Productivity sites with treatments of bole-only harvesting (OM1), whole-tree harvesting (OM2), forest-floor removal (OM3), minimal compaction (C0), light compaction (C1) and heavier compaction (C2)……17 Figure 1.4. Organic-matter removal treatments consist of bole-only harvesting (BOH), whole-tree harvesting (WTH) and forest-floor removal (FFR). The BOH treatment removes only logs. The WTH treatment removes logs, bark, branches, needles and crowns. The FFR treatment removes all aboveground biomass including the forest floor, thereby exposing the mineral soil.…………………………………………...…17 Figure 2.1. Pictures showing visual symptoms of (A) western gall rust (Endocronartium harknessii), (B) root-disease symptoms (mainly Armillaria ostoyae), (C) foliar disease Lophodermella concolor, and (D) foliar disease Elytroderma deformans.…………………………………………………………………………….28 Figure 2.2. In the Interior Douglas Fir (IDF) zone, whole-tree harvest (WTH) plots had significantly higher dead or dying trees compared to the forest-floor removal (FFR) plots (p = 0.0176). The bole-only harvest (BOH) plots were not significantly different from either the WTH or FFR plots.…………………………...………………………28 Figure 2.3. In the Sub Boreal Spruce (SBS) zone, total soil carbon (kg/ha) was negatively related to the number of dead or dying trees (R2 = 0.21, p = 0.0143). A linear best-fit line is shown.…………………………………………………………………….……29 Figure 2.4. Bulk density (g/cm3) is positively related to healthy trees in the Interior Douglas Fir (IDF) zone but negatively related in the Sub Boreal Spruce (SBS) zone. Sites are colour coded to visualize differences among sites. Linear best-fit lines and R2 values are displayed.……………………………………………………….…………29 Figure 2.5. Bulk density (g/cm3) is not related to healthy trees when both BEC zones are considered together. Sites are colour coded to visualize differences among sites. Linear best-fit line and R2 value are displayed.………………………………………30 Figure 3.1. Treatment-plot averages of disease occurrence were positively related tree growth at O’Connor Lake (circle), Dairy Creek (cross), and Topley (triangles). Average tree volume (cm3) was positively related to the percent of living trees with western gall rust occurrence (A), total number of disease occurrences (B), and the percent of living trees with foliar disease (C). Tree height increment (m/5 years) was positively related to the percent of living trees with western gall rust occurrence (D),  xi the percent of living trees with foliar disease (E), and the percent of living trees with root-disease symptoms (F).…………………………………………………………40 Figure 3.2. Random-group averages of disease occurrence were positively related tree growth at O’Connor Lake (circle), Black Pines (squares) and Log Lake (diamonds). Average tree volume (cm3) was positively related to the percent of living trees with root-disease symptoms (A). Height increment (m/5 yrs) was positively related to the total number of disease occurrence (B) and the percent of living trees with root-disease symptoms (C).……………………………………………………………………...…41 Figure 3.3. The balance of resource allocation to growth and secondary metabolites differs under optimum nutrition (left) and nutrient limitation (right). Arrow size represents relative amounts and is not based on real data. Secondary metabolites are related to constitutive and induced tree defence against disease and insect attack. Stand productivity is a product of both tree health (disease and insect occurrence) and tree growth.……………………………………………………………………………...…42 Figure 4.1. Images at Black Pines, plot 4, in the original red-green-blue (RGB) format before manipulations (left), excess greenness (EG) values (middle) and green chromatic coordinate (GCC) values (right). EG and GCC images are grey-scale representations with lighter hues being higher values indicating a higher proportion of green light being reflected.……………………………………………………………52 Figure 4.2. An image displaying the excess greenness representation for Black Pines plot 1. Analysis was conducted on regions of interest (ROIs) at the individual tree crown scale (left) and plot scale (right). Red polygons depict the ROIs used……………….52 Figure 4.3. Relationships between the percentage of trees with healthy vigour and green chromatic coordinate (GCC) and excess greenness (EG) were significantly positive: GCCplot (p < 0.0001), GCCtree (p < 0.0001) and EGplot (p = 0.0083). Sites are colour coded and linear-regression equations are shown.……………………….……53 Figure 4.4. Relationships between tree mortality and green chromatic coordinate (GCC) and excess greenness (EG) were significantly negative: GCCplot (p = 0.0002) and GCCtree (p < 0.0001). Sites are colour coded and linear-regression equations are shown.……………………………………………………………………………...…54 Figure 4.5. Relationships between foliar-disease occurrence and green chromatic coordinate (GCC) and excess greenness (EG) were significantly negative: EGplot (p = 0.0001), EGtree (p = 0.0029), GCCplot (p = 0.0334) and GCCtree (p = 0.0478). Sites are colour coded and linear-regression equations are shown.………………..……55-56 Figure 4.6. Relationships between root-disease occurrence and green chromatic coordinate (GCC) and excess greenness (EG) were significantly negative: EGplot (p < 0.0001) and GCCplot (p = 0.0083). Sites are colour coded and linear-regression equations are shown.……………………………………………………………….…56 Figure 4.7. Relationships between western-gall-rust occurrence and green chromatic coordinate (GCC) and excess greenness (EG) were significant. The relationship with EGtree (p = 0.0002) was negative, while with GCCtree (p < 0.0001) and GCCplot (p < 0.0001) were positive. Sites are colour coded and linear-regression equations are shown.………………………………………………………………………...………57  xii Figure 4.8. Relationships between the percentage of trees with healthy vigour and green chromatic coordinate calculated at a plot scale (GCCplot) at high, medium and low ranges of leaf-area index (LAI).………………………………………………………58 Figure 4.9. Relationships between the percentage of trees with healthy vigour and green chromatic coordinate calculated at a plot scale (GCCplot) at high, medium and low ranges of foliar nitrogen (%).…………………………………………………………58 Figure 5.1. A conceptual diagram of relationships tested in this chapter. Forest-floor removal treatments could influence topsoil-nutrient content, ectomycorrhizae relative abundance, Suillus relative abundance, and foliar-nutrient concentrations. Theses variables, in turn, interact tree growth and health.……………………………………76 Figure 5.2. Crown cover (%) based on TIPSY model (black) and from collected data at each site: Black Pines (BP, red), Dairy Creek (DC, orange), O’Connor Lake (OC, yellow), Log Lake (LL, green), Skulow Lake (SL, blue) and Topley (TO, purple).………………………………………………………………………..………77 Figure 6.1. Lodgepole pine height (A) and average tree volume (B) at ages 1, 3, 5, 10, 15 and 20 averaged for each site: Black Pines (red), Dairy Creek (orange), O’Connor Lake (yellow), Log Lake (green), Skulow Lake (blue) and Topley (purple).…...……83 Figure 6.2. Discriminant analysis using pre-harvest data of net merchantable volume (m3), topsoil carbon (kg/ha), topsoil N (kg/ha), topsoil mineralizable N (kg/ha), topsoil P (kg/ha) and topsoil K (kg/ha) averaged per plot data to differentiate sites: Black Pines (red), Dairy Creek (orange), O’Connor Lake (yellow), Log Lake (green), Skulow Lake (blue) and Topley (purple). Normal 50% contour lines delineate the groups.…………………………………………………………………………...……85 Figure 6.3. Discriminant analysis using current (year 15 and 20) data of total stem volume (m3), topsoil carbon (kg/ha), topsoil N (kg/ha), topsoil mineralizable N (kg/ha), topsoil P (kg/ha) and topsoil K (kg/ha) averaged per plot data to differentiate sites: Black Pines (red), Dairy Creek (orange), O’Connor Lake (yellow), Log Lake (green), Skulow Lake (blue) and Topley (purple). Normal 50% contour lines delineate the groups.………………………………………………………………………….…86 Figure 6.4. Canopy closure (%) based on TIPSY model (red) and real data (blue) from each site.………………………………………………………………………………88 Figure 6.5. Average tree height (m) based on TIPSY model (red) and real data (blue) from each site.……………………………………………………………………...…89 Figure 6.6. Stand density (stems/ha) based on TIPSY model (red) and real data (blue) from each site.……………………………………………………………………...…90     xiii  LIST OF ABBREVIATIONS BC – British Columbia BOH – bole-only harvest CC – canopy closure or crown closure (%) EG – excess greenness EGplot – excess greenness calculated on the spatial scale of a treatment plot EGtree – excess greenness calculated on the spatial scale of individual tree crowns FFR – forest-floor removal GCC – green chromatic coordinate GCCplot – green chromatic coordinate calculated on the spatial scale of a treatment plot GCCtree – green chromatic coordinate calculated on the spatial scale of individual tree crowns ha – hectare kg – kilogram LAI – leaf area index LTSP – long-term soil productivity PCA - principal components analysis TASS – tree and stand simulator TIPSY - table interpolation program for stand yields  US – United States WTH – whole-tree harvest   xiv  ACKNOWLEDGEMENTS I appreciate the support and feedback from my supervisory committee (Cindy Prescott, Bill Chapman, Marty Kranabetter, and Sue Grayston). Bill Chapman initiated this project and put the funding in place to carry out the research. Without him, this research would not have been possible. Cindy Prescott provided many patient hours helping me organize ideas and editing written work. Marty Kranabetter’s help with data and data analysis was invaluable. Sue Grayston was a dependable sounding board and encourager of interesting ideas. I am also grateful for the input on specific details of this thesis provided by Nicholas Coops, Tom Veblen, Richard Hamlin, Shaw Mansfield, Bill Mohn, Shannon Berch, Kendra Mitchell, Jacynthe Masse, and Jon Leff. I also thank Peter Ott for statistical advice.   I owe particular thanks to the British Columbian Ministry of Forests, Lands and Natural Resource Operations and its employees that established and monitored the Long-Term Soil Productivity sties over the last 20 years. Their hard work and forward thinking enabled the current testing of critical questions on how forest management influences stand condition using standardized research facilities. This research would not have been possible without the financial support from the National Science and Engineering Research Council (NSERC), the BC Ministry of Forests, Lands and Natural Resource Operations, and the University of British Columbia.  I would also like to thank my family, friends, fellow graduate students and Hayes Zirnhelt for their continued moral support. Thank you for your contributions towards my PhD experience.     1 Chapter 1. Introduction This chapter provides a background on the literature relevant to this area of study. From this foundation, the thesis objectives and organization are stated, providing an overview of the scope of and approach to this project. Finally, there is a description of the study sites and species.   1.1 Background Forests provide humanity with many goods and services. Products from the forest sector include lumber, pulp and paper, raw logs, and bioenergy fuel sources (Barnes 2015). In British Columbia (BC) the forest sector accounted for 30% of the total manufacturing sales, 35% of total merchandise exports, and 60,700 jobs in 2014 (Barnes 2015). In addition to these industrial products, forests provide the services of regeneration after disturbance, primary production, nutrient cycling, carbon storage (climate regulation), natural hazard regulation (e.g. floods), insect and pathogen regulation, wildlife habitat, temperature regulation, erosion control, water retention, recreation, food (e.g. berries and mushrooms), biodiversity, and rangeland for cattle (Turner et al. 2013 and references therein). These services are estimated to have a much larger monetary value than the industrial products sold (Costanza et al. 1997). Forests also have intrinsic value by contributing to happiness, peace of mind, recreation, aesthetics, culture, spiritual health, and support of biodiversity. 1.1.1 Forest health A primary goal of forest management is to ensure forest operations do not reduce short- or long-term productivity, while maintaining other values. This is becoming more challenging with increased demand for forest resources, increased use of forested land, and increased external stresses on forests. Declines in forest health have become more widespread and severe (van Mantgem et al. 2005) and are projected to increase by 2% per year (Mickler 1996). Forest health is a term used in many contexts to represent many levels of forest attributes. In this thesis the more traditional use of the term ‘forest health’ is used to encompass pest occurrence, tree mortality, and loss of tree vigour. Forest health encompasses the broad context and meaning of this subject; whereas, ‘stand health’ or ‘tree health’ describe the measures used in this research. Over the last 100 years in North America, the occurrence of Swiss needle cast, several root-rot pathogens, fusiform rust, southern pine beetle, western spruce budworm and eastern spruce budworm have increased (Schowalter and Turchin 1993; Wickman et al. 1993; Castello et al.  2 1995). In the face of these stresses, it is necessary to understand the factors that determine forest resilience and how disturbance during logging activities may influence forest resilience. Forest resilience to disturbance depends on soil characteristics that vary greatly among locations (Page-Dumroese et al. 2000). Assessment of the effect of forest management on forest productivity has focused on measures of tree growth, with measures of forest health and ecosystem processes being secondary. This has created a significant gap in our understanding of how management influences productivity by changing forest health and ecosystem function.   Damaging insects and disease (collectively pests) influence large areas of forest globally (van Lierop et al. 2015). In the United States (US), forest pests kill 84.6 million cubic meters of timber annually, accounting for 25% of the total annual harvest (Johnston and Crossley 2002). In BC, forest pests kill approximately 25% of the annual-allowable cut (BC Ministry of Forest and Range 2011). This is twelve-times more mortality than caused by forest fires (BC Ministry of Forest and Range 2011). In addition to tree mortality, forest pests are estimated to reduce total annual harvest by an additional 25% due to reduced growth rates (Mickler 1996). In loblolly and slash pine forests, 47 million US dollars worth of timber is lost annually to fusiform rust alone (Mickler 1996). Cascading ecological impacts can result from severe forest pest outbreaks. For example, the reduction of whitebark pine (Pinus albicaulis) forests due to the exotic fungal pathogen white pine blister rust (Cronartium ribicola) impacts species such as the Clark's nutcracker (McKinney et al. 2009) and the grizzly bear (Reinhart et al. 2001).   More intensive monitoring of forest health has been recommended because of its significant impact on economic and ecological values (Woods and Coates 2013). There are three main ways to monitor forest health: ground-based visual assessments, aerial sketch mapping and image analysis. Ground-based assessments are the most detailed because pest species can be precisely identified, but they are time consuming and expensive over large areas. The sketch-mapping technique captures visual signs of pest occurrence by a person looking out the window of a small plane, which is then recorded onto a paper map. Large areas can be covered with this technique, but there are challenges with consistency among observers and precisely defining the boundaries of affected areas. Analysis of images captured from planes or satellites can cover large areas efficiently, methods can be standardized and interpretations can be conducted subjectively as well  3 as objectively. For example, a measure of plant condition can be determined from the images by calculating the relative proportions of reflected wavelengths, called spectral-reflectance indices.  Trees that currently have fast growth rates and are healthy may not necessarily be well defended against future insect or disease attack. In British Columbia, forest companies are required to ensure that the logged area meet free-to-grow standards after they have logged publically owned land (Forest and Range Practices Act 2002). Free-to-grow standards are primarily based on stem density and tree height estimated 10 to 20 years after harvesting (Forest Practices Branch 2014). If free-to-grow standards are met, then the responsibility of managing that land reverts back to the government with the assumption that productivity will remain until the next harvest (Forest and Range Practices Act 2002). Plant physiology suggests that trees that are currently growing fast may not be well protected against future pest attack. In all plants, there is a continuous and dynamic allocation of resources to critical life functions: growth, defence, reproduction, storage and maintenance (Herms and Mattson 1992). In situations where resources are abundant and stress is low, resources will preferentially be allocated to growth (Herms and Mattson 1992). If pest occurrence is low in these situations, it is possible to have good growth and healthy trees. However, these trees may be poorly defended against future pest attack because few resources are being allocated to defence (Herms and Mattson 1992). Although the relationships between growth and health can therefore provide interesting insight into this physiological process, it is difficult to test in British Columbia because of the current system of forest management that results in little post free-growing data on both growth and health at the same sites.  Defence against pests could be enhanced when tree growth is limited by nutrients. In situations where nutrients limit growth, photosynthesis still occurs creating ‘extra’ photosynthate that can be allocated to defence against pests (Herms and Mattson 1992). This provides a mechanism to explain why plants with growth-limiting nutrient status can have higher chemical defense (Donaldson et al. 2006, Osier and Lindroth 2006, Sampedro et al. 2011) and anatomical defense (Moreira et al. 2008), relative to plants with adequate nutrients. It also explains why in a southern US loblolly pine plantation after fertilization and weed control was conducted to increase tree growth southern pine coneworm occurrence also increased (Nowak and Berisford 2000). Understanding tree nutrition in relation to tree growth and resistance to pest attacks is therefore a  4 key area of research because of the physiological processes that regulate both growth and defence (Herms and Mattson 1992; Andersson et al. 2000).  The omission of forest health issues in monitoring forest productivity may cause predictive models to overestimate timber production. In BC, growth-and-yield models use tree growth data collected at stand-development monitoring plots to predict future timber supply. This information is used to inform sustainable annual harvest levels. The most common growth-and-yield platform in BC is the Table Interpolation Program for Stand Yields (TIPSY) interface that is based on the Tree And Stand Simulator (TASS) model (Mitchell et al. 2000). The input data for these models are based on tree growth, stand density and site index variables. Monitoring plots were omitted from the database if disease or insect damage is high (Woods and Coates 2013). This model assumes a period of low-stable mortality, with limited damage from pests, to the dominant trees growing within the plantations (Skovsgaard and Vanclay 2008; Puettmann et al. 2009). The validity of these monitoring practices and model assumptions have been challenged with data from lodgepole-pine stands in the Okanagan region having five-times lower tree density than the TIPSY model predicted (Woods and Coates 2013). The discrepancy between model predictions and actual stand densities was primarily related to pest damage to dominant trees (Woods and Coates 2013). 1.1.2 Monitoring forest productivity The effects of forest management activities on forest productivity are also monitored primarily with measures of tree growth. Organic-matter removal and soil compaction during harvesting are two management actions that have been predicted to affect forest productivity by reducing soil nutrient pools, water infiltration and root growth (Powers et al. 1990; Sayer 2006; Powers 2006; Veteli et al. 2006; Condron et al. 2010). The Long-Term Soil Productivity (LTSP) projects monitors the effect of organic-matter removal and soil compaction on forest productivity at 120 sites throughout North America (Powers 2006). Stand productivity is monitored on these sites by measuring tree growth, soil nutrients and foliar nutrients at these sites (Powers 2006) with no standardized measures or published results on forest health.  Organic-matter removal can influence stand productivity via changes to microclimate, water relations, nutrient pools, nutrient availability, and the soil microbial community. Organic-matter  5 removal or displacement can occur during disposal of logging residue, site preparation, road construction, and inadvertently during timber harvesting activity.  In some parts of the world biomass-to-bioenergy projects remove certain elements of organic residue and those types of activities are in the early stages of development in Canada. Logging residue disposal and biomass to bioenergy projects may remove more fine-organic material (branches, foliage, crowns, slash, and bark) compared to just removing logs. The fine-organic matter can contain higher concentrations of certain nutrients. Mechanical site preparation techniques remove or displace the forest floor to enhance planting sites. Microclimate has the largest influence on stand productivity at the early stages of stand development (Figure 1.1). By removing organic matter, soil temperature and moisture become more variable, which can either increase or decrease tree growth depending on the climate. In cold areas, organic-matter removal can increase productivity by increasing soil temperature (Laiho and Prescott 2004). For example, in Alaska, the accumulation of forest floor reduced productivity likely due to low temperatures and reduced aeration (Van Cleve and Dyrness 1983). In dry climates, however, organic-matter removal can reduce productivity by reducing soil moisture content (Ginter et al. 1979; Kamaluddin et al. 2005; Sayer 2006).   Nutrient supply becomes most important around the time of canopy closure when nutrient demands are high. Organic-matter removal directly removes nutrients, but changes to nutrient availability are more complex. Relative to bole-only harvest, whole-tree harvest and forest-floor removal can increase nutrient removal by 50 to 400% (Mann et al. 1988; Smith et al. 1986). This is a large increase in nutrient removal relative to bole-only harvest, but is not necessarily a large a proportion of the total ecosystem nutrient pool. For example, the total nitrogen pool would be reduced by 4 to 13% during whole-tree harvesting and by 7 to 21% during forest-floor removal (Table 1.1). In a 450-year-old Douglas-fir forest, the total phosphorus pool would be reduced by 43% during whole-tree harvesting and by 76% during forest-floor removal (Table 1.2). In the same forest, the total calcium and potassium pools would be reduced by approximately 20% during whole-tree harvesting and by approximately 30% during forest-floor removal (Table 1.2). The availability of nutrients may be more important to tree nutrition than total pool yet is difficult to assess because of the influence of multiple dynamic factors: ectomycorrhizae abundance and composition, nutrient turnover rates, root structure and abundance, nutrient forms, atmospheric  6 inputs, translocation within trees, throughfall, stem flow, weathering rates, climate, soil fauna (shredders), litter quality and leaching.   Organic-matter removal can also influence forest productivity indirectly through changes to the soil microbial community. At six LTSP sites in BC planted with lodgepole pine, the soil microbial community structure and diversity was significantly different between organic-matter removal treatments 15 to 20 years after harvesting (Hartmann et al. 2012). Ectomycorrhizal fungi were among the most influenced by organic-matter removal treatments (Hartmann et al. 2012) having implications for forest productivity. Ectomycorrhizal fungi are tree mutualists that provide nutrients in exchange for carbon. Ectomycorrhizal abundance is generally reduced by organic-matter removal and soil disturbance leading to the prediction that stand productivity would be reduced (Perry et al. 1989). However, Suillus species of ectomycorrhizae are adapted to disturbance and can enhance nitrogen acquisition. Nitrogen-fixing bacteria harboured within Suillus species (Paul et al. 2013) can fix nitrogen at similar rates to alder (Alnus) species (Paul et al. 2007), allowing lodgepole pine to acquire adequate amounts of nitrogen, even when growing in gravel pits (Chapman and Paul 2012). Although this nitrogen supply is necessary for growth, it is typically associated with a high-energy cost relative to accessing labile nitrogen pools. The relationships between the fungal community, treatments, tree growth, and tree health at LTSP sites have not been fully explored.   Soil compaction increases soil bulk density by decreasing pore space, which can influence tree growth through changing soil moisture, nutrient availability, root growth, and soil aeration. Soil compaction mainly occurs on roads, landings and skid trails, but the entire harvested area can also be affected (Grigal 2000). Roads, landings and skid trails can cover between 1.5 to 24% of the harvested area with an average of approximately 8% (Grigal 2000). The severity of disruption from compaction varies depending on logging equipment, topography, soil moisture, soil texture, and road specifications (Grigal 2000). Management practices that require more frequent entry to the site with heavy equipment can compact the soil if used incorrectly.  Data collected during the first 20 years following treatment initiation and planting on the LTSP sites suggest that treatment effects on tree growth are small. Considering 42 LTSP sites across  7 North America, Fleming et al. (2006) found that productivity of 5-year-old stands did not differ between whole-tree harvest and bole-only harvesting plots. The 10-year summary of the LTSP sites also showed generally weak effects, except for reduced productivity on whole-tree harvested plots with severe compaction (Ponder et al. 2012). However, in North Carolina and Louisiana, even forest-floor removal and heavy compaction had no significant effect on 10-year-old loblolly pine volume (Sanchez et al. 2006). Kranabetter et al. (2006) found no strong effect of forest-floor removal on tree growth or foliar nutrition of 12-year-old lodgepole pine and interior spruce when there was no compaction. However, tree growth was reduced when forest-floor removal was combined with both light and severe soil compaction (Kranabetter et al. 2006). There were no significant differences in tree height or diameter between whole-tree and bole-only harvested plots in 10- to 14-year-old US northern hardwood forests (n = 13, Roxby and Howard 2013). It has been hypothesized that the weak treatment effects are due to these stands having not yet reached canopy closure (Thiffault et al. 2011; Ponder et al. 2012). Treatment effects are predicted to become apparent after canopy closure when nutrient demand is highest (Figure 1.1; Thiffault et al. 2011; Ponder et al. 2012).   Results from the first 20 years of the LTSP project have demonstrated site- and species-specificity of effects of soil disturbance on productivity. The effect of organic-matter removal treatments on tree growth depends on soil fertility, tree species, stand age, and climate. Whole-tree harvesting reduced productivity of 10-year-old Douglas-fir stands on poor soils, but not on rich soils (Compton and Cole 1991). Similarly, whole-tree thinning removals were identified as having greater risk or causing nutrient limitations on coarse-textured low fertility soils compared to fine-textured high fertility soils (Page-Dumroese et al. 2010b). Although treatments had relatively small effects on conifer species, forest-floor removal treatments reduced 5-year-old aspen biomass by approximately 50% on loamy sand soils in northern Minnesota (Stone and Elioff 1998). Treatment differences became apparent in Norway spruce trees only after 10 years and in Scots pine trees after 15 years (Egnell and Leijon 1999), with the main differences occurring between ages 8 and 12 years (Egnell 2011). After 24 years, bole-only harvesting plots had 20% more wood biomass than the whole-tree harvest plots (Egnell and Valinger 2003). Forest-floor removal increased tree growth in Mediterranean climates, but reduced tree growth in warm-humid productive climates thought to be related to temperature and nutrient effects  8 (Fleming et al. 2006).   The effect of soil compaction on tree growth depends on soil texture and forest-floor removal. On coarse-textured soils, compaction can increase seedling growth and survival by increasing moisture retention, reducing vegetation competition, and increasing late growing-season soil temperatures (Fleming et al. 2006). Similarly, moderate compaction of coarse-textured soils increased growth and survival of lodgepole pine and Douglas-fir seedlings when litter and organic matter remained on the site (Tan et al. 2009). The growth differences in these 2- and 3-year-old seedlings were attributed to increased water-holding capacity after compaction (Tan et al. 2009). Alternatively, compaction can reduce tree growth on fine-textured soils (Page-Dumroese et al. 2006). Compaction on fine-textured soils reduces stand growth through limiting the movement of air, water, nutrients, and roots (Brussard and van Faasen 1994; Thibodeau et al. 2000; Grigal 2000; Bulmer and Simpson 2005; Page-Dumroese et al. 2006; von Wilpert and Schaffer 2006). Soil compaction can restrict root growth leading to decreased water and nutrient adsorption (Nambiar and Sands 1992; Sheriff and Nambiar 1995). Growth reductions induced by compaction have been recorded to persist in Pinus ponderosa (Helms et al. 1986), Pseudotsuga menziesii (Stewart et al. 1988) and Pinus radiata (Firth and Murphy 1989). On fine-textured soils, the negative consequences of compaction due to altered soil physical properties are not easily repaired, remain for relatively long time periods and are of high certainty (Grigal 2000). The effect of compaction on tree growth is most severe when the forest floor is removed (Kranabetter et al. 2006; Ponder et al. 2012). 1.1.3 Management effects of forest health It is unknown whether the LTSP treatments influence forest pest occurrence, although it is possible for changes to organic matter and soil porosity to influence forest pest occurrence (Fellin 1980; Cheng and Igarashi 1987; Takahashi 1994; Zhong and van der Kamp 1999; Mori et al. 2004; Munck and Stanosz 2008; Oblinger et al. 2011). Organic-matter removal can influence pest occurrence directly by changing pest abundance, pest habitat, and pest resources; and indirectly by changing host resistance and pest antagonists. Whole-tree harvesting and forest-floor removal could reduce pests by directly removing individuals, their offspring and their habitat. Coarse woody debris (CWD) can influence bark-beetle occurrence by changing their food supply,  9 feeding behaviour, shelter, strength of competition, vulnerability to predation, reproductive success, oviposition, and dispersal (Fellin 1980). Large amounts of fresh logging slash or thinning debris can increase the occurrence of bark and engraver beetles by providing suitable substrate for adults to lay eggs (Fellin 1980).   Mortality from root pathogens could be influenced by CWD removal. Mori et al. (2004) considered germination and survival of Abies mariesii, Abies veitchii, Picea jezoensis var. hondoensis and Tsuga diversifolia seedlings on different substrates in an old-growth forest. Although germination occurred on all substrates, survival was significantly higher on woody debris in mid-decay classes, resulting in older seedlings occurring more commonly growing on woody debris (Mori et al. 2004). Similarly, significantly more Abies sachalinensis seedlings were found growing on woody debris than on the forest floor (Takahashi et al. 2000). The seedlings that did occur on the forest floor were stunted and rarely survived to grow taller than 15 cm (Takahashi et al. 2000). This phenomenon is likely related to the high concentration of root pathogens in the forest floor. The pathogen Racodium therryanum is abundant in the forest floor and A horizon, but is scarce in CWD and mineral soil (Cheng and Igarashi 1987; Takahashi 1994). In this way, woody debris may provide ‘safe sites’ for seedling establishment, especially in the first year of growth (Mori et al. 2004). Zhong and van der Kamp (1999) also found seed mortality from fungal pathogens (mainly black mould, Rhizoctonia sp.) to be highest on the forest floor and much lower on CWD.   These phenomena suggest that forest-floor removal could reduce pathogen occurrence; however, pathogen antagonists have to be considered. Contact with the forest floor was observed to reduce Diplodia shoot blight inoculum on decomposing branches (Oblinger et al. 2011). Similarly, foliar pathogen occurrence was reduced when litter was touching the soil surface in red pine nurseries (Munck and Stanosz 2008). This could be related to antagonistic decomposer fungi colonizing the resource and outcompeting the pathogenic fungi. Organic-matter removal could reduce the abundance of saprotrophic fungi, such as Hypholoma, which are antagonistic to pathogens like Armillaria root disease (Chapman et al. 2004). Antagonistic nematodes were more effective at reducing pine weevil abundance in seedlings growing in the forest floor compared to seedlings on bare mineral soil (Williams et al. 2013).  10  Soil compaction can influence pest occurrence. Armillaria root rot increased with more frequent harvesting entries (Kile 2000), possibly due to soil compaction and tree damage during harvesting (Jurskis 2005). A greenhouse trial with American chestnut seedlings tested the role of soil compaction on the occurrence of Phytophthora root rot (Rhoades et al. 2003). Soil moisture and compaction significantly increased seedling mortality due to Phytophthora root rot with the highest mortality in wet compacted soils (Rhoades et al. 2003).    1.2 Thesis objectives and organization The effect of soil disturbance on tree growth is complex but relatively well understood; yet the effects on other indicators of forest health are rarely studied. Other indicators of forest health include pest occurrence (Jurskis 2005), leaf area, and leaf pigment (McLaughlin and Wimmer 1999; Makinde and Salami 2013). Integrating research on tree growth and forest health is necessary to optimize forest productivity because pest occurrence reduces forest productivity and because processes that control tree growth also control tree defence (Herms and Mattson 1992; Andersson et al. 2000). Existing trials that monitor the effect of management on long-term forest productivity provide an opportunity to integrate forest health into forest-productivity monitoring more rapidly than starting new trials.   The overall objective of this research is to integrate measures of forest health into monitoring forest productivity after soil disturbance during harvesting. The benefit of this information is not to reduce pest occurrence, but to better be able to predict timber supply and effects of forest management activities. This thesis is organized into four main research chapters that ask: 1. Do soil-disturbance treatments significantly influence stand-health metrics? 2. What is the relationship between measures of tree growth and stand health? 3. Can spectral-reflectance indices monitor ground-based stand-health metrics?	4. How does forest-floor removal influence soil nutrient pools, ectomycorrhizal fungi and foliar nutrients? Do these variables relate to measures of tree growth and health? And how close are these sites to canopy closure when nutrient limitations are predicted to become apparent? 	 11  These questions are addressed by using six Long-Term Soil Productivity (LTSP) sites in British Columbia, Canada planted with lodgepole pine 15 to 20 years ago. The LTSP project was created to determine how organic-matter removal and soil compaction affect forest productivity. The research presented here contributes directly to this collaboration by providing novel data on forest health and integrating it with existing data on tree growth, soil chemistry, fungal community and foliar nutrients. The concluding chapter provides a synthesis that links the research results to larger questions in forest management related to forest resilience, context-dependency of management practices and future timber supply.  1.3 Study sites and species This research was conducted on six long-term soil productivity (LTSP) sites in the interior of British Columbia, Canada. The majority (71%) of harvesting in BC occurs in the interior of the province (Barnes 2015). Three sites are located in the Sub-Boreal Spruce (SBS) biogeoclimatic ecosystem classification (BEC) zone and three in the Interior Douglas-Fir (IDF) BEC zone (Figure 1.2). At the time of this study, the trees at the SBS sites were approximately 20 years old and the trees at the IDF sites 15 years old. Log Lake, Skulow Lake and Topley are in the SBS and Black Pines, Dairy Creek and O’Connor Lake are in the IDF (Figure 1.2). Mean annual temperature (MAT) ranged from 2 to 4.9 ˚C and mean annual precipitation (MAP) ranged from 418 to 738 mm (Table 1.3). Elevations ranged from 785 to 1180 m above sea level and net merchantable volume (NMV; a measure of site productive capacity) ranged from 142 to 448 m3/ha (Table 1.3). The three IDF sites have Brunisolic Gray Luvisols with a silt loam texture (Table 1.3). The three SBS sites have greater variability and geographic distribution (Table 1.3 and Figure 1.2). The Skulow Lake and Topley sites have Orthic Gray Luvisols with loam textures and the Log Lake site has a Gleyed Humo-Ferric Podzol with a silt loam texture (Table 1.3). All six sites are located on morainal blanket landforms and hemimor humus forms (Berch et al. 2010). Berch et al. (2010) and Holcomb (1996) provide full site descriptions and establishment reports.   The LTSP sites have a full-factorial and blocked design with three levels of organic-matter removal and three levels of soil compaction, resulting in nine treatment plots per block (Figure  12 1.3). Organic-matter removal treatments consist of bole-only harvesting (OM1), whole-tree harvesting (OM2), and whole-tree harvesting plus forest-floor removal (OM3). In the bole-only harvesting treatment, boles were removed but tree crowns, branches, needles, felled woody debris, herbaceous understory, and the forest floor were retained (Figure 1.4). This is most similar to ‘conventional’ ground-based harvesting techniques used in the BC Interior. In the whole-tree harvest treatment, whole-trees were removed (bole, branches, crowns and needles) but herbaceous understory, and forest floor (including well decomposed woody debris) were retained (Figure 1.4). In the forest-floor removal treatment, whole trees were harvested and all aboveground biomass including the forest floor was removed resulting in exposed mineral soil (Figure 1.4). Soil-compaction treatments consist of no soil compaction (C0), light soil compaction resulting in a 2 cm surface elevation drop (C1), and heavy soil compaction resulting in a 4 cm surface elevation drop (C2). Each treatment plot is 40 by 70 m (0.28 ha) with 100 core research trees, of two species, planted in ten by ten blocks and marked with aluminum tags. Initial planting density was approximately 1600 stems per hectare with infilling of natural seedlings being discouraged (Holcomb 1996). Currently these sites have approximately 1100 to 1300 stems per hectare stocking density (Chapter 6).   At all six sites, containerized lodgepole pine (Pinus contorta Douglas ex Loudon) seedlings were planted (Holcomb 1996). Lodgepole pine is a tall slender tree that grows in most of BC’s interior (Parish 1994). It is a two-needle pine with cones of various shapes from 2 to 4 cm in length (Parish 1994). Of all trees planted in BC, 53% are lodgepole pine (Weaver 2013). Lodgepole pine is a diploxylon member of the genus Pinus, which are able to rapidly colonize after disturbance even when nutrients or water are limited (Ricklefs 1990).   13 1.4 Tables and figures  Table 1.1. Nitrogen amount (kg/ha) and percent (%) distribution within ecosystem components in Douglas-fir stands that are 36, 47 and 450 years old in Washington and Oregon.   36 years1 47 years2 450 years1 450 years3 Component N % N N % N N % N N % N Live trees 288 8.7 605 4.1 502 9.5 566 9.9 Understory 6 0.2 5 0.0 58 1.1 14 0.2 CWD 14 0.4 74 0.5 132 2.5 - - Total removed in WTH 308 9.3 684 4.6 692 13.1 580 10.1          Forest floor 161 4.9 453 3.1 434 8.2 445 7.8 Total removed in FFR 469 14.2 1137 7.7 1126 21.3 1025 17.9          Roots 32 1.0 205 1.4 162 3.1 140 2.4 Mineral soil 2809 84.9 13143 89.8 4300 81.1 4560 79.7 1 Modified from Table 7.2 in Johnson et al. (1982). 2 Modified from Table 23 in Ares et al. (2007). 3 Modified from the Appendix in Cole and Rapp (1980).    Table 1.2. Nutrients pools (kg/ha) and percentages calculated in each ecosystem components of a 450-year-old Douglas-fir forest in Oregon, USA (Cole and Rapp 1980). Component N % N P % P Ca % Ca K  % K Foliage and branches 217 3.8 50 25.5 286 8 141 14.3 Boles 349 6.1 36 18.4 401 11.2 48 4.9 Total removed in WTH 566 9.9 86 43.9 687 19.2 189 19.2          Understory vegetation 14 0.2 2 1 2 0.1 10 1 Forest litter layer 445 7.8 62 31.6 619 17.3 80 8.1 Total removed in FFR 1025 17.9 150 76.5 1308 36.6 279 28.3          Roots 140 2.5 12 6.1 225 6.3 50 5.1 Soil rooting zone 4560 79.7 34 17.4 2040 57.1 660 66.7        14 Table 1.3. Black Pines, Dairy Creek and O’Connor Lake sites were located in the Interior Douglas Fir biogeoclimactic ecosystem classification (BEC) zone. Log Lake, Skulow Lake and Topley sites were located in the Sub Boreal Spruce BEC zone. Information at each site is provided: latitude (˚), longitude (˚), elevation (m), year planted, mean annual temperature (MAT), mean annual precipitation (MAP; Wang et al. 2012), soil texture, and dominant soil classification (Berch et al. 2010).  Interior Douglas Fir Zone  Black Pines Dairy Creek O’Connor Lat. (°) 50 56 25 N 50 51 10 N 50 53 40 N Long. (°) 120 17 45 W 120 25 10 W 120 21 10 W Elevation (m) 1180 1150 1075 Year planted 1998 1997 1999 MAT (˚C) 4.9 4.5 4.7 MAP (mm) 434 453 418 Soil texture silt loam silt loam silt loam Dominant soil    classification Brunisolic Gray Luvisol Brunisolic Gray Luvisol Brunisolic Gray Luvisol    Sub Boreal Spruce Zone   Log Lake Skulow Lake Topley Lat. (°)  54 21 58 N 52 18 57 N 54 36 35 N Long. (°)  122 36 42 W 121 54 54 W 126 18 39 W Elevation (m)  785 1050 1100 Year planted  1991 1993 1992 MAT (˚C)  2.7 3.5 2 MAP (mm)  738 518 532 Soil texture  silt loam/loam loam loam/clay loam Dominant soil    classification  Gleyed Humo-Ferric Podzol Orthic Gray Luvisol Orthic Gray Luvisol      15  Figure 1.1. Relative importance of nutrients and microclimate (moisture and temperature) on stand productivity changes with time through stand development (Thiffault et al. 2011).           16  Figure 1.2. Map of Biogeoclimactic Ecosystem Classification (BEC) zones in British Columbia with approximate locations of the six long-term soil productivity sites used in this study marked with black crosses. Sites named Topley, Skulow Lake and Log Lake are located in the Sub Boreal Spruce (SBS) zone and sites named Black Pines, Dairy Creek and O’Connor Lake are located in the Interior Douglas Fir (IDF) zone.     17  Figure 1.3. Plot design of the Long-Term Soil Productivity sites with treatments of bole-only harvesting (OM1), whole-tree harvesting (OM2), forest-floor removal (OM3), minimal compaction (C0), light compaction (C1) and heavier compaction (C2).    Figure 1.4. Organic-matter removal treatments consist of bole-only harvesting (BOH), whole-tree harvesting (WTH) and forest-floor removal (FFR). The BOH treatment removes only logs. The WTH treatment removes logs, bark, branches, needles and crowns. The FFR treatment removes all aboveground biomass including the forest floor, thereby exposing the mineral soil.    18 Chapter 2. Effect of soil disturbance on stand health  2.1 Synopsis • Although stand health influences stand productivity, the LTSP project does not routinely monitor stand health nor are there any published results on health issues at these instillations.  • Foliar disease (Lophodermella concolor and/or Elytroderma deformans), western gall rust (Endocronartium harknessii) and root disease (primarily Armillaria ostoyae) were found at all six sites.  • Within a site 25-66% of the trees had healthy vigour, 4-23% were dead or dying, 12-54% had high or medium foliar-disease severity, 2-64% had galls from western gall rust, and 5-21% had root-disease symptoms. • Organic-matter removal and soil compaction treatment effects on stand-health metrics differed among sites. • When soil carbon and bulk density were used as measures of organic-matter removal and soil-compaction treatments, respectively, we found stands with more soil carbon were generally healthier.  2.2 Introduction Although the term ‘stand health’ is commonly used, it has many different meanings. The traditional or utilitarian view of stand health is a forest free from damaging insects and diseases, has vigorous growth, is able to withstand some level of stress, and processes such as nutrient cycling and stand development occur at characteristic rates (Kimmins 2004). Alternatively, the holistic or environmental view of stand health is a mature productive forest community with an uneven canopy, many large trees, dead trees, abundance of dead wood and decaying logs and the presence of rare or specialist indicator species (Kimmins 2004). The forest stands studied here are in an early seral stage, and may never reach a late seral community due to logging. Therefore, this chapter, and the rest of the thesis, use the traditional definition referring to damaging insect and disease occurrence and overall tree vigour.    19 Stand health is a desirable forest attribute that forest management is legally mandated to maintain in British Columbia (1993 Forest Practices Code of British Columbia). There is significant research and monitoring of stand health, yet it is rarely considered in relation to forest productivity or management options. This is surprising given that tree growth (wood accumulation) and death or disease (wood loss) both contribute to forest productivity. In young managed forests that have the goal of biomass production, high levels of insect or disease damage, high mortality rates, and poor growth would be considered unhealthy. There is no published data on stand health at any of the LTSP sites, yet the LTSP sites present a perfect opportunity to monitor stand health as part of forest productivity.   This chapter documents stand health measured as damaging insect and disease occurrence, tree mortality rates, and overall tree vigour. These measures of stand health are then tested for treatment effects. This chapter addresses three main lines of questioning: 1. What are the main damaging insects and diseases? What are their occurrences? Is this occurrence within a ‘normal’ range?  2. Do organic-matter removal (OM) or soil compaction (C) treatments significantly affect stand-health metrics or their response to treatments?  3. Can soil carbon or bulk density be used to measure the influence of organic-matter removal and soil compaction treatment on stand health?   2.3 Methods In June 2013, preliminary site visits were conducted with the principal investigators of each LTSP site to become familiar with the sites and to identify forest pests. Stand health surveys were conducted from July to September 2013. The 100-core research trees in each plot were identified using the aluminum marking tags. The tree’s leader, needles, branches, stem, and root collar were inspected for macroscopic symptoms of pest occurrence. Pest identity, presence and severity were recorded. Foliar-disease severity was recorded based on the percent of the foliage affected. High severity was greater than 50% of the foliage affected, medium severity was 25 to 50% and low severity was less than 25%. Resinosis and lesions near the root collar were both considered as visual symptoms of root disease (Morrison et al. 1991). These symptoms could also be indications of Warren’s root collar weevil presence; however, this species was not identified at  20 these sites. Pictures and samples of representative pest occurrences and severities were taken. Pest identification and classification were summarized based on conifer damage agent and condition codes (http://www.for.gov.bc.ca/hfp/health/fhdata/FS747%20conifer%2020111003.pdf).   The presence of pathogenic fungal species in the soil was tested using pyrosequencing data from Hartmann et al. (2012). Sequences of fungal genes identified from soil samples at these sites were downloaded from Supplementary Data 2 file (Hartmann et al. 20121). Briefly, soils were collected in the same way as for nutrient analysis. Then deoxyribonucleic acids (DNA) were extracted using the FastDNA SPIN Kit and quantified using the Quant-iT Pico-Green kit (Hartmann et al. 2012). The 454-pyrosequencing GS-FLX Titanium technology was used to amplify and sequence the eukaryotic ribosomal hypervariable region of the internal transcribed spacers (ITS2) (Hartman et al. 2012). Raw sequences were submitted to national center for biotechnology information’s (NCBI) database using the basic local alignment search tool (BLAST). Operational taxonomic units (OTU) were clustered to 97% similarity and a representative sequence selected. Representative OTUs that did not match the UNITE database to 75% were removed (8% of the OTUs) because these are likely due to errors. Taxonomic information to species level was collected from UNITE. The species list generated by this search was inspected for known root pathogens of lodgepole pine.  To assess overall tree condition, a vigour classification rating was assigned to each tree ranging from 1 to 5 based on Newsome and Perry (2002). Vigour Class 1 indicates a tree with vigorous growth and is in good condition with no visual indication of stress. A Class 2 tree has a moderate growth rate and is in fair condition with some visual indication of stress, which may include minor defects. A Class 3 tree is in poor condition with obvious visual signs of severe stress, may have major defects, and has a poor growth rate. A Class 4 tree is moribund and 5 is dead.   Five metrics of stand health were developed from the stand health survey to test for treatment effects: the number of healthy trees (Class 1), the number of dead or dying trees (Classes 4 and 5), the number of trees with high or medium foliar disease, the number of trees with galls from                                                 1 Data available at www.nature.com.ezproxy.library.ubc.ca/ismej/journal/v6/n12/suppinfo/ismej201284s1.html  21 western gall rust and number of trees with root-disease symptoms. Although other pests occurred at some sites, foliar disease, western gall rust and root-disease symptoms were used for analysis because they occurred at all sites enabling comparisons and were the most abundant.  The response of each stand health metric to treatment levels was quantified using response ratios that quantify treatment effects by comparing a treatment mean relative to a reference mean. Response ratios were calculated by,   response ratio = treatment/reference   where reference is the bole-only no compaction treatment (OM1C0) and treatment is all other treatment plots with additional soil disturbance (Thiffault et al. 2011). Response ratios greater than one indicate a measure increased after additional soil disturbance and less than one indicate a measure decreased after additional soil disturbance. Bole-only no compaction plots were used as the reference to test the influence of additional levels of soil disturbance on stand-health metrics, as has been done for tree growth (McBride et al. 2011; Efroymson et al. 2013; Thiffault et al. 2011). No galls from western gall rust were visible in the reference plot at O’Connor Lake (Table A1) making it impossible to calculate the response ratio. Therefore, one western gall rust occurrence was added to each treatment plot. Response ratios are well suited to the LTSP sites where the magnitude of change varies between sites (Hedges et al. 1999; Fleming et al. 2006; Thiffault et al. 2011; Ponder et al. 2012).    Significant differences among treatments were tested using ANOVAs within each BEC zone. Organic-matter removal and soil-compaction treatments were tested separately. Tukey Kramer post hoc tests were used to determine how treatments differed.   Treatment effects were also tested using mineral-soil fine fraction bulk density (g/cm3) and soil carbon (kg/ha) in the top 20 cm of mineral soil and the forest floor2, which are direct measures of soil compaction and organic-matter removal treatments respectively (Hazlett et al. 2014). Fine                                                 2 Soil bulk density and total carbon originate from the BC Ministry of Forests, Lands and Natural Resource Operations and are from year 10 in the IDF and 15 in the SBS.   22 fraction bulk density was used for comparisons because these soils have varying levels of coarse fragments. Mineral-soil fine fraction bulk density (hereafter bulk density) is mineral-soil mass divided by mineral-soil volume of the fine fraction (< 2 mm). Total soil carbon (kg/ha) was calculated by,  = [CM (g/kg) * BDM (g/cm3) * TM (cm) * 1^5] + [CF (g/kg) * BDF (g/cm3) * TF (cm) * 1^5]  where C is the concentration of carbon, BD is bulk density, T is layer thickness and 1^5 is for unit conversion. Subscript M is for the mineral soil layer and F is for the forest floor layer.   Linear least squares regressions were used to compare soil carbon and bulk density to stand-health metrics. Thresholds and non-linear relationships were visually assessed using the plotted data. All analyses were conducted in JMP version 11 with results considered significant when p < 0.05 (SAS Institute Inc. 2012).  2.4 Results At all six sites, foliar disease (Lophodermella concolor and/or Elytroderma deformans), western gall rust (Endocronartium harknessii) and root-disease symptoms were present (Table 2.1 and Figure 2.1). The most common root disease in this region is Armillaria solidipes previously A. ostoyae (Morrison et al. 1991) and its presence has been confirmed at Skulow Lake (unpublished data). In addition to these common disease found at all sites some additional forest pests were found at specific sites. An insect defoliator occurred at Black Pines; Northern twig pitch nodule moths (Petrova albicapitana) occurred at O’Connor Lake and Skulow Lake; Sequoia pitch moths (Synanthedon sequoia) occurred at Log Lake and Skulow Lake; comandra blister rust (Cronartium comandrae) occurred at Skulow Lake and Topley; and stalactiform blister rust (Cronartium stalactiforme or coleosporioides) occurred at Log Lake and Skulow Lake. Mortality caused by forest pests including Cronartium comandrae, some severe cases of Elytroderma deformans, and western gall rust on stems was observed. The percentage of trees per site that were healthy (Class 1) ranged from 25-66%, were dead or dying (Classes 4, 5 and 6) ranged from 4-23%, had high or medium foliar-disease severity ranged from 12-54%, had galls from western gall rust ranged from 2-64%, and had root-disease symptoms ranged from 5-21% (Table 2.1).  23  No known species of root pathogens were detected within the genetic sequence data. There were four ‘putative’ fungal pathogen species: Phacidiopycnis washingtonensis, Phaeoannellomyces elegans, Phoma radicina, and Volutella colletotrichoides. None of these species are pathogenic to lodgepole pine. For example, P. washingtonensis is pathogenic to pome fruits and V. colletotrichoides is pathogenic to legumes.   In the IDF, WTH plots had significantly higher dead or dying trees (p = 0.0176; Figure 2.2) and response ratios of dead or dying trees (p = 0.0074) compared to FFR plots. Soil-compaction treatments did not significantly affect any stand-health metrics. In the SBS, there were no significant treatment effects.   Total soil carbon ranged from 20,640 to 92,637 (kg/ha) in the IDF zone and 18,942 to 83,117 (kg/ha) in the SBS zone (Table A2). In both the IDF and SBS, stands with more soil carbon were generally healthier. In the IDF zone, soil carbon was negatively related to foliar disease (R2 = 0.22, p = 0.0136). In the SBS, soil carbon was negatively related to dead or dying trees (R2 = 0.21, p = 0.0143; Figure 2.3) and to root disease (R2 = 0.16, p = 0.0382).   Bulk density ranged from 0.85 to 1.15 (g/cm3) in the IDF zone and from 0.98 to 1.46 (g/cm3) in the SBS zone (Table A2). Relationships between stand health and soil bulk density differed between BEC zones. In the IDF, bulk density was negatively related to dead or dying trees (R2 = 0.25, p = 0.0083) and positively related to gall rust (R2 = 0.17, p = 0.0352). In the SBS, bulk density was negatively related to gall rust (R2 = 0.21, p = 0.0175) and response ratio of health trees (R2 = 0.31, p = 0.0027). This apparent opposite direction of bulk density in relation to stand health in the IDF and SBS may be due to BEC zones covering a different range of soil bulk densities (Figure 2.4). In general bulk density is lower in the IDF compared to the SBS. When the two BEC zones are considered together, there is no significant relationship between bulk density and healthy trees (Figure 2.5).  There does not appear to be any obvious thresholds or non-linear relationships in the plotted data of soil carbon and bulk density compared to stand health (Figure 2.3 and 2.4).  24  2.5 Discussion Variable treatment response among sites suggests that the response of forest productivity to soil disturbance is context-dependent. In the first 15 years of the LTSP study, treatment effects on tree growth were insignificant or varied between sites (Powers et al. 2005; Kranabetter et al. 2006; Fleming et al. 2006; Sanchez et al. 2006a; Scott and Dean 2006; Ponder et al. 2012; Holub et al. 2013). Some of this variability has been taken to suggest that, in general, forest stands are resilient to soil disturbance during logging (Ponder et al. 2012). However, it has also been found that variable response depends on some site characteristic. For example, on fine-textured soils, compaction reduces tree growth whereas on coarse-textured loose soils, compaction can increase tree growth (Ponder et al. 2012). Also unproductive sites with low phosphorus concentrations are more sensitive to biomass removal (Scott and Dean 2006) whereas deep nutrient rich soils are resilient to biomass removal (Holub et al. 2013). Independent of whether treatment response variability is driven by stochasticity or context-dependency, these results again highlight the rarity of one management activity producing one outcome when applied to multiple sites (Ponder et al. 2012; Vance et al. 2014; Messier et al. 2013).  The rates of pest occurrence documented here are consistent with other recent surveys of mid-aged lodgepole pine stands in British Columbia. Rates of pest occurrence documented in our study are similar to those estimated by Heineman et al. (2010) in 66 mid-aged lodgepole pine stands in BC. The sites considered in this and the Heineman et al. (2010) paper overlapped in geographic range with the Heineman et al. (2010) sites extending further south and east. Western gall rust was found at all sites in both studies, affecting 2 - 64% of the trees in our study and 3 -74% of the trees in the Heineman et al. (2010) study. Log Lake had the highest western gall rust occurrence likely due to this site receiving 200 mm more annual precipitation than the other sites (Table 2.1) and western gall rust requires moisture for spore dispersal (Allen et al. 1996). We found 12 - 54% of the trees have high or medium foliar disease; Heineman et al. (2010) found 1 - 50% of the trees have pine needle cast (Lophodermella concolor). At our sites, 5 - 21% of the trees showed signs of root disease, similar to the 2 - 23% of trees affected by the root disease Armillaria ostoyae by Heineman et al. (2010). The inability to detect root pathogens from the pyrosequencing data could be related to technical or biological limitations. These data indicate  25 current levels of pest occurrence that can be used as baselines for comparisons with other forest types and for monitoring changes over time.    All measures of treatment response are relative to a pre-determined baseline. Appropriate selection and consistent use of this baseline is necessary for comparing results among studies (Vance et al. 2014). Within the LTSP experimental design, the bole-only harvesting with no compaction (OM1C0) treatment is used as the baseline because it represents the least amount of soil disturbance. Results from these comparisons can assess how additional levels of soil disturbance influence stand characteristics and can inform appropriate best practices in areas that will be harvested (Thiffault et al. 2014). Baselines may also include stands with natural disturbances or un-harvested stands (Vance et al. 2014) with results indicating how harvesting influences stand characteristics relative to unmanaged stands.   In the IDF, forest-floor removal plots having higher survival may be related to microclimate conditions or control of competing vegetation. Organic-matter removal can benefit trees in cold climates by increasing thermal conductivity thereby increasing soil temperatures or harm trees in dry climates by removing the mulch layer allowing the soil to dry more quickly (Zabowski et al. 2000; Roberts et al. 2005). The effect of changed microclimate on tree survial may be short-lived because the relative importance of microclimate to seedling productivity diminishes with tree size throughout stand development (Thiffault et al. 2011). Therefore, it could be that since the IDF sites are younger they are still expressing differences in microclimate. The forest-floor removal plots on the SBS sites, which are 5 years older and closer to canopy closure, did not always have higher survival.   At Black Pines, survival was lowest on the whole-tree harvested plots, indicating slash removal does not have the same microclimate benefit as forest-floor removal. It is possible that slash removal has detrimental effects to seedling survival by reducing frost protection (Fellin 1980; Fleming et al. 2006).   In conclusion, this study contributes to the ongoing research at the LTSP sites by adding data on stand health. Our findings highlight the prevalence of site-specific responses to management  26 practices and possible context-dependencies. The stand health data provides a baseline to compare with other sites and future research on forest pest occurrence. It also allows for comparisons on tree growth, tree nutrition, soil properties and forest pest occurrence. By integrating stand-health metrics into measurement and assessment of forest productivity it will be possible to shift towards forest management that considers overall ‘ecosystem health’ (Page-Dumroese 2010a) rather than simply considering tree growth.      27 2.6 Tables and figures  Table 2.1. Average number and percentage of healthy trees, dead or dying trees, tree with high or medium foliar-disease severity, trees with galls from western gall rust, and root-disease symptoms at each site. Three sites are in the Interior Douglas Fir (IDF) zone and three are in the Sub Boreal Spruce (SBS) zone. Percentages of healthy and dead or dying trees are calculated from the total number of trees planted per site (900) and percentages of disease occurrence are calculated from the number of living trees per site. Site Healthy trees Dead or dying Foliar disease1 Gall rust Root disease Black Pines 480 (56%) 40 (4%) 106 (12%) 44 (5%) 41 (5%) Dairy Creek 318 (41%) 121 (13%) 364 (47%) 15 (2%) 69 (9%) O'Connor 195 (28%) 210 (23%) 376 (54%) 13 (2%) 45 (7%) Log Lake 524 (66%) 100 (11%) 254 (32%) 511 (64%) 61 (8%) Skulow Lake 191 (25%) 133 (15%) 394 (51%)2 53 (7%) 161 (21%) Topley 555 (64%) 39 (4%) 308 (36%) 327 (38%) 47 (5%) 1 The foliar disease, Lophodermella concolor, was present at all sites. 2 The foliar disease, Elytroderma deformans, was also present at Skulow Lake.         28  Figure 2.1. Pictures showing visual symptoms of (A) western gall rust (Endocronartium harknessii), (B) root disease (mainly Armillaria ostoyae), (C) the foliar disease Lophodermella concolor, and (D) the foliar disease Elytroderma deformans.    Figure 2.2. In the Interior Douglas Fir (IDF) zone, whole-tree harvest (WTH) plots had significantly higher dead or dying trees compared to the forest-floor removal (FFR) plots (p = 0.0176). The bole-only harvest (BOH) plots were not significantly different from either the WTH or FFR plots.     29  Figure 2.3. In the Sub Boreal Spruce (SBS) zone, total soil carbon (kg/ha) was negatively related to the number of dead or dying trees (R2 = 0.21, p = 0.0143). A linear best-fit line is shown.    Figure 2.4. Bulk density (g/cm3) is positively related to healthy trees in the Interior Douglas Fir (IDF) zone but negatively related in the Sub Boreal Spruce (SBS) zone. Sites are colour coded to visualize differences among sites. Linear best-fit lines and R2 values are displayed.    30  Figure 2.5. Bulk density (g/cm3) is not related to healthy trees when both BEC zones are considered together. Sites are colour coded to visualize differences among sites. Linear best-fit line and R2 value are displayed.     31 Chapter 3. Relationships between tree growth and stand health 3.1 Synopsis • Forest management in British Columbia (BC), Canada assumes that fast growing stands will have minimal mortality to the dominant trees. • This assumption has recently been challenged, yet there is limited data of stand health and growth at the same sites.  • Here we explore the relationships between lodgepole pine growth and disease occurrence using six Long-Term Soil Productivity (LTSP) installations in BC.  • Treatment plots and random groups that had higher average tree growth also had higher average disease occurrence.  • These findings again challenge the assumption that fast-growing plantations are free of disease, having implications for growth-and-yield models, sustainable harvesting levels, and predicting future timber supply.   3.2 Introduction Tree growth is the main metric used to monitor forest productivity and predict future timber supply. This approach, and the absence of disease occurrence data collected in conjunction with tree-growth data, is based on the assumption that fast-growing stands rarely have significant insect or disease damage (Skovsgaard and Vanclay 2008; Puettmann et al. 2009). This assumption has been challenged by recent findings suggesting that current predictive models overestimate wood production because of higher than expected insect and disease damage (Woods and Coates 2013). Contrary to predictions, damage was sustained in dominant and co-dominant trees rather than in suppressed trees (Woods and Coates 2013). Although tree growth and disease occurrence are dependent on each other and profoundly influence long-term forest productivity (Andersson et al. 2000), surprisingly few studies simultaneously monitor these metrics or compare them on the same plots.   Although it is assumed that fast growing stands are healthy, ecological theory states that the relationship between plant growth and defence depends on resource availability. The assumption that fast-growing stands are healthier is largely based on the idea that vigorously growing trees  32 would be able to ‘out-grow’ insect or disease attacks (Andersson et al. 2000). This could occur in situations with minimal pest pressure and ample resources; however, there is a physiological trade-off of between resource allocation to growth and defence. The relationship between plant growth and health is outlined in growth-differentiation balance (GDB) framework known as “The Dilemma of Plants: To Grow or Defend” (Herms and Mattson 1992). This theory predicts with ample resources plants will allocate most photosynthates to growth, however, under growth limiting conditions photosynthesis still occurs with these photosynthates being available for defence against disease. This phenomenon is evident in the literature showing that growth limiting nutrient conditions increase plant chemical defense (Donaldson et al. 2006; Osier and Lindroth 2006; Sampedro et al. 2011) and anatomical defense (Moreira et al. 2008). Simultaneous monitoring of tree growth and health in well-designed field trials would better enable us to disentangle the relationships between these critical variables.  Existing long-term experiments monitoring forest productivity are a significant resource for testing the relationship between tree growth and disease occurrence. The Long-Term Soil Productivity (LTSP) project monitors the effect of soil compaction and organic-matter removal on stand productivity across North America (Powers 2006). Stand productivity is monitored using measures of tree growth and nutrient status (Powers 2006). Using the stand health data collected in Chapter 3 of this thesis, it is now possible to compare stand health and growth. This chapter examines the relationship between tree growth and disease occurrence at the six study sites.   3.3 Methods The six measures of stand health developed in Chapter 2 are used here for comparisons with lodgepole pine growth. Stand-health metrics used in this chapter include dead or dying trees, foliar disease, gall rust occurrence, root disease and total disease all indicators of forest health.   Lodgpole pine growth was measured by height (m), height increment (m/5yrs), average tree volume (cm3), total tree volume (cm3) and average tree volume increment (cm3/5yrs). Tree  33 height3 was measured with a pole every five years and height increment (m/5yrs) calculated by subtracting the two most recent height measurements. Individual tree volume was calculated as,  volume = EXP(-2.15873+1.86671*LN(Dia)+1.18486*LN(Ht)+1.49899*(h/Ht)),  where Dia is tree diameter (cm), Ht is tree height (cm) and h is the height (cm) above ground level at which diameter was measured (Kovats 1977). Here we used diameter at breast height measurements so h = 130 cm above ground level. Average tree volume (cm3) is the plot average of each trees volume. Total tree volume (cm3) is the sum of all the tree volumes per plot. Volume increment (cm3/5yrs) was calculated by subtracting two most recent average volume measurements. Trees that died during the 5-year period (resulting in negative increment values) were removed from the analysis.  Relationships between measures of tree growth and disease occurrence were tested using simple linear least-squares regressions. As the sites differed in environmental conditions and tree age, tree growth and disease occurrence were compared within each site separately. Relationships were tested using treatment-plot averages as well as random groupings of trees within each site (n = 9). Live trees were assigned to random groups using a random-number generator. Tree growth and disease occurrence were then averaged within these random groups. For both the treatment-plot averages and random-group averages, a total of 150 relationships were tested (5 tree-growth metrics * 5 disease metrics * 6 sites). Two-sided goodness-of-fit tests were used to test whether the number of significant positive relationships significantly differed from the number of significant negative relationships (H0: n[+] = n[-] and HA: n[+] ≠ n[-]).  To avoid over-estimating significant relationships by using multiple co-linear tree-growth metrics, we replicated the above-mentioned analysis using only height increment (m/5yrs) and average volume (cm3). These two tree-growth measures were the least related and capture both incremental and cumulative growth. For this analysis, a total of 60 relationships were tested (2 tree-growth metrics * 5 disease metrics * 6 sites).                                                   3 Tree height and diameter data originate from the BC Ministry of Forests, Lands and Natural Resource Operations.  34 Positive relationships between measures of tree growth and disease occurrence would indicate that groups of trees with higher growth rates or larger volumes also had higher disease occurrence. Comparing results from treatment-plot averages with those from random-group averages enables us to explore the possible influence of the treatments themselves on this relationship. For example, significant relationships when using treatment-plot averages but no significant relationships when using random-group averages would suggest that the treatments themselves influence both tree growth and disease occurrence. Results were considered significant when p < 0.05. All analyses were conducted in JMP version 11 (SAS Institute Inc. 2012).   3.4 Results Using treatment-plot averages, there were 28 significant relationships between tree growth and disease occurrence (Table 3.1). Therefore, 19% of the relationships were significant (28 out of 150 cases). Of the 28 significant relationships, 26 were positive (93%; Table 3.1). The goodness-of-fit test indicated significantly more positive relationships than negative relationships (p < 0.0001).  The relationship between tree growth and disease occurrence independent of treatments effects was compared using averages of random groups. There were 12 significant relationships out of the 150 possible cases (8%), 11 of which were positive (92%; Table 3.2). The goodness-of-fit test indicated significantly more positive relationships than negative relationships (p = 0.003).   Using only the tree-growth measures of height increment and average volume to avoid co-linearity, the overall results were the same. Using treatment plots, 10 cases out of 60 cases were significant (17%). Significantly more of these had positive relationships (9 out of 10, p = 0.0107; Figure 3.1). Using random groups, 6 cases out of 60 cases were significant (10%). Significantly more of these had positive relationships (6 out of 6, p = 0.0156; Figure 3.2).  3.5 Discussion Our finding that treatment plots and random groups with larger lodgepole pine trees generally  35 also had more disease concurs with Woods and Coates’s (2013) finding of substantial insect and disease occurrence in mid-aged plantations in BC with larger trees being targeted. Some of these relationships appear to be driven by a few high values for disease occurrence (Figs. 1A and 1C –1F). The results would be more robust if the distribution of disease occurrence had been more even among plots. Nevertheless, it is striking to observe that the few plots with high disease occurrence also had the highest growth rates and cumulative growth. This situation may apply to large areas of forestland in British Columbia, as lodgepole pine seedlings have been planted on 3.8 million hectares (53% of all second-growth forests; BC Ministry of Forests, Mines and Lands 2010; Weaver 2013), and levels of disease are comparable with previous studies (Reid et al. 2015). Long-term monitoring of both health and growth is warranted, with no omission of trees, plots or sites where damaging agents are present.   It is unlikely that disease occurrence increased tree growth; therefore, faster-growing trees must have become more diseased. It could be that faster-growing trees have greater likelihood of encountering disease. This could be the situation for root disease, where a larger root system would come into contact with more disease inoculum, but is a less satisfying explanation for positive relationships with foliar-disease severity, which was based on the percentage of foliage affected.   The reduced number of positive relationships when using random groups compared with treatment plots suggests that the treatments themselves may influence the relationship between tree growth and disease occurrence. The organic-matter removal and soil-compaction treatments could directly influence disease populations, habitat, and substrate and indirectly influence disease through effects on disease antagonists and host tree susceptibility. For example, disease inoculum that occurs on needles and branches could be reduced by whole-tree harvesting, which removes these materials; however, could be increased by forest-floor removal because contact with the forest floor reduced inoculum amounts on the branches and needles (Oblinger et al. 2011).  Another example involving organic-matter removal is that changes to the amount of woody debris could alter the abundance of saprotrophic fungi, such as Hypholoma, which are antagonistic to Armillaria root disease (Chapman et al. 2004). Soil compaction has been linked to  36 disease occurrence, with the damaging fungal pathogen Phytophthora being highest on seedlings in wet compacted soils (Rhoades et al. 2003).   Another possible mechanism for the observed positive relationship between tree growth and disease occurrence could be that higher mortality caused by disease would thin the stands, allowing them to grow faster. This, however, was not the case at our sites because plots with higher disease occurrence have not yet experienced higher rates of overall mortality. The only relationship that was significant was at O’Connor Lake where plots with more root disease had less overall mortality (R2 = 0.67, p = 0.0071).   It could also be that in faster-growing trees, more resources are allocated to growth rather than to defence against disease. Within all plants there is a continually changing allocation of resources to vital processes of growth, defence, reproduction, storage and maintenance (Herms and Mattson 1992). Resource allocation varies according to plant species, developmental stage, environmental conditions, competition, occurrence of damaging agents and resource limitation (Herms and Mattson 1992; Boege and Marquis 2006; Boege et al. 2007). Nutritional differences, in particular, may be the most likely variable that differed between groups of trees with higher growth, which could influence allocation of resources to growth and defence (Figure 3). In all plants, photosynthesis can continue even when growth is constrained by nutrient availability (Herms and Mattson 1992); the ‘extra’ photosynthates produced under conditions of nutrient-limited growth are used in the production of secondary metabolites such as monoterpenes and phenols (Chishaki and Horiguchi 1997; Wallis et al. 2008). These secondary metabolites inhibit growth and spore-germination of common fungal pathogens associated with pine (Wallis et al. 2008; Eckhardt et al. 2009) and can also reduce damage from herbivorous insects (Bryant et al. 1987). Nutrient limitations have also been found to increase the amount of chemical (Sampedro et al. 2011) and anatomical defence strategies in pine (Moreira et al. 2008). Therefore, any intervention to increase tree growth may simultaneously cause trees to be less defended against damaging insects and disease.  The results in this chapter suggest that investment of resources and time into increasing early growth alone may result in higher disease occurrence. The tree improvement program in British  37 Columbia (BC) started in the 1960s with the aim of improving tree growth (Xie and Yanchuk 2003). This program has succeeded in increasing short-term tree growth by up to 30% (Xie and Yanchuk 2003). Using growth-and-yield models, these short-term growth gains were scaled up to a predicted 15 to 25% gains over natural regeneration over one rotation period (BC Ministry of Forests, Mines and Lands 2010). However, these models were found to overestimate biomass production by 5 times in 10 to 30 year old stands because they did not accurately account for losses due to damaging insects and diseases (Woods and Coates 2013). BC’s tree improvement program now selects for characteristics of pest resistance and recovery in addition to those of tree growth (Yanchuk 2009). Between 1988 and 2007, 75% of the harvested or burned area was planted with seedlings from the tree improvement program (B.C. Ministry of Forests, Mines and Lands 2010). The legacy of this selection is yet to be fully determined.   Results thus support the contention of Andersson et al. (2000) that the central aim of forest ecosystem research and forest management should be the optimization of both tree growth and pest resistance, because pest resistance, in part, depends on processes involved in tree growth. Interactions among nutrient availability, tree growth and tree health warrant further examination to determine the conditions that promote both growth and health, and optimize long-term forest productivity. Disease occurrence was measured only once in this trial, so it is not possible to determine when the disease arrived, how it progressed, or how growth changed after disease onset. This knowledge gap is widespread, given the lack of standardized forest-health data collected in conjunction with tree-growth data. Tree health, in addition to tree growth, should be frequently monitored in silvicultural trials to examine the effects of growth-enhancing silvicultural interventions on subsequent tree health and long-term stand productivity.      38  3.6 Tables and figures  Table 3.1. Tree growth in relation to tree health based on treatment-plot averages.  Site Growth metric Health metric Relationship R2 p-value Dairy Creek Average volume Gall rust Positive 0.46 0.0452  Height Total disease Positive 0.50 0.0335  Total volume Dead or dying Negative 0.58 0.0172 O’Connor Average volume Foliar disease Positive 0.62 0.0113  Average volume Gall rust Positive 0.79 0.0013  Average volume Total disease Positive 0.79 0.0013  Height Foliar disease Positive 0.49 0.0364  Height Gall rust Positive 0.90 0.0001  Height Root disease Positive 0.53 0.0256  Height Total disease Positive 0.70 0.0052  Height increment Foliar disease Positive 0.47 0.0408  Height increment Gall rust Positive 0.91 <0.0001  Height increment Root disease Positive 0.55 0.0223  Height increment Total disease Positive 0.68 0.0060  Total volume Foliar disease Positive 0.52 0.0294  Total volume Gall rust Positive 0.82 0.0008  Total volume Root disease Positive 0.47 0.0421  Total volume Total disease Positive 0.71 0.0046  Volume increment Foliar disease Positive 0.62 0.0113  Volume increment Gall rust Positive 0.79 0.0014  Volume increment Total disease Positive 0.79 0.0013 Skulow Total volume Total disease Positive 0.47 0.0430 Topley Average volume Total disease Positive 0.45 0.0467  Height Gall rust Positive 0.47 0.0409  Height increment Root disease Negative 0.45 0.0491  Total volume Gall rust Positive 0.50 0.0318  Total volume Total disease Positive 0.58 0.0173   Volume increment Total disease Positive 0.47 0.0400          39  Table 3.2. Tree growth in relation to tree health based on random-group averages. Site Growth metric Health metric Relationship R2 p-value Black Pines Average volume Root disease Positive 0.56 0.0199  Height Root disease Positive 0.55 0.0217  Volume increment Root disease Positive 0.58 0.0171 Dairy Creek Height Dead or dying Positive 0.47 0.0427 Log Lake Height increment Total disease Positive 0.46 0.0443  Total volume Dead or dying Negative 0.59 0.0158 O'Connor Lake Average volume Root disease Positive 0.86 0.0003  Height Root disease Positive 0.81 0.001  Height increment Root disease Positive 0.77 0.0019  Height increment Total disease Positive 0.47 0.0413  Total volume Root disease Positive 0.76 0.0023  Volume increment Root disease Positive 0.85 0.0004        40  Figure 3.1. Treatment-plot averages of disease occurrence were positively related to tree growth at O’Connor Lake (circle), Dairy Creek (cross), and Topley (triangles). Average tree volume (cm3) was positively related to the percent of living trees with western gall rust occurrence (A), total number of disease occurrences (B), and the percent of living trees with foliar disease (C). Tree height increment (m/5 years) was positively related to the percent of living trees with western gall rust occurrence (D), the percent of living trees with foliar disease (E), and the percent of living trees with root-disease symptoms (F).   41   Figure 3.2. Random-group averages of disease occurrence were positively related to tree growth at O’Connor Lake (circle), Black Pines (squares) and Log Lake (diamonds). Average tree volume (cm3) was positively related to the percent of living trees with root-disease symptoms (A). Height increment (m/5 yrs) was positively related to the total number of disease occurrence (B) and the percent of living trees with root-disease symptoms (C).   42  Figure 3.3. The balance of resource allocation to growth and secondary metabolites differs under optimum nutrition (left) and nutrient limitation (right). Arrow size represents relative amounts and is not based on real data. Secondary metabolites are related to constitutive and induced tree defence against disease and insect attack. Stand productivity is a product of both tree health (disease and insect occurrence) and tree growth.     43 Chapter 4. Spectral reflectance and stand health 4.1 Synopsis To test the feasibility of using two spectral-reflectance indices to assess differences in stand-health metrics in 54 plots of lodgepole pine in British Columbia, Canada. Spectral-reflectance indices of excess greenness (EG) and green chromatic coordinate (GCC) were calculated over two spatial scales from colour images captured from a Cessna T210. EG and GCC were then compared to five ground-based metrics of stand health: vigour, mortality, foliar disease, western gall rust, and root disease. Spectral-reflectance indices did detect the ground-based metrics of stand-health, except western gall rust occurrence. These relationships were weaker for small and large tree classes, but were unaffected by foliar nitrogen concentration. Stand-health metrics accounted for 36.5 to 60.9% of the variability in spectral-reflectance values. EG and GCC values calculated at both tree and plot scales were significantly related to stand-health metrics. EG and GCC values could be used to set thresholds below which ground-checks of stand health would be warranted.  4.2 Introduction The amount of damage to forests caused by insects and disease is predicted to increase annually (Mickler 1996). In mid-rotation stands in British Columbia (BC), Woods and Coates (2013) found much greater mortality than predicted by growth-and-yield models mainly caused by insects and disease. As this has implications for merchantable volumes at harvest, they recommended more intensive monitoring of forest health. Ground assessments of forest health over large areas such as BC are time consuming and expensive. Small planes are used in BC to survey visible outbreaks of forest pests over the entire province (The Aerial Overview Survey). This sketch-mapping technique captures what the observer can see from the plane window and record on a paper map. Consistency in identifying pest occurrence and severity levels among observers is critical but difficult. Analysis of images captured from planes or satellites can provide a more standardized and objective method for detecting changes to forest condition.   From either satellite or aerial images, the proportions of reflected wavelengths are used to calculate spectral-reflectance indices, which provide an indication of plant condition. The most common spectral-reflectance index is the normalized difference vegetation index (NDVI;  44 Soudani et al. 2012). NDVI uses the longer red- and near-infrared wavelengths, which have less atmospheric distortion, and so are preferable when using satellite images (Nijland et al. 2014). When using aerial images, spectral reflectance indices such as Excess Greenness (EG) and Green Chromatic Coordinate (GCC) which use shorter (red, green and blue) wavelengths, can be used and have been found to outperform NDVI (Nijland et al. 2014).   Both EG and GCC indicate the proportion of green light reflected (relative to other wavelengths), and have been used to assess gross primary productivity of forests (Saitoh et al. 2012) and to monitor forest phenology changes associated with climate change (Richardson et al. 2007; Sonnentag et al. 2012; Petach et al. 2014; Keenan et al. 2014; Yang et al. 2014). The use of EG and GCC for monitoring plant condition (stress or health) is based on the assumption that healthy vegetation will reflect a higher proportion of green light than will unhealthy vegetation. The amount of reflected green light will also depend on productivity and foliar nitrogen (Dillen et al. 2012). EG and GCC have successfully indicated plant condition in agricultural and laboratory settings (e.g. Bacci et al. 1998; Nijland et al. 2014). They have also been used to identify damage classes in branches of declining Norway spruce trees (Ruth et al. 1991). However, the utility of these indices for monitoring stand health on operationally relevant scales needs to be assessed.   The spatial scale captured by images and the region of interest (ROI) used to calculate spectral-reflectance indices range from individual leaves to forest canopies, depending on objectives. Studies that test and compare the utility of multiple spectral-reflectance indices often take place in laboratories and use individual leaves or plants (e.g. Nijland et al. 2014; Woebbecke et al. 1995). In contrast, when monitoring spring green-up of deciduous stands, images of the entire canopy are captured and representative areas are selected as the ROI (Richardson et al. 2007; Yang et al. 2014; Petach et al. 2014).   Here we assess the utility of aerial imagery for monitoring stand health by comparing EG and GCC values with ground-based metrics of stand health collected on the same plots. Our specific questions are: 1. How well do EG or GCC values relate to ground-based metrics of stand health (vigour, mortality and disease occurrence)?   45 2. How does stand productivity and foliar nitrogen concentrations influence the relationships between spectral-reflectance indices and stand-health metrics? 3. How much of the variability in EG and GCC values can be attributed stand-health metrics?  4. Are analyses at the individual tree scale or at the plot scale more useful for assessing stand-health metrics?  5. Are there thresholds in EG or GCC values that indicate a loss of stand health that could be used to prioritize areas for ground checks?  4.3 Methods The company Terrasaurus captured images from a Cessna T210 Centurion II Survey Aircraft with a photogrammetric medium format aerial colour camera between July and November 2013, except at Topley, which was photographed in August 2014. These images were georeferenced, mosaicked, and aerially triangulated. Plot 9 at Dairy Creek was not photographed, and therefore was not included in this analysis.   Red-green-blue images were loaded into the software program ENvironment for Visualizing Images (ENVI) 4.8 for analysis (Exelis Visual Information Solutions). Using the ‘band math’ tool excess greenness (EG) and green chromatic coordinate (GCC) indices were calculated by the equations,   EG = 2G-(R+B), and  GCC = G/(R+G+B)   respectively, where G is green light, R is red light and B is blue light (Nijland et al. 2014). Larger EG and GCC values indicated that a higher proportion of green light is being reflected, suggesting healthier trees (lighter hues in Figure 4.1). Green chromatic coordinate is also referred to as chromaticity or abbreviated to gcc (for example, Bacci et al. 1998). When first developed, the excess-greenness index did not have a name but was referred to as 2g-r-b: the short hand of its formula (Woebbecke et al. 1995).  46 EG and GCC values were calculated over two regions of interest (ROI) covering different spatial scales. First, ROIs were created over ten individual tree crowns near the center of each treatment plot: hereafter EGtree or GCCtree (Figure 4.2). Analysis of individual tree crowns minimizes background noise from the non-tree vegetation and ground cover. Second, ROIs were created over the entire plot to capture differences in overall stand condition: hereafter EGplot or GCCplot (Figure 4.2). Plot ROIs could be expected to have higher variability but could be more useful for forest management in assessing stand condition.   To determine how EG and GCC values relate to stand health and to determine how much variability in EG and GCC values is accounted for by stand health, ground-based visual forest-health surveys were conducted in the summer of 2013. From this survey, five forest-health metrics were developed: the percent of living trees with healthy vigour (healthy vigour), the number of dead or dying trees (mortality), the percent of living trees with high or medium foliar disease, the percent of living trees with galls from western gall rust (Endocronartium harknessii) and the percent of living trees with root-disease symptoms (Chapter 2).   Stand productivity was measured using Leaf-Area Index (LAI) collected with hemi-view photos during the summer of 2013 using a Nikon Coolpix5400 digital camera and a full (360˚) fisheye Zoom Nikkor ED lens. Five photos were taken in each plot: at the four corners and the centre. LAI was calculated in Gap Light Analyzer version 2.0 (Frazer et al. 1999) by,  LAI = ln(P(θ))cos(θ)/G(θ),    where θ is the zenith angle of view P(θ) is the gap fraction and G(θ) is the fraction of foliage projected on the plane parallel to the ground (Bréda 2003). The number of azimuth regions was set to 16 and zenith regions to 10. All analyses used LAI ring 5, which integrates over zenith angles 0 to 75˚ (Frazer et al. 1999). Foliar samples were collected from 15 trees per plot and bulked together into 3 sub-samples (Berch et al. 2010). Foliage was dried and total nitrogen analyzed by combustion at the BC Ministry of Environment’s Laboratory in Victoria, BC  (Kalra and Maynard 1991).   47 Linear regressions were used to determine the direction of any significant relationships between EG and GCC, and stand-health metrics. To determine the influence of stand productivity and foliar nitrogen concentrations on the relationship between spectral reflectance and stand health, we plotted this relationship along gradients of stand productivity or foliar nitrogen. Linear models were used to determine the relative importance of explanatory variables on response variables. In these models EGtree, EGplot, GCCtree and GCCplot were the response variables and healthy vigour, mortality, foliar disease, and root disease were the predictor variables. Western gall rust was not included in the models because its occurrence was not accurately captured by the EG or GCC when all plots were considered (Figure 4.7). The relative importance of each predictor variable was determined using the R2 contribution averaged over orderings among regressors (Lindeman et al. 1980) from the relaimpo package. All analyses were conducted in RStudio (RStudio Team 2015).   4.4 Results EG and GCC values were both positively related to the percent of trees with healthy vigour when considering all plots (Figure 4.3), supporting the assumption that healthy vegetation reflects a higher proportion of green light than does unhealthy vegetation. Also as expected, the EG and GCC values were negatively related to mortality (Figure 4.4), foliar disease (Figure 4.5) and root disease (Figure 4.6). The same relationships were found when calculated for each site separately (Table 4.1). The one exception was mortality being positively related to EGplot at O’Connor Lake (Table 4.1). Spectral-reflectance indices were not able to reliably capture western gall rust occurrence when all sites were considered. Western gall rust was negatively related to EGtree but positively related to GCCtree and GCCplot (Figure 4.7). Within individual sites, however, the relationship between western gall rust occurrence and spectral-reflectance values was always negative, as expected (Table 4.1).   Relationships among spectral-reflectance indices and stand-health metrics were weaker at low and high LAI ranges compared to the medium LAI range (Figure 4.8), but varied little among foliar nitrogen ranges (Figure 4.9).   48 Stand-health metrics (healthy vigour, mortality, and foliar and root disease occurrence) explained 36.5 to 60.9% of the variability in EG and GCC values (Table 4.2). Foliar disease occurrence explained the largest proportion of the variability in EGtree values (21.8%) compared to the other stand health variables that explained 1.7% to 7.2% (Table 4.2). Both foliar disease and root disease explained a large proportion of EGplot values (18.4% and 24.9% respectively) compared to the measures of overall health and mortality that explained 5.8% and 0.9% respectively (Table 4.2). The largest proportions of GCCtree and GCCplot metrics were explained by healthy vigour (23.7% and 37.7% respectively) and mortality (16.2% and 11.4%, respectively; Table 4.2). Of the four spectral reflectance indices, stand-health metrics contributed to the largest proportion of variability in GCCplot (60.9%, Table 4.2).  EG and GCC values calculated at both tree and plot scales were significantly related to stand-health metrics (Table 4.1). However, a higher percent of variability in EG and GCC values was explained when they were calculated at the plot scale (Table 4.2).   It might be possible to set threshold values for EG and GCC depending on stand-health goals for specific scenarios using linear-regression equations based on ground-truthed data. For example, GCCplot values could be used to set thresholds for healthy vigour because of the strong relationship between these variables (Figure 4.3). Using our data, the equation explaining the significant relationship between GCCplot and healthy vigour is,   Y = 0.3839 + 0.000439 * X  where Y is GCCplot and X is healthy vigour (Figure 4.3). It is then simple to choose a level of healthy vigour that is ‘acceptable’ and calculate the corresponding GCCplot value. For example, if the goal was to identify stands with lower than 60% of the trees having healthy vigour, then the EG threshold value would be 0.41 (Y = 0.3839 + 0.000439 * 60). Indeed it appears that above GCCplot of 0.41 there would be little reason to be concerned with any of the stand-health metrics (Figures 4.3 to 4.7). However, GCCplot values below 0.41 could indicate a forest health issue (Figures 4.3 to 4.7).    49 4.5 Discussion Our findings indicate that both EG or GCC indices calculated from aerial images at two spatial scales could be useful for monitoring stand health because ground-based measures of healthier trees related to higher EG and GCC values. GCC values were most strongly related to overall healthy vigour and tree mortality, while EG values were most strongly related to foliar and root disease occurrence. Spectral-reflectance values calculated at both individual-tree and plot scales were correlated with stand-health metrics. However, stand-health metrics explained a higher proportion of the variability in both EG and GCC values when they were calculated at a plot scale.  Excess greenness was most strongly related to foliar disease metrics and so may be a particularly effective technique for monitoring foliar disease. Identification of foliar disease from sketch-map ocular surveys is difficult until damage is quite severe so can be visually seen by the observer, so EG has potential to be a tool for early detection of foliar disease. This needs to be tested further with sites over a gradient of foliar disease infection. It is also possible that this technique, that uses readily available and inexpensive equipment, could be integrated into ocular surveys such as the aerial-overview survey program currently used in British Columbia.    Western gall rust occurrence was not accurately accounted for by EG and GCC values when comparing all sites. This could be explained by the low occurrence of gall rust at all sites except Log Lake and Topley. When regressions were tested at Log Lake and Topley individually western gall rust occurrence was negatively related to EGplot (Table 4.1) indicating that at sites with high gall rust occurrence detection could be possible. Another possible reason for a lower response of greenness to gall rust is related to moisture. Gall rust occurrence is linked to precipitation because spore release requires moisture (Forest Health Protection 2011; Yukon Energy, Mines and Resources 2015). Within the range of precipitations in our study, sites with higher precipitation have higher productivity. Therefore, wetter sites could have both more productive trees and more gall rust. However, we did not have low moisture sites with higher levels of gall rust, which would have allowed us to test this possibility. It is not surprising that, of the diseases considered here, western gall rust occurrence was the most complicated to detect.   50 Uses of threshold values are common with spectral-reflectance indices. Spectral-reflectance values have been identified that related to plant-stress levels (Bacci et al. 1998) and spring green-up levels (e.g. Richardson et al. 2007; Keenan et al. (2014). Keenan et al. (2014) found that a threshold-crossing approach was more effective at estimating phenology compared to a curve-fitting approach. We found that strong relationships between spectral-reflectance indices and stand-health metrics can be used to calculate EG or GCC values that appear to correspond to acceptable levels of forest health. Above these values the stands should be healthy and below them ground-checks to determine the causes of low values could be recommended. For example, GCCplot values greater than 0.41 were generally associated with high vigour, low mortality and low disease occurrences, so would not require checking. However, GCCplot values below 0.41 would indicate ground-checks to determine any health issues are recommended. Further testing of thresholds is recommended with the goal of establishing critical values of GCC or EG for use in monitoring forest health over large land areas.  The relatively weak relationships among spectral-reflectance indices and stand-health metrics at low LAI may indicate that at low productivity sites variables other than disease (drought, nutrients, frost, etc.) are also influencing GCC. Spectral-reflectance indices were strongly related to stand-health metrics at all ranges of foliar nitrogen concentrations, indicating that foliar nitrogen does not impact the ability of greenness indices to detect differences in stand health.   In conclusion, spectral-reflectance indices were related to stand-health metrics as expected by the assumption that healthy vegetation reflects a higher proportion of green light than does unhealthy vegetation. GCC best represented differences in healthy vigour and mortality, whereas EG best represented differences in foliar disease occurrence. EG and GCC values at both individual-tree and plot scales were significantly related to stand-health metrics. Thresholds for GCC and EG values can be developed using linear-regression equations that are based on ground-truthed data.   51 4.6 Tables and figures  Table 4.1. Significant relationships among tree-health and spectral-reflectance variables calculated at each site separately. Site Health Spectral R2 p-value Direction Black Pines Healthy vigour EGtree 0.6 0.0134 Positive   GCCtree 0.62 0.0121 Positive  Mortality EGtree 0.48 0.0374 Negative   GCCtree 0.5 0.0342 Negative  Root disease EGplot 0.48 0.0374 Negative Dairy Creek Mortality GCCplot 0.77 0.0044 Negative  Foliar disease EGtree 0.62 0.0204 Negative O'Connor Lake Mortality EGplot 0.48 0.0401 Positive  Root disease EGplot 0.59 0.015 Negative   EGtree 0.62 0.0112 Negative  Gall rust EGtree 0.6 0.0138 Negative   EGplot 0.62 0.0113 Negative Skulow  Root disease EGplot 0.84 0.0005 Negative   GCCplot 0.84 0.0005 Negative Topley Gall rust EGplot 0.63 0.0105 Negative   Table 4.2. The percent of variability in EGtree, EGplot, GCCtree and GCCplot explained by healthy vigour, mortality, foliar-disease occurrence, and root-disease occurrence calculated using all 54 plots. Total amount of variability is summarized at the bottom.   EGtree EGplot GCCtree GCCplot Healthy vigour 7.2% 5.8% 23.7% 37.7% Mortality 5.8% 0.9% 16.2% 11.4% Foliar disease 21.8% 18.4% 4.3% 4.8% Root disease 1.7% 24.9% 2.7% 7.0% TOTAL 36.5% 50.0% 46.8% 60.9%  52  Figure 4.1. Images at Black Pines, plot 4, in the original red-green-blue (RGB) format before manipulations (left), excess greenness (EG) values (middle) and green chromatic coordinate (GCC) values (right). EG and GCC images are grey-scale representations with lighter hues being higher values indicating a higher proportion of green light being reflected.    Figure 4.2. An image displaying the excess greenness representation for Black Pines plot 1. Analysis was conducted on regions of interest (ROIs) at the individual tree crown scale (left) and plot scale (right). Red polygons depict the ROIs used.  53  Figure 4.3. Relationships between the percentage of trees with healthy vigour and green chromatic coordinate (GCC) and excess greenness (EG) were significantly positive: GCCplot (p < 0.0001), GCCtree (p < 0.0001) and EGplot (p = 0.0083). Sites are colour coded and linear-regression equations are shown.    54  Figure 4.4. Relationships between tree mortality and green chromatic coordinate (GCC) and excess greenness (EG) were significantly negative: GCCplot (p = 0.0002) and GCCtree (p < 0.0001). Sites are colour coded and linear-regression equations are shown.  55  Figure 4.5. Relationships between foliar-disease occurrence and green chromatic coordinate (GCC) and excess greenness (EG) were significantly negative: EGplot (p = 0.0001), EGtree (p =  56 0.0029), GCCplot (p = 0.0334) and GCCtree (p = 0.0478). Sites are colour coded and linear-regression equations are shown.  Figure 4.6. Relationships between root-disease occurrence and green chromatic coordinate (GCC) and excess greenness (EG) were significantly negative: EGplot (p < 0.0001) and GCCplot (p = 0.0083). Sites are colour coded and linear-regression equations are shown.  57  Figure 4.7. Relationships between western-gall-rust occurrence and green chromatic coordinate (GCC) and excess greenness (EG) were significant. The relationship with EGtree (p = 0.0002) was negative, while with GCCtree (p < 0.0001) and GCCplot (p < 0.0001) were positive. Sites are colour coded and linear-regression equations are shown.  58  Figure 4.8. Relationships between the percentage of trees with healthy vigour and green chromatic coordinate calculated at a plot scale (GCCplot) at high, medium and low ranges of leaf-area index (LAI).   Figure 4.9. Relationships between the percentage of trees with healthy vigour and green chromatic coordinate calculated at a plot scale (GCCplot) at high, medium and low ranges of foliar nitrogen (%).   59  Chapter 5. Nutrient contribution of the forest floor 5.1 Synopsis Forest-floor removal is predicted to reduce forest productivity because of the importance of this soil layer as a nutrient pool and substrate for mycorrhizal fungi. The first 5-20 years of results from LTSP forest-floor removal experiments have been mixed and predict that reductions in productivity are more likely when stands reach canopy closure. Independent of forest-floor removal experiments, it has been found that Suillus species of ectomycorrhizal fungi provide pine trees with atmospherically fixed nitrogen and are adapted to survive in low nutrient soils. This chapter integrates data on canopy closure, topsoil-nutrient content, foliar-nutrient concentrations, nutrient deficiencies, total ectomycorrhizal abundance, Suillus species abundance, tree growth and tree health at these six long-term soil productivity (LTSP) sites to determine the nutritional impacts of forest-floor removal. Three sites are in the Interior Douglas Fir (IDF) zone and three in the Sub Boreal Spruce (SBS) zone. In both zones, forest-floor removal plots had smaller nutrient pools in the top 20 cm of soil. In the IDF zone this trend was consistent among sites, but in the SBS zone varied among sites. Foliar-nutrient concentrations were less affected by forest-floor removal. In the SBS zone, foliar P and K concentrations were reduced and this trend was consistent among sites. In the IDF zone, Suillus species abundance was higher in the forest-floor removal plots with this trend varying among sites. In the SBS zone, ectomycorrhizal abundance was lower and Suillus species abundance higher in the forest-floor removal plots with this trend being consistent among sites.   5.2 Introduction Removal or displacement of the forest floor during harvest-related activities became a concern for long-term forest productivity because of the forest floor’s critical role in nutrient availability, its sensitivity to harvesting (Tiedemann et al. 2000) and its long recovery time (Sayer 2006). The forest floor consists of the litter and humus or organic layers that form on top of the mineral soil. During forestry operations, the forest floor is removed to create roads, skid trails, landings and is sometimes displaced during site preparation treatments such as screefing or scarifying. Forest-floor removal has the potential to reduce productivity by removing nutrients and changing the  60 ectomycorrhizae community (Jurgensen et al. 1997; Page-Dumroese and Jurgensen 2006; Powers 2006).   The amount of nutrients stored in the forest floor depends on the age and type of forest, forest floor depth, as well as the nutrient concentrations. In a variety of temperate coniferous forests, the forest floor contains 3.1 to 8.2% of the total nitrogen pool (Johnson et al. 1982; Ares et al. 2007; Cole and Rapp 1980). In a 450-year-old Douglas-fir forest in Oregon, the ‘forest litter layer’ contained 31.6% of the phosphorus pool, 17.3% of the calcium pool and 8.1% of the potassium pool (Cole and Rapp 1980). Forest-floor removal could reduce site productivity by impacting the ectomycorrhizal fungi community, which is involved in nutrient acquisition (Jurgensen et al. 1997; Page-Dumroese et al. 2010b). In western forests, active ectomycorrhizal root tips are often most abundant in the forest floor (Harvey et al. 1986) and forest-floor removal is generally thought to reduce ectomycorrhizal species abundance (Harvey et al. 1980, Perry et al. 1989). However, ectomycorrhizal species in the Suillus genus are adapted to disturbances that remove the forest floor and can provide lodgepole pine growing in gravel pits with adequate nitrogen (Paul et al. 2007; Chapman and Paul 2012; Paul et al. 2013).  Regardless of the total soil nutrient pool, nutrient limitations are equally dependent on demands of the plants for nutrients. Nutrient requirements can vary greatly between seedlings and mature trees (Mahendrappa et al. 1986). Nutrients generally do not limit trees until canopy closure, when leaf area and soil nutrient demands are at their maximum (Thiffault et al. 2011; Mason et al. 2012; Ponder et al. 2012). Canopy closure can be measured using hemi-view photography and computer programs that delineate sky from vegetation.   Forest-floor removal experiments primarily monitor forest productivity by measuring tree growth; however, the complex effects and interactions among variables related to nutrition and forest productivity are rarely studied. The effects of forest-floor removal on tree growth have been mixed, depending on tree species, location and tree age (Van Cleve and Dyrness 1983;  61 Stone and Elioff 1998; Fleming et al. 2006; Kranabetter et al. 2006; Sanchez et al. 2006; Ponder et al. 2012). Foliar-nutrient concentrations ten years after forest-floor removal were lower in the Boreal Great Lakes region, not affected in the Western Montane region, and higher at two Mediterranean sites (Ponder et al. 2012). There were also positive, neutral and negative relationships between foliar-nutrient concentrations and stand productivity, depending on region (Ponder et al. 2012). Gomez et al. (2002) considered N mineralization, N uptake, foliar-nutrient status, and tree growth at three sites with different soil textures. They found forest-floor removal reduced N uptake in clay and loam soils but increased N uptake in sandy loam soils (Gomez et al. 2002). Even with this significant progress, several questions remain unanswered.   This study explores the nutritional impacts of forest-floor removal on regenerating lodgepole-pine stands at six long-term soil productivity (LTSP) sites in British Columbia, Canada (Figure 1). Specifically, this study tests five main questions:  1. Are these sites nearing canopy closure?  2. How much (size and percent) of the soil nutrient pool was removed during forest-floor removal? 3. Do foliar-nutrient concentrations indicate deficiencies? If so, is this a dilution effect from increased growth?  4. Are topsoil-nutrient content, foliar-nutrient concentrations or fungal abundance significantly different in forest-floor removal plots compared to bole-only harvest plots? 5. How do nutrition variables relate to tree growth and health variables at each site?  Hemi-view photography measured canopy closure to determine the stage of stand development. Nutrient amount and percentage removed during forest-floor removal were calculated. To determine whether or not nutrient deficiencies existed, foliar-nutrient concentrations were measured and compared to known deficiency levels. The effects of forest floor-removal on current topsoil-nutrient content, ectomycorrhizal abundance, Suillus abundance, and foliar-nutrient concentrations were tested relative to bole-only harvesting. Finally, how nutrient variables relate to tree growth and health was explored.   62 5.3 Methods Canopy closure Canopy closure was used to measure stand development. Canopy closure was measured using hemi-view photography during the summer of 2013 using a Nikon Coolpix5400 digital camera and a full (360˚) fisheye Zoom Nikkor ED lens. Five photos were taken in each plot: at the four corners and the centre. These photos were uploaded into Gap Light Analyzer version 2.0 (Frazer et al. 1999) and canopy openness calculated. Canopy closure is the inverse of canopy openness, which measures the percentage of sky showing in the canopy. As canopies are rarely 100% closed, it was necessary to determine what values would indicate closed canopies. This was accomplished by using the Table Interpolation Program for Standard Yields (TIPSY) interface that is based on the Tree And Stand Simulator (TASS) model (Mitchell et al. 2000). This is the most common growth-and-yield model used in British Columbia. TIPSY version 4.3 was used to predict canopy closure with input data of pure stands of lodgepole pine planted at a stocking rate of 1600 stems/ha on sites with an average slope of 3% and site index of 17. Models were generated in the SBS and IDF zones.  Soil nutrients The amount and percent of nutrients removed during forest-floor removal was calculated with chemical and physical properties measured in both the forest-floor and mineral-soil layers the year before harvesting. Forest-floor and mineral soil was collected at 15 locations within each plot and bulked into three samples for analysis. Samples were oven-dried and sent to the BC Ministry of Environment’s Technical Services Section Laboratory in Victoria, BC for chemical analysis of the concentrations (mg/kg) of carbon (C), nitrogen (N), mineralizable N (minrl N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg) and sulfur (S). Nutrient content (kg/ha) was calculated for both the forest floor and top 20 cm of mineral soil by,  = C (kg/kg) * BD (kg/cm3) * V (cm3/ha),  where C is nutrient concentration, BD is bulk density, V is the volume of soil per hectare. Volume of the forest floor per hectare was calculated using measurements of forest-floor thickness at each plot. Bulk density of the forest floor was calculated from the mass of the forest  63 floor taken from a known sample volume. Volume of the mineral soil was calculated for the top 20 cm of soil. Bulk density of the mineral soil was calculated by dividing fine-fraction mass (< 2 mm) by the total sample volume. This provides a measure of nutrient content in the soil fine fraction to account for differences in the amounts of coarse fragments among sites. The amount of nutrients removed during forest-floor removal is considered to be the nutrient content of the forest floor prior to harvest and the percent removed during forest-floor removal is considered to be the nutrient content of the forest floor divided by the sum of nutrient content in the forest-floor and mineral soil-layers (hereafter referred to as topsoil).   The same soil chemical and physical measurements were collected at years 15 in the IDF sites and 20 in the SBS sites as a measure of ‘current’ nutrient content. Methodology and calculations were the same as described above. This current-nutrient content data was used to test for treatment effects and for relationships between tree health and growth.   Foliar-nutrient concentration When trees started dormancy (September to October), foliar samples were collected from 15 trees per plot and bulked together into three sub-samples (Berch et al. 2010). Foliage was dried, milled and analyzed for concentrations (%) of N, P, K, Ca, Mg and S at the BC Ministry of Environment’s Technical Services Section Laboratory in Victoria, BC. For each sample, 100 needles were weighed to estimate average needle mass (mg). Deficiency levels were based on those reported in Brockley (2001). When nutrient deficiencies were detected, dilution effects were considered by determining how LAI and needle weight relate to foliar-nutrient concentrations. Negative relationships would indicate a dilution effect.  Soil fungi Data on the relative abundance of fungal genes in the mineral soil and forest floor were obtained from Hartmann et al. (20124). Genetic material was collected from the same soil samples used for the nutrient analysis described above. Deoxyribonucleic acids (DNA) were extracted from these soil samples using the FastDNA SPIN Kit and quantified using the Quant-iT Pico-Green kit.                                                 4 Data available at www.nature.com.ezproxy.library.ubc.ca/ismej/journal/v6/n12/suppinfo/ismej201284s1.html  64 Genes were amplified and the internal transcribed spacers (ITS2) sequenced using the 454-pyrosequencing GS-FLX Titanium technology (Hartman et al. 2012). Relative abundance of fungal species or groups of species is calculated as the proportion of genes from a species relative to the total number of genes in that sample.  Relative abundance of putative ectomycorrhizal species and species in the Suillus genus were calculated within each sample. These relative abundances were averaged by plot. As there were no samples for the forest-floor layer in the forest-floor removal plots, these average values for the forest-floor and mineral-soil layer were summed together to allow for statistical analysis among treatment plots.   Stand health Stand health was surveyed between July and September 2013 (Reid et al. 2015). Every tree leader, needles, branches, stem, and root collar were inspected for macroscopic symptoms of pest occurrence. Pest identity, presence, and severity were recorded. Foliar disease, western gall rust and root-disease symptoms were used for analysis because they occurred at all sites. Foliar-disease severity was high if greater than 50% of the foliage was affected, medium if 25 to 50% of the foliage was affected and low if less than 25% of the foliage was affected. Resinosis and lesions near the root collar were both considered symptoms of root disease (Morrison et al. 1991). Every tree was given a vigour-classification rating from 1 to 5 (Newsome and Perry 2002). Class 1 is a tree with vigorous growth, good condition, and no visual indication of stress. Class 2 is a tree with moderate growth rate, fair condition, and some visual indication of stress (minor defects). Class 3 is a tree in poor condition with obvious visual signs of severe stress, may have major defects, and has a poor growth rate. A Class 4 tree is moribund and a Class 5 tree is dead. Missing trees were assumed to be dead. Five metrics of stand health were developed from the forest-health survey: the number of (1) healthy trees - Class 1, (2) dead or dying trees - Classes 4 and 5, (3) trees with high or medium foliar disease, (4) trees with galls from western gall rust and (5) trees with root-disease symptoms.  Tree growth and stand productivity  65 Tree height (m), total tree volume (cm3), and Leaf-Area Index (LAI) were used to measure lodgepole pine growth. Tree height and diameter were measured5 when IDF sites were 15 years old and SBS sites were 20 years old. Tree height was measured with a pole and diameter with a measuring tape at breast height (130 cm). Tree volume (cm3) was calculated for each individual tree by,  = EXP(-2.15873+1.86671*LN(Dia)+1.18486*LN(Ht)+1.49899*(h/Ht)),  where Dia is tree diameter (cm), Ht is tree height (cm) and h is the height (cm) above ground level at which diameter was measured (Kovats 1977). Here diameter at breast height was used, so h = 130 cm above ground level. Total-tree volume (cm3) is the sum of all the individual tree volumes per plot. LAI was assessed based on the same hemi-view photos used to calculate canopy closure. LAI was calculated in Gap Light Analyzer version 2.0 (Frazer et al. 1999) by,  = ln(P(θ))cos(θ)/G(θ),  where θ is the zenith angle of view P(θ) is the gap fraction and G(θ) is the fraction of foliage projected on the plane parallel to the ground (Bréda 2003). The number of azimuth regions was set to 16 and zenith regions to 10. All analyses used LAI ring 5, which integrates over zenith angles 0 to 75˚ (Frazer et al. 1999).  Analysis Significant differences in topsoil-nutrient content, ectomycorrhizal abundance, Suillus abundance, and foliar-nutrient concentrations between forest-floor removal plots and bole-only harvest plots were tested using factorial GLMs. This analysis tests the effect of forest-floor removal relative to bole-only harvest on topsoil-nutrient content, foliar-nutrient concentrations, and fungal-community relative abundance. Soil-compaction level and site are included as covariates and their interactions with organic-matter removal tested. Significant interaction terms indicate that soil-compaction level or site influences the effect of forest-floor removal on nutrient                                                 5 Tree height and diameter data originate from the BC Ministry of Forests, Lands and Natural Resource Operations.  66 variables. Non-significant interaction terms indicate that the effect of forest-floor removal on nutrient variables is consistent among soil-compaction level or sites. Analysis was done with plot averages within each BEC zone.   To determine how nutrient variables (soil, foliar, and fungal) relate to tree growth and health, a correlation matrix was generated with correlation-coefficient (r) values calculated at each site. The directions of these relationships are compared among sites for each nutrient-tree combination. Analyses were conducted in JMP 11 with significance considered at p < 0.05 (SAS Institute Inc. 2012).   5.4 Results  Average canopy closure ranged from 50 - 78% among sites (Table 5.2). TIPSY predicts the highest canopy closure values for these stands to be 80%, occurring when stands are between 20 and 25 years old (Figure 5.2). Log Lake and Topley had the highest canopy-closure values (Table 5.2) and are the closest to canopy closure (Figure 5.2).   For many soil nutrients, the greatest amounts (kg/ha) removed as a result of forest-floor removal were at O’Connor Lake and Topley (Table 5.3). The highest proportions (%) of soil nutrients removed as a result of forest-floor removal were at O’Connor Lake and Log Lake (Table 5.4). The exception being that the highest proportions (%) of soil P and K being removed as a result of forest-floor removal at Skulow and Topley (Table 5.4).   Foliar nitrogen concentrations were at least slightly deficient at all sites except Dairy Creek where nitrogen was not deficient (Table 5.5). Foliar phosphorus concentrations were slightly to moderately deficient at O’Connor Lake, Skulow Lake and Topley (Table 5.5). Foliar potassium concentrations were slightly deficient at Topley, but adequate at all other sites (Table 5.5). Foliar calcium and magnesium concentrations were adequate (Table 5.5). Foliar sulfur concentrations indicated deficiencies at all sites (Table 5.5). No dilution effects were detected by linear regressions with needle mass or LAI. The only significant relationship was at O’Connor Lake, where foliar sulfur concentrations positively related to LAI (R2 = 0.47, p = 0.0412).    67 In the IDF zone, forest-floor removal plots had less soil nitrogen, phosphorus, potassium, sulfur, calcium, and magnesium content (kg/ha) compared to the bole-only harvest plots (Table 5.6). This trend was consistent among soil-compaction levels and sites, indicated by the lack of a significant interaction terms (Table 5.6). In the SBS zone, forest-floor removal influenced soil nitrogen, potassium, sulfur, and calcium content (kg/ha) compared to the bole-only harvest plots; however, site influenced all these trends (Table 5.6). In both zones, forest-floor removal significantly affected foliar Ca and Mg concentrations, but these effects were site specific (Table 5.7). In the SBS, forest-floor removal significantly reduced foliar P and K concentrations, and this effect was consistent among sites (Table 5.7). In the IDF, forest-floor removal increased Suillus species relative abundance, but this effect varied among sites (Table 5.8). In the SBS, forest-floor removal reduced ectomycorrhizae and increases Suillus species relative abundance, and was consistent among sites (Table 5.8).   The relationships between nutrient variables and tree condition (growth and health) varied among sites (Table 5.9-5.11). The exception being foliar Mg concentrations were always positively related to tree height (m) and root disease occurrence (Table 5.10).   5.5 Discussion Although nutrient content in the forest-floor removal plots was lower than in the bole-only harvest plots, this has not translated into widespread foliar nutrient deficiencies or loss of productivity. This could be related to most sites not yet reaching canopy closure. In the SBS zone, foliar P and K concentrations were reduced in the forest-floor removal plots relative to the bole-only harvest plots. These sites were also closest to canopy closure. This evidence holds open the possibility that treatment effects may become more pronounced after canopy closure (Thiffault et al. 2011; Ponder et al. 2012). However, the mechanism to explain this trend, that nutrients only become limiting after canopy closure, was not strongly supported by the data because there were nutrient deficiencies at all sites regardless of canopy closure. There are obviously more questions related to nutrient availability and the influence of forest-floor removal on availability that could be tested.    68 The lack of effect of forest-floor removal on foliar N concentrations and forest productivity may be attributable to additional nitrogen fixation. On the same sites, stable isotope analysis using 15N indicates that trees growing on the forest-floor removal plots derive more nitrogen from atmospheric sources (unpublished data). Although abundance of mycorrhizal fungi was generally lower in forest-floor removal plots, Suillus abundance was increased. The increase in Suillus species abundance in the forest-floor removal plots could be allowing lodgepole pine trees to access similar amounts of N although the soil N pool has been reduced. Suillus species are disturbance-adapted species that allow the early successional species of lodgepole pine to rapidly colonize areas with low nutrient content (Chapman and Paul 2012). Diazotrophic bacteria that fix atmospheric nitrogen reside within Suillus fungi allowing fixed nitrogen to become available to the trees (Paul et al. 2013). As Suillus are mainly associated with pine trees, this mechanism cannot explain why there are no treatment effects for other tree species at the LTSP sites. Additional N-fixation could also be associated with increased Alnus abundance in the forest-floor removal plots at some sites (personal observations).   Additional nitrogen fixation in the forest-floor removal plots does not explain what might be compensating for other nutrients. One possibility is that forest-floor removal allows the mineral soil to warm up faster in the spring (Kranabetter and Chapman 1999; Zabowski et al. 2000; Kamaluddin et al. 2005; Roberts et al. 2005). Warmer soil temperatures could increase nutrient availability by increasing decomposition rates (Kranabetter and Chapman 1999), the growing season, or rooting depth. This mechanism would be most pronounced in cold wet climates. In dry-warm climates, forest-floor removal could reduce productivity by increased temperatures leading to drier soils (Kamaluddin et al. 2005). These effects related to microclimate condition would have the largest influence on tree growth within the first 5 to 10 years (Thiffault et al. 2011).   The relationships between nutrient variables and tree-condition variables (growth and health) varied among sites. This is consistent with the site-specific effects of forest-floor removal on tree growth (Ponder et al. 2012) and health (Reid et al. 2015).    69  5.6 Tables and figures  Table 5.1. Foliar-nutrient concentrations indicating adequate (green), slight to moderate deficiency (yellow), moderate to severe deficiency (orange) and severe deficiency  (red) for tree growth (Brockley 2001).  Nutrient Deficiency rating Foliar conc. (%) N Severe < 1.00  Moderate to severe 1.00 - 1.15  Slight to moderate  1.15 - 1.35  Adequate > 1.35 P Severe < 0.08  Moderate to severe 0.08-0.10  Slight to moderate  0.10-0.12  Adequate > 0.12 K Severe < 0.30  Moderate to severe 0.30 - 0.35  Slight to moderate  0.35 - 0.40  Adequate > 0.40 Ca Severe < 0.06  Moderate to severe 0.06 - 0.08  Slight to moderate  0.08 - 0.10  Adequate > 0.10 Mg Severe < 0.04  Moderate to severe 0.04 - 0.06  Slight to moderate  0.06 - 0.08  Adequate > 0.08 S Severe < 0.06  Moderate to severe 0.06 - 0.08  Slight to moderate  0.08 - 0.10  Adequate > 0.10  Table 5.2. Mean and standard deviation (std dev) values of canopy closure calculated at each site. Sorted from highest to lowest canopy-closure values.  Site Mean Std Dev Log Lake 78 2.8 Topley 74 4.4 Dairy Creek 69 2.0 Black Pines 63 11.5 O'Connor Lake 52 6.2 Skulow Lake 50 2.5    70 Table 5.3. The amount (kg/ha) of carbon (C), nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg) and sulfur (S) in the forest floor during the pre-harvest soil survey in the forest-floor removal plots. Darker shading indicates higher numbers.  BEC Site C N P S K Ca Mg IDF Black Pines 20742 648 5 64 52 780 39  Dairy Creek 29446 894 9 86 56 684 51  O'Connor 31586 1256 6 120 89 1332 79 SBS Log Lake 24593 742 7 84 56 447 32  Skulow 16572 527 5 42 70 229 83  Topley 37235 1063 7 121 92 453 45   Table 5.4. The proportion (%) of carbon (C), nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg) and sulfur (S) removed in the forest floor relative to the topsoil in the forest-floor removal plots. Darker shading indicates higher numbers.  BEC Site C N P S K Ca Mg IDF Black Pines 40 28 3 30 15 22 18  Dairy Creek 45 34 3 42 15 23 18  O'Connor 47 36 4 45 20 28 21 SBS Log Lake 54 38 9 56 39 41 31  Skulow 45 27 15 35 46 12 8  Topley 46 35 39 41 42 17 10   Table 5.5. Foliar-nutrient concentrations (%) calculated on a dry mass basis with colour coding for nutrient deficiency ratings (Table 3). When nutrients did indicate deficiencies, there was no evidence for dilution effects (no dilution = ND). Site N (%) P (%) K (%) Ca (%) Mg (%) S (%) Black Pines 1.21 ND 0.13 0.54 0.22 0.09 0.09 ND Dairy Creek 1.38 0.14 0.60 0.23 0.12 0.10 ND Log Lake 1.27 ND 0.14 0.45 0.19 0.10 0.10 ND O'Connor Lake 1.25 ND 0.12 ND 0.61 0.27 0.10 0.08 ND* Skulow Lake 1.16 ND 0.12 ND 0.41 0.23 0.16 0.09 ND Topley 1.07 ND 0.12 ND 0.40 ND 0.21 0.08 0.08 ND * Positive relationship with LAI (R2 = 0.47, p = 0.0412).    71 Table 5.6. Effects of organic-matter removal (OMR; forest-floor removal plots versus bole-only harvest plots), soil compaction (COMP), and site on topsoil-nutrient content (kg/ha). Analyses were conducted with plot averages within each BEC zone. Bold font indicates significance considered at p < 0.05.  BEC Factor N P K S Ca Mg IDF         OMR <.0001 0.0011 0.0021 <.0001 0.0129 0.0472  COMP 0.0305 0.2514 0.0842 0.0529 0.214 0.0053  Site 0.3402 <.0001 0.0353 0.0271 0.0847 0.0019  OM*COMP 0.3075 0.1179 0.2374 0.4296 0.3044 0.1851  OM*Site 0.4264 0.112 0.373 0.4027 0.785 0.6801 SBS         OMR <.0001 0.0516 <.0001 <.0001 0.0068 0.1214  COMP 0.141 0.4885 0.2799 0.0181 0.2711 0.6161  Site <.0001 <.0001 <.0001 <.0001 <.0001 <.0001  OM*COMP 0.0027 0.0587 0.0689 0.0045 0.029 0.002  OM*Site 0.0008 0.7141 0.0253 0.0019 0.0311 0.0112   Table 5.7. Effects of organic-matter removal (OMR; forest-floor removal plots versus bole-only harvest plots), soil compaction (COMP), and site on foliar-nutrient concentration (%). Analyses were conducted with plot averages within each BEC zone. Bold font indicates significance considered at p < 0.05.  BEC Factor Fol N Fol P Fol K Fol S Fol Ca Fol Mg IDF         OMR 0.8334 0.7834 0.7626 0.1517 0.0004 0.0003  COMP 0.3269 0.3968 0.0184 0.1267 0.8522 0.3143  Site 0.026 0.0032 0.0012 0.0165 <.0001 <.0001  OM*COMP 0.9295 0.896 0.0824 0.581 0.0253 0.2587  OM*Site 0.7152 0.4571 0.0036 0.4976 <.0001 0.0185 SBS         OMR 0.1886 0.022 0.0236 0.8845 0.0013 0.0461  COMP 0.1963 0.6103 0.0905 0.0866 0.2323 0.0278  Site <.0001 <.0001 0.0021 <.0001 0.0001 <.0001  OM*COMP 0.0569 0.4122 0.337 0.4724 0.1895 0.7075   OM*Site 0.0121 0.5065 0.0739 0.3432 0.0309 0.0482   72 Table 5.8. Effects of organic-matter removal (OMR; forest-floor removal plots versus bole-only harvest plots), soil compaction (COMP), and site on ectomycorrhizal (ecto) and Suillus relative abundance. Analyses were conducted with plot averages within each BEC zone. Bold font indicates significance considered at p < 0.05.  BEC Factor Ecto.  Suillus spp. IDF     OMR 0.1789 <.0001  COMP 0.2984 0.3572  Site <.0001 <.0001  OM*COMP 0.2527 0.0026  OM*Site 0.3609 0.0051 SBS     OMR 0.0075 <.0001  COMP 0.1901 0.0224  Site 0.0048 0.291  OM*COMP 0.3011 0.1011   OM*Site 0.1682 0.2028     73 Table 5.9. Correlation-coefficient (r) values for the relationships of soil-nutrient content (kg/ha) with tree growth and health variables are summarized. Green shading indicates correlation-coefficient values above zero and red shading indicates correlation-coefficient values below zero. Relationships were tested within each site to determine consistency among sites. Nutrient Site Height Tot. vol. Healthy Dead Foliar Gall rust Root N (kg/ha)          Black Pines -0.09 0.11 -0.68 0.91 -0.05 -0.66 0.48  Dairy Creek 0.29 0.00 -0.80 0.51 -0.04 0.49 -0.02  Log Lake 0.16 0.16 -0.10 0.90 -0.36 -0.65 -0.33  O'Connor Lake -0.65 -0.76 0.09 0.90 -0.56 -0.64 -0.64  Skulow Lake -0.23 -0.57 -0.01 -0.14 0.53 -0.34 -0.86  Topley -0.57 -0.27 -0.40 -0.49 0.40 -0.78 0.72 P (kg/ha)          Black Pines 0.05 0.16 -0.88 0.78 0.26 -0.40 0.64  Dairy Creek -0.19 -0.28 -0.63 0.56 -0.43 0.35 0.26  Log Lake 0.51 0.61 0.30 0.48 -0.51 -0.63 0.00  O'Connor Lake -0.24 -0.24 -0.25 0.26 -0.02 -0.11 -0.16  Skulow Lake -0.13 -0.22 -0.38 0.32 0.33 0.00 -0.43  Topley 0.26 0.55 -0.83 -0.22 0.92 0.06 0.00 K (kg/ha)          Black Pines -0.39 -0.21 -0.52 0.66 -0.47 -0.85 0.15  Dairy Creek 0.02 0.22 -0.43 0.09 -0.22 0.59 -0.06  Log Lake -0.02 -0.01 -0.22 0.89 -0.36 -0.63 -0.38  O'Connor Lake -0.73 -0.79 0.19 0.73 -0.60 -0.62 -0.48  Skulow Lake -0.01 -0.43 -0.28 0.31 0.62 -0.63 -0.94  Topley -0.44 -0.08 -0.56 -0.50 0.61 -0.65 0.57 S (kg/ha)          Black Pines -0.10 0.08 -0.63 0.88 -0.13 -0.67 0.46  Dairy Creek 0.25 -0.01 -0.78 0.51 -0.09 0.50 0.00  Log Lake -0.13 -0.07 -0.11 0.86 -0.52 -0.74 -0.30  O'Connor Lake -0.42 -0.56 -0.10 0.85 -0.39 -0.43 -0.54  Skulow Lake 0.26 -0.27 0.29 0.16 0.47 -0.52 -0.95  Topley -0.52 -0.31 -0.31 -0.47 0.28 -0.75 0.86 Ca (kg/ha)          Black Pines -0.40 -0.22 -0.69 0.78 -0.29 -0.83 0.20  Dairy Creek 0.54 0.50 -0.48 0.06 0.10 0.77 -0.18  Log Lake 0.04 0.07 -0.05 0.91 -0.47 -0.69 -0.34  O'Connor Lake -0.36 -0.43 -0.07 0.59 -0.35 -0.22 -0.19  Skulow Lake -0.54 -0.43 0.47 -0.87 -0.32 0.34 0.16  Topley -0.76 -0.65 0.11 -0.36 -0.13 -0.81 0.69 Mg (kg/ha)          Black Pines -0.49 -0.32 -0.48 0.66 -0.53 -0.88 0.04  Dairy Creek 0.43 0.52 -0.44 -0.05 0.17 0.59 -0.47  Log Lake 0.07 0.02 -0.28 0.91 -0.21 -0.40 -0.56  O'Connor Lake 0.08 -0.06 -0.20 0.54 -0.17 0.18 0.18  Skulow Lake -0.54 -0.24 0.27 -0.38 -0.65 0.94 0.48  Topley -0.56 -0.43 0.16 -0.38 -0.25 -0.71 0.79    74 Table 5.10. Correlation-coefficient (r) values for the relationships of foliar-nutrient concentration (%) with tree growth and health variables are summarized. Green shading indicates correlation-coefficient values above zero and red shading indicates correlation-coefficient values below zero. Relationships were tested within each site to determine consistency or differences among sites. Nutrient Site Height Tot. vol. Healthy Dead Foliar Gall rust Root Fol N            Black Pines 0.44 0.43 -0.07 0.53 0.29 0.16 0.60  Dairy Creek -0.11 0.61 0.33 -0.78 0.24 -0.01 -0.90  Log Lake -0.34 -0.39 -0.62 0.80 -0.05 -0.66 -0.16  O'Connor Lake 0.26 0.13 -0.38 0.31 0.31 -0.08 -0.53  Skulow Lake 0.63 0.38 -0.19 0.64 0.27 -0.69 -0.21  Topley 0.55 0.55 0.09 0.57 -0.43 0.22 -0.25 Fol P            Black Pines 0.14 -0.01 0.04 0.10 0.05 0.22 0.23  Dairy Creek 0.60 0.85 0.21 -0.68 0.57 0.34 -0.67  Log Lake -0.68 -0.57 -0.28 0.50 -0.45 -0.74 0.05  O'Connor Lake 0.09 -0.08 -0.19 0.57 -0.02 -0.18 -0.51  Skulow Lake 0.48 0.16 -0.28 0.38 0.63 -0.94 -0.52  Topley 0.45 0.71 -0.49 0.23 0.18 -0.10 0.16 Fol K            Black Pines 0.06 -0.09 0.21 -0.03 -0.09 0.20 0.06  Dairy Creek -0.37 0.53 0.76 -0.88 -0.13 0.04 -0.41  Log Lake -0.41 -0.54 -0.67 0.60 0.25 -0.52 0.05  O'Connor Lake -0.39 -0.54 0.30 0.75 -0.59 -0.49 -0.44  Skulow Lake 0.16 -0.16 -0.22 0.26 0.42 -0.79 -0.54  Topley 0.25 0.45 0.19 -0.52 0.05 0.52 -0.33 Fol S            Black Pines 0.55 0.49 0.40 0.17 0.23 0.42 0.38  Dairy Creek 0.48 0.86 0.35 -0.84 0.56 0.28 -0.66  Log Lake -0.66 -0.69 -0.48 -0.54 0.43 0.48 0.09  O'Connor Lake 0.75 0.64 -0.74 -0.01 0.72 0.51 -0.01  Skulow Lake 0.62 0.38 -0.38 0.85 0.34 -0.63 -0.27  Topley 0.44 0.41 0.14 -0.08 -0.41 0.03 0.71 Fol Ca            Black Pines 0.74 0.62 0.28 -0.32 0.80 0.89 0.32  Dairy Creek -0.57 -0.66 0.00 0.58 -0.73 -0.12 0.76  Log Lake -0.54 -0.49 -0.51 0.09 -0.03 -0.01 -0.23  O'Connor Lake 0.95 0.97 -0.49 -0.67 0.75 0.97 0.82  Skulow Lake 0.22 0.32 0.25 0.42 -0.75 0.46 0.57  Topley 0.40 0.19 0.80 0.12 -0.86 0.56 -0.25 Fol Mg            Black Pines 0.66 0.39 0.22 -0.46 0.68 0.98 0.34  Dairy Creek 0.58 0.66 0.56 -0.67 0.43 0.30 0.02  Log Lake 0.64 0.56 -0.03 0.33 0.14 -0.61 0.43  O'Connor Lake 0.98 0.96 -0.63 -0.50 0.79 0.92 0.63  Skulow Lake 0.45 0.80 -0.31 -0.22 0.43 -0.31 0.54  Topley 0.09 -0.20 0.75 0.30 -0.96 0.08 0.13   75 Table 5.11. Correlation-coefficient (r) values for the relationships of fungal-species relative abundance with tree growth and health variables are summarized. Green shading indicates correlation-coefficient values above zero and red shading indicates correlation-coefficient values below zero. Relationships were tested within each site to determine consistency or differences among sites. Nutrient Site Height Tot. vol. Healthy Dead Foliar Gall rust Root Ecto          Black Pines 0.69 0.85 0.09 0.68 0.31 0.04 0.72  Dairy Creek -0.77 -0.46 0.26 0.34 -0.90 -0.11 0.49  Log Lake -0.03 -0.11 -0.33 0.96 -0.16 -0.32 -0.67  O'Connor Lake -0.73 -0.65 0.64 0.11 -0.59 -0.63 -0.22  Skulow Lake 0.84 0.81 0.36 -0.22 0.47 -0.40 0.09  Topley 0.54 0.73 -0.88 0.05 0.72 0.02 0.33 Suillus spp          Black Pines -0.09 -0.16 0.59 -0.56 0.02 0.39 -0.57  Dairy Creek -0.52 0.22 0.92 -0.60 -0.43 -0.07 0.09  Log Lake 0.46 0.49 0.50 -0.32 -0.08 0.70 -0.42  O'Connor Lake 0.57 0.72 -0.05 -0.95 0.53 0.62 0.69  Skulow Lake -0.20 0.27 0.09 -0.70 -0.25 0.37 0.80  Topley 0.60 0.24 0.27 0.48 -0.48 0.46 0.00         76  Figure 5.1. A conceptual diagram of relationships tested in this chapter. Forest-floor removal treatments could influence topsoil-nutrient content, ectomycorrhizae relative abundance, Suillus relative abundance, and foliar-nutrient concentrations. Theses variables, in turn, interact tree growth and health.        77 Data	typeBPDCLLOLSLTIPSYTOMean(Crown	Cover	(%))	&	Crown	Cover	(%)	vs.	AgeAge10 20 30 40Crown	Cover	(%)406080 Figure 5.2. Crown cover (%) based on TIPSY model outputs (black) and from collected data at each site: Black Pines (BP, red), Dairy Creek (DC, orange), O’Connor Lake (OC, yellow), Log Lake (LL, green), Skulow Lake (SL, blue) and Topley (TO, purple).    78 Chapter 6. Conclusion and synthesis This thesis integrated forest health into monitoring and managing for forest productivity using six Long-Term Soil Productivity (LTSP) sites in the interior of British Columbia, Canada. This research collaborated with the BC Ministry of Forests, Lands and Natural Resource Operations and the LTSP network that has over 120 sites throughout North America. The main findings of this research include: 1. Soil-disturbance treatments had site-specific effects on ground-based measures of forest health: healthy vigour, mortality, foliar disease, stem rusts, and root disease.  2. Treatment plots and random groups with larger trees also had higher disease occurrence.  3. The spectral-reflectance index of green chromatic coordinate correlated well with healthy vigour when calculated over entire treatment plots. 4. Forest-floor removal reduced topsoil-nutrient content and ectomycorrhizal abundance, but increased Suillus species abundance relative to bole-only harvest plots.  5. Forest-floor removal reduced foliar P and K concentrations relative to bole-only harvest plots only at the SBS zone, which had the highest canopy closure.   Contributing novel data on forest health has provided a baseline for future research and suggested that soil disturbance can influence forest-health metrics. The abundance of pests identified in this research can be used as a baseline to track changes over time and how pest occurrence now may influence future biomass production. The observed ranges of pest occurrence here may be applicable to a larger area of lodgepole pine forests in BC because they correspond to those documented in Heinemann et al. (2010). The forest-health metrics developed in this research, including pest occurrence, differed among treatment plots often with greater magnitude than differences in tree growth. This suggests that forest health may be more ‘sensitive’ to soil-disturbance treatments than tree growth. These differences were site-specific, suggesting these effects are context-dependent.   At all sites, there was evidence in conflict with the current management paradigm that fast-growing trees will be healthy, or at least not have more pest issues compared to slow-growing trees. If this trend is widespread, the last 50 years of management based on tree growth could impact future forest health and productivity. A much larger coordinated effort to  79 comprehensively test this relationship is needed before any conclusive implications are drawn. Nutrition may be related to this relationship given that in Chapter 5 foliar nutrients were positively related to growth but had mixed relationships with health. It is possible that better nutrition makes trees grow faster but also makes them more susceptible to disease attack. This could be related to a physiological trade-off between allocating resources to growth versus defence (Herms and Mattson 1992) or that more nutritious vegetation is simply more desirable to pests.   The increased need for standardized objective forest-health monitoring makes the use of spectral-reflectance indices attractive. These remote-sensing techniques are rapidly becoming cheaper and more accessible (Nijland et al. 2014). These methods could compliment ground-based surveys by providing more spatial coverage and compliment sketch-map surveys by providing more consistent results between investigators. This research found that indices of excess greenness and green chromatic coordinate correlated well with ground-based measures of overall tree vigour and foliar disease. Gall rust occurrence was not reliably detected most likely due to the confounding signal of healthy trees at sites with higher precipitations due to the live history of gall rust. The most appropriate spatial scale used to calculate spectral-reflectance indices will depend on the question and end use of the information. Calculating indices on each individual tree crown have less variability and more replication, which can be useful when testing subtle treatment differences that need to be focused on individual trees. This analysis is more time-consuming than calculating indices over treatment plots. If the objective is to compare overall forest condition between different areas or through time, it might be more efficient and to calculate spectral-reflectance over larger ‘plot’ or ‘stand’ scales. The best indicator of overall health was green chromatic coordinate calculated over the entire plot.   The somewhat counterintuitive findings that, in general, the LTSP sites have been resilient to forest-floor removal may be related to processes that compensate for nutrient removal or stand-development dynamics. There was evidence that forest-floor removal reduces the topsoil nutrient pools at most sites. There is also a suggestion that nitrogen-fixing species become more abundant in the forest-floor removal plots. We found that Suillus species of ectomycorrhizae were more abundant in the forest-floor removal plots. These species are disturbance-adapted and are  80 associated with nitrogen-fixing bacteria that can provide lodgepole pine with sufficient amounts of nitrogen even in the poorest soil (Paul et al. 2006; Paul et al. 2007; Chapman and Paul 2012; Paul et al. 2013). The hypothesis that treatment effects will become apparent after canopy closure was not contradicted by these results, but remains unresolved. Foliar-nutrient concentrations were lower in the forest-floor removal plot only in the SBS zone, which was closest to canopy closure. The explanation for this phenomenon is that nutrients are not generally limiting before canopy closure (Thiffault et al. 2011; Mason et al. 2012). From this explanation, it would be predicted that the indications of any nutrient deficiencies for that particular site (dependent on parent material) would also only become apparent after canopy closure. The foliar nutrient data examined in this research indicated nutrient deficiencies at all sites, no matter how close they were to canopy closure. This is likely related to the complex issue of nutrient availability rather than nutrient pools.  6.1 Synthesis This section brings together datasets presented from all research chapters to present new analyses that explore ecosystem resilience, context-dependencies, and validity of predictive models.   Resilience of forest productivity to disturbance during logging This research can be used to comment on the resistance and resilience of forest productivity (growth and health) to one soil disturbance event during harvesting. Resistance is defined as the capacity of an ecosystem to absorb disturbances and remain relatively unchanged (Holling 1973; Carpenter et al. 200; Folke et al. 2004). Resistance, in terms of forest management, has been thought of as the influence of system characteristics on the magnitude of disturbance effects (DeRose and Long 2014). Resistance can be measured by the degree of change occurring after different levels of disturbance (DeRose and Long 2014; Standish et al. 2014). Ecological resilience is related to the amount of disturbance an ecosystem can withstand without being pushed into an alternative stable state (Holling 1973; Holling 1996). To measure ecological resilience directly it would be necessary to push a system into an alternate stable state to quantify the amount of disturbance needed (Standish et al. 2014). For example, changing a forest system to a bog system or a grassland system. Engineering resilience is generally defined as the rate of recovery of an ecosystem component after disturbance (Gunderson et al. 2010). It can be  81 measured by the trajectory of recovery of certain ecosystem attributes over time (Gunderson et al. 2010; Standish et al. 2014).   When considering ecosystem resilience and resistance it is necessary to define resilience of what ecosystem properties, to what type and magnitude of disturbance, during what temporal scale and over what spatial scale (Carpenter et al. 2001; DeRose and Long 2014). For the LTSP network, the time scale would be approximately 20 years and the spatial scale could range from continental to individual sites. There are existing data that could be used to consider the trajectory of wood biomass recovery and the resistance of tree growth and health to the soil-disturbance treatments specifically or harvesting in general. ‘Sensitive’ sites can be identified as those with low resilience or resistance to disturbance.  The resistance of tree growth and health to pulse disturbances of organic-matter removal and soil compaction during logging can be tested relative to a baseline (bole-only harvesting with no compaction) using response ratios. Response ratios (RRs) are calculated by,   RR = treatment / reference,  where reference are values at the bole-only and no-compaction treatment plots, and treatment are values at all other treatment plots (Thiffault et al. 2011). Response-ratio values greater than one indicate the measure increased after additional soil disturbance, and response-ratio values less than one indicate the measure decreased after additional soil disturbance. Bole-only harvesting and no-compaction plots are used as the ‘reference’ case to test the influence of additional levels of soil disturbance on the measure of interest (Thiffault et al. 2011). At O’Connor Lake, one gall-rust occurrence was added to all treatment plots to enable response ratio calculations that were not possible otherwise, because no visible signs of gall rust were detected on the reference plot. Response ratios are well suited to the LTSP sites, where the magnitude of change varies between sites (Hedges et al. 1999; Fleming et al. 2006; Thiffault et al. 2011; Ponder et al. 2012).  Concerns of loss of productivity would be valid if treatments reduced growth and healthy vigour (RR < 1) and increased death and total disease (RR > 1). The average response ratio of all eight- 82 treatment plots at all six sites show that there is generally a positive response of total stem volume and disease, but a negative response to healthy vigour (Table 6.1). Treatments had no effect on tree height and only slightly increased dead or dying trees relative to bole-only harvesting with no compaction (Table 6.1). These overall trends are site-specific (Table 6.1). Skulow Lake had the lowest resilience in terms of tree height, total volume and healthy vigour (smallest RRs; Table 6.1).   Table 6.1. Average response ratio for tree height, total stem volume, healthy vigour, dead or dying trees and total-tree disease calculated from all eight-treatment plots. Values greater than one indicate an increase after treatments and values less than one indicate a decrease after treatments relative to bole-only harvesting with no compaction. BEC Site Height Tot. vol. Healthy vigour Dead or dying Total disease IDF        Black Pines 1.01 0.96 0.91 1.54 1.07  Dairy Creek 1.00 1.28 1.60 0.38 0.82  O'Connor Lake 1.02 1.20 0.72 0.78 1.72 SBS        Log Lake 1.05 1.21 1.19 0.77 1.15  Skulow Lake 0.79 0.78 0.60 1.15 1.01  Topley 1.09 1.27 0.83 1.50 1.30  All sites 1.00 1.12 0.97 1.02 1.18   Context-dependence  This thesis again highlights that effects of management on forest productivity are context-dependent (Puettmann et al. 2009; Page-Dumroese et al. 2010a; Ponder et al. 2012). Each site differs in inherent site quality making the trajectory of forest productivity recovery also vary (Figure 6.1). Datasets on topsoil-nutrient content (current and pre-harvest), foliar nutrition, slash and soil physical properties were summarized for the six LTSP sites. This data was used to develop hypotheses on why certain sites may have different response of tree growth or health to soil-disturbance treatments.     83  Figure 6.1. Lodgepole pine height (A) and average tree volume (B) at ages 1, 3, 5, 10, 15 and 20 averaged for each site: Black Pines (red), Dairy Creek (orange), O’Connor Lake (yellow), Log Lake (green), Skulow Lake (blue) and Topley (purple).  Skulow Lake may be the least resilient site to organic-matter removal because of the poor soil fertility and low amounts of soil carbon and slash. Skulow Lake had the smallest topsoil-carbon content prior to harvest and was the least productive site indicated by net merchantable volume (Table 6.2). Skulow Lake still has the lowest values for topsoil carbon content (kg/ha), but also has the least amount of topsoil N and mineralizable N content (Table 6.2). In addition to low soil nutrition, Skulow Lake also had the least amount of slash left on site after harvesting and currently has the lowest forest-floor mass (Table 6.2). These characteristics of low soil fertility and carbon content have previously been suggested as indicators of site sensitivity to biomass harvesting (Page-Dumroese et al. 2010b; Thiffault et al. 2011; Hazlett 2014). The finer textured soils at Skulow Lake could also make this site more susceptible to the negative effects of soil compaction. The sensitivity of fine-texture soils to soil compaction during harvesting has been previously documented (Page-Dumroese et al. 2006; Kimsey et al. 2011; Ponder et al. 2012).      84 Table 6.2. Data of pre-harvest topsoil-nutrient content (kg/ha), current topsoil-nutrient content (kg/ha), and organic-matter properties summarized for each of the six sites: Black Pines (BP), Dairy Creek (DC), O’Connor Lake (OC), Log Lake (LL), Skulow Lake (SL) and Topley (TO). Category Measure BP DC OC LL SL TO Pre-harvest topsoil-nutrient content (kg/ha)       C  53532 63523 59505 49325 36289 85336  N 2263 2591 2993 2026 2046 3366  Minrl N 67 81 82 34 44 77  P 207 263 129 78 31 21  S 203 202 223 166 121 320  Ca 3409 2880 4068 1127 1888 3225  K 363 328 407 149 153 238  Mg 211 281 332 103 1077 615 Current topsoil-nutrient content (kg/ha)        C 58903 47034 39509 39804 32183 63171  N 2190 2025 1990 1530 1463 2635  Minrl N 52 58 68 40 31 60  P 210 252 155 71 39 23  S 240 184 198 85 132 230  Ca 3713 2689 3133 970 2014 3482  K 354 332 414 90 146 147  Mg 225 276 286 95 1157 649 Organic-material        Total slash 24.93 19.92 15.03 15.19 6.93 8.94  Forest-floor bulk density 0.10 0.10 0.08 0.11 0.08 0.13  Forest-floor mass 4.51 4.70 3.52 3.81 2.73 5.63   To determine if pre-harvest data could differentiate sites, and therefore be used to determine site sensitivity, discriminant analysis was used with pre-harvest data of net merchantable volume (m3), topsoil carbon (kg/ha), topsoil N (kg/ha), topsoil mineralizable N (kg/ha), topsoil P (kg/ha) and topsoil K (kg/ha) averaged per plot. Results suggest that these sites were significantly different before harvesting occurred (p < 0.0001; Wilks’ Lambda approximate F-test6). Axis 1 was related to topsoil P content and differentiated the three IDF sites (Black Pines, Dairy Creek and O’Connor Lake; Figure 6.2). Axis 2 was related to topsoil C and N content and differentiated the three SBS sites (Log Lake, Topley and Skulow Lake; Figure 6.2). Using the same analysis with current data (year 15 for the IDF and year 20 for the SBS) the sites still significantly differed from each other (p < 0.0001), but the IDF sites have become more similar and axis 2 is more strongly related to tree volume (m3), which is highest at Log Lake (Figure 6.3).                                                 6 Analysis was conducted in JMP 11 (SAS Institute Inc. 2012).  85 -12-10-8-6-4-202Canonical2 Black	PinesDairy	CreekLog	Lake O'Connor	LakeSkulow	LakeTopley-4 -2 0 2 4 6 8 10Canonical1SiteBlack	PinesDairy	CreekLog	LakeO'Connor	LakeSkulow	LakeTopley Figure 6.2. Discriminant analysis using pre-harvest data of net merchantable volume (m3), topsoil carbon (kg/ha), topsoil N (kg/ha), topsoil mineralizable N (kg/ha), topsoil P (kg/ha) and topsoil K (kg/ha) averaged per plot data to differentiate sites: Black Pines (red), Dairy Creek (orange), O’Connor Lake (yellow), Log Lake (green), Skulow Lake (blue) and Topley (purple). Normal 50% contour lines delineate the groups.       86 -6-4-202468Canonical2Black PinesDairy CreekLog LakeO'Connor LakeSkulow LakeTopley-6 -4 -2 0 2 4 6Canonical1SiteBlack PinesDairy CreekLog LakeO'Connor LakeSkulow LakeTopley Figure 6.3. Discriminant analysis using current (year 15 and 20) data of total stem volume (m3), topsoil carbon (kg/ha), topsoil N (kg/ha), topsoil mineralizable N (kg/ha), topsoil P (kg/ha) and topsoil K (kg/ha) averaged per plot data to differentiate sites: Black Pines (red), Dairy Creek (orange), O’Connor Lake (yellow), Log Lake (green), Skulow Lake (blue) and Topley (purple). Normal 50% contour lines delineate the groups.   Growth-and-yield model  In light of the site differences in resilience and the minimal impact of a one-time soil disturbance event during logging at these sites, it is interesting to determine the validity of predictive growth-and-yield models. Comparing projected productivity with real data helps to determine confidence in model outputs that are used to determine future timber supply and annual allowable harvest rates. The main growth-and-yield model used in British Columbia is the Tree And Stand Simulator (TASS) model with the Table Interpolation Program for Stand Yields (TIPSY) interface (Mitchell et al. 2000). Losses to biomass production are accounted for using two Operational Adjustment Factors (OAFs). OAF values range from 0 to 1, with a value of 0 indicating all wood production is lost and a value of 1 indicating no loss in wood production. The first OAF is a constant multiplier with a default value of 0.85 (15% loss of productivity at all  87 ages) that accounts for unproductive areas (4%), holes caused by irregular stand development (4%), endemic disease (4%) and other random risk factors (3%; Stearns-Smith 2013). The second OAF is an incremental multiplier that is indexed to a base age of 100 with a default value of 0.95 that accounts for decay, waste, breakage and forest health impacts that manifest slowly over time (Sterns-Smith 2013).   The six LTSP sites in this study were submitted to TIPSY and model results compared to 15- to 20-year-old lodgepole pine growth and canopy-closure data. TIPSY version 4.3 was downloaded7 on December 15th, 2015. Input data for the six LTSP sites included region, average 4% slope, and a pure stand of lodgepole pine planted at 1600 stems/ha. Site-index values for these sites were collected from the Berch et al. (2010) establishment report (Table. 6.3). Default OAFs values were used. Other than Skulow Lake, which was lower than predicted, TIPSY output data was a good approximation of canopy closure and mean tree height (Figure 6.4 and 6.5). The model underestimated mean tree height except at Skulow Lake (Figure 6.5), but overestimated stem density especially in the IDF (Figure 6.6). Some of these inaccuracies in the model outputs could be related to the difficulty of accurately measuring site index from height by age curves of older trees on the sites. It is likely that the site index values estimated for these sites are too low. TIPSY is very sensitive to site index input values, so this could have caused some of the underestimations of tree height.   Table 6.3. Site-index values for Douglas-fir (Fd), white spruce (Sx) and lodgepole pine (Pl) based on pre-harvest data collected at each site and the value used for the TIPSY models.  Site Site index Entered to TIPSY Log Lake Fd – 19.2 Sx – 18.5 19.0 Skulow Lake Pl - 17.9 17.9 Topley Pl - 15.7 15.7 Dairy Creek Pl – 18 Fd – 18.1 18.0 Black Pines Pl - 14.7 Fd – 15.6 14.7 O'Connor Lake Pl - 14.3 Fd – 16.8 14.3                                                  7 https://www.for.gov.bc.ca/hts/growth/download/download.html  88 Data typeRealTIPSYBlack Pines Dairy Creek O'Connor LakeSiteAge0 5 10 15 20 25Canopy closure (%)0204060800 5 10 15 20 25 0 5 10 15 20 25  Log Lake Skulow Lake TopleySiteAge0 5 10 15 20 25Canopy closure (%)0204060800 5 10 15 20 25 0 5 10 15 20 25 Figure 6.4. Canopy closure (%) based on TIPSY model (red) and real data (blue) from each site.        89 Data typeRealTIPSYBlack Pines Dairy Creek O'Connor LakeSiteAge0 5 10 15 20 25Tree height (m)0.02.55.07.510.00 5 10 15 20 25 0 5 10 15 20 25  Data typeLog Lake Skulow Lake TopleySiteAge0 5 10 15 20 25Tree height (m)0.02.55.07.510.00 5 10 15 20 25 0 5 10 15 20 25 Figure 6.5. Average tree height (m) based on TIPSY model (red) and real data (blue) from each site.        90 Data typeRealTIPSYBlack Pines Dairy Creek O'Connor LakeSiteAge0 5 10 15 20 25Stem density (#/ha)800900100011001200130014000 5 10 15 20 25 0 5 10 15 20 25  Log Lake Skulow Lake TopleySiteAge0 5 10 15 20 25Stem density (#/ha)800900100011001200130014000 5 10 15 20 25 0 5 10 15 20 25 Figure 6.6. Stand density (stems/ha) based on TIPSY model (red) and real data (blue) from each site.   6.2 Future research directions The evidence that groups of faster-growing trees also had higher disease occurrence highlights the gap in knowledge on how management interventions to increase tree growth impact future tree health. For example, in British Columbia the Tree Improvement Program has successfully increased short-term growth of planted seedlings, yet the long-term health of these seedlings is unknown. The Tree Improvement Program started in the 1960s with the goal of improving tree growth through selecting fast-growing trees and breeding them in nurseries in order to increase timber production (Xie and Yanchuk 2003). This program has been very successful. Short-term seedling growth has been increased by 30% relative to unselected seedlings (Xie and Yanchuk  91 2003). Using growth-and-yield models, these short-term gains in tree growth are projected to result in 15% - 20% more wood at harvest (BC Ministry of Forests, Mines and Lands 2010). Approximately 75% of the harvested or burnt area has been planted with these ‘plus’ seedlings (The State of BC’s Forests 2010) and is legal required on publicly owned lands (Silviculture Practices Regulations Part 2, Division 1 of the Forest Practices Code).   Plant physiology would suggest that these ‘plus’ trees might allocate fewer resources to defence (Herms and Mattson 1992). Inevitability, the effect of selecting fast-growing trees on tree defence is a complex and context-dependent issue that can only be tested with rigorous, thoughtful and objective studies. In the past, the Tree Improvement Program considered tree resistance to pests in reaction to damaging pest outbreaks (Yanchuk 2012). This reactionary approach, of finding genetic resistance to a pest after the damage has already occurred, can aid regeneration but does not increase the resistance of the forest to large outbreaks. In the past five years, characteristics of pest resistance and recovery are being selected for in addition to fast growth (Yanchuk 2009).   One limitation of the LTSP program is that it cannot be used to identify or learn more about sites that will be more susceptible to productivity loss after disturbance (sensitive sites). This is because locations for the LTSP sites were selected to areas with ‘average’ characteristics of a productive forest. Therefore, low productivity areas or those with poor site nutrition were not selected for these studies. This site selection procedure is useful from a management point-of-view, but further research on sensitive sites could be useful as these sites are at the highest risk of being negatively impacted by harvesting.   A major success of the LTSP project has been to identify regional, species and site trends in sensitivity to soil disturbance. Using the varied response of stand condition to treatments to test context-dependency is a worthwhile continued area of research. This information is critical for determining place-based best-management practices. Continued effort on this research topic is needed because the relationships will change over time (Figure 1.1). Also, new indicators of stand productivity, including stand health, have not yet been fully investigated for context-dependencies.   92  Bibliography  Allen, E.A., Morrison, D.J., and Wallis, G. 1996. Common tree diseases of British Columbia. Natural Resources Canada, Pacific Forestry Centre. Victoria, BC.  Andersson, F.O., Ågren, G.I., and Führer, E. 2000. Sustainable tree biomass production. Forest Ecology and Management 132:51-62.  Ares, A., Terry, T.A., Piatek, K.B., Harrison, R.B., Miller, R.E., Flaming, B.L., Licata, C.W., Strahm, B.D., Harrington, C.A., Meade, R., Anderson, H.W., Brodie, L.C., and Kraft, J.M. 2007. The fall river long-term site productivity study in coastal Washington: site characteristics, methods, and biomass and carbon and nitrogen stores before and after harvest. US Department of Agriculture. Pacific Northwest Research Station. Portland, OR. General Technical Report 691:85. Bacci, L., De Vincenzi, M., Rapi, B., Arca, B., and Benincasa, F. 1998. Two methods for the analysis of colorimetric components applied to plant stress monitoring. Computers and Electronics in Agriculture 19:167-186.  Barnes, A. 2015. 2014 Economic state of the BC Forest Sector. Ministry of Forests, Lands, and Natural Resource Operations. Competitiveness and Innovation Branch. https://http://www.for.gov.bc.ca/ftp/het/external/!publish/web/economic-state/Economic-State-of-BC-Forest-Sector-2014.pdf. Accessed October 2015. BC Ministry of Forests, Lands and Natural Resouce Operations. 2010. The State of British Columbia’s Forests, 3rd ed., Victoria, BC.  BC Ministry of Forests and Range. 2011. Introduction to Forest Health. Forest Practices Branch. Victoria, BC. <http://www.for.gov.bc.ca/hfp/health/FHintro/FHINTRO.HTM>. Accessed September 2014. Berch, S., Curran, M., Chapman, W.K., Dube, S., Hope, G., Kabzems, R., Kranabetter, J.M., and Hannam, K.D. 2010. Long-term soil productivity study (LTSP): The effects of soil compaction and organic matter retention on long-term soil productivity in British Columbia. BC Ministry of Forests and Range. Victoria, BC.  Boege, K., Dirzo, R., Siemens, D., and Brown, P. 2007. Ontogenetic switches from plant resistance to tolerance: minimizing costs with age? Ecology Letters 10:177-187.   93 Boege, K. and Marquis, R.J. 2006. Plant quality and predation risk mediated by plant ontogeny: consequences for herbivores and plants. Oikos 115:559-572.  Forest Practices Branch. 2014. Reference Guide for FDP Stocking Standards. Ministry of Forests and Range. https://www.for.gov.bc.ca/hfp/silviculture/stocking_stds.htm. Accessed June 2014. Bréda, N.J.J. 2003. Ground-based measurements of leaf area index: a review of methods, instruments and current controversies. Journal of Experimental Botany 54:2403-2417.  Brussard, L. and van Faasen, H.C. 1994. Effects of compaction on soil biota and soil biological processes. Pages 215-235 in Soane, B.D. and van Ouwerkerk, C., editors. Soil compaction and crop production. Elsevier, Amsterdam. Bryant, J.P., Clausen, T.P., Reichardt, P.B., McCarthy, M.C., and Werner, R.A. 1987. Effect of nitrogen fertilization upon the secondary chemistry and nutritional value of quaking aspen (Populus tremuloides Michx.) leaves for the large aspen tortrix (Choristoneura conflictana (Wlaker)). Oecologia 73:513-517.  Bulmer, C.E. and Simpson, D.G. 2005. Soil compaction and water content as factors affecting growth of lodgepole pine seedlings on sandy clay soil. Canadian Journal of Soil Science 85:667-679.  Castello, J.D., Leopold, D.J., and Smallidge, P.J. 1995. Pathogens, patterns, and processes in forest ecosystems. BioScience 45:16-24.  Chapman, B.K., Xiao, G., and Myer, S. 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. and Paul, L. 2012. Evidence that Northern pioneering pines with tuberculate mycorrhizae are unaffected by varying soil nitrogen levels. Microbial Ecology:DOI 10.1007/s00248-00012-00076-00240.  Cheng, D. and Igarashi, T. 1987. Fungi associated with natural regeneration of Picea jezoensis Carr. in seed stage-Their distribution on forest floors and pathogenicity to the seeds. Research Bulletins of the College Experiment Forests 44:175-188.  Chishaki, N. and Horiguchi, T. 1997. Responses of secondary metabolism in plants to nutrient deficiency. Plant nutrition - for sustainable food production and environment. p.341-345.   94 Cole, D.W. and Rapp, M. 1980. Chapter 6. Elemental cycling in forest ecosystems. in Reichle, D.E., editor. Dynamic properties of forest ecosystems. Cambridge University Press, Malta. Compton, J.E. and Cole, D.W. 1991. Impact of harvest intensity on growth and nutrition of successive rotations of Douglas-fir. Pages 151-161 in Dyck, W.J. and Mees, C.A., editors. Long-term field trials to assess environmental impacts of harvesting. Forest Research Institute, Rotorua, New Zealand, FRI Bulletin 161. Condron, L., Stark, C., O'Callaghan, M., Clinton, P., and Huang, Z. 2010. Chapter 4. The role of microbial communities in the formation and decomposition of soil organic matter. in Dixon, G.R. and Tilston, E.L., editors. Soil microbiology and sustainable crop production. Springer, New York. Costanza, R., d'Arge, R., de Groot, R., Farber, S., Grasso, M., Hannon, B., Limburg, K., Naeem, S., O'Neill, R.V., Paruelo, J., Raskin, R.G., Sutton, P., and van der Belt, M. 1997. The value of the world's ecosystem services and natural capital. Nature 387:253-260.  Dillen, S.Y., de Beeck, M.O., Hufkens, K., Buonanduci, M., and Phillips, N.G. 2012. Seasonal patterns of foliar reflectance in relation to photosynthetic capacity and color index in two co-occurring tree species, Quercus rubra and Betula papyrifera. Agricultural and Forest Meteorology 160:60-68.  Donaldson, J.R., Kruger, E.L., and Lindroth, R.L. 2006. Competition- and resource- mediated tradeoffs between growth and defensive chemistry in trembling aspen (Populus tremuloides). New Phytologist 169:561-570.  Eckhardt, L.G., Menard, R.D., and Gray, E.D. 2009. Effects of oleoresins and monoterpenes on in vitro growth of fungi associated with pine decline in the Southern United States. Forest Pathology 39:157-167.  Efroymson, R.A., Dale, V.H., Kline, K.L., McBride, A.C., Bielicki, J.M., Smith, R.L., Parish, E.S., Schweizer, P.E., and Shaw, D.M. 2013. Environmental indicators of biofuel sustainability: what about context? Environmental Management 51:291-306.  Egnell, G. 2011. Is the productivity decline in Norway spruce following whole-tree harvesting in the final felling in boreal Sweden permanent or temporary? Forest Ecology and Management 261:148-153.   95 Egnell, G. and Leijon, B. 1999. Survival and growth of planted seedlings of Pinus sylvestris and Picea abies after different levels of biomass removal in clear-felling. Scandinavian Journal of Forest Research 14:303-311.  Egnell, G. and Valinger, E. 2003. Survival, growth, and growth allocation of planted Scots pine trees after different levels of biomass removal in clear-felling. Forest Ecology and Management 177:65-74.  Exelis Visual Information Solutions. 2004. ENVI User's Guide. Research Systems Inc., Boulder, Colorado. Fellin, D.G. 1980a. A review of some relationships of harvesting, residue management and fire to forest insects and disease. Pages 335-414 in Environmental consequences of timber harvesting in rocky mountain coniferous forests. USDA Forest Service General Technical Report INT-90, Missoula, Montana. Firth, J. and Murphy, B. 1989. Skid trails and their effects on the growth and management of young Pinus radiata. New Zealand Journal of Forest Science 19:22-28.  Fleming, R.L., Powers, R.F., Foster, N.W., Kranabetter, J.M., Scott, D.A., Ponder, F.J., Berch, S., Chapman, B.K., Kabzems, R.D., Ludovici, K.H., Morris, D.M., Page-Dumroese, D., Sanborn, P.T., Sanchez, F.G., Stone, D.M., and Tiarks, A.E. 2006. Effects of organic matter removal, soil compaction, and vegetation control on 5-year seedling performance: a regional comparison of Long-Term Soil Productivity sites. Canadian Journal of Forest Research 36:529-550.  Forest Health Protection. 2011. Western gall rust: pine-to-pine rust of branches and stems. Forest Service, Rocky Mountain Region. 2 pages. Forest and Range Practices Act. 2002.Chapter 69. Victoria, BC, Canada.  Frazer, G.W., Canham, C.D., and Lertzman, K.P. 1999. Gap Light Analyzer (GLA): Imaging software to extract canopy structure and gap light transmission indices from true-colour fisheye photographs, users manual and program documentation., Simon Fraser University, Burnaby, British Columbia.  Ginter, D.L., McLeod, K.W., and Sherrod, C. 1979. Water stress in longleaf pine induced by litter removal. Forest Ecology and Management 2:13-20.  Grigal, D.F. 2000. Effects of extensive forest management on soil productivity. Forest Ecology and Management 138:167-185.   96 Hartmann, M., Howes, C.G., VanInsberghe, D., Yu, H., Bachar, D., Christen, R., Nilsson, R.H., Hallam, S.J., and Mohn, W.W. 2012. Significant and persistent impact of timber harvesting on soil microbial communities in Northern coniferous forests. International Society for Microbial Ecology 6:2199-2218.  Hazlett, P.W., Morris, D.M., and Fleming, R.L. 2014. Effects of biomass removals on site carbon and nutrients and jack pine growth in boreal forests. Soil Science Society of America Journal, North American Forest Soils Conference Proceedings. DOI: 10.2136/sssaj2013.08.0372nafsc. Hedges, L.V., Gurevitch, J., and Curtis, P.S. 1999. The meta-analysis of response ratios in experimental ecology. Ecology 80:1150-1156.  Heineman, J.L., Sachs, D.L., Mather, W.J., and Simard, S. 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.  Helms, J.A., Hipkin, C., and Alexander, E.B. 1986. Effects of soil compaction on height growth of a California ponderosa pine plantation. Western Journal of Applied Forestry 1:104-108.  Herms, D.A. and Mattson, M.J. 1992. The dilemma of plants: to grow or defend. Quarterly Review of Biology 67:283-335.  Holocmb, R.W. 1996. The Long-Term Soil Productivity Study in British Columbia. Canadian Forest Service and B.C. Ministry of Forests. Victoria, B.C. FRDA Rep. 256.  Holub, S.M., Terry, T.A., Harrington, C.A., Harrison, R.B., and Meade, R. 2013. Tree growth ten years after residual biomass removal, soil compaction, tillage, and competing vegetation control in a highly-productive Douglas-fir plantation. Forest Ecology and Management 305:60-66.  Johnson, D.W., Cole, D.W., Bledsoe, C.S., Cromack, B., Edmonds, R.L., Gessel, S.P., Grier, B.N., Richards, B.N., and Vogt, K.A. 1982. Nutrient cycling in forests of the Pacific Northwest. Pages 186-232  Analysis of coniferous forest ecosystems in the Western United States. Academic Press, New York. Johnston, J.M. and Crossley Jr., D.A. 2002. Forest recovery in the southeast US: soil ecology as an essential component of ecosystem management. Forest Ecology and Management 155:187-203.   97 Jurgensen, M.F., Harvey, A.E., Graham, R.T., Page-Dumroese, D.S., Tonn, J.R., Larsen, M.J., and Jain, T.B. 1997. Impacts of timber harvesting on soil organic matter, nitrogen, productivity, and health of inland northwest forests. Forest Science 43:234-251.  Jurskis, V. 2005. Eucalypt decline in Australia, and a general concept of tree decline and dieback. Forest Ecology and Management 215:1-20.  Kamaluddin, M., Chang, S.X., Curran, M.P., and Zwiazek, J.J. 2005. Soil compaction and forest floor removal affect early growth and physiology of lodgepole pine and Douglas-fir in British Columbia. Forest Science 51:513-521.  Keenan, T.F., Darby, B., Felts, E., Sonnentag, O., Friedl, M.A., Hufkens, K., O'Keefe, J., Klosterman, S., Munger, J.W., Toomey, M., and Richardson, A.D. 2014. Tracking forest phenology and seasonal physiology using digital repeat photography: a critical assessment. Ecological Applications 24:1478-1489.  Kile, G.A. 2000. Woody root rots of eucalypts. Pages 293-206 in Keane, P.J., Kile, G.A., Podger, F.D. and Brown, B.N., editors. Diseases and pathogens of eucalypts. CSIRO, Melborne. Kimmins, J.P. 2004. Emulating natural forest disturbances: what does this mean. Pages 8-28 in Perera, A.H., Buse, L.J. and Weber, M.G., editors. Emulating natural forest landscape disturbances: concepts and applications. Columbia University Press, New York. Kalra, Y.P., Maynard, D.G. 1991. Methods manual for forest soil and plant analysis. Forestry Canada. Northern Forestry Centre, Edmonton. Informational Report, NOR-X-319E, p. 116.  Kovats, M. 1977. Estimating juvenille tree volumes for provenance and progeny testing. Canadian Journal of Forest Research 7:335-342.  Kranabetter, J.M., Sanborn, P., Chapman, B.K., and Dube, S. 2006. The contrasting response to soil disturbance between lodgepole pine and hybrid white spruce in subboreal forests. Soil Science Society of America Journal 70:1591-1599.  Laiho, R. and Prescott, C.E. 2004. Decay and nutrient dynamics of coarse woody debris in northern coniferous forests: a synthesis. Canadian Journal of Forest Research 34:763-777.  Lindeman, R.H., Merenda, P.F., and Gold, R.Z. 1980. Introduction to bivariate and multivariate analysis. Scott, Foresman, Glenview, IL.  Makinde, E.O. and Salami, A.T. 2013. Remote sensing of vegetation stress and indicators.in Global Geospatial Conference, Addis Ababa, Ethiopia.  98 Mann, L.K., Johnson, D.W., West, D.C., Cole, D.W., Hornbeck, J.W., Martin, C.W., and Rickerk, H. 1988. Effect of whole-tree and stem-only clearcutting on post-harvest hydrologic losses, nutrient capital and regrowth. Forest Science 34:412-428.  McBride, A.C., Dale, V.H., Baskaran, L.M., Downing, M.E., Eaton, L.M., Efroymson, R.A., Garten Jr., C.T., Kline, K.L., Jager, H.I., Mulholland, P.J., Parish, E.S., Schweizer, P.E., and Storey, J.M. 2011. Indicators to support environmental sustainability of bioenergy systems. Ecological Indicators 11:1277-1289.  McKinney, S.T., Fiedler, C.E., and Tomback, D.F. 2009. Invasive pathogen threatens bird-pine mutualism: implications for sustaining a high-elevation ecosystem. Ecological Applications 19:597-607.  McLaughlin, S.B. and Wimmer, R. 1999. Tansley Review no. 104 calcium physiology and terrestrial ecosystem processes. New Phytologist 142:373-417.  Messier, C., Puettmann, K.J., and Coates, K.D. 2013. Managing forests as complex adaptive systems: building resilience to the challenge of global change. Routledge, New York.  Mickler, R.A. 1996. Southern pine forests of North America. Pages 2-57 in Fox, S. and Mickler, R.A., editors. Impact of air pollutants on souther pine forests. Springer, New York. Mitchell, K.J., Stone, M., Grout, S.E., Di Luca, M., Nigh, G.D., Goudie, J.W., Stone, J.N., Nussbaum, A.F., Yanchuk, A.D., Stearns-Smith, S., and Brockley, R. 2000. Table Interpolation program for standard yields (TIPSY) version 4.3. Research Branch, British Columbia Forest and Range, Victoria, BC. Moreira, X., Sampedro, L., Zas, R., and Solla, A. 2008. Alterations of the resin canal system of Pinus pinaster seedlings after fertilization of a healthy and of a Hylobius abietis attacked stand. Trees 22:771-777.  Mori, A., Mizumachi, E., Osono, T., and Doi, Y. 2004. Substrate-associated seedling recruitment and establishment of major conifer species in an old-growth subalpine forest in central Japan. Forest Ecology and Management 196:287-297.  Morrison, D.J., Merler, H., and Norris, D. 1991. Detection, recognition and management of Armillaria and Phellinus root diseases in the southern interior of British Columbia. BC Ministry of Forests, Research Branch. FRDA Report 179. Victoria, BC.  Munck, I.A. and Stanosz, G.R. 2008. Excised shoots of top-pruned red pine seedlings, a source of inoculum of the Diplodia shoot blight pathogen. Forest Pathology 38:196-202.   99 Nambiar, E.K.S. and Sands, R. 1992. Effects of compaction and simulated root channels in the subsoil on root development, water uptake and growth of radiata pine. Tree Physiology 10:297-306.  Newsome, T.A. and Perry, J.L. 2002. Stand-tending and rehabilitation treatment options for 36-year-old, height-repressed lodgepole pine. BC Ministry of Forests, Research Branch. Technical Report 007. Victoria, BC.  Nijland, W., de Jongb, R., de Jongc, S.M., Wulderd, M.A., Batera, C.W., and Coops, N.C. 2014. Monitoring plant condition and phenology using infrared sensitive consumer grade digital cameras. Agricultural and Forest Meteorology 184:98-106.  Nowak, J.T. and Berisford, C.W. 2000. Effects of intensive forest management practices on insect infestation levels and loblolly pine growth. Journal of Economic Entomology 92:336-341.  Oblinger, B.W., Smith, D.R., and Stanosz, G.R. 2011. Red pine harvest debris as a potential source of inoculum of Diplodia shoot blight pathogens. Forest Ecology and Management 262:663-670.  Osier, T.L. and Lindroth, R.L. 2006. Genotype and environment determine allocation to and costs of resistance in quaking aspen. Oecologia 148:293-303.  Page-Dumroese, D., Jurgensen, M., Elliot, W., Rice, T., Nesser, J., Collins, T., and Meurisse, R. 2000. Soil quality standards and guidelines for forest sustainability in northwestern North America. Forest Ecology and Management 138:445-462.  Page-Dumroese, D., Jurgensen, M., Neary, D., Curran, M., and Trettin, C. 2010a. Soil quality is fundamental to ensuring healthy forests. US Department of Agriculture and Forest Service. Pacific Northwest Research Station. General Technical Report 802 Part 1:27-36. Page-Dumroese, D., Jurgensen, M., and Terry, T. 2010b. Maintaining soil productivity during forest or biomass-to-energy thinning harvests in the western United States. Western Journal of Applied Forestry 25:5-11.  Page-Dumroese, D. and Jurgensen, M. 2006. Soil carbon and nitrogen pools in mid- to late-successional forest stands of the northwestern USA: Potential impact of fire. Canadian Journal of Forest Research 36:2270-2284.  Page-Dumroese, D., Jurgensen, M., Tiarks, A.E., Ponder, F.J., Sanchez, F.G., Fleming, R.L., Kranabetter, J.M., Powers, R.F., Stone, D.M., Elioff, J.D., and Scott, D.A. 2006. Soil  100 physical property changes at the North American Long-Term Soil Productivity study sites: 1 and 5 years after compaction. Canadian Journal of Forest Research 36:551-564.  Parish, R. 1994. Tree book: Learning to recognize trees of British Columbia. Canadian Forest Service, Victoria.  Paul, L.R., Chapman, B.K., and Chanway, C.P. 2007. Nitrogen fixation associated with Suillus tomentosus tuberculate ectomycorrhizae on Pinus contorta var. latifolia. Annals of Botany:1-9. 10.1093/aob/mcm061. Paul, L.R., Chapman, B.K., and Chanway, C.P. 2013. Diazotrophic bacteria reside inside Suillus tomentosus/Pinus contorta tuberculate ectomycorrhizae. Botany 91:48-52.  Perry, D.A., Amaranthus, M.P., Borchers, J.G., Borchers, S.L., and Brainerd, R.E. 1989. Bootstrapping in ecosystems: internal interactions largely determine productivity and stability in biological systems with strong positive feedback. BioScience 39:230-237.  Petach, A.R., Toomey, M., Aubrecht, D.M., and Richardson, A.D. 2014. Monitoring vegetation phenology using an infrared-enabled security camera. Agricultural and Forest Meteorology 195-196:143-151.  Ponder, F.J., Fleming, R.L., Berch, S., Busse, M.D., Elioff, J.D., Hazlett, P.W., Kabzems, R.D., Kranabetter, J.M., Morris, D.M., Page-Dumroese, D., Palik, B.J., Powers, R.F., Sanchez, F.G., Scott, D.A., Stagg, R.H., Stone, D.M., Young, D.H., Zhang, J., Ludovici, K.H., McKenney, D.W., Mossa, D.S., Sanborn, P.T., and Voldseth, R.A. 2012. Effects of organic matter removal, soil compaction and vegetation control on 10th year biomass and foliar nutrition: LTSP continent-wide comparisons. Forest Ecology and Management 278:35-54.  Powers, R.F. 2006. Long-Term Soil Productivity: genesis of the concept and principles behind the program. Canadian Journal of Forest Research 36:519-528.  Powers, R.F., Alban, D.H., Miller, R.E., Tiarks, A.E., Wells, C.G., Avers, P.E., Cline, R.G., Fitzgerald, R.O., and Loftus Jr., N.S. 1990. Sustaining site productivity in North American forests: problems and prospects. Pages 49-79 in Seventh North American Forest Soil Conference on Sustained Productivity of Forest Soils., Faculty of Forestry, University of British Columbia, Vancouver, BC.  101 Powers, R.F., Scott, D.A., Sanchez, F.G., Voldseth, R., Page-Dumroese, D., Elioff, J.D., and Stone, D.M. 2005. The North American long-term soil productivity experiment: Findings from the first decade of research. Forest Ecology and Management 220:31-50.  Puettmann, K.J., Coates, K.D., and Messier, C. 2009. A Critique of Silviculture: Managing for Complexity. Island Press, Washington, DC.  Reid, A.M., Chapman, B.K., Kranabetter, J.M., and Prescott, C.E. 2015. Response of lodgepole pine health to soil-disturbance treatments in British Columbia, Canada. Canadian Journal of Forest Research 45:1045-1055. DOI: 10.1139/cjfr-2015-0029. Reinhart, D.P., Haroldson, M.A., Mattson, D.J., and Gunther, K.A. 2001. Effects of exotic species on Yellowstone's grizzly bears. Western North American Naturalist 61:277-288.  Rhoades, C.C., Brosi, S.L., Dattilo, A.J., and Vincelli, P. 2003. Effects of soil compaction and moisture on incidence of phytophthora root rot on American chestnut (Castanea dentata) seedlings. Forest Ecology and Management 184:47-54.  Richardson, A.D., Jenkins, J.P., Braswell, B.H., Hollinger, D.Y., Ollinger, S.V., and Smith, M.L. 2007. Use of digital webcam images to track spring green-up in a deciduous broadleaf forest. Oecologia 152:323-334.  Ricklefs, R.E. 1990. Ecology. Freeman, New York. 896. Roberts, S.D., Harrington, C.A., and Terry, T.A. 2005. Harvest residue and competing vegetation affect soil moisture, soil temperature, N availability, and Douglas-fir seedling growth. Forest Ecology and Management 205:333-350.  Roxby, G.E. and Howard, T.E. 2013. Whole-tree harvesting and site productivity: Twenty-nine northern hardwood sites in central New Hampshire and western Maine. Forest Ecology and Management 293:114-121.  Ruth, B., Hoque, E., Weisel, B., and Hutzler, P.J.S. 1991. Reflectance and fluorescence parameters of needles of norway spruce affected by forest decline. Remote Sensing of Environment 38:35-44.  R Studio Team. 2015. RStudio: Integrated Development for R. RStudio, Inc., Boston, MA. Saitoh, T.M., Nagai, S., Saigusa, N., Kobayashi, H., Suzuki, R., Nasahara, K.N., and Muraoka, H. 2012. Assessing the use of camera-based indices for characterizing canopy phenology in relation to gross primary production in a deciduous broad-leaved and an evergreen coniferous forest in Japan. Ecological Informatics 11:45-54.   102 Sampedro, L., Moreira, X., and Zas, R. 2011. Costs of constitutive and herbivore-induced chemical defenses in pine trees emerge only under low resources availability. Journal of Ecology 99:818-827.  Sanchez, F.G., Scott, D.A., and Lundovici, K.H. 2006a. Negligible effects of severe organic matter removal and soil compaction on loblolly pine growth over 10 years. Forest Ecology and Management 227:145-154.  SAS. Institute Inc. 2012. JMP, Version 10. Cary, N.C. Sayer, E.J. 2006. Using experimental manipulation to assess the roles of leaf litter in the functioning of forest ecosystems. Biological Reviews 81:1-31.  Schowalter, T.D. and Turchin, P. 1993. Southern pine beetle infestation development: interaction between pine and hardwood basal areas. Forest Science 39:201-210.  Scott, D.A. and Dean, T.J. 2006. Energy trade-offs between intensive biomass utilization, site productivity loss, and ameliorative treatments in loblolly pine plantations. Biomass Bioenergy 30:1001-1010.  Sheriff, D.W. and Nambiar, E.K.S. 1995. Effect of subsoil compaction and three densities of stimulated root channels in the subsoil on growth, carbon gain and water uptake of Pinus radiata. Australian Journal of Plant Physiology 22:1001-1013.  Skovsgaard, J.P. and Vanclay, J.K. 2008. Forest site productivity: a review of the evolution of dendrometric concepts for even-aged stands. Forestry 81:13-31.  Smith, C.T., McCormack, M.L., Hornbeck, J.W., and Martin, C.W. 1986. Nutrient and biomass removals from a red spruce-balsam fir whole-tree harvest. Canadian Journal of Forest Research 16:381-388.  Sonnentag, O., Hufkens, K., Teshera-Sterne, C., Young, A.M., Friedl, M., Braswell, B.H., Milliman, T., O'Keefe, J., and Richardson, A.D. 2012. Digital repeat photography for phenological research in forest ecosystems. Agricultural and Forest Meteorology 152:159-177.  Soudani, K., Hmimina, G., Delpierre, N., Pontailler, J.-Y., Aubinet, M., Bonal, D., Caquet, B., De Grandcourt, A., Burban, B., Flechard, C., Guyon, D., Grainer, A., Gross, P., Heinesh, B., Longdoz, B., Loustau, D., Moureaux, C., Ourcival, J.-M., Rambal, S., Saint André, L., and Dufrêne, E. 2012. Ground-based network of NDVI measurements for tracking  103 temporal dynamics of canopy structure and vegetation phenology in different biomes. Remote Sensing of Environment 123:234-245.  Stearns-Smith, S., 2013. Procedures for initializing TIPSY with silviculture survey data, Version 1. Stearns-Smith and Associates. Stewart, R., Froelich, H., and Olsen, E. 1988. Soil compaction: an economic model. Western Journal of Applied Forestry 3:20-22.  Stone, D.M. and Elioff, J.D. 1998. Soil properties and aspen developent five years after compaction and forest floor removal. Canadian Journal of Soil Science 78:51-58.  Takahashi, K. 1994. Effect of size structure, forest floor type and disturbance regime on tree species composition in a coniferous forest in Japan. Journal of Ecology 82:769-773.  Takahashi, M., Sakai, Y., Ootomo, R., and Shiozaki, M. 2000. Establishment of tree seedlings and water-soluble nutrients in coarse woody debris in an old-growth Picea-Abies forest in Hokkaido, northern Japan. Canadian Journal of Forest Research 30:1148-1155.  Tan, X., Curran, M., Chang, S.X., and Maynard, D.G. 2009. Early growth responses of lodgepole pine and Douglas-fir to soil compaction, organic matter removal, and rehabilitation treatments in Southeastern British Columbia. Forest Science 55:210-220.  Thibodeau, L., Raymond, P., Camire, C., and Munson, A.D. 2000. Impact of precommercial thinning in balsam fir stands on soil nitrogen dynamics, microbial biomass, decomposition and foliar nutrition. Canadian Journal of Forest Research 30:229-238.  Thiffault, E., Barrette, J., Paré, D., Titus, B.D., Keys, K., Morris, D.M., and Hope, G. 2014. Developing and validating indicators of site suitability for forest harvesting residue removal. Ecological Indicators 43:1-18.  Thiffault, E., Hannam, K.D., Paré, D., Titus, B.D., Hazlett, P.W., Maynard, D.G., and Brais, S. 2011. Effects of forest biomass harvesting on soil productivity in boreal and temperate forests - A review. . Environmental Review 19:278-309.  Turner, M.G., Donato, D.C., and Romme, W.H. 2013. Consequences of spatial heterogeneity for ecosystem services in changing forest landscapes: priorities for future research. Landscape Ecology 28:1081-1097.  Van Cleve, K. and Dyrness, C.T. 1983. Introduction and overview of a multidisciplinary research project: The structure and function of a black spruce (Picea mariana) forest in relation to other fire-affected taiga ecosystems. Canadian Journal of Forest Research 13:695-702.   104 van Lierop, P., Lindquist, E., Sathyapala, S., and Franceschini, G. 2015. Global forest area disturbance from fire, insect pests, diseases and severe weather events. Forest Ecology and Management 352:78-88.  van Mantgem, P.J., Stephenson, N.L., Byrne, J.C., Daniels, L.D., Franklin, J.F., Fulé, P.Z., Harmon, M.E., Larson, A.J., Smith, J.M., Taylor, A.H., and Veblen, T.T. 2005. Widespread increase of tree mortality rates in the Western United States. Science 323:521-524.  Vance, E.D., Aust, W.M., Strahm, B.D., Froese, R.E., Harrison, R.B., and Morris, L.A. 2014. Biomass harvesting and soil productivity: Is the science meeting our policy needs? Soil Science Society of America Journal 78:S95-S104.  Veteli, T.O., Koricheva, J., Niemelä, and Kellomäki, S. 2006. Effects of forest managment on the abundance of insect pests on Scots pine. Forest Ecology and Management 231:214-217.  von Wilpert, K. and Schaffer, J. 2006. Ecological effects of soil compaction and initial recovery dynamics: A preliminary study. European Journal of Forest Research 125:129-138.  Wallis, C., Eyles, A., Chorbadjian, R., Gardener, B.M., Hansen, R., Cipollini, D., Herms, D.A., and Bonello, P. 2008. Systemic induction of phloem secondary metabolism and its relationship to resistance to a canker pathogen in Austrian pine. New Phytologist 177:767-778.  Weaver, D. 2013. Table 2.4 Seedlings planted on crown land in 2012/2013 by forest region. BC Ministry of Forests and Range. Forest Practices Branch. http://www.for.gov.bc.ca/hfp/silviculture/statistics/2012-13.htm. Accessed November 2013. Wickman, B.E., Mason, R.P., and Swetnam, T.W. 1993. Searching for long-term patterns of forest insect outbreaks.in Individuals, populations, and patterns, Norwich, England. Williams, C.D., Dillion, A.B., Girling, R.D., and Griffin, C.T. 2013. Organic soils promote the efficacy of entomopathogenic nematodes, with different foraging strategies, in the control of a major forest pest: A meta-analysis of field trial data. Biological Control 65:357-364.  Woebbecke, D.M., Meyer, G.E., Von Bargen, K., and Mortensen, D.A. 1995. Color indices for weed identification under various soil residue and lighting conditions. American Society of Agricultural Engineers 38:259-269.   105 Woods, A. and Coates, K.D. 2013. Are biotic disturbance agents challenging basic tenets of growth and yield and sustainable forest management? Forestry 86:543-554. doi:10.1093/forestry/cpt026. Xie, C.-Y. and Yanchuk, A.D. 2003. Breeding values of parental trees, genetic worth of seed orchard seedlots, and yields of improved stocks in British Columbia. Western Journal of Applied Forestry 18:88-100.  Yanchuk, A.D. 2009. Lodgepole pine, Western white pine (interior), Ponderosa pine, Broadleaves (interior). BC Ministry of Forests. http://www.for.gov.bc.ca/hre/forgen/interior/pine.htm - 1. Accessed March 2014. Yang, X., Tang, J., and Mustard, J.F. 2014. Beyond leaf color: comparing camera-based phenological metrics with leaf biochemical, biophysical, and spectral properties throughout the growing season of a temperate deciduous forest. Journal of Geophysical Research: Biogeosciences 119:181-191.  Yukon Energy, Mines and Resources. 2015. Western gall rusts: Yukon forest health - forest insect and disease. Forest Management Branch.  Zabowski, D., Java, B., Scherer, G., Everett, R.L., and Ottmar, R. 2000. Timber harvesting residue treatment: Part 1. Responses of conifer seedlings, soils and microclimate. Forest Ecology and Management 126:25-34.  Zhong, J. and van der Kamp, B.J. 1999. Pathology of conifer seed and timing of germination in high-elevation subalpine fir and Engelmann spruce forests of the southern interior of British Columbia. Canadian Journal of Forest Research 29:187-193.     106  Appendix A: Data tables  Table A1. Stand-health metrics of the number of healthy trees, dead or dying trees, trees with foliar disease, trees with western gall rust occurrence and trees with symptoms of root disease within each treatment plot at the Interior Douglas-Fir (IDF) and the Sub-Boreal Spruce (SBS) zones. Treatments consist of bole-only harvesting (OM1), whole-tree harvesting (OM2), and whole-tree harvesting plus forest-floor removal (OM3). Soil-compaction treatments consist of no soil compaction (C0), light soil compaction (C1), and heavy soil compaction (C2). Site Treatment Healthy trees Dead or dying Foliar disease Gall rust Root disease Black Pines OM1C0 58 3 12 5 3  OM1C1 56 3 15 9 10  OM1C2 51 5 9 1 7  OM2C0 56 5 9 0 2  OM2C1 43 11 22 4 6  OM2C2 44 8 10 6 2  OM3C0 59 1 6 4 1  OM3C1 51 2 13 7 5  OM3C2 62 2 10 8 5 Dairy Creek OM1C0 23 30 42 1 16  OM1C1 30 10 45 6 9  OM1C2 24 17 52 0 3  OM2C0 44 13 15 1 12  OM2C1 40 19 25 1 5  OM2C2 17 24 43 2 1  OM3C0 42 2 66 1 8  OM3C1 56 5 31 2 9  OM3C2 42 1 45 1 6 O'Connor Lake OM1C0 29 29 27 1* 2  OM1C1 32 20 26 1* 5  OM1C2 27 37 23 1* 4  OM2C0 6 36 43 1* 1  OM2C1 2 38 57 1* 1  OM2C2 36 18 33 2* 5  OM3C0 3 14 83 5* 6  OM3C1 45 4 29 3* 9  OM3C2 15 14 55 7* 12           107 Continued… Site Treatment Healthy trees Dead or dying Foliar disease Gall rust Root disease Log Lake OM1C0 50 14 26 44 11  OM1C1 38 15 40 62 4  OM1C2 62 18 12 46 4  OM2C0 66 4 34 67 6  OM2C1 58 10 36 52 3  OM2C2 60 13 31 67 6  OM3C0 77 9 24 54 11  OM3C1 66 9 22 65 5  OM3C2 47 8 29 54 11 Skulow Lake OM1C0 33 13 54 4 9  OM1C1 19 18 51 3 9  OM1C2 20 11 64 6 12  OM2C0 9 21 55 4 7  OM2C1 14 22 11 7 24  OM2C2 21 10 54 2 23  OM3C0 28 8 36 8 24  OM3C1 27 15 18 13 23  OM3C2 20 15 51 6 30 Topley OM1C0 73 3 28 28 4  OM1C1 50 5 49 36 3  OM1C2 66 2 27 33 9  OM2C0 47 7 26 24 13  OM2C1 42 3 53 43 6  OM2C2 43 3 75 42 5  OM3C0 80 1 28 46 1  OM3C1 84 5 5 38 3  OM3C2 70 10 17 37 3 * indicates that one gall rust occurrence was added for analysis         108 Table A2. Averaged mineral-soil fine-fraction bulk density (g/cm3) and total-soil carbon (kg/ha) within each treatment plot at the Interior Douglas-Fir (IDF) and Sub-Boreal Spruce (SBS) zones. Organic-matter removal treatments consist of bole-only harvesting (OM1), whole-tree harvesting (OM2), and whole-tree harvesting plus forest-floor removal (OM3). Soil-compaction treatments consist of no soil compaction (C0), light soil compaction (C1), and heavy soil compaction (C2). BEC Site Treatment Bulk density Soil carbon IDF Black Pines OM1C0 0.86 51359   OM1C1 0.98 58554   OM1C2 1.05 92637   OM2C0 0.90 75017   OM2C1 1.02 64378   OM2C2 1.11 74877   OM3C0 1.08 37416   OM3C1 1.15 42225   OM3C2 1.13 33667  Dairy Creek OM1C0 0.85 48567   OM1C1 0.96 69363   OM1C2 1.01 63952   OM2C0 0.91 74254   OM2C1 1.01 51269   OM2C2 1.00 45644   OM3C0 0.97 20640   OM3C1 1.08 20879   OM3C2 1.09 28735  O'Connor Lake OM1C0 0.87 40742   OM1C1 0.94 48911   OM1C2 0.94 46840   OM2C0 0.87 43960   OM2C1 1.02 45091   OM2C2 1.06 44225   OM3C0 1.03 27668   OM3C1 1.08 27489   OM3C2 1.00 30652               109 Continued…  BEC Site Treatment Bulk density Soil carbon SBS Log Lake OM1C0 0.98 39858   OM1C1 1.33 48289   OM1C2 1.17 58938   OM2C0 1.04 34921   OM2C1 1.27 53374   OM2C2 1.19 48495   OM3C0 1.06 29688   OM3C1 1.19 18942   OM3C2 1.20 25728  Skulow Lake OM1C0 1.24 35823   OM1C1 1.26 36024   OM1C2 1.36 39951   OM2C0 1.27 34451   OM2C1 1.39 30574   OM2C2 1.41 39633   OM3C0 1.30 26605   OM3C1 1.37 26142   OM3C2 1.46 20447  Topley OM1C0 1.24 78306   OM1C1 1.41 65925   OM1C2 1.24 79450   OM2C0 1.12 78755   OM2C1 1.45 77870   OM2C2 1.29 83117   OM3C0 1.38 36580   OM3C1 1.38 38070   OM3C2 1.31 30464    

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0306863/manifest

Comment

Related Items